Updated schematic for a popular DIY linear bench power supply

[Aramp]

Hi everybody,

I have recently purchased a popular power supply DIY kit on Aliexpress. It's available on eBay, Aliexpress and Amazon. If you google something like "0-30V 0-5A Adjustable Power Supply Constant Voltage And Current Diy", you'll find all the links. The attached ZIP file contains the image of the empty board, so that you can easily identify it.

The power supply is rather decent, and could serve well on the bench. Howerver, there are two problems with it.

The first problem is that it may not work, because the linear regulators used are not supplemented with dumping capacitors as per the datasheet, and will most probably start oscillating. And the second, more general problem is that the schematic supplied with them is for an old version of the power supply, so if something does not work, it's very hard to debug the problem. That's exactly what happened to me: it did'nt work, and figuring out the reason was a headache, because the schematic was incosistent with the board.

So I went on to draw the updated one, and that's where I figured the problem with the capacitors, and as you may guess, it was in the part that was changed. In doing so, I have also found the original Chinese copy in PDF format, as the one on the sites is translated into English with automatic translation and is converted into a low resolution raster image, so it's sometimes difficult to discern. The alignment instructions that are provided are also poorly translated, so I had to redo the translation and correct it for proper English.

In the attachement there is a drawing of the new version of the schematic, with all the new replaced components, as well as the original high-quality PDF drawing with proper translation from Chinese that comes with the boards when you purchase them. Please note that the C20 and C21 capacitors are not supplied and there are no places on the board for them. As I mentioned above, they are needed to stop the regulators from oscillating. I have soldered them to the legs of regulators before soldering them into their places. I have tried to retain the parts labeling close to the one found in the original shematic and on the silk print.

The ZIP file also contains a proper translation of adjustment instructions. The adjustment is actually rather simple.

[moffy]

Just a quick note the 3DF20 in the redrawn schematic is shown as PNP instead of NPN.

[Aramp]

Just a quick note the 3DF20 in the redrawn schematic is shown as PNP instead of NPN.

Thank you! I have fixed that.

[p.larner]

I built the same a while ago,that used two transfo's,do you just use one?,there is a yt vid of a fella building this,also his website has the schematics.

[Aramp]

I've used one transformer, happened to have one scrapped from an old transistor amplifier. They are often sold cheaper than a separate transformer. Tried to find vids on youtube when run into the problem, found a couple, none had my problem and was unable to find the schematic. So had to draw the schematic myself. It also helped to understand how it works, which is a good thing.

[Kleinstein]

As drawn for the new version the voltage feedback is from the wrong side of the ampmeter. The regulator will not compensate for the drop at the ampmeter.

[dietert1]

I found the kit at mercadolivre.com.br for 149 R$ (about 30 €). That version appears to have the 7812 and 7912 voltage regulators for the opamps. The main problem with this kit is the difficulty to get a proper mains transformer. Then the meters are missing and the enclosure.
Considering that i got OWON P4305 supplies for 140 €, i think that kit only makes sense for learning purposes. The OWON has a toroidal transformer and it includes digital control of its settings and readings of actual current and voltage via RS232. Its main disadvantage is the missing analog (fast) overcurrent limiter.

Regards, Dieter

[Aramp]

As drawn for the new version the voltage feedback is from the wrong side of the ampmeter. The regulator will not compensate for the drop at the ampmeter.

You are definitely right. Overall, there are several changes, some are good, some are not that good. Judging from the original Chinese forum where this was first published, the guy seemed to be focused on turning the new design into a business of selling the kits, hence he didn't publish the schematic and didn't disclose what exactly he has changed.

[Aramp]

Considering that i got OWON P4305 supplies for 140 €, i think that kit only makes sense for learning purposes.

Dieter, agree with you, unless there is a transformer laying around that you want to utilise, which was exactly my case. The second reason that made me  opt for DIY is the sheer weight of linear power supplies. You see, there are parts of the world where shipping gets very expensive with weight, and I live in that kind of country. And finally, unless very restricted in place, one wants to have at least a couple of power supplies anyway.

[Kleinstein]

It is a bit problematic of having a kit that needs an extra, relatively special transformer. Without a schematics it would be a bit tricky to adjust to a slightly different transformer (e.g. lower power or different voltage). The principle circuit is quite flexible and the same PCB could be used with different voltage ranges with more minimal changes in the parts.

[Aramp]

It is a bit problematic of having a kit that needs an extra, relatively special transformer. Without a schematics it would be a bit tricky to adjust to a slightly different transformer (e.g. lower power or different voltage). The principle circuit is quite flexible and the same PCB could be used with different voltage ranges with more minimal changes in the parts.

I have used a typical transformer that was widely used in many power amplifier designes. Opamp based amplifiers normally need bipolar source, and often the rails for low-power opamps and high power push-pull stages are separated. For the purpose of this board you would need a low-current bipolar winding for 15V and then a high-power winding that outputs voltage that depends on the output power of the amplifier. I was lucky enough that the transformer I've used had additional taps in the high-power secondary. However, in most cases it is realtively easy to get out the iron and add taps, or just wind a layer over the existing coil even without removing the iron to get the additional lower voltages. I have done that several times.

[p.larner]

i have a big transfo with about 10 or so taps ranging from 3v to 50v can i use taps from that for the cct control as well as the main power?,or will doing that cause a short?,i have a small 18v transfo but no room to fit it,using just the main big transfo would be better if poss for both the main power and control,is that dooable?.

[Kleinstein]

The type of circuit needs the supply for the control (7812 and 7912 in the circuit) and the  main power for the output isolated from each other. So one needs at least 2 separate secondary windings. If really needed one could reduce the +-12 V to a lower value. 12 V may be handy for the relays.  The control part only needs relatively low power (e.g. 2 W) - it is mainly the relays and if used a digital display part.
Unless one finds a suitable transformer using 2 transformer is a real option. An alternative maybe using a torroidal main transformer and add an extra secondary - here it could be easier with a lower voltage and lower voltage relays.

[enut11]

Hi @Aramp
Thanks for posting this power supply kit. After a frustrating time with another Chinese kit (30v/3A, red pcb), I am looking forward to a better experience and this kit seems to fit the bill.
I only need about 1A-2A so could do with just one output transistor and one relay. I will think about it when the kit arrives.

BTW, I found a small error on your re-drawn circuit diagram. You have the emitters of Q2 and Q4 joined. Q2 emitter only goes to R23 (1K) EDIT: and the bases of the output transistors.
Maybe let out a puff of smoke if the error was also on the actual PCB
enut11

[Aramp]

Hi enut11

Hi @Aramp
Thanks for posting this power supply kit. After a frustrating time with another Chinese kit (30v/3A, red pcb), I am looking forward to a better experience and this kit seems to fit the bill.
I only need about 1A-2A so could do with just one output transistor and one relay. I will think about it when the kit arrives.

I've also got the red kit, and I am playing with it right now. The one I have posted here is clearly better because of automated voltage switching and separate balanced power for the opamps independent of the main AC supply. It requires a more sophisticated transformer, though.

As to the "red" kit, I am thinking of adding the same switching circuit to it. Otherwise, it's a very clever design. BTW, its original schematic comes from the Oct 1978 issue of "Practical Electronics" journal, which can be easily found online. At the time, opamps were expensive, and the author did amazing work minimizing their number. And that's why the Chinese like the schematic: they also want everything cheap. However, while the author tried to save dozens of dollars in the 1978 currency value, today, our Chinese friends are saving cents. The problem I have now with the "red" kit is that I cannot get more than 23 Volts from it, no matter how much the input AC voltage is. It seems to be the result of the driver transistor being fake and not providing enough current/amplification or having a wrong bias. Will figure it out.

BTW, I found a small error on your re-drawn circuit diagram. You have the emitters of Q2 and Q4 joined. Q2 emitter only goes to R23 (1K).

I have just double-checked the schematic that I have posted and found nothing wrong there. The emitter of Q2 is not connected to the emitter of Q4. It goes to the bases of both Q4 and Q5 (they basically are connected in parallel for double the power output). Q2 works as an emitter follower, providing enough power to drive the powerful transistors out of the outputs of U1a and U1b, which control current and voltage respectively. Without it, the opamps may have a hard time providing enough base current for two transistors.     

[enut11]

@Aramp
This is what I found in your updated circuit. Emitters of Q2 and Q4 are connected??

[enut11]

@Aramp
This is probably not the thread to discuss 'red kit' problems but I found the output is a little lower than the opamp positive rail.
In my case I had main filter cap at 34v but was powering the opamps from a 24v regulator resulting in a max output of around 22v volt.
enut11

[Aramp]

Enut11,

Indeed, you are right! There is a dot that should not be there that skipped my attention. Fixed and updated the ZIP file.

As to the "red kit," indeed, that's offtopic, but the problem I have is probably related to the opamp not delivering enough output voltage for some reason at 25 volts region. Even when the rail-to-rail voltage is close to 36 volts, and the non-inverting input has a higher potential than the inverting one, the output does not get higher than 30 volts in relation to -5.2 volts negative rail (which is some 25 volts from the ground). I am investigating the issue further. It could be an out-of-spec opamp; they probably get them cheap from factories.

[enut11]

@Aramp
Suggest we continue discussion of 'red PS kit' on another thread or revive this one:
https://www.eevblog.com/forum/projects/hobbyist-current-limited-power-supplykit/msg4676974/#msg4676974
Post #3 leads on to the red kit mods.

[enut11]

The 'green PCB' 30v/5A kit arrived and I spent yesterday populating the very crowded the board. Only one problem so far, a component with no legs! Fortunately I had a spare.
EDIT: be aware that some of the cheapest kits do not have output transistors.

[enut11]

Suggest you save pictures of the unpopulated PCB to assist with any future troubleshooting

[enut11]

First run with the 'green' power supply.
Cold start low voltage with no load. Output is rising and never seems to settle down even after 50 min. Vertical scale is 16mV. This is not good enough for my needs.

The voltage reference is a TL431 IC configured for 2.5v output. It sits between 2 voltage regulators, not ideal even though they are not running hot. The 431 could be a fake chip.

I will have a look at fitting a better Vref...

[enut11]

Found a problem with @Aramp's wiring for the voltage pot. To be fair, it was also wrong on the PCB.
As shown, the voltage pot presented a variable load on the TL431 reference. This explains why the PS never settles down at low voltages - see above.
Fixed by reversing the wiper and hot leads to the voltage pot.

[enut11]

A repeat of the above test (low volts, no load) shows stability is reached in less than 2 min so the TL431 reference chip is ok and working well.

Having reversed the hot and wiper leads of the Vpot to the PCB now provides a constant load to the Vref line.

Next step is to fit a larger heat-sink to the output transistor and to do some load tests.

[Aramp]

enun11,

You are right; that's how the original schematic and the PCB were designed. Interestingly, the pot for the current limiting circuit is placed properly, while the current drift is much less important. I'll change my device as well, as I have experienced some voltage drift and did not think there could be a design flaw. Now, I will probably leave the schematic as is. Otherwise, people will be perplexed by the difference between the PCB and the schematic. I will only add a link to this forum in the doc inside the zip file so that people who download the file from now on will be able to find this discussion.

BTW, I have noticed the efficiency of current limiting in this power supply compared to the "red" one. Shortening the output is the prescribed way to set the current limit for this circuit. There are no sparks, nothing. The current is immediately dropped when the output is shorted, and you can control it with the pot. With the "red" one, you get some serious sparks due to a significant spike in current.  I have seen many people replacing the output transistor with a powerful one (myself included) and placing it on a good heatsink before even trying shorting. Others tried and blew almost all active components.