0-30 Vdc Stabilized Power Supply

[Sallala1]

Hi!

Finally I built the circuit.
It started working and I was very happy.
Without a load, I can adjust the voltage from 0-31,8V, tested with a multimeter.
I use an old 32VAC transformer. The DC voltage is 46V at the rectifier.
The negative supply is -5V exactly.
I use OPA445 it's max supply +-45V  but the absolute maximum rating at +-50V.


My happyness gone when I tryed to test the circuit under a load  :'(
I started to test the circuit, with an 1 Ohm resistance at 2-3V.

1. I adjusted the voltage
2. cureant adjsutment potentiometer at the center
3 turn on the corcuit. Everithing seems to be ok.
4. I tryed to adjust the voltage and something happend:
Current supply mode led flash ones, and the output is jumped to 13V

Now, I can adjust the voltage only from 13V-31,8V DC.
The DC voltage at the rectifier decreased to 42V

What happend?
Where can I find the bug?

 

[Sallala1]

OK, I unsoldier the 2n3055 and the circuit now 0-31,5V again.
I'm happy, that my opa455"s  alive

The 2n3055 have a split between it's collector and base.

But what happend? It's a very big problem, when the 2n3055 go wrong, the output jump to 13v isn't it?
If I use the supply for an expensive circuit to supply 5V for example, and the 3055-> die.

Now, I start to find another one in my boxes....

 

[Sallala1]

Ehh, I solved my problem.
1. I measured wrong the 2n3055. But is still dead. There is a split between it's emitter and base.
2. I think it was wrong before I soldered it. (I have got three 2n3055 at home. I remember, once I test one of them. Maybe this was, beacuse I remember it's red painted text)

OK, now I try to start testing the circuit under heavy load.

 

[Spagbol]

Sallala, before you blow up another transistor......

If you test with a 1 Ohm resistor at 2-3 volts as you say, the 2N3055 will have to disspate a lot of power! (~100W!!) You really need to use more than 1 2n3055 (like in the cct in the thread for the 5A version) and they will need a very effective heatsink!

 

[audioguru2]

How can a 2N3055 in a metal TO-3 case split? ???
Maybe you are using a 2N3055 in a plastic case, that can't dissipate up to 60W like one in a metal case.

It failed very quickly. I expected it to take a few seconds to overheat before failing.
Don't you have the 2N3055 bolted to a huge heatsink with some thermal grease?

Maybe your circuit's current regulator doesn't work. It detects the high current and lights the LED but maybe D9 is open so it doesn't pull the voltage down.

View attachment 37624

 

[Sallala1]

Thank you for the reply Audioguru!

I have a very poor English knowledge.
I mean:
split = not conduct  :

Of yourse, I use a metal case 2n3055, as the picture that you attached.

Spagbol: thank you, I know that, and use a heatshink with a fan. But it' true, not a good resistor value for the start

I'am go to eat something, and after start a correct exact test with oscilloscope.

The fast tests with an 50W 12V bulb:
- Circuit can drive more than 4Amps at ~11V
- The curent regulator regulate ~4,8A at the maximum potentiometer turn
- With a 2-3Amp load, the voltage jump 10-20mV if I remove the load
(Now, I used only one 2n3055)

I made many photos whlie I building.
I made small modification on the original PCB according to Audioguru's posts. (Q2 emitter, dual 2n3055, recifier diode bridge, BD139 feet layout. I will upload everithyng soon.

Thank you!
Gábor Salamon

 

[audioguru2]

- The current regulator regulates ~4,8A at the maximum potentiometer turn.

There is your problem.
With your 46V supply (your transformer's voltage is too high) and 11V output at 4.8A, the power dissipated in the 2N3055 is 157W!

The pot shouldn't give 4.8A anyway.
The 11.2V reference at the output of U1 is applied to the voltage divider of R18 (56k), P2 (10k) and R17 (33 ohms) producing 1.7V max at P2. 1.7V across R7 (0.47 ohms) is only 3.6A and should be the max that the current regulator can supply.
Maybe your pot is 14k instead of 10k.

 

[Sallala1]

I measured the DC voltage again. It's 42VDC on C1.
(I measured the 46V before I put the IC's into the panel. Maybe this makes the difference?)

My pots are 10K. (as the label says)
U1 output is just 10,56V for me. (I sed BZX 5v6 as D8)
U(D8) is 5,25V
Something wrong with my reference?


I'am sorry. My old RFT scope dead. I must to go to the service :-S
I can't measure the ripple now.  :-[

 

[audioguru2]

I measured the DC voltage again. It's 42VDC on C1.
(I measured the 46V before I put the IC's into the panel. Maybe this makes the difference?)

42VDC when loaded is normal for your 32V transformer. The 30V transformer that we recommend would produce 40VDC and reduce max heating in the output transistor by only 6W.

My pots are 10K. (as the label says)

Measure the current-regulator's pot. Your 4.8A max output indicates that its value is more than 14k, or the resistance of R7 (0.47 ohms) is too low.

U1 output is just 10,56V for me. (I used BZX 5v6 as D8)

The BZX79C5V6 zener D8 is rated from 5.2V to 6V when the circuit uses a 1k resistor for R4. If you used the original 4.7k for R4, its voltage will be a little lower. The U1 opamp doubles its voltage.

 

[Sallala1]

Ok, I will unsoldier the potentiometers, and measure they.
My R4 is 1k, as according to the modified part list.

I promised the pcb plans.
Here they are:


My finished PCB's picture.


My modified soldier side plan


My modified pcb side


My finished circuit, and my transformator. This time, with only one power transistor.


My finished circuit, closer look.


PCB side, with the bypass capacitors for the OPAMP's. You can see, I use a shortcut for the powertransistor's emitter resistance. I will replace this with an 0,1 Ohm R, if I place a second 2n3055


The PCB plans are contain the "Q1's emitter" modification, and ready for bridge rectifier. They are draw for the BD139 transistors and ready for two 2n3055. Contain two fuse.

Sallala

 

[Spagbol]

What's the Q2 emitter modification you mention? I've skimmed this thread, maybe I missed something?

I've built the Op-amp bits with a pair of TL082s with their supply rail clamped at 22v by a zener. So far so good, but not sure how to change the output stage to give it some voltage gain?

Audioguru: thanks for the suggestion for temp sensing with a diode, will have a go at this once I get the PS working.

 

[Sallala1]

Spagbol:

Oh, I missed. You need to modify the Q1's emitter, not Q2.
Now I will correct my post.

(at Q1, you know, what to modify?)

 

[Spagbol]

Aha, suppose I must read thru the thread again. Nice photos by the way, I've never made a PCB before but maybe I will try it.

 

[Sallala1]

T1's emitter connected to the input voltage's null point. This is wrong!
T1's emitter must to connect to the output's 0V, otherwise it doesn't make it's job.

 

[Sallala1]

Twentyone century.
Nevermind!
I plan and made my PCB-s with my hand

1. plan the connections on a papper
2. put the paper to the pcb, and sign the holes with a sharp prod
3. clean the brass plate well
4. now, you can see the place of holes on the brass plate, draw the pcb with water resist ink. Or you can use special pen. Water resist ink makes coherent surface, but hard to draw. You can draw easyer with a pen, but the surface will be ragged after the etchwork.
5. put the pcb into Ferric chloride (Fe2Cl3) solute in hot water, with face down
6. 15...30min and your pcb are ready
7. clean the pcb

That's it!
Take care, to read usage instructions for the ferric chloride!!!

 

[Alun]

Why bother with all of that when you can use a laser printer and magazine paper as shown here.

 

[Spagbol]

Not sure I understand about Q1s emitter, looks ok in original cct to me. It is a safeguard against there being no -5v rail isn't it? So when there's no -5v the transistor (Q1) conducts and pulls U2 to ground so there's no drive to the output transistors??

 

[Sallala1]

As I read in the topic:
Q1 for pull down the output voltage, if you decrease the voltage adjust potentiometer.

The current flows throught R7, Q1's emitter voltage increased, and T1 won't conduct->U2 can drive T2

If thers no current flow througth R7, then Q1's emiter go down, it's base higher than it's base->T1 conduct->pull down the output throught D10.

But this is not clear to me. Wait for others' answer!

 

[Sallala1]

Why bother with all of that when you can use a laser printer and magazine paper as shown here.

Great tipp Alun!
Thank you, I will try this.

But you know, to print something to a leser printer, you must to draw it with a computer. ;-)

 

[audioguru2]

I am sorry that I made an error with the emitter of Q1. The original schematic is correct.
I corrected my mistake in a post a while ago.
It is not a serious error. The connection location for the emitter could be at the output's 0V terminal without harm and Q1 will still do its job of shorting to ground the output of U2 when the negative supply quickly disappears when the mains is turned off.