Electronic constant current DC load

[jwrelectro]

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.

[SeanB]

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.

[Fennec]

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.

[SeanB]

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.

[steve30]

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 .

[IanJ]

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! 

[steve30]

IanJ, that looks really good .

Here's a photo of my stripboard version  .

[MJLorton]

Brilliant post and information Ian. I really appreciate the contribution.

[MJLorton]

IanJ, that looks really good .

Here's a photo of my stripboard version  .

Thanks for your post Steve.

[steve30]

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?

[jwrelectro]

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.

[dr_p]

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

[SeanB]

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.

[IanJ]

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.

[Monkeh]

You could also just drive the FETs separately. One op-amp per FET, give each FET a resistor as feedback to the op-amp.