MJLorton Solar Power and Electronic Measurement Equipment Forum
Youtube Video Episodes => YouTube Video Episodes => Topic started by: MJLorton on September 24, 2012, 08:42:26 AM
-
Hi Folks,
Here is the "final" prototype circuit that I'll be using for my first build.
I will use this as a test to see how well it works and do any troubleshooting / learning from here....
Electronic DC Load #5 - Final circuit before build:
http://www.youtube.com/watch?v=cvc0XsB7LMk
Link to Kibi's "Dave Jones' dummy load with added features" :
http://mjlorton.com/forum/index.php?topic=29.msg92#msg92
Dave Jones / EEVblog video with the original post and design by Dave Jones:
http://www.youtube.com/watch?v=8xX2SVcItOA
WHAT IS AN OP-AMP?:
http://talkingelectronics.com/projects/OP-AMP/OP-AMP-1.html
My components:
* Infineon MOSFET N-channel 200V 13.5A TO-220 - BUZ31L H (I will be using two of these in parallel)
* Vishay 534 Series Pot with 6.34mm shaft, 50K
* Vishay 534 Series Pot with 6.34mm shaft, 5K
* 1 x 10 k Ohm resistor
* 1 x 10 turn 10K trim pot
* Arcol HS50 Al house wirewound high power resistor,1R (1 ohm) 50W - HS50 1R J
* National Semiconductor Quad op amp,LM324N 1MHz DIP14
* 1uF Capacitor
Cheers,
Martin.
-
Hi Martin
Thanks for the info. Below is my circuit on a breadboard inspired by you and your videos. For my op-amps i use a LM358 as its what i had kicking around and only contains 2 op-amps. Still need to get some 10 turn pots though. I also notice without the 1 ohm load resistor the adjustment isn't as nice as it is with it in the circuit.
Thanks again, Mark
(http://i37.servimg.com/u/f37/11/77/76/68/img_0913.jpg)
-
The ten turn pot it is a bit costly ;-)
Martin what is the specifications ?
Max Load in Ampere ?
Min & Max input voltage ?
-
Hi Martin,
This is the same post I put on your YouTube channel:
I completely understand if you do not want to do this but I was hoping you might wait a couple of days before you start your final build. I have the same E-MOSFET that you have and have gotten most of my components to build the el-load. I will start a project tonight and looking over your diagram I think I have a few suggestions. Since I will include my reasoning I think it will take multiple posts to this thread. I will include pics and maybe a link to videos.
I have never built a circuit with a MOSFET but like you I like to experiment and learn new things. Some of things I will try and address in the next couple of posts: LM324N, the 1 uF capacitor, adding build out resistors to the two FET gates to prevent problems with the phase margin, the 10k-ohm voltage divider, the 10 k opamp load resistor, and
current hogging.
As I mentioned before I really like your attitude of exploration and your YouTube channel. One negative of watching Dave Jones and your videos it is costing me money. I order the TekPower TP4000ZC and it seems to work including connection to a PC via USB adapter. John
-
Hi Martin
Thanks for the info. Below is my circuit on a breadboard inspired by you and your videos. For my op-amps i use a LM358 as its what i had kicking around and only contains 2 op-amps. Still need to get some 10 turn pots though. I also notice without the 1 ohm load resistor the adjustment isn't as nice as it is with it in the circuit.
Thanks again, Mark
(http://i37.servimg.com/u/f37/11/77/76/68/img_0913.jpg)
Hi Mark,
Thanks for the great post. Interesting to hear your feedback about the adjustment without the 1 Ohm resister as well.
I look forward to following your progress.
Cheers,
Martin.
-
The ten turn pot it is a bit costly ;-)
Martin what is the specifications ?
Max Load in Ampere ?
Min & Max input voltage ?
Hello Kiriakos,
The maximum load on mine will be limited to 2 amp. The FETs can handle 13.5 amp as long as the total power is within 90 watt.
The FETs are rated for 200 v but again as long as power is within 90 watt. Mine will operate from about 0.5 volt to 30 volt.
I plan to build another one that will do about 5-10 amp at 12-15 volt.
Cheers,
Martin.
-
Hi Martin,
This is the same post I put on your YouTube channel:
I completely understand if you do not want to do this but I was hoping you might wait a couple of days before you start your final build. I have the same E-MOSFET that you have and have gotten most of my components to build the el-load. I will start a project tonight and looking over your diagram I think I have a few suggestions. Since I will include my reasoning I think it will take multiple posts to this thread. I will include pics and maybe a link to videos.
I have never built a circuit with a MOSFET but like you I like to experiment and learn new things. Some of things I will try and address in the next couple of posts: LM324N, the 1 uF capacitor, adding build out resistors to the two FET gates to prevent problems with the phase margin, the 10k-ohm voltage divider, the 10 k opamp load resistor, and
current hogging.
As I mentioned before I really like your attitude of exploration and your YouTube channel. One negative of watching Dave Jones and your videos it is costing me money. I order the TekPower TP4000ZC and it seems to work including connection to a PC via USB adapter. John
Hi John,
Thanks for the post and I look forward to your input and seeing your progress. It will take at least a week before I have time to start my build.
I'm glad to hear the TekPower is working for you.
Thanks and cheers,
Martin.
-
Hello Kiriakos,
I plan to build another one that will do about 5-10 amp at 12-15 volt.
Cheers,
Martin.
If you do one at 17A 12V it would be more worthwhile because it will have the potentials to stress those batteries found in the UPS power supply too.
Which usually are 7Ah or 17Ah. ;)
-
Hello Kiriakos,
I plan to build another one that will do about 5-10 amp at 12-15 volt.
Cheers,
Martin.
If you do one at 17A 12V it would be more worthwhile because it will have the potentials to stress those batteries found in the UPS power supply too.
Which usually are 7Ah or 17Ah. ;)
Yes...good point. I'll look at some FETs in the TO-3 package for that option.
-
More FET's in parallel in a TO247 or TO220 package will be easier to assemble though, and will be easier to find a heatsink for. I have some massive power supplies for old minicomputers that use a massive heatsink with TO3 devices on it. I have drilled the layout a few times for TO3 devices, it is hard to do accurately.
-
I have drilled the layout a few times for TO3 devices, it is hard to do accurately.
To do Work on aluminum is an art by it self, I am master to it, but you need and professional quality tools too.
Even the quality of the drill bits, and the rotation speed, is tremendously important factors.
http://hackedgadgets.com/2010/09/04/diy-resistors-decade-box/ (http://hackedgadgets.com/2010/09/04/diy-resistors-decade-box/)
-
More FET's in parallel in a TO247 or TO220 package will be easier to assemble though, and will be easier to find a heatsink for. I have some massive power supplies for old minicomputers that use a massive heatsink with TO3 devices on it. I have drilled the layout a few times for TO3 devices, it is hard to do accurately.
Thanks, valuable feedback....plus I have found that the TO3 packages are many times more expensive than the TO220 packages.
-
I have drilled the layout a few times for TO3 devices, it is hard to do accurately.
To do Work on aluminum is an art by it self, I am master to it, but you need and professional quality tools too.
Even the quality of the drill bits, and the rotation speed, is tremendously important factors.
http://hackedgadgets.com/2010/09/04/diy-resistors-decade-box/ (http://hackedgadgets.com/2010/09/04/diy-resistors-decade-box/)
Very nice project Kiriakos...I need to add it to my list of tools too!
Cheers,
Martin.
-
Hi Martin,
Just an update on my el-load project. I have been working on the design (Dave's and yours) for the last couple of days and it is nearing completion. It is Friday night here in California and I think I will be able to post most of the design information this weekend. I am waiting on the chassis and a digital meter which should be here in a few days. Basic features (that might change) are: 0 to 5 amps at 12 volts, digital readout of current with set feature and on-off output control, heatsink tunnel with fan, two E-MOSTFETs (the ones you selected) for load control, only need one quad opamp, and some other goodies. I plan to put the control circuits on a PC board and the basic design can support much higher voltages and currents... given a lot more E-MOSFETs.
John
-
Hi Everyone, I am new to the forum and an amateur at electronics (but I am a very experienced software engineer). I watched Dave's and Martin's DC load videos and decided to build one as well. I have it working successfully on a breadboard, although I still need to do some more in depth measurements with the scope across a broader range of voltages and current. I built my load circuit using the same op amp, mosfet and power resistor as Martin. I also used a 10-turn pot that is nearly identical as well. My original attempt powered the op amp at 12V but I ran a voltage divider and another voltage follower to drop the voltage to the pot and comparator up to 11V. The circuit did work but I did not have much granularity in the pot - less than one turn and I was over 1 amp! Yikes! Today I tweaked the circuit to be more like Martin's last video on the subject by building up an LM317 with a couple of resistors to give me a 6V supply to the load circuit. I also added a pair of 10K resistors to form a fixed voltage divider just before the comparator op amp and another 10K resistor to ground just before the mosfet's gate. Under this config, the 10-turn pot has a lot of fine-grained control, which is nice. I am only using one 50K, 10-turn pot, not the coarse and fine controls. I do not believe I need a fine control. I have only tested it up to 1.5 amps for a very short time since I'm on a breadboard and I don't want a meltdown - the mosfet gets hot fast as I near one amp of constant current. My load is connected to a separate bench top power supply much like Martin does in his videos. Pretty cool project! Thanks for all of the great information!
-
Hi Martin,
Just an update on my el-load project. I have been working on the design (Dave's and yours) for the last couple of days and it is nearing completion. It is Friday night here in California and I think I will be able to post most of the design information this weekend. I am waiting on the chassis and a digital meter which should be here in a few days. Basic features (that might change) are: 0 to 5 amps at 12 volts, digital readout of current with set feature and on-off output control, heatsink tunnel with fan, two E-MOSTFETs (the ones you selected) for load control, only need one quad opamp, and some other goodies. I plan to put the control circuits on a PC board and the basic design can support much higher voltages and currents... given a lot more E-MOSFETs.
John
Hi John,
Excellent, I look forward to seeing how it turns out. It's great seeing everyone's own tweaks and adaptations for their needs.
Thanks for the post.
Cheers,
Martin.
-
Hi Everyone, I am new to the forum and an amateur at electronics (but I am a very experienced software engineer). I watched Dave's and Martin's DC load videos and decided to build one as well. I have it working successfully on a breadboard, although I still need to do some more in depth measurements with the scope across a broader range of voltages and current. I built my load circuit using the same op amp, mosfet and power resistor as Martin. I also used a 10-turn pot that is nearly identical as well. My original attempt powered the op amp at 12V but I ran a voltage divider and another voltage follower to drop the voltage to the pot and comparator up to 11V. The circuit did work but I did not have much granularity in the pot - less than one turn and I was over 1 amp! Yikes! Today I tweaked the circuit to be more like Martin's last video on the subject by building up an LM317 with a couple of resistors to give me a 6V supply to the load circuit. I also added a pair of 10K resistors to form a fixed voltage divider just before the comparator op amp and another 10K resistor to ground just before the mosfet's gate. Under this config, the 10-turn pot has a lot of fine-grained control, which is nice. I am only using one 50K, 10-turn pot, not the coarse and fine controls. I do not believe I need a fine control. I have only tested it up to 1.5 amps for a very short time since I'm on a breadboard and I don't want a meltdown - the mosfet gets hot fast as I near one amp of constant current. My load is connected to a separate bench top power supply much like Martin does in his videos. Pretty cool project! Thanks for all of the great information!
Hi Wade and welcome to the forum.
Thanks very much for your post and the information on your project. I look forward to hearing how it progresses...do post some pictures if you have the time.
Cheers,
Martin.
-
The case and digital panel meter should be here today.
There will be several posts on the design and fabrication of the el-load.
I thought I would start with the homemade heatsink tunnel. It is composed of two 100mm x 100mm x 25mm aluminum heatsinks that are taped in their centers for a TO-220 mounting screw. There is a ball bearing 80mm x 80mm fan for forced air-cooling and it operates on 12V DC at a rated 70 mA, ( I measured just 55 mA), 24.7 CFM, 22 dBa noise. Part number is 3110GL-B4W-B19 and I got mine at Jameco Electronics for $5.95 USD. I pounded into the heatsink fins eight threaded standoffs so I could attach the fan to the two heatsinks and then attach the whole module to the chassis. I got a couple of small rectangular aluminum pieces for the top and bottom, which with the heatsinks form the air channel. The fan is so quite I am going to let it run all the time while the unit is turned on. Each of the two E-MOSFETs are attached to their respective heatsink and one of them has a thermistor to monitor the temperature. The heatsink tunnel's performance is very nice. With the fan turned off and an ambient temperature of 25 degrees C (from now on 25C) and a load current of 5.0A the temperature of the transistor cases where at 102C in 5 minutes. Same measurements with the fan on the transistor case temperatures were 62.5C. I am using the transistor Martin selected, BUZ31LH. Maximum operating temperature for these devices is 150C but there total power dissipating at this high temperature drops to zero. I decided to limit their temperature so I could easily get 5 amps load current at 12 VDC. I also experimented with blowing air over just the channel and associated heatsink fins and then the channel plus the exterior fins. Using the fan and blowing air over all of the fins the temperature was 66.8C. Blowing air only through the channel the temperature was 60C. I think the reason for this is I didn't have a container around the external fins and so some of the air flow escaped. Also directing all the air flow through the channel added additional cooling to the transistors which are the heat source. My heatsink will be enclosed in a metal case and therefore no physical connection to the outside. This is important because the metal tab on the TO-220 case is electrically connected to the drain and therefore high voltages could appear on the heatsink. I did used an electrical insulator so no voltage is on the heatsink but the metal tab still presents a hazard. As an example if you have a 30 to 40 volt solar cell panel connected to this device there may be a shock hazard. These heatsink tunnels are basic modules and you could add more for higher load currents. I have tested this one to 6 Amps and 12 volts with no problems and the total cost including transistors and thermistor was around $20.00 USD. I will be adding more posts over the next few days and include a schematic diagram of el-load. Hope you find this material helpful but I probably don't know what I am talking about but then again that never stopped me before.
Below are a few pictures of the heatsink tunnel...
-
This is the second post on my el-load. In this post I will describe over temperature protection. As mentioned in the previous post I installed a thermistor in the heatsink tunnel. It is physically clamped to one of the E-MOSFETs. The circuit could support monitoring both transistors but I decided to just monitor one. A nice thing about evaluating the temperature of the device is that it is determined by both the current (IL) and VDS of the transistors. Also it can detect if the fan fails. A negative... there's always a negative is the thermal inertia of the system. The BUZ31LH E-MOSFET's can operate at a maximum temp of 150C. At this high temp the devices dissipation rating drops to zero. Looking at the devices de-rating curve I wanted to limit the temperature of the transistor's case to a maximum of 100C.
The thermistor is shown in one of the pictures below and I soldered a couple of blue wires with heat shrink to make the probe. I purchased this part from Jameco, part number 207037 for $0.59 USD. The thermistor can operate up to 150C, tolerance is +/- 10 % and nominal resistance is 10K-ohms at 25C and 637 ohms at 100C.
Please refer to the schematic section of the Over Temperature circuit, (OTC) shown in the picture below. The entire design of the complete el-load uses one inexpensive Quad CMOS opamp, the MCP6004. It works very well, is stable driving the E-MOSFETs and is a power supply rail-to-rail output device. It also has the same pin out as the LM324 series devices. I got my MCP6004 devices for about $0.50 USD each from Mouser.com
U1D is running on a regulated supply voltage of +5V. The OTC circuit is a voltage comparator with hysteresis (meaning two trip/threshold points). The opamp is running positive feedback via R4 and is comparing one of the two reference voltages at pin 12. The thermistor TH1 is bonded to the E-MOSFET in the heatsink tunnel. R9 and TH1 form a voltage divider and send a voltage level related to the temperature of the transistor to pin 13. You could use a potentiometer in place of R9 if you wanted to calibrate the temperature voltage. R1, R2, and R4 set the two reference voltages that correspond to 90C and 100C. These two temperatures could also be easily changed with different values of those three resistors. D1 is a red led that indicates when over temperature has occurred. R5 is just the current limiting resistor for the led and to not overload U1D. R11 connects the U1D output voltage to the gate of a control E-MOSFET and the 270-ohm value was selected to minimize capacitive loading of the opamp and make sure it is stable (phase margin).
Ok so what is the basic operation of this circuit? When the temperature of the E-MOSFETs is below 100C the led is off and the system operates normally. If however you select a load current (IL) and the load voltage is too high the temperature of the E-MOSFETs will start rising. At a 100% power overload of the system I noted a 1C increase per second. When the temperature reaches 100C the led will light and a positive control voltage is feed from R11 to an overload E-MOSFET. That FET, not shown in this section of the diagram will pull the control line of the system low and turn off the load E-MOSFETs. The temperature will then start dropping but because of the two trip points it will not restore or turn off the led until the temperature drops to 90C. This is important to prevent erratic operation of the OTC. Just like your home heating system has two trip points so you are not cycling the furnace off and on rapidly. When the 90C temp is reached the OTC will turn off the led and restore load current to the device under test (DUT) automatically. Therefore you have some protection of the system if left unattended. I am planning on putting the red led directly above the load current control so the user can easily see when over temp occurs. Another option for this circuit is to have a buzzer installed so you have both visual and aural notification of a problem. Total cost of this circuit is around $2.00 USD. A much better but more costly option is to use a micro-controller. A few more posts to come with a complete circuit diagram of the entire system. Hope this is helpful. John
-
Thermal paste between the mosfet and the sensor will make it respond faster, as it will have better heat transfer from the tab to it. A thin blob around it before clamping will do.
-
SeanB,
Thanks for the suggestion and I will give it a try and see if it speeds up the reading.
-
Hi John,
Thanks very much for the great posts and progress...I'm wiser after reading about your additions.
Cheers,
Martin.
-
Here is the third post on my el-load. In this post I will discuss a simple section of the el-load. This section is the control voltage stage which sets the load current (IL). As mentioned earlier this is basically Dave Jones' circuit. Please refer to the schematic diagram picture below. RV1 is a 50 k-ohm, 10-turn potentiometer (pot) that I already had in my parts box. The pot forms a simple voltage divider and it can output a voltage from 0 to 5 volts on its wiper (pin 2). The voltage supplied to this pot is from a 5 volt regulated power supply. The voltage to this pot needs to be regulated because if it changes so does the value of IL. Most ohmic values for the pot, say from 5k-ohms to 100k-ohms should work in this circuit. The selected voltage at the wiper of the pot is applied to pin 3 of U1A which is 1/4 of the MCP6004. This stage is a voltage follower used as a buffer and its gain is set to unity by the wire connecting pin 2 to pin 1, (100% negative feedback). The output of the opamp at pin 1 drives an adjustable voltage divider circuit of R3 and RV3. This scales and limits the control voltage from 0 to 2.5 volts, In my circuit the 2.5 volts equals 25 amp of IL. Obviously my E-MOSFET transistor would not be happy and quickly die if I tried to get 12.5 amps from each of them. That is why I have RV3 which can set the current limit on the system. I have it currently set to 550 mV maximum which equals a maximum IL of 5.5 amps. RV3 is also a 10-turn pot that will be located on the circuit board and should not be changed once an optimum/safe value is determined for the system. R3 has a second function which is to protect the opamp when the crowbar E-MOSFET shuts down the IL, (discussed in a following post). The opamp's short circuit output current is typical +/- 23 mA but I don't like shorting opamp outputs to ground. Therefore the U1A opamp will always see at least 10k-ohms at it output thanks to R3.
-
Here is the fourth post on my el-load. In this post I will go over the error opamp and output transistors. Please refer to the schematic diagram shown below. U1B is the error/driver opamp which is 1/4 of the MCP6004, (yea the 4 in that number stands for 4 opamps (quad). Pin 5 of the opamp receives the control voltage from the previous stage and that voltage determines the load current (IL). The opamp is wired to compare the control voltage to the voltage drop across the current sense resistor R8. This opamp has a stated minimum Avol or Aol, (open loop voltage gain) of 88 dB or stated another way 25,000. Since this is a differential amplifier (pins 5 and 6) the difference signal between those two pins gets amplified by at least a factor of 25,000. So for finite output voltages from 0 to 5 volts the input differential voltage would be almost zero. So as a first approximation just think that the opamp will try and output a voltage at pin 7 to make the voltage at pin 6 equal the voltage at pin 5. Let's say we want an IL of 5 amps with a load voltage of 12 volts. Note that R8 is in series with the load circuit so the load current passes through R8 and using Ohm's Law 5 amps flowing through R8 a 0.1 ohm resistor will give rise to a voltage drop of 500 mV (5 amps x 0.1 ohms = 500 mV). Therefore we would set our control voltage to apply 500 mV to pin 5 of the error opamp. This means at power up the opamp sees 500 mV on pin 5 and 0 volts on pin 6. The opamp would then see this error and adjust its output voltage at pin 7 such that the voltage at R8 would equal 500 mV and thus pin 5 and pin 6 would have approximately equal voltages. What is the output voltage at pin 7? It has to be the 500 mVs across R8 but it must also add the threshold voltage of the transistors Q1 and Q2 plus a little more to get the 5 amps to flow through the output. The BUZ 31LH E-MOSFETs have a typical threshold voltage of: 1.2 v min, 1.6 v typ, 2 v max. I measured 4 of these fets and found them all very close to 1.555 volts, (great a good match). So back to our example the voltage would need to be a little over 1.555v + 0.5 v. So the voltage at pin 7 is a bit over 2 volts. If you monitored the pin 7 voltage as you turned up your voltage control pot you would almost immediately see the voltage jump to a little over 1.5 volts and then adjust up more as you dialed in higher control voltages.
The two E-MOSFETs are in parallel and if they are well matched they should divide the IL equally with half of the 5 amps flowing through Q1 and the other half through Q2. Let's assume that when we set the IL to 5 amps each transistor has 2.5 amps of drain current. The power dissipation of one transistor is then VDS x ID. VDS is the voltage drop across the transistor from Drain to Source. We said we were doing 5 amps at 12 volts do the VDS is 12v - .5v = 11.5 volts and then multiple by ID of 2.5 amps would equal 28.8 watts. The 0.5 volts in the previous equation was from the voltage drop across the sense resistor R8. Each transistor is dissipating approximately 29 watts for a total of 58 watts. My sense resistor R8 is a 5 watt 0.1 ohm. The power the resistor must handle at 5 amps is the current squared 25 time the resistor's ohmic value of 0.1 ohm which equals 2.5 watts.
So what happens if the IL current tries to change? If it tries to increase that will drop a larger voltage across R8 which will increase the voltage a pin 6 of the opamp. The opamp sees that this voltage is greater than the control voltage at pin 5 and since the larger voltage is on the negative (inverting) input the opamp will reduce it output voltage which in turn will increase the channel resistance between Drain and Source of the fets. This action will lower the current back to the set value. If the IL tries to decrease the voltage on pin 6 will drop and the opamp will see that a larger voltage is on pin 5 the positive (non-inverting) terminal and will increase its output voltage and thus lower the channel resistance of the fets and this will increase IL back to the set value. To recap the two transistors are acting as variable resistors in parallel and the opamp is varying the resistance from the Source to the Drain so that the current matches the value related to the control voltage.
Not shown on this diagram is the thermistor which is monitoring the temperature of Q1. The purpose of R6 and R7 is to make the circuit stable when the opamp is driving a capacitive load. From the data sheet the input capacitance of the BUZ31LH worst case is 1600 pF. Using the diagram and design formula from the MCP6004 I determined I needed 250 ohms of resistance. I chose a standard value of 270 ohms for the build out resistors R6 and R7. The circuit on a breadboard and with a rats nest of wires seems to be very stable... we shall see. Check out the rats nest and breadboards for the completed circuit in the picture below. As a side note I have not put in any discharge resistors from the gates of the fets to ground. Initial tests seem to show they may not be need more on that topic later. So far I am really liking this quad opamp (MCP6004). Hope this helps and that I didn't make any major mistakes. More to come, John
-
Here is the fifth post on my el-load. In this post I will discuss the complete el-load excluding the +12v and +5 volt power supply. I will cover that dual voltage power supply and the case later. Please refer to the schematic diagram in the picture below. In previous posts I covered the Over Temp Stage (OTC) in the lower left-hand of the diagram. I also covered the voltage control and output stages that appear across the top of the diagram. In this post I will cover the switching, crowbar, line driver, load on LED, and panel meter. Sorry the switch symbol is so large but that's what was available. I am just learning the open source KICAD program so there are a few things that I will need to change. I hope to also use KICAD to make the circuit board artwork.
OK let's talk about the switching. It looks a bit involved but is actually pretty simple and inexpensive. The four symbol circles represent actually just one rotary switch that cost $3.50 USD. It is a three position, four pole, non-shorting switch. This switching operation could be done with electronic switching or relays but this approach is simple and cheap. One pole is section S1A at the top of the diagram and it has three positions; 1, 2, and 3. A little confusing but starting from position 1 and going to position 3 is moving the panel knob clockwise on this diagram it is counter-clockwise. The three positions of the switch are OFF - SET - ON. In the OFF position the unit is turned off... duh. In the SET position the unit is turned on but the load is disabled. In this position you can set the current using the front panel LCD meter without the load being in operation or even connected. This is helpful because it is hard to know initially where the 10-turn RV1 is set. Moving to position 3 (ON) enables the load current and reads its value on the LCD meter. At anytime you wish to disable the output without turning off the unit, you would just move the switch from ON to SET. Then make any changes to the external circuit and then switch back to ON position to restore operation. Also as a little safety feature you can never turn the unit completely on without going through the SET mode. That way you can see the current you are about to allow before it actually happens.
Looking at section S1A you can see that it connects the green Load On led to the output of U1D. This means that the ON LED D2 will only turn on in this position and that the output of U1D must be in a low voltage state. If however there is a fault with the temperature U1D's output will be high and not allow the Green diode D2 to be on while the Temp red LED D1 is on. If the output of U1D is high, near the 5 volt rail it will bias Q3 the Crowbar E-MOSFET to saturate and pull the control line going to pin 5 of U1B near zero volts. This will turn off Q1 and Q2 and therefore IL will go to zero. Why use such a large FET for Q3. They are cheap and I had one. It is also nice for repair to have to only deal with one type. You could of course use a very low power device for Q3.
Switch section S1B grounds the input to pin 5 of U1B in the OFF and SET position so that IL is zero. In position 3 the switch couples the control voltage to pin 5 of U1B for normal operation in the ON mode.
Switch section S1C in the OFF position grounds the input to both the panel meter and the line driver U1C. I like grounding these input until the unit powers up. In the SET position section S1C connects the control voltage from the wiper of RV3 to both the line driver and panel meter. This allow you to view and set the IL current before enabling the output. When section S1C is in the ON position the voltage (related to IL) across R8 the sense resistor is sent to the line driver and panel meter for monitoring.
Section S1D is the AC power section and connects to the power supply. In the OFF position the AC line is broken and the unit is off. In the other two positions the power supply is turned on. This rotary switch control the 120 volts AC and this switch has that rating and is also fully enclosed and has a plastic shaft for safety. If you decide to use a rotary switch make sure it can handle and is designed to operate AC line voltage. For example the one I selected cannot handle the 220 V provided in other locations.
I have a close-up of this rotary switch in a picture below and it hasn't yet been wired for the AC line but all the other connection have been made for testing. I like using ribbon cable to make the wire management neat. So far I have been running test to see if when switching the system, it had transients or instability. Below is a oscope picture showing the turn off and turn on of the system and this is at a low sweep speed and looks very clean with no overshoot. I also used much faster sweep speeds and still looks good with a slight ramp up.
In this post a couple of comments on the line driver and panel meter. The panel meter is under $20.00 USD and is 3 and 1/2 digital back light LCD. It can be programmed with solder jumpers for AC/DC current and voltages. It also includes several ranges and you can set the decimal point. Right now I have it set to DC volts on the 2 V range with the decimal point 2-digits from the right. Therefore it is reading out in amps with a resolution of 10 mA. For example if the display is 12.50 that would represent 12.50 amps of IL. I am not too happy with the back light, pretty dim but it does the job, I guess. I was going to use the light from the panel meter to indicate when the unit was on but decided to add a red power on led in the power supply section. The purpose for the line driver U1C is to provide either the SET voltage or the voltage related to IL to a remote voltmeter. The back of the unit will have two banana jacks at 3/4-inch spacing so you could plug in a dual banana jack connected to a DMM. This way you could read down to milli-amps or for easy monitoring. For example have the unit under the bench and still easily view the IL value. R12 is another 270 resistor and will limit the opamp current to a safe value so that if you accidentally short the remote wires together the opamp will not be harms and since U1C is a buffer you will not upset the operation of the el-load. I plan to do a PC board, final case assembly, and performance testing but this may take a week or two. Hope this helps, John
-
Me again... Here is the sixth post on the el-load. In this post I will go over the power supply for the el-load, (not much here). If you were not using a 12 v DC fan then the only power needed would be at 5-volt DC and very low current. This power supply is not battery power and is designed to operate on 120 v AC 60 Hz.
Please refer to the schematic shown below and I have also included a picture of the physical circuit. As you can see on the left side of the schematic are the 3 wires coming from an AC outlet, (a plug will be added). Wiring here in California, USA is the hot line is usually colored black with the neutral being white and the ground is green. There is a fuse yet to be determined, once I measure operating power requirements of the circuits. There are two wires after the fuse that are labeled TO S1D. This is the main power on switch located on the other schematic and it is the 4th pole of that rotary switch.
T1 has a secondary winding that is center tapped, (12.6 v CT). I wanted a CT so that the differential voltage across the input and output of the 7805, 5-volt regulator would be reasonable and not generate unnecessary heat. The four rectifier diodes are 1N4007's and are what I had in my parts box. They are rated at 1000 v PIV and IDiode of 1 Amp. The transformer is rated at 2 A so a bit of overkill.
The 4 diodes act like a full-wave bridge and deliver a peak voltage of around 18 volts to the main filter capacitor C3 for the 12 volt side. This capacitor is for storing energy and filtering the ripple. C5 is for high freq noise suppression and giving stability to the 3-port voltage regulator 7812. The 7812 is a 12 v, 1A linear regulator and I only need a little current so again overkill. C7 on the output is also noise suppression and for stability. There are other capacitors in the system for filtering and decoupling close to the active devices.
Note that the supply for the 7805 is from the center tap and therefore improves efficiency because the voltage to the input of that regulator is only half of the 18 volts peak. Added to the output of the 7805 is R13 and D7 for a power on indicator. . Oh I forgot there are two other resistor in the photo (not in the schematic) and were used for testing. Again not much to this circuit. Hope this helps, John.
-
Thanks very much John. Great posts.
-
Hi Martin,
Hope these post have some useful information on the el-load. Just an update, I have been trying to learn KICAD so that I can layout a circuit board for this project. Today I think I may know just enough to start the layout. I hope to have a completed circuit board in about a week. I will be doing my own PC board fabrication and since I do not have any experience in silk-screening , it will be just a plain circuit board. I also think I have all the case and external parts to finish the build. Hopefully more to follow. John
-
Hi Martin,
Hope these post have some useful information on the el-load. Just an update, I have been trying to learn KICAD so that I can layout a circuit board for this project. Today I think I may know just enough to start the layout. I hope to have a completed circuit board in about a week. I will be doing my own PC board fabrication and since I do not have any experience in silk-screening , it will be just a plain circuit board. I also think I have all the case and external parts to finish the build. Hopefully more to follow. John
John...you are a good man, hats off for all the effort.
I was about to consider fabricating my own PCB too...but I might leave that for my second build. Look forward to seeing yours.
Cheers,
Martin.
-
I do not have any experience in silk-screening , it will be just a plain circuit board.
I've just done my first PCB in house. Silkscreen with the same method as copper traces - thermotransfer. Not perfect but quite good IMO.
(http://imageshack.us/a/img141/7613/outea.th.jpg) (http://imageshack.us/photo/my-images/141/outea.jpg/)
-
Nice job arekm! I have done several pcb's using a laser printer and photo positive boards. Your outlines and lettering on the top side look good and I have never done that process. What did you use to create the artwork for the traces? John
-
The PCB was designed in Sprint Layout software (http://www.abacom-online.de/uk/html/sprint-layout.html) by someone else. I've done the rest. Copper layer and silkscreen layer were printed on a laser printer on separate pages.
The whole thermotransfer idea is to transfer toner (laser printer) printed on slick paper to laminate, then do etching (toner is immune to sodium persulfate B327). The same method is used for silkscreen - just transfer toner and silkscreen done.
Some video (sorry, polish subtitles but you will get the idea) - http://www.youtube.com/watch?v=vgHFet8ci-I
Note: I'm using unmodified lamination device (Lervia KH4410) instead of iron (clothes iron) used on this video.
Etching done in home made device http://www.youtube.com/watch?v=Ui0G9cptFcs
The whole process is like 30-45 minutes per pcb (not counting drilling).
Drilling is easier after etching - small etched points in pads make drilling easier.
Edit: final product (well, case is missing) http://imageshack.us/photo/my-images/6/img8288wn.jpg/
-
Thanks for the links arekm. John
-
The picture below is the first pass at the el-load pcb. The resolution is not very good because it is a screen capture of the print preview of the artwork. This picture is also just the traces of the board with no lettering. The board size is a little less than 5 x 3 inches. I was happy that so far there are no jumpers in the layout. This board contains all the controller circuitry and the dual voltage power supply. Many of the pads around the edges are for the wiring connections of the off-board components. I tried to bring those connections out to the appropriate edge of the board to reduce the wiring lengths. I had thought of designing the board so that the rotary switch was solder directly to the board but I think the board would either require being double sided or have many jumpers to accomplish the task. I also didn't like having the AC 120 V on the board. John
-
Again...Arekm and John...great posts. It certainly inspires me to give the PCB manufacture process a bash.
Cheers,
Martin.
-
Count me in, about getting one PCB too. ;-)
-
I completed a second version of the artwork using KiCAD and from that artwork made the pcb for the el-load. Looking at the pictures below, in the first one you can see the exposure lamp, artwork, and photo-sensitive printed circuit board (pcb). The pcb is exposed to the light for 6 minutes using the timer show in the right side of the unit. The next picture shows the board placed in a glass tray with a photo developer and then rocked back and forth for approximately 3 minutes to remove the exposed resist. The next picture shows the developed board ready for etching. The next picture shows the ferric chloride solution used to remove the exposed copper from the board. The final picture shows the etched board before the resist is removed, sheared to size and drilled. This system is a positive photo-fabrication process and the total time from start to finish is under 20 minutes. More to come.... John
-
The above is what I like to avoid, and why my vote was in favor of a mass-order PCB solution.
The older boys haves much less patience to do this things.
-
I understand where you are coming from. I am one of the older boys too. :) The reason I like this process occasionally is to really test out the design and if it fails I can have another version in an hour or two.
-
I was making PCB at the age of 18, and actually with just a waterproof pen, that makes it a totally hand made process. :)
But now days if you invest in a laser printer everything becomes more automated.
The point is that I do not design experimental circuits.
Now days I prefer to buy a ready Kit and to just assembly it.
-
Please refer to the pictures below of the el-load control board. In the first picture you can see the back of the board and the board is shown trimmed, drilled, and tin plated. The second and third pictures show the front and side of the control board populated with components. There are still a lot of pads that will allow external wiring connections later. The board is larger than required but I like the extra room on a first prototype and also the size helps keep the external wiring distributed across the front panel of the case. The three TO-220 cases are the two voltage regulators and the one crowbar E-MOSFET. There is no high current connections on this board and none of the TO-220 devices require heatsinks. All the resistors are 1% metal film, 1/4 watt. All the capacitors are over-rated for working voltages and you can see power supply decoupling capacitors distributed across the board. More to come... John
-
Please refer to the pictures below on my el-load. In the first picture I have set the basic parts into the case to determine proper mounting locations. In the upper left corner is a 12.6-V CT, 2-A shielded power transformer which is way over rated but I am think of future additions. In the lower left corner is the control pcb on metal stand-offs. In the right side area is the heatsink tunnel or load module. In the second picture is a preliminary front panel layout. I am trying to make sure that the panel components do not interfere with the other chassis components, (need to think in 3D). I had hoped to use a larger input air filter, (shown on right side) but settled on this smaller one and hope it doesn't impact the air flow in too negative of a way. More to come... John
-
Looking very neat!!
-
Should not be a problem, limiting the airflow a lot. Still you will find the case will let in air via slots. You might want to turn the fan around and put the filter on the back of the case, letting the air out via the sides of the case and sucking the air past the heatsink after it is filtered by the filter. The fan will be more effective then with laminar flow into it.
-
Brilliant work John. Thanks for the great posts.
-
I finally got round to building one of these yesterday. I'm sure it will help me with my power supply projects.
I used an IRF540 MOSFET and managed to get adjustment from a few milliamps right up to 3.5A when it was connected up to a 5V source.
I got a lot of oscillation though. There were some funky patterns on the oscilloscope :D. To fix this I moved the capacitor to the drain of the FET.
I think I might try some different MOSFETs as well as this is a bit of a learning experience for me as I have never used FETs before.
-
I finally got round to building one of these yesterday. I'm sure it will help me with my power supply projects.
I used an IRF540 MOSFET and managed to get adjustment from a few milliamps right up to 3.5A when it was connected up to a 5V source.
I got a lot of oscillation though. There were some funky patterns on the oscilloscope :D. To fix this I moved the capacitor to the drain of the FET.
I think I might try some different MOSFETs as well as this is a bit of a learning experience for me as I have never used FETs before.
Hats off to you Steve and enjoy the tinkering until it does the job for you.
If you have a moment please post some pictures....we'd love to see it.
Cheers,
Martin.
-
I'll get the gear out later and do some photos of the oscilloscope. That would be cool.
I will note that the IRF540 requires higher voltages on the gate to be 'turned on', so I powered the circuit off a 24V PSU rather than just 5V.
Anyway, it worked nicely, so I thing I might try and make a nice hand drawn PCB for it.
-
Good Day Everyone,
Hi all, BTFdev is a newbie here.
First Martin, great job on providing further analysis and research on this subject. I found this very educational and valuable. I do have a few questions for the experts, please refer to the schematic diagram Circuit2.jpg posted by Martin.
1. With the combination of 5K/50K pots (R4/R2), can you please tell me what is the min and max voltage from R2 (Pin 2) ?
2. After passing the 1st Opamp, what is the min and max voltage at LM324 Pin1?
3. How about the 10K pot (R3)? What is the min and max voltage from R3?
4. After passing the 2nd Opamp, what is the min and max voltage at LM324 Pin7?
Thanks in advance.
Team BTF
-
OK, I will try and answer your questions. I am using a different opamp so Martin can jump in if I make any mistakes. If you don't mind I will add some of my reasoning behind the answers.
1) The output of the three port regulator LM7806 is +6-volts. so with R4 wiper at its pin 3 of the pot the full 6-volts is applied to R2. So with R2's wiper (Pin 2 at Pin1 of that pot) the maximum voltage out of pin 2 to the opamp's pin 3 is 6-volts. With R4's wiper at pin 1 of the pot the maximum voltage out is 6v (50 kohms / (5 kohms + 50 kohms)) = 5.45-volts. The minimum voltage out of the R2 pot is 0-volts no matter where the wiper on R4 is set.
2) The 1st opamp is a voltage follower with 100% negative feedback so the gain is unity (1). Therefore the voltage at pin 1 of the opamp is the same as the R2 wiper at pin 2 of the pot.
3) The maximum voltage at R3's wiper is 6-votls and the minimum is 0-volts. For this to be true for the 6-volts the wiper of R4 must be at pin 3 of that pot.
4) The voltage at pin 7 is a little more difficult to determine because of VGS(threshold) which can vary from transistor to transistor. I am using the same transistor as Martin and the VGS(threshold) is around 1.55 volts. So the output voltage at pin 7 would be from around 0-volts to around 6 volts less the internal drop of the opamp. You would not want to run the system at this level because what is happening is you are hitting the power supply rail which in this circuit is shown to be 6 volts. Just as an example let's say the internal drop of the opamp is 200 mV. So the max output would be 6v - 200 mV = 5.8 volts. If the VGS(Threshold) is say 1.5 volts then the voltage across R5 would be 5.8v - (1.5v + 1.5 v) = 2.8 v and therefore equates to an ILoad of 2.8 Amps. Please note that in this circuit arrangement you are not getting the full adjustment range out of R3 if you have the other pots set to obtain maximum control voltage.
Hope this helps... John
Good Day Everyone,
Hi all, BTFdev is a newbie here.
First Martin, great job on providing further analysis and research on this subject. I found this very educational and valuable. I do have a few questions for the experts, please refer to the schematic diagram Circuit2.jpg posted by Martin.
1. With the combination of 5K/50K pots (R4/R2), can you please tell me what is the min and max voltage from R2 (Pin 2) ?
2. After passing the 1st Opamp, what is the min and max voltage at LM324 Pin1?
3. How about the 10K pot (R3)? What is the min and max voltage from R3?
4. After passing the 2nd Opamp, what is the min and max voltage at LM324 Pin7?
Thanks in advance.
Team BTF
-
I finally got round to taking some pictures, and decided to write it all up in a web page.
Here it is:
http://stevecoates.net/cc_load/
-
I finally got round to taking some pictures, and decided to write it all up in a web page.
Here it is:
http://stevecoates.net/cc_load/
Brilliant Steve. Thanks for the post.
-
Thanks for the replies and inputs from our questions.
After even more Googling and Youtubing, we have more understanding of this dummy load concept. We also did build one prototype ourselves and now researching a better logic level MOSFET.
One thing we noticed from our research is that, some dummy loads are powered by separate battery but we also seen some powered entirely by the load. Any pros and cons for these two approaches? So in general for people who interested in building their own dummy load, here are a few things to consider:
- What to use to power your rail?
- What is the Voltage/Source to feed your opamp
- What is the voltage needed to feed your MOSFET to induce its function in ohmic and saturated region?
We saw some designs featuring voltage/current display and cooling fans, you also need to consider how are you going to supply power to these peripherals.
Looking forward for more info on this subject.
Team BTF.
-
Separate power means you can have a wide input voltage range, from the max of the pass element to where the sense resistor is dropping 90% of the voltage. Single source power limits you to the minimum needed for the controller to the max voltage rating of the control, typically 8-20V then. A separate power supply also allows you to have cooling fans to make the heatsink more efficient as a forced air cooler is both cooler running and a lot smaller.
-
Here is my take on Dave's CC load:
It has current and voltage displays and the base is doubled in plexi not to accidentally short the PCB.
I added a series 6A diode to the FET to protect against wrong polarity. Power resistor is 0R33, 5W, 10% with the worst thermal stability ever but I don't care. This means that 0.1A shows correctly on the panel meter but at 1.2A the resistor heats up and increases in value, dropping more voltage, thus the panel meter shows 1.6A and thins go downhill from there. Every part used is scrapped from somewhere else, so I used what I had.
Max current through it is short of 4A. Max drain current of the FET is 7A. Op-amp is power with 9V and is enough for 6.7A (displayed on the panel meter, tested on a lead-acid). The TO92 5V regulator is for the panel meters, but it was not necessary, they have their own LM1117-3.3.
http://www.ebay.com/itm/Mini-red-LED-Digital-Voltage-Panel-Meter-DC-4-5V-30V-gauge-/261031174906?pt=LH_DefaultDomain_0&hash=item3cc6aba2fa
These panel meters are very nice - I only have to remove a diode from the PCB and supply 4.5-30V for power and 0-30V to be measured. Internal resistor divider is 330K vs. 12K + 22K pot. On the one measuring current I replaced 330K with 50K pot and set it up to measure the voltage on the power resistor. I was in a hurry so I didn't investigate how to move the the decimal dot, but it's doable. For 2$ shipped you can't go wrong with these.
It's powered from an external 20V wall adapter.
For a better one I would use 0R1 5% resistor 5W, 10-30A diode and higher current FET, use a PC CPU heat sink and fan.
-
Nice meters, I should buy a whole lot of them....... Nice setup there, nice and compact.
-
> and higher current FET, use a PC CPU heat sink and fan.
You should find a FET with a much lower RDS_on than more current. Your 2SK2749 has a RDS_on around 1,6Ω. At 2A you have 6,4W amount of heat to dissipated. That's much.
The IRF3710 for axample is a Monster with 23mΩ . At 2A you only have 92mW at the FET. 6,3 Watts under.
And this 6A reverse Diode you don't need, most MOSFETs have a build in high current reverse one. Much enough for your lill board.
This 0R1 5W resistor is not the best idea. At 2-3A it goes glow. Use 25W or more, like Martin. You feel better then, belive me.
The best, you don't need those large heatsinks or PC cooler. You can still use your old one.
-
You should find a FET with a much lower RDS_on than more current.
not surprisingly these two go hand in hand. One 50A FET will have low Rds_on by design otherwise it will dissipate a lot of power.
The IRF3710 for axample is a Monster with 23mΩ . At 2A you only have 92mW at the FET.
you got this wrong: Rds_on is the FET is fully open, and the FET + resistor are the only components eating the power. There won't be 2 A flowing through there, but much more: 12V / ( 23mΩ+ 0.33Ω ) = 34A
The FET will never get to work in on-off state, it's Rds will never come close to those 23mΩ so it doesn't matter how low it is.
If you want to pull 2A from a 12V supply, you HAVE to drop 24W somewhere. You can choose a huge power resistor(also higher value) and a low Rds_on FET, thus dropping most power on the resistor. Alternatively you can use a lower power resistor (3-5W, but lower value) and a higher power(!) FET and drop more power on the FET. Either way it'll work, but I find FETs cheaper to come by - old ATX PSUs etc.
And this 6A reverse Diode you don't need, most MOSFETs have a build in high current reverse one.
The diode I used is in series with the FET and the power resistor, it's used to protect against reverse connection when the internal diode would be forward biased and subject to the whole power dissipation. It's not paralleled with the FET.
This 0R1 5W resistor is not the best idea. At 2-3A it goes glow. Use 25W or more, like Martin. You feel better then, belive me.
3A on a 0.1Ω dissipates 3*3*0.1=0.9W. What are you talking about?!
The best, you don't need those large heatsinks or PC cooler. You can still use your old one.
as I've said before, you have to dissipate the power somewhere. Either FET or resistor, so cooling stays the same, sorry.
-
The FET dissipates all of the power, the low on resistance allows it to be controlled into the linerar region, if the on resistance is dissipating all the load then the FETY is full on, and there is no longer control of the current.
As to power in the sense resistor you want the resistor to run as cool as possible, as the voltage drop across it changes with temperature. A hot resistor is not the same value as a cold one by a measurable amount.
-
So you don't have a PWM mode ?
-
I know several others have added PWM (Pulse Width Modulation) to their el-loads. I haven't thought much about it but I was wondering what the reasoning for adding that mode is to an el-load. Always happy to learn something new but this is a load that is to dissipate power so why worry about the improved efficiency of PWM in regulating current? Also I would think it is a negative feature in this application because of the noise created on the load circuit. Chopping up the current flow with a switching action doesn't seem as good as using a linear (variable resistance) control.
-
PWM on a load is rather silly to me, you want to have heat there, not as low a loss as possible. That is why you have a load. PWM is useful if you are wanting to do transient regulation testing of a power supply, though there too a fast ability to change setpoints are much better, or a simple high speed switch with a non inductive fixed resistor to introduce a step change will work better.
-
The problem with MOSFETs is, they are made for switching mode. Maybe 5% of all MOSFETs are useful in linear option as well. Bipol Transistor works much better at this Point. Look like audio amps.
2nd, the higher the thermal dissipation loss / temperature, the easier peaks can destroy the FET.
PWM and a higher power resistance, so I would build. Or maybe with bipolar Transistors.
Check out google / Datasheets for "power MOSFET SOA / FBSOA" linear operation.
-
most mosfets are used as switches, so are optimised for that. But they do have a usable linear range, and with a high voltage one you can get quite a good show out of it at the bottom end. Not much current wise, but then again at that end just using ones from the same batch in parallel with each having a load resistor it works out quite well. If you use a resistor network to add the drops together you can get reasonable current measurement.
-
I finally replaced my IRF540 MOSFET with a logic level one. I have used an RFP12N10L. Seems to do the job nicely, and the circuit can be powered off a PP3 battery. I built it up on a few bits of stripboard, so I'll post a photo later :).
-
Hi,
My own load is capable of 25vdc @ 10A.
Based on a couple of IXTK46N50L (TO-264 500v, 46A) these are N-type Linear Power Mosfets. Like has been mentioned here, conventional mosfets can be used but you would be limiting the range of voltages/loads the Dummy Load would be able to cover due to running the mosfet in it's linear mode. The IXTK mosfets overcome these limitations by extending the transistors’ FBSOA. I.E. ETI (Electro-Thermal-Instability) as a result of positive feedback within the mosfet when used in linear mode.
The op-amp I use is a LM8272 which is ideal as it is able to drive large capacitive loads such as power mosfets. I've used other op-amps and under certain situations you can get oscillations. The 8272 helps greatly.
I also use a 35W 0.27ohm shunt resistor (TO263 package). For the rating it's extremely small.
http://datasheets.globalspec.com/ds/4939/Newarkelement14/500BDB5E-D915-4061-B619-0721D61A5163
My load runs off a 12v Lipo battery which is housed inside the load, and also has a wee 12vdc fan incorporated which pushes air across the mosfets & shunt resistor. The current consumtion is 115mA which includes the fan.
The large heatsink houses the electronics and gets quite warm at full load (250w dissipation), but well within limits. The fan really, really helps.
Still to do once I get around to it.......Add a small P/B so I can check the battery voltage on the LCD.
Ian.
PS. Sorry for the squigly resistor symbology...........25yrs of doing it that way! ;)
(http://www.ianjohnston.com/content/images/stories/IanJ/DummyLoad2/IMG_3844.JPG)
(http://www.ianjohnston.com/content/images/stories/IanJ/DummyLoad2/schematic.JPG)
-
IanJ, that looks really good :).
Here's a photo of my stripboard version 8).
-
Brilliant post and information Ian. I really appreciate the contribution.
-
IanJ, that looks really good :).
Here's a photo of my stripboard version 8).
Thanks for your post Steve.
-
Just wondering, before I try this, if I put FETs in parallel, do I just connect the gate to the gate, source to source, and drain to drain? Do I need to add or change any other components?
-
steve30, Others can probably give you better comments but from what I have read you can usually just parallel the drains of the fets. If this is a MOSFET output stage then the additional gate of the second FET will double the input capacitance that the driver circuit sees so you should probably use two separate gate resistors for load isolation. If there is excessive current hogging then maybe Source equalizing resistors might be needed. Watch out for instability on the output circuit.
-
Just wondering, before I try this, if I put FETs in parallel, do I just connect the gate to the gate, source to source, and drain to drain? Do I need to add or change any other components?
Intensive internet searching reveals unexpected results: http://lmgtfy.com/?q=parallel+mosfets
-
You can but there are caveats. Basically higher capacitance, current sharing and voltage variation between units. If doing it use a gate resistor per device ( not to limit current but to prevent oscillation of the devices, a mosfet will quite happily oscillate at 100MHz using the stray capacitance and inductance of it's leads if they are longish) of around 100-220 ohms, and a low value source resistor per device ( really only needs to drop 100mV at the current per device, it promotes current sharing between devices to be more even) and most importantly to use devices all from the same batch, preferably taken sequentially from a reel as they then are very likely to come from the same area on the slice so will have near identical parameters. Mixing devices will be bad, as you will find differences in threshold voltage causes one device to carry most of the current.
-
Ferrite beads (on the gate) can also be an alternative solution to damp ringing.
Here's a good wee article:-
http://www.ohm.com.tr/doc/Microsemi---Eliminating-Parasitic-Oscillation-between-Parallel-MOSFETs.pdf
Ian.
-
You could also just drive the FETs separately. One op-amp per FET, give each FET a resistor as feedback to the op-amp.
-
Hi Folks,
With great guidance from my "design consultant" based in Canada...and hoping I follow the steps and execute correctly....here is the start of the build of my enclosure for my DC load.
I'll post HD pictures once complete....I've just dashed out of the garage to do this.....
-
Continued....
-
Very nice Martin, nice work. I see where you have been shopping for tools........
-
Ok...a little more progress...
I should have used rubber cement to attached the 1:1 CAD drawing to the enclosure....but could not find any in the three shops I went to. I used contact adhesive, not the best idea, but it worked ok...just needed a little solvent to remove.
The drilled holes for the potentiometers are slightly out...again...I should have used a punch or pilot hole...next time ;-)....but there's a little wiggle room to move one up and the other down to have them almost perfectly aligned.
-
Very nice Martin, nice work. I see where you have been shopping for tools........
Thanks....not Game!
-
No, Makro...... Looks like a little from Servistar as well, along with some from Midas.
Next time use Pritt, it works well, and just takes a hour or two to dry enough to hold firm enough to do the damage through it. Best if left overnight, but is easy to remove afterwards with hot water and a little soap.
love the nice drawings there, very nice. Nice case as well. With the pots I would suggest buying a taper reamer, i saw one and bought it from Builders, but you should be able to get them from many places. Try Buco, they are getting a reasonable range of tools, though it is rather store dependant.
-
Hi all,
Taper Reamers are great as are the Step Drill Bit...........great for drilling panels, especially in the pedestal drill, and they will deburr as they drill.
(http://www.ianjohnston.com/various/stepped.jpg)
Ian.
-
Plus they are self centring as well. Those there are pretty well used I see.
Another thing useful is a hole saw set.
-
Gents....thanks for the valuable input....I had some of my holes go a little off center....I do have a larger step drill (too big) and will pop out now to get some smaller ones. (and hopefully post your package Sean!)
-
Looks very nice so far Martin.
-
Looks very nice so far Martin.
Thanks Sean...hard and time consuming work for someone like me not used to doing this kind of thing... :o
-
Learning to do cabinetwork neatly is a long learning curve. I am still learning......
-
Advice taken from several folks....thanks!
I invested in some step drills....what a difference!
Another day at it and a little more progress....
-
Where did you get the stepped drills and for how much.
Nice with the masking of the hole edges, but where is the heatsink compound under the Mosfet and sense resistor.
Like the captive nut for holding it down, next project purchase for you is a set of cheap metric taps and dies, so that you can learn how to thread the holes to make hidden fastenings and blind and low profile metalwork.
-
Hi,
Per SeanB, you should isolate everything on the heatsink......it might work perfect on the bench but then one day you sit the heatsink on the case of a PSU you are testing and BANG!
A full mica insulating kit (with heatsink compound) under the mosfet would be good, but be careful to make sure it is tight and flat against the heatsink......I've seen the plastic top hats (they form part of the insulating kits) melt because the heat can transfer away.
Ian.
-
The drill bits came from Builders Warehouse...just over R1000 for the three.
The FET and Resistor were just placed there for orientation...if you look close enough at the resistor in the final picture it has heatsink compound underneath it. The case of the resistor is isolated from the terminals (I have just checked to confirm that) so no need for mica there.
I do have mica strips for the under side of the FET. I was going to use a piece on the top, under a washer to isolate...and do something to isolate the shank of the bolt...
Yes...metric taps for the next project... :P
-
Hello all
During the holidays our team also got our own dummy load built.
Vishay IRL520 logic level MOSFET
Texas Instruments LM324N Op-Amp
Hopesun PM-2 Ammeter
Bourns 3540S-1-103 10-turn 10K pot
Riedon PF2470 TO-247 power thin film resistors
We have captured some pictures and they are on our website photo journal page.
http://www.binarytaskforce.com/photocontact-album-13 (http://www.binarytaskforce.com/photocontact-album-13)
Cheers, Team BTF
-
I managed to have some chill time today and continued to tinker and I'm just about have the enclosure complete...
I hope to have the assembly and wiring complete this weekend.
-
Like the safety sandals there Martin.........
Though those vent holes were part of the moulding of the case, well done with getting them all so nice.
-
Like the safety sandals there Martin.........
Though those vent holes were part of the moulding of the case, well done with getting them all so nice.
Yes, carbon fibre reinforced! :P
...part of the moulding....not...courtesy of some great CAD drawings and some painstaking drilling :o
-
Ok...almost there!!!!
Started assembly...had hoped to complete it this weekend but...I had still not decided how I was going to power the panel meters....
I decided in the end to go with a DC to DC converter which I have now ordered form RS so I should have it all together next weekend.
-
Looking good there Martin.
-
Finished at last!
Here is the link to the final video: https://www.youtube.com/watch?v=afM1aujgAF8
I'll post some pics here and follow-up with the CAD drawings, final circuit and parts list soon.
-
Looks excellent, very well crafted. Well done.
-
Hello all
Previously mentioned before, during the holidays we got our own dummy load built.
Vishay IRL520 logic level MOSFET
Texas Instruments LM324N Op-Amp
Hopesun PM-2 Ammeter
Bourns 3540S-1-103 10-turn 10K pot
Riedon PF2470 TO-247 power thin film resistors
Provide a quick write up here.
http://www.binarytaskforce.com/Weblog/2013/02/08/dummy-load/ (http://www.binarytaskforce.com/Weblog/2013/02/08/dummy-load/)
Cheers, Team BTF
-
Martin I know you put a lot of work into that el-load and good job! You mentioned in the video that with your max current limit set you did not get to use the full range of your course current adjustment control. May I make a suggestion to look at my full schematic in this thread and see how I implemented this feature. I think you may only need to move a couple of wires on your unit so that the max current limit comes after the course current adjustment and not before it. This way you will get the full range of that control no mater where the max current limit is set because you are then simply scaling the control voltage to the opamp. I haven't seen your updated diagram so I might not be correct. Anyway great job and that was a good detailed review in the video.
-
Martin I know you put a lot of work into that el-load and good job! You mentioned in the video that with your max current limit set you did not get to use the full range of your course current adjustment control. May I make a suggestion to look at my full schematic in this thread and see how I implemented this feature. I think you may only need to move a couple of wires on your unit so that the max current limit comes after the course current adjustment and not before it. This way you will get the full range of that control no mater where the max current limit is set because you are then simply scaling the control voltage to the opamp. I haven't seen your updated diagram so I might not be correct. Anyway great job and that was a good detailed review in the video.
Thanks very much for the feedback John. Heavens...I read your post twice (bit of a head cold...) thinking "what have I done..." but now I realise what you are saying and you are correct...I'm "wasting" many turns on the coarse adjustment pot.
I will investigate soon!
Cheers,
Martin.
-
Looks excellent, very well crafted. Well done.
Thanks Billy.
-
Here are the CAD drawings of the enclosure.
Thanks very much to Yves (forum id: Yttrium) for doing the enclosure design work and providing great guidance along the way.
Please note that these drawings are for non commercial use only.
Parts list:
* Infineon MOSFET N-channel 200V 13.5A TO-220 - BUZ31L H
* Vishay 534 Series Pot with 6.34mm shaft, 50K
* Vishay 534 Series Pot with 6.34mm shaft, 5K
* 10 turn 10 k trim potentiometer
* 1uF 50 volt electrolytic capacitor
* Arcol HS50 Al house wirewound high power resistor,1R (1 ohm) 50W - HS50 1R J
* National Semiconductor Quad op amp, LM324N 1MHz DIP14
* LM7806 voltage regulator
* TRACOPOWER, TMA1212S, unregulated DC-DC,12V 1W - (isolated power for the panel meters)
Cheers,
Martin.
-
Very nicely done!
-
Martin I know you put a lot of work into that el-load and good job! You mentioned in the video that with your max current limit set you did not get to use the full range of your course current adjustment control. May I make a suggestion to look at my full schematic in this thread and see how I implemented this feature. I think you may only need to move a couple of wires on your unit so that the max current limit comes after the course current adjustment and not before it. This way you will get the full range of that control no mater where the max current limit is set because you are then simply scaling the control voltage to the opamp. I haven't seen your updated diagram so I might not be correct. Anyway great job and that was a good detailed review in the video.
Hi John,
I've just posted the basic circuit diagram. Can you please take a look and point me to where I might have made a boo boo regarding the coarse current adjustment.
Cheers,
Martin.
-
R4 should be the limit pot, with R2 being the coarse control, then R3 being the fine control. If you only have a 2k coarse control use it in the place of R2, and use a 1k ten turn preset as the limit pot in the place of R4. this means the maximum voltage available to the opamp as a reference is then going to be a voltage in the full turn range of R4.
What are the functions on the pots you have? It will also be a lot better to run the LM325 from the input side of the egulator rather than the output. It will be able to drive the mosfet better, as it then has a higher supply voltage to work with.
-
Hi Martin,
I looked at your schematic and I think it should work, so I may have misinterpreted your video. I am assuming that R4 is your fine adjust with R2 as your coarse adjust. R3 is your maximum current limit. Hope I got that right. Putting two potentiometers in series for fine and coarse adjustments does work but there are some problems with linearity and total control. For example when you have the R2 coarse pot's wiper at or near the pot's pin 3 you lose control of the fine adjustment pot R4. With the fine pot's wiper at its pin 1 the coarse control goes to about 10% less than the total control voltage, (which is normal for this configuration). You should also jumper pin2 and pin 3 on R4 with a wire. If and when pots fail a loss of wiper contact is common. With that jumper the unit would still work at a slightly lower maximum current until you replaced the pot. Without the jumper the unit would stop working as a load.
I assume the way you are setting the upper current limit is to set pot R4's wiper to pin 3 and R2's wiper to pin 1 then adjust R3 for maximum allowable current of the unit. If you are not getting the proper range from the coarse pot, as a troubleshooting check, set the fine control's wiper to its pot pin 3 and see that you have full range over the coarse pot adjustment. If you do not have the proper control range you probably have a wiring error in your circuit because the diagram looks all right.
There are several better ways to implement a fine/course adjustment to reduce or get rid of the problems I mentioned. A dual ganged pot in place of the fine control is very good or you can see how chrisw957 did another method using Dave's basic circuit. I think chrisw957 is using single turn pots.
http://www.sleepyrobot.com/wp-content/uploads/2011/05/currentload.pdf
Let me know if I can be of any further help.
-
R4 should be the limit pot, with R2 being the coarse control, then R3 being the fine control. If you only have a 2k coarse control use it in the place of R2, and use a 1k ten turn preset as the limit pot in the place of R4. this means the maximum voltage available to the opamp as a reference is then going to be a voltage in the full turn range of R4.
What are the functions on the pots you have? It will also be a lot better to run the LM325 from the input side of the egulator rather than the output. It will be able to drive the mosfet better, as it then has a higher supply voltage to work with.
Thanks Sean, I see what you are saying. I don't really want to undo this circuit at this point as it's fulfilling it's function for me at this point in time. I will pop it on a bread board to test what you have noted ready for the next one I put together.
R4 - Fine
R2 - Coarse
R3 - Limit
Cheers.
-
Hi Martin,
There are several better ways to implement a fine/course adjustment to reduce or get rid of the problems I mentioned. A dual ganged pot in place of the fine control is very good or you can see how chrisw957 did another method using Dave's basic circuit. I think chrisw957 is using single turn pots.
http://www.sleepyrobot.com/wp-content/uploads/2011/05/currentload.pdf
Let me know if I can be of any further help.
Hi John,
Thanks very much for your input. I had a look at Chrisw957's circuit and I will certainly test his method on a breadboard too.
Just to note on my diagram:
R4 - Fine
R2 - Coarse
R3 - Limit
Cheers,
Martin.
-
Q1: How do you wire the amp meter into this circuit.
Q2: How do you identify the pins on the variable resistors.
-
Hello mrbearuk,
If you go back a few pages I posted a circuit diagram. The ammeter is in series with the + positive input and pin 2 of the FET. The two variable resistors are labelled with the pin outs. There should be a similar diagram / symbol on your variable resister / 10 turn pot that corresponds with the circuit diagram.
Hope that helps.
Cheers,
Martin.
-
Hello,
I'm need to build DC-load with will be able to dissipate 150W of heat.
I assume following requirements:
- maximal voltage (30V)
- maximal current (5A)
My question is does someone of You create PCB for similar device ?
In posts at this topic i found some projects, but there was dedicated to smaller power and was more compact.
My device should just have controller PCB with output for transistors (no transistor or heatsink on PCB itself)
My device will use computer CPU cooler.
I'm just asking if someone create similar project and will share (to not do same job twice) if not i will create by my own and if everything will work share project.
Regards
-
Any design will do. Just replace the power transistor (BJT or MOSFET) with a bigger one, or better yet, use multiple transistors. I think it's decent to keep it under 50W per TO220 and 100W per TO3 or TO247. You may also increase the value of the sense resistor to dissipate some power there.
My second dummy load handles 15A, 45V or 250W (whatever comes first). I think I can squeeze 3-400W if air cooled.
I don't have a complete schematic but I drew the power side, since that's where the magic happens. There's an error: the 0R1 5W resistors are actually 2 in parallel for each FET, so effectively 0R05 10W for each transistor.
(https://lh3.googleusercontent.com/-ZOHZDvHsoek/Ujt5gAZAVPI/AAAAAAAAC8o/t13pSVOf7VU/w694-h609-no/DUMMY+LOAD.jpg)
And this is how it looks like, complete with range switch (x1, x10, x100) and external modulation input (BNC):
(https://lh4.googleusercontent.com/-XTlj5Gz-HCc/UkjCw-8nrtI/AAAAAAAAC90/6KsvSY1n_IA/w588-h528-no/IMG_2820.jpg)
I tested this with an ATX power supply and maxed out at 18A from the 12V rail, so in excess of 200W.
So what I'm saying is use any design, but beef up the power side, and it will work.
-
How many of you have built this and would you change anything ????
I just started designing this and it's not done ..
Whats your thoughts ???
(http://i87.photobucket.com/albums/k133/nucklehead_2006/IMG_20131113_0722401_zps454cbdc7.jpg)
-
Hi Martin, Well finally I get to make a start on the constant current load, first project for 2014, /I have two led display units which I intended to put in it, however looking on e-Bay I see a backlit LCD unit which does DC volts and amps, with that nice pale ice blue back light, so is there any special criteria, or spec for the metering, I'll not be adding the data meter sockets on mine as I don't have or have the need for such a multimeter. This is a nice project, and it's nice to see everyone ,making one to suit their own specific needs ;)
Paul
-
Hi Martin, Well finally I get to make a start on the constant current load, first project for 2014, /I have two led display units which I intended to put in it, however looking on e-Bay I see a backlit LCD unit which does DC volts and amps, with that nice pale ice blue back light, so is there any special criteria, or spec for the metering, I'll not be adding the data meter sockets on mine as I don't have or have the need for such a multimeter. This is a nice project, and it's nice to see everyone ,making one to suit their own specific needs ;)
Paul
Hi Paul,
No , if the meter has the measurement range and resolution to match the DC load you are building it should work fine. Hope you have fun with the build.
Cheers,
Martin.
-
Why is the main 10-turn pot 50K ???
Can't it be some other value ??
-
Why is the main 10-turn pot 50K ???
Can't it be some other value ??
Yes, its just a voltage divider, so any value ought to do the trick.
-
Where can I get the case and heatsink? I use CPC to order my parts but cant find them there.
thank you.
-
For mosfet, I assume IRL540N (LogicLevel) is just fine?
Ax
-
Martin, from where did you get that enclosure from? I can only find closed Hammond cases (=not removable from one side).
Ax
-
Hi Ax,
I got it from RS Components in South Africa....I've just gone online to check and they don't seem to stock it any more. I'll see if I can dig up a part number...
Cheers.
-
Thanks. Found it, built it.
I found out that the highest current load is 2.4A. I used LM324 with IRL540N. Am I right if I assume it's because LM324 has top rail voltage Vdd-1.5. For 2.4A load gate voltage is 4.7V which is all LM324 can drive using 6V Vdd.
Hmmm should I use higher Vdd for the controlling circuit to get 3A or so...?
Ax.
-
Hmmm... I opened IRL540N datasheet and 4.7V is enough to drive approx 45A. I can't understand what limits the current as low as 2.4A.
btw: potentiometer has "dead zone" beyond 4.5V due to LM324 Vdd-1.5V issue. It is more or less cosmetic issue but can be fixed by two resistors to lower 6V.
Ax.
Edit: easy fix for potentiometer can be done with voltage divider of two resistors: 4k7 and 10k.
-
Are you realy sure the MOSFET's Vgs is at 4.7V? Because I'm thinking maybe you're measuring the voltage from GND to Gate, not from Source to Gate. What I mean is there is a voltage drop on that sense resistor, and as current increases, in can be substantial.
-
Martin, from where did you get that enclosure from? I can only find closed Hammond cases (=not removable from one side).
Ax
I've seen them on E-Bay, however there even expensive on there too,£55:00p UK that's a lot of money. :o
-
Hi, before anything I would like to thank in advance for the helpful videos! I'm building the de load project. I'm using the LM358 and the mosfet is a IRFZ44N. I would like the circuit to draw at least 3A over a maximum of 15V. On the circuit the mosfet gets very hot and I'm wondering how hot is normal. Supplying the circuit with 9V, drawing 500mA over 6V gets the mosfet build up about 80° C rapidly. (Without sink). What would happen at higher voltages or amps? Is it the wrong mosfet? The data sheet is very similar. On the oscilloscope reading the gate I get oscillations above 300ma.
-
0.5A at 6V is 3W of power, and this has to be dissipated somewhere, and in this case that place is the MOSFET> 15V at 3A means you need to dump 45W of heat, so the heatsink is definitely needed, as without the heatsink the Mosfet is only really going to be able to dissipate 1W of power at best.
-
Hi Folks / hi Martin,
I would like to built Martins electronic load by myself but I do have trouble
in getting the BUZ31LH.
All places / shops gave me the answer that this thing will not be produced
anymore and is already sold out.
Does anybody know a replacement for the BUZ31LH?
And if yes, do I have to change the circuit at some point?
Thank you very much in advance for your support!!!!
And thank you Martin for your wonderfull youtube chanel. It brought me back
to the electronic hobby.
Regards from Germany
Ronald
-
I'm using the IRFP240 (TO-247) as a replacement in my smallish test unit, which is available for just 1,40€ a pop from Reichelt. It worked with the circuit shown in http://www.mjlorton.com/forum/index.php?action=dlattach;topic=106.0;attach=264;image and it also does work fine with halved source resistors (so 2x 1 Ohm in parallel). It even does this with two MOSFETS and the LM324 part doubled, which is already very powerful.
For the final unit I'm however thinking about using the IRFP250MPBF (from Mouser, no sources in Germany I guess) and again smaller resistance values. Dunno if it works to the specs I'm aiming for, but you'll never know until you've tried... ;D
-
Thank you very much for your reply!!!!