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

[Ughadoo]

Roomi, as audioguru and indulis already said, there is no problem to exchange a large capacitor to some smaller types in parallel.

ESR and parasitic inductance will be lower with a parallel capacitor assembly, but this will not affect its performance (in fact, if this was a switching power supply, it could even improve the performance).

Of course, price and space will be different, but it's better than not assembly the project due to a impossible to find piece.

 

[Roomi]

Many many thanks to all u friends. I am gonna buy several pieces.. 

 

[TooExpensive]

Hi,
this is my first post here, so please dont eat or beat me.

I WAS willing to build either the oiginal or the improved PSU. I even bought some parts for the project today. But after coming home i made the mistake reading the two threats. The one for the original design with incredible 78 pages and this one with 22. A huge input and an outstanding example of endurance. Good work, especially by kain and audioguru. This runs for 2 years and still postings are comming in this proves the interest in this circuit. But after reading all this i am missing a summary. It seems that the double sided layout works out very fine. Maybe the autors can add it to the original projects page to make it easier building it. I find it very confusing to search every bit of information from the posts. Maybe some member could post the recent parts list/schematics/layout?

Bye, keep it going!

 

[audioguru2]

Hi TooExpensive,
Welcome to our forum. ;D
The threads are long because they should be divided: Theory about why the original circuit doesn't work, Max ratings of parts, Substitute parts, Heatsinks, Adding meters, Troubleshooting etc.

Somebody should post a new project about what they have done when making theirs.

 

[TooExpensive]

Hi there,

I have researched a bit and found this project worth building, it is of usable quality and as a big plus the layout is available so Ill just give it a try.
Since I have a 24V/210VA transformer available this is my starting point. I want to use the layout of the original project of course with the benefits of higher output current. U1-3 will stay TL081, I have them already. 3x 2N3055, each with emitter resistors of 0,1R-0,27R/min. 2W (all will get the same value). Q1/Q2 will be TIP31A or BD241 depending what I can find in my boxes or what is available at the store.
If I got it right the absolute max. current is despite the power dropped over the output transistors depending on the voltage over R7 which is 1,41V (I hope I remember this value right?!) for full current limit with current potentiometer set to max. current? My simple mind says: lets drop R7, add another 2N3055 and get more current safely. I know this will most likely not work out...why...an even better: can it be done and how can it be done?
Anyways for the parts consider the parts list here:

 

[audioguru2]

Since I have a 24V/210VA transformer available this is my starting point.
U1-3 will stay TL081, I have them already.

The 24VAC will be about 25.5VAC without a load. Then the peak voltage will be 36V and the unregulated voltage will be 34.7VDC. Most of the opamps use the positive unregulated voltage plus the negative 5.6V supply for a total of 40.3VDC. The absolute max supply voltage for a TL081 opamp is only 36V. There are a few inexpensive opamps rated at 44V available such as the MC34071AP and the TLE2141CP which would be more reliable for your project.
The losses in the circuit will produce a max regulated voltage output of only about 25VDC at full load because the transformer's voltage is too low for 30VDC. The 25VDC might have high ripple. A 30VAC transformer and high voltage opamps are needed for 30VDC at full load.
The power rating of your transformer is fine for a load of 5A. 

3x 2N3055, each with emitter resistors of 0,1R-0,27R/min. 2W (all will get the same value). Q1/Q2 will be TIP31A or BD241 depending what I can find in my boxes or what is available at the store.

Yes, but use 0.33 ohm/2W emitter resistors. People reported that the 2N3055 transistors have a wide range of current gain and the higher resistance allows them share the current better.

If I got it right the absolute max. current is despite the power dropped over the output transistors depending on the voltage over R7 which is 1,41V (I hope I remember this value right?!) for full current limit with current potentiometer set to max. current?

People found the current too high since the pot has a very wide tolerance. A 0.27 ohm resistor for R7 will produce a max current of up to 7A which will cause the 3 2N3055 transistors, Q2 and your transformer to melt if the output voltage is low. People add a preset trimpot in series with the adjustment pot for calibration to exactly 5.0A for the max current They also add a trimpot in series with the voltage-adjust pot to calibrate the max voltage which is also too high..

One thing i cant figure out about it: C10 10uF/50V. I cant find it anywhere...what did I miss? Is it a good addition to the original layout, whereever it will go ?

There isn't a C10. It was a mistake.

Audioguru, guru now! (as everyone can see, theory is important!)

Some people think so. I hope that I have helped you.

 

[TooExpensive]

Thank you for your fast reply,
I have all missing parts together now (mostly high power resistors). Except the OPamps MC34072 was not in stock the other n/a. Ill figure out about this problem later. Tomorrow will be a corrosive day (hopefully). Ill keep you updated (maybe with some pictures from my 640x480 webcam, who knows...)
Good night.

 

[audioguru2]

Except the OPamps MC34072 was not in stock

The MC34072 is a dual opamp. The project needs single opamps like the MC34071 which might be in stock.
Lots of online electronic supply stores have Motorola (now ON Semi) and Texas Instruments opamps that are rated for 44V.

 

[Roomi]

I have planned to use maxim digital pot instead of the two pots  voltage contrl and current cntrl(10K).But the maximum supply voltage of these maxim's ics is only 5V-6V).The datashhet also indicates, exceeding the voltage at any pin greater than suplly voltage may damage them.Is there any solution and has some body made the ckt with digital pot ?I am eagerly looking for it because it will fly out the mechanical problem and control will be precision.    

 

[audioguru2]

Hi Roomi,
This project has a voltage reference at the voltage adjust pot of 11.2VDC. Then the output amplifier boosts it up to 30V with its gain determined by the ratio of R12 and R11.

Reduce the 11.2V reference voltage down to 5V with a voltage divider, adjust it with a digital pot then increase the gain of the output amplifier so that it still has a 30VDC output.

The current setting pot has a max of 1.7V across it. A digital pot would be fine without making any changes.

 

[Roomi]

Hi Audioguru,
A lot of thnks..I have another idea.If i use maxim's 2.5V shunt voltage reference instead of D8 then 5V should be obtained on the outout of U1 which doesn't require any voltage divider.What do u think..??? ???

Is there any need of finer adjustment pot(1K) if digital pot is used..??? ???

Also i want to display the max output  current on 7segment display rather than marking on the body i.e.when i change the current pot shaft the max output current will be displayed on the display.Can u  please post a ckt here for this purpose or figure out a point to make it possible..?? ???

Again many many thnks to all u friends specially Audioguru and Ante..... ;D ;D

 

[audioguru2]

If i use maxim's 2.5V shunt voltage reference instead of D8 then 5V should be obtained on the outout of U1 which doesn't require any voltage divider.What do u think..??? ???

I thought about using a 78L05 to replace the zener diode and the opamp, then also power the digital pots from it. You could add a trimpot in the gain-determining parts for calibration.

Is there any need of finer adjustment pot(1K) if digital pot is used..??? ???

Maybe there are enough steps and the resolution is good enough for you. You won't be able to smoothly sweep the voltage up or down. As I mentioned above, a trimpot is needed for calibration.

Also i want to display the max output

 

[TooExpensive]

Hi,
as promised I want to post the fruit(s) of my efforts. Thanks again to everybody in the front lines for developing the circuit to this state.
The image quality is very low but its the best I can offer, therefore I have not tried to make an image of the scope line when testing the circuit.
Please keep in mind that its an experimental arrangement for testing purposes.
While building the board I noticed that the board layout is in some places a little unsuitable for this circuit, its very tight and the main conductor pathes are a bit small, but nothing serious. There is always a soltion.
The main facts: The values are always checked by oscilloscope/AC/0,1V-div/2ms-div to get no measurable disortion at the output clamps while driving my selfmade restistor network (pic.5)
Smoothened by 22000uF extern plus 4700uF on board.
Transformer ratings: 630VA 2x29,4V running free (original 1x26,3V off two paralleled windings to which I added 3 turns each)
The smaller transformer on pic.4 is 250VA 1x24V for the first tests but I want to have two outputs.
I exchanged the opamps to LM741 which runs at its supply voltage limits but nothing gets warmer than it shoulds an everything is stable.
What do I get for my money: max Volt out without load: 32,4V
Shortened output current at max. Volts and max. current: 8,3A
Maximum stable output Voltage with 4,3 Ohms attached while turning the Volt. Pot. up and the Curr. Pot. set to max. 27,6V pure flatness on the scope line at given settings while current rises to 6,2A.
Voltage set to 30V without load and turning up the current under load gives 6,25A and the clamp voltage drops to 27,8V again flatline on the scope.
R7 is two paralleled 0,47Ohms/5W but will be exchanges against 0,22Ohms/20W, they get very hot.
Of course bigger heatsink for the output transistors are needed, but its just the testsetup.
Alltogether I hope you will agree to the rating 25V/6A which was my target. It could deliver more by lifting the transformer voltage and using better rated opamps, maybe worth a thought.
Bye for now, and dont fry much power resistors.

PS: on the first test R16 smoked a bit because I exchanged Base with Emitter on one 2N3055 but nothing happened to the rest of the circuit. Man, I hate mounting TO-3 in the late evening hours .

25Vpsu.rar

 

[audioguru2]

Hi Too Expensive,
You probably can't measure it because the waveform will be distorted but what do you think is the modified transformer's voltage rating when your project has a 5ADC load?

How many milli-volts DC does the output voltsge drop from no load to 5A load at 25VDC?

 

[TooExpensive]

Hi,
I can measure the output disortion, what I did was after my mind to determin the point when the circuit becomes instable or the supply voltage drops too much. When I set the potentiometers beyond this point when driving low resistance loads it shows in form of fast growing spikes on the screen which also shows in dc setting. Of course I cant get this resolution in dc. I cant say how much the output voltage drops at 25V/5A because i didnt measure it yet, but I can say for 30V preset at 4,3Ohms. I measured 6,25A and the clamp voltage dropped to 27,8V. I try to keep that in mind to measure it. Sure it is an important stability point.
Can you repeat the question about the transformer please in other words, my english is not good enough to give you the right answer.
Bye for now.

 

[audioguru2]

Your temporary wires are too long and have too much resistance to properly measure the regulation of your circuit. A wire with only 0.002 ohms will create a voltage drop of 10mV at 5A.

I think your "spikes" are ripple since the U2 opamp is saturated up at the positive supply.

Transformer voltage ratings are at full rated current. We don't know the voltage rating of your transformer because you modified it.

 

[TooExpensive]

Hi again,
now I understand the question about the transformer:
Without mod.: free: 26,33V
loaded circuit with 25V preset/4,3Ohm: 25,7V
modded: free: 28,96V
loaded: 28,3V
I got these values from my records and I could also read the Voltage after the rectifier/smoothing sinks on the big transformer from 37,8V to 34,4V. Should I think about shottky rectifier? Or will thicker wires do the trick...Ill see.
In the original thread the BD249 for output transistor is mentioned...is this usable...are modifications needed to do this?
Bye again.

 

[Roomi]

Is there any advantage using  darlington power transistor like DTS4075A(30Amp) instead of 2N3055 ?

 

[audioguru2]

Hi Roomi,
Datasheetarchive.com doesn't list the darlington. It would overheat by itself so at least two are needed. The driver transistor and a couple of resistors might not be needed.

 

[mvs sarma]

Your temporary wires are too long and have too much resistance to properly measure the regulation of your circuit. A wire with only 0.002 ohms will create a voltage drop of 10mV at 5A...............

Hi audioguru

you may recollect the type of power supplies that had two more wires(4 wire method) and the lead drop was taken care of and the voltage is corrected at the load-- i don't mean that we can then use thin wires and take the reisk of fire due to I^2*R loss . in specific case where voltage drop is to be minimised, in addition to proper guage wirs, this remote votage monitoring and correction maybe adopte for this supply also.

regards

Sarma