Convert 0-30V 3A PSU to 5A or more

[audioguru2]

Hi Chester,
Welcome to our forum.
On your website I like your propclock and big sparks. Good luck with your 15KW sine-wave inverter project.

This 30V/5A lab power supply:
If your custom-wound transformer produces 28VAC with a light load, then 44V-rated opamps are almost at their limit. Just hope that your mains voltage doesn't increase 5% like mine does at times. The calculations are as follows:
1) The peak voltage of 28VAC is 28 X 1.414 = 39.6V.
2) The rectifier bridge has 2 diodes conducting in series with a resulting voltage drop of 1.4V, so the positive voltage is 38.2VDC.
3) Add the negative supply voltage of 5.6V and the total is 43.8VDC.
You still have a 200mV margin of safety!
The voltage will be slightly over 44V if your 5.6V zener is 5% high.

I don't think that you will get 30VDC/5A regulated though, as follows:
1) Fully loaded, the transformer's output (600VA!) will probably drop to about 27VAC RMS, but its peak voltage of 38.2V will drop further during momentary rectifier current of 25A down to about 37.2V.
So the transformer's output sine-wave will have flattened peaks.
2) The rectifier bridge will probably have a voltage drop of 2.2V when conducting momentary current of 25A, and the voltage is now 35.0V.
3) A 10,000uF capacitor will probably have about 1Vp-p of ripple when fed with full-wave rectification and having a 5A load, so now the voltage is down to about 34.0VDC.
4) The three 2n3055 transistors may have a b-e voltage of 1.2+V with a collector current of 1.67A each, plus the voltage drop across the 0.1 ohm emitter resistors of 0.17V drops the output voltage to 32.6V.
5) The TIP31A for Q2 could have a b-e drop of 0.8V with a collector current of 125mA, dropping the output voltage to 31.8V.
6) If it is only 100 ohms, R15 will further drop the voltage 0.2V down to 31.6V.
7) The saturation voltage of U2 could be as high as 1.6V, dropping the output voltage down to exactly 30.0V. Just makes it? Nope, I'm not finished yet.
8) 5A through R7 (0.27 ohms) will drop the output voltage down to 28.65V. Not enough.

You might get lucky and have a lower voltage drop on your transformer and transistors, but then the maximum voltage output will probably have lots of ripple each time your 'fridge or AC turns on.

Caution! Do not use the negative offset adjustment as shown for U2 on our schematics, with your OP177 opamps. Their offset voltage control must connect to the positive supply, instead of the negative supply like most other opamps.
They have a fantastically low offset voltage anyway, so you don't need to adjust it.
The OP177 has a voltage gain that is 10 or more times less than most other opamps, so your project's regulation and ripple will suffer.

If you want 30VDC/5A that is well-regulated and has low ripple, then maybe you should re-consider using a 30VAC transformer and OPA445AP high-voltage and high-gain opamps.

 

[ante1]

Hi Chester,

Welcome to this forum.
I had a quick look at your site

 

[audioguru2]

Hi Ante,
Good point!

If your power supply isn't reliable with its patrts operating well within their ratings, it is dangerous to your health and to your load.

 

[Hilo90mhz]

Thanks for all the help everyone

Good luck with your 15KW sine-wave inverter project.

Thanks, I really need it... trying to switch 350Amps at 48v average is no joke

Caution! Do not use the negative offset adjustment as shown for U2 on our schematics, with your OP177 opamps. Their offset voltage control must connect to the positive supply, instead of the negative supply like most other opamps.

yup, I discovered that by comparing pinouts.. and designed my PCB accordingly.

Just one thing about the OP177, remember when or if they emit smoke so can anything connected to the output also be doing.

Good point.. I assumed I could do some stress testing after it was completed and if it works it works... but youre right, if im running this close to the maximum limits... anything could happen. Id still like to try

Maybe ill do some destructive testing on the OP177 to see just how high they will go, anyone for leaving this PS on for 4 days at a time? and then subjecting it to pulsed inductive loads... or something like that.

Since you are winding you secondary, why not make multiple outputs to experiment with?

I actually plan to build two of these supplies.. Using the same transformer core, but I wont build the second until I have the first working reliably, and maybe a fixed 12V and 5V supply also... who knows.

Chester

 

[Hilo90mhz]

Wound the transformer, etched and assembled the PCB, scrounged for resistors..

Just got it working a few minutes ago...

The transformer puts out 26.06V AC unloaded and 25.78V AC loaded with 1.2A (on output of regulator)

After the rectifiers = 35.16v DC unloaded and 33.47v loaded with 1.2A

Regulator puts out 28.99V DC unloaded, 28.95v with a 1.2A load.

The 1.2A load is just a halogen bulb I happened to have handy... Soon ill find a better load to test the full capacity, also this was just a test setup... its not in a case or anything yet, and im only using one 2n3055 right now.

Ive ordered samples of the OPA445..

Ive noticed that the current limit just acts like a cutoff switch.. Id assumed it would just lower the voltage to that current, but it seems more like full off full on type of thing.. maybe its just really sensitive? Is this how most variable current PS work? The current limit is just like a cutoff? so you just use the voltage knob to do both? I can see it might be more usefull this way, especially if you have no voltage meters, because the output is either at the set voltage, or its at close to zero and the current limit LED is lit.

I really appreciate all the time you guys have spent revising and improving this circuit.. All the hard work already done


Chester

 

[audioguru2]

Hi Chester,
Your rectifiers have a higher voltage drop than I expected. Aren't you using a 10A or more rectifier bridge module?

With the regulator set to 29.99V it loses voltage regulation when its output drops to 28.95V with a 1.2A load. I'll bet that the output has a lot of ripple.

The current regulator isn't working properly. It should be linear and accurate, not just a switch. At first I thought that the incandescent light bulb load was showing its changing resistance with its temperature, since it will draw nearly the same current with its voltage turned down. Catch-22: Current is too high, cut the voltage down. Still too high, cut voltage down even more, etc.

Analysing the current regulator I see an error on the "Smarter" schematic: Pin 7 of U1 should connect to the positive unregulated rail (rectifiers output), not to the output of U3. So the current regulator was dragging-down the positive supply for the voltage reference.

 

[ante1]

OOPS!

I have missed that one big time! :-[ Sorry for the inconvenience. Please let me know if there are any more errors and I will edit and repost a correct circuit guys!

 

[Hilo90mhz]

Thanks for spoting the error Audioguru!, I thought that connection looked funny.. but didn't bother to check it against the original schematic..

The bridge rectifier is a junkbox model CM3501 = 35A 100V It has a nice aluminum case for bolting to a heatsink, measured at .436 volt drop per diode

With the regulator set to 29.99V it loses voltage regulation when its output drops to 28.95V with a 1.2A load. I'll bet that the output has a lot of ripple.

Did you read my reply wrong? or am I missing something?, I was only getting 5 thousandths of a volt difference from no load to 1.2A load. Which could have by the leads going from the regulator output to my load... ( I was measuring on the load, not the regulator output)

I did notice the bulbs almost constant current draw over different voltages.. makes a nice test load in that respect... must be the simplest constant current device there is.

Ive fixed the V+ connection for U1 and the current regulation works great now..

I dont know what the best way to measure ripple on the output is.. My fluke multimeter reads zero AC voltage on the output, with or without a 1.2A load.

 

[audioguru2]

Hi Chester,
I'm glad that the current regulator works great now.
Yeah, I'm sorry that I mis-read your post, but maybe you did too. The voltage regulation is actually quite good, dropping 40mV (not 5mV as you say) and I bet that it is much better if you measure right at the regulator's output.
I'll also bet that you used 1W zener diodes that have their voltage rated at 50mA, didn't you? D8 has only 1.2mA through it, so should be rated at such a low current. A 1W zener will reduce the regulator's max. output voltage. Maybe we should reduce the value of R4 to 1K, to increase the zener current and get it up over its voltage knee.
Never mind the rectifiers' voltage drop. It will take a 'scope to see whether they switch slowly so that the peak voltage from the transformer has past before they begin conducting.
What value is your C1? The ripple rejection is excellent if you can't measure even 1mV, and you can say that it is "unmeasureable".

Do you like the current regulation that is created by an ordinary lightbulb? But it is very slow to react. A lighbulb causes a huge current surge until it has warmed-up.

 

[Hilo90mhz]

Hey.. ya, sorry about that, i got my decimal places mixed up ;D

Im actually using 5.1v 1W zeners, with a regular diode in series, using the foreward voltage drop to bring it up to about 5.6v..

I have a 20mhz DSO, I dont think they are so slow that the peak is missed.. the sine just gets a flat top.

My C1 is 10,000 uf..

I like the lightbulb current regulation just because its sooo simple.. nothing impressive about its performance.

-- still waiting on the HV OP amps..

Chester

 

[audioguru2]

Hi Chester,
I am amazed that your re-wound microwave oven transformer produces flat tops on its sine-wave output when the rectifiers conduct. I thought that it would be powerful enough so that its voltage wouldn't sag like that. Maybe your rectfier current is causing your electrical utility's mains voltage to have flat tops!

I still think that your Zener diodes are operating at a voltage that is much less than their rating because your maximum unloaded output voltage is lower than expected. The range of the voltage adjustment pot should be able to set the output voltage well into saturation of the output circuit. If your 10K resistors R5 and R6 are exactly the same then the output voltage of U1 should be exactly 11.2V, if your 5.1V zener diode plus 0.6V regular diode add to 5.6V. If the output gain-setting resistors R11 and R12 also have an accurate ratio then I'll bet that your U1 has an output voltage of only 9V plus change.

I knew it! About 30 years ago I studied the voltage curves of zener diodes with changes of current through them. Today I found an old data book but I can't find those curves on Google. Datasheets don't have curves anymore, just like skinny women.
Our project, with R4 being 4.7K, has the voltage reference zener diode operating at only 1.2mA.
A 1W zener diode such as the 5.6V 1N4734A is rated and measured at a current of 45mA. At 1.2mA it just barely conducts and is about only 4.2V (the normally vertical voltage line has curved to be nearly diagonal, like an ordinary resistor). Your 5.1V zener is even worse!
A 500mW zener diode such as the 5.6V BZY79C5V6 (available worldwide) is rated and measured at a current of 5mA, where it performs very well. It also performs poorly at 1.2mA (5.3V) but is much better than a 1W zener diode.

Therefore we should change R4 to 1K (for a zener current of 5.6mA), and specify using a BZY79C5V6 zener diodes for D7 and D8.
Edited Nov.6/04: in bold.

 

[Kain]

I have built dual output 3A PSU - the original chematic. When not used on extremas it's ok. Now that I'm trying to build the 5A PSU I need really large heatsinks since I want it to be as stable as possible. Now the trouble is that this kind of heatsinks aren't easy to find. I was planning to use 2 of those instead on 2 walls of the box, so basically those h/s will be as part of the box. I'm interested in 2 flat h/s sized 150x250mm or more. Any suggestions where I could purchase something similar or same? I live in California, US btw. If samples are available from somewhere I would really appreciate it ;D

 

[audioguru2]

Hi Kain,
Don't worry about stability, if your heatsinks don't work well then the output transistors will be destroyed!
If the project is powering a low voltage or accidently shorted load at 5A, the output transistors must dissipate about 200W.
You were talking about using a flat heatsink. But it must have big, exposed fins like the picture in the article, and even bigger.

My Newarkinone catalog has 11 pages of heatsinks, some with power ratings and temperature rise. Digikey is another large supplier. Order on the web or by phone.

 

[ante1]

Hi Kain,

If you want to put some extra money in the project you could go for water-cooling. The psu will not be very mobile but it will be immortal for sure! ;D

 

[Kain]

Ante, I've been thinking about this but it's kind of pointless to put water cooling system considering that the air cooling h/s with large enough surface area would do it fine. I have built water cooling for my PC though, but the reason is overclock, noise and staility. It has been working for 1.5 years so far with no leaks and problems (crosses fingers). About the immortality of the project - if I make a water cooled setup and it happens to leak somewhere it will be extremely mortal ;D Oh, btw here is a picture of the front pannel of my PSU - almost finished. Only the writings and knobs for the potentiometers are missing

View attachment 36064

 

[ante1]

Kain,

The PSU looks good, I can see it

 

[BOHDA]

Hello Ante
I see last observations about zenner diodes and I have to put a question-In this case can not be used TL431?
About the PCB-Is anyone who made a new PCB for PS?The old one is a little too big and can not support curents above 3A.

 

[audioguru2]

Hi Bohda,
A TL431 is much better than a zener diode with an ordinary resistor. Since D8 is operating inside the constant current feedback loop of U1 in our project, this combination is approximately equal to the performance of a TL431, when the zener and resistor are the parts as discussed before.

A few people have made a new PCB but have not reported back again.

 

[audioguru2]

Hi Guys,
I am sorry that I made a typo error in a previous post about the zener diodes.
D8 is the main voltage reference in this project and should be a BZY79C5V6, and D7 is the voltage regulator for the negative supply and can also be the same.
I have edited and corrected the error.

 

[Kain]

Ok, since nobody posted any PCB for this project, I decided to finish the job. The board I have worked on so far would be double layer board since I was trying to make it smaller, I really hope you don't mind. I can do single layer too but it would be about the same size as the original. The one I am about to finish is 100/60 mm.