My Variable DC Power Supply (30V/1A)

[Strada916]

Hello Ladies and Gents!

I stopped a little bit with the construction because of the situation at work, but now I'm about to continue. I don't have all the new components yet so I'd like to ask a couple of questions in hope that somebody can shed some light.

1. If I do everything according to blankfield's suggestion how many Amps of output can I expect from the power supply? How can that be estimated?
It will largely be limited by the transformers output VA. If you are using 5W resistors P=I^2*R, 5 = I^2*0.1 transpose the formula for I = sqr(5/0.1) = 7A each trans, may want to beef the shunt resistor to a 20W

2. Is there any reason to consider changing the LM324N integrated circuit with something better or it's good and won't burn?
324 seems to be ok

3. What value should the three resistors be? The ones that divide the load equally on the three transistors? Do they need to be also 5 W or smaller value is also OK? I'm just wondering because space is small in the case.
Start with 5W see what happens
4. How about diodes? Should I buy something better than the default ones?
Why?

[birrbert]

Hi Strada916!

Thanks for the info! The transformer is second hand. It was used in an audio amplifier and it can handle 4-5 Amps easily without getting hot (the guy who made it told me).

Now, I'm a bit confused about the shunt resistor. blankfield wrote that I can reduce it to 5W, but you are saying that I need a 20W. What's the truth?
The value of the shunt resistor depends on the value of the three other resistors that divide the load equally on the three transistors? Sorry about the noobish questions; I honestly don't mean to offend anybody; I just want to get a clear explanation how it works.

Regarding diodes, I asked that question only because those which come in the package I'm not sure how many Amps can they handle. I will double check and just make sure that they can handle at least 5 Amps.

edit:
Question nr. 2 was about the LM324N. I read the description of the J31 kit and even though Google Translate does a terrible job in translating from Polish to English I found out that the upper limit of this IC is 33 V. This is actually incorrect, because the datasheet says 32 V. What's the problem with that? My transformer is 25 VAC, but after the diodes I'm getting 36 VDC (measured with multimeter). This 36 V is about right if we multiply 25 with 1.41 (I'm still learning what this factor number is). I have to confirm that with measuring at the legs of the IC, but still, even though 4 extra Volts might not seem too much, I think I need to be aware of that. Plan A to substitute LM324N with something that can handle higher Voltage or plan B would be to get a lower Voltage transformer, e.g. 20-22 VAC, I guess.

[SeanB]

Actually you can replace R5 with a short and use one of the other resistors for current limit. It will save on the one resistor. Use 5W units, and as good a tolerance as you can get, 5% or better, and take the connections for current limit from any one of them. As they share the current you can use one for sense. They will work for up to 5A per resistor, or 15A total without problem, though you will be better off using a lower current like 1A per resistor to reduce the change in resistance with heating.

The 324 will do fine, it is a good unit in this case, simple and robust. The bridge rectifier can be upgraded, a 25A unit will run cooler and not need a big heatsink at all.

[Strada916]

If the diodes you have are 1N4001 or simular then yes get some higher current ones. Also the above circuit will be happy with 15A flowing, however the shunt resistor may melt due to heat at that current. Since you transformer can only output 5A. I would not worry. I see you concern about 324 as your transformer has a higher output voltage. Maybe put some diodes in series to drop the voltage for the IC only.

P = I^2*R this formula is your friend 

20W = I^2*0.1
rearrange to find I, I = sqrt(20/0.1) = 14Amps but you always under rate by 50% So there is plenty of head room.

1.41 = sqrt of 2, its derived from the sin wave being positive humps and no longer going negative .

I think the 3 X 2N3055 is over kill, two to be safe, you could even get away with just the one. as the transistor can handle it. Although it would get fairly hot at 5A on its own.

[SeanB]

The 3 2N3055 is actually fine for 5A, as the SOA of the transistors is pretty poor, and 3 sharing will have enough SOA for it to survive a short at full output indefinitely with a big enough heatsink. As well the gain of the transistors will be higher at lower collector current, though you are always able to use better more modern transistors with higher voltage capacity and gain, using these will still be better with 3, as the heat can then be transferred from the dies easier with 3 separate paths.

At 15A a 2N3055 is only guaranteed to have a gain more than 1, it can go up to 10 or more at 3A, reducing the base current required from the driver.

[birrbert]

Reading your replies I started to get the hang of it.

To start with I'm going to stick to the original J31 scheme as much as I can, meaning that I will install the shunt resistor, with the change that I'll use a 0.1 Ohm/20 Watt instead of 1 Ohm/5 Watt. This is simpler for me and if it doesn't work out then I'll do SeanB's method of "short and use one of the other resistors for current limit".

I will upgrade the bridge rectifier.

Thanks for the tips about resistor tolerance! In the case of power resistors (5W and above) I couldn't find lower tolerance than 5%. I will use cemented carbon film resistors. All the other resistors I will swap with 1% tolerance metal film resistors (I have a friend who sells them at the price of 1 Cent/piece).

tbc

[Monkeh]

Thanks for the tips about resistor tolerance! In the case of power resistors (5W and above) I couldn't find lower tolerance than 5%.

http://uk.farnell.com/vishay-dale/lvr05r1000fe73/resistor-precision-0-1-ohm-1/dp/1108091

Ignore the tempco, it lies.

[birrbert]

Hi Monkeh! Thanks for the heads up, though at first I will go for the cheaper version. Well, these ones are not badly expensive, but I have to pay a lot for shipping. We'll see. By the way, temperature coefficient is explained in the datasheet.

The good news for today is that I received a big heat-sink with 2N3055H transistors on it. I was lucky to find this and be able to buy it for a low price. Plus, I got a couple of cool bridge rectifiers. Check out photos by clicking the numbers below!

1 | 2 | 3 | 4 | 5 | 6 | 7.

Only God knows how I will be able to attach this beast to the back of the case. Dimensions: 24 cm long, 10 cm high, 4.5 cm deep.

[Monkeh]

By the way, temperature coefficient is explained in the datasheet.

I know, that's why I said ignore the site.

[Strada916]

What are the two transistors in the centre of the heat sink?

[Monkeh]

What are the two transistors in the centre of the heat sink?

http://www.svntc.com/TPDF/121.pdf

[Strada916]

What are the two transistors in the centre of the heat sink?

http://www.svntc.com/TPDF/121.pdf

wow.  Since they are already on the heat sink can you use them instead of the BD139?

[SeanB]

Yes, you will just need one extra TO3 and one TO66 isolation kit for them and the other 2N3055 that is not isolated. As well remove all the devices, clean the heatsinks and the bases of them and re-apply the heatsink compound and mount them again, the old compound probably has dried out by now.

[Adrian]

Congratulations! 

[birrbert]

Thanks! Mersi!

But, to be honest, I couldn't finish it without your continuous help. So, thanks guys! I really hope that I can complete the unit in a reasonable amount of time.

edit:
What kind of heat-sink compound do you recommend? I found too many types (with silicone, without silicone, silver based, etc).

edit2:
I went for some very cheap silicone based compound. During next week I'll be getting all the new stuff I bought: a new J-31 kit, insulators, thermal compound, isopropyl alcohol, capacitors, resistors, etc. The list is long, but as they arrive, I'll get to work. For TO-66 I found for a good price on eBay, but that might take a bit longer to arrive to me than all the others.