Banggood PSU Enhancements

[Cliff Matthews]

Putting in numbers for visual effect..
Here's how 25v might look across this lonesome 2SD1047 (silently dying at just 2 amps)

This is why an 18vac transformer makes better sense
*Numbers plucked from Step 8 at http://www.instructables.com/id/Superb-Lab-Power-Supply/.

[Calambres]

...the trimmer do nothing to set the zero volts...

Issue solved! I changed U2 for a real Ti TL071CN and a real Ti UA741CP I had in my parts bin and problem solved! Now the trimmers work like a charm in both circuits.

It is weird how those chinese Ti TL081CP, obviously fake, do not seem to have a valid circuit for OFFSET NULL 1 and 2 (pins 1 and 5) but the rest of the circuit seem to be at least compatible.

I'll try with a set of true-non-fake LM741 (not "C")

[Cliff Matthews]

It is weird how those chinese Ti TL081CP, obviously fake, do not seem to have a valid circuit for OFFSET NULL..

Hi Calambres, good to be aware of. For those buying cheap kits, can you post a close-up photo of these fakes? Thanks!

[Calambres]

It has been quite difficult to take a shot of such a faintly printed chip, but I've managed to get a couple of pics.

The one with a single chip isf the fake Ti TL081CP that were bundled with the kit and the other is a comparison with a real legit Ti TL082CP.

[Cliff Matthews]

I bought 2 of these kits and all 6 op-amps had the pin-1 dot, and compared with those bought locally the letters are not sharp.

[Calambres]

If you look closely at some photos of these kits from diffrent sources you'll see different versions of the PCB and slight differences in the components.
I was unlucky with the OpAmps I got, but that is not a big inonvenience as I originally put IC sockets and replacing the chips is quite easy and cheap. Besides, the TL081 was never a good option for this circuit so I'm heading to my local components store to get some LM741.

[Kleinstein]

Especially if using a relative high transformer voltage ( e.g. >18 V) it is a good idea to use sockets for the OPs, as the supply voltage might be outside the specified range of the TL081.

[Calambres]

The D1047 power transistor in one of my boards went south! 

No short circuit, no mishaps, no overpower, no zap!, no magic smoke... just normal operation 10V @ 1.5A 

I think I got a "very chinese" sample. First was my issue with the Op Amps and now this!
Now to find a replacement for this tranny. That reference is not available locally.
I'm going to try with a good old 2N3055.

[Kleinstein]

The 2N3055 is a little slower, but the circuit should still work with it. The TIP3055 would be a more similar case - but maybe to weak.  A TIP35 would be a little more robust and similar case and might thus work better with the same heat sink.

The heat for the transistor is highers at low output voltage. If the heat sink is not very large 1.5 A and 10 V output could already be the limit for an SD1047.

[Cliff Matthews]

That Chinesium transistor failed under where it should have.. 36Vrail-10Vout=26Vdrop@1.5a=39W (about 80C over ambient).

[patacon]

Hi guys,
Thanks all of you for the fruitful discussion about these PSUs. I stumbled upon this Forum today while searching for info about these PSUs
(un) Fortunatey I had purchased one of these kits a few days ago. As I am total beginner I thought it would be a nice first project.
After assembling it I tested it and it worked ( supplied from 0 to 30 DCvolts output) for 5 minuts. Then output plunged to 0.4 volts and never came back, even if I tried several times. As I dont have  anyone to ask for advice nor much experince I started troubleshooting by trial and error:  checked every component several times, tested resistances, capacitances, all diodes and transistors for shorts, etc etc but still dont get what  the reason of the failure is. I also replaced all transistors but no luck.
As a final take,  I bought a second set and built it all over again. This one never yielded more than 0.4 volts at PSU output
My best wild ass guess ( see schem with voltage readings attached) is that the output at U2 is so low that  Q2 and q4 are never triggered , but again I am a total newbee and this could be all wrong. In anycase I couldn't tell why U2 AND u1 might be failing...

I have read above in  this thread  the valuable advice given to Soligen and Wrongtwins but I am not capable of solving my issues with that info

In anycase , all I want is to learn what the problem is and why this happened. I dont care too much about the PSU performance.I just want to learn about electronics.  So if you guys could give me a hand to spot the issues I'd more than appreciate it.
Thanks in advance.

Remarks:

• Input is provided by a 220 VAC to 24  VAC- 10W transformer- Actual output is 26 VAC.
  all other Voltage readings in red fonts at the schem are measured  by inserting negative  multimeter lead into location #4 (negative PSU output) and positive lead into the spots marked with a small red dot.

R2 and R15 get very hot, almost burning but I havent gotten any "smoke" from ICs.
As a last trial I had inserted  a 100 ohm resistor in series to one of the outputs of the secondary of the AC-AC transformer which reduced input voltage for PSU to 16 VAC but this didnt work ( probably some component is already damaged...)
[/list]

[Audioguru]

As discussed a few pages ago, the TL081 opamps will fail soon because the total supply is too high for them.
Your 26VAC has a peak of 36.8V and the rectifiers reduce the positive opamp supply to +35.4V. The negative supply should be -5,6V so the total opamp supply is 35.4V + 5,6V= 41.0V but the absolute maximum allowed supply for the TL081 is only 36V. That is why I suggested using 44V opamps in this project many years ago.

Your output voltage is too low because Q1 is always turned on reducing it. Q1 is supposed to have -5.6V on R14 to keep it turned off but your -5.6V is much too low at -0.7V or maybe it is very wrong at +0.7V. Maybe because D5, D6, D7, C2 or C3 is connected backwards, so check them.

[Calambres]

That Chinesium transistor failed under where it should have.. 36Vrail-10Vout=26Vdrop@1.5a=39W (about 80C over ambient).

The transistor was not hot when it failed. Besides, it had a very generous heat sink attached. I guess it was defective.
As I previously said, I've got two of these circuits and the other circuit is working fine with different loads and voltages.

By the way... when it failed I was feeding the circuit with 18V AC 

R2 and R15 get very hot, almost burning but I havent gotten any "smoke" from ICs.

In my failing board, R15 was VERY hot when the D1047 transistor failed!

[patacon]

Your output voltage is too low because Q1 is always turned on reducing it. Q1 is supposed to have -5.6V on R14 to keep it turned off but your -5.6V is much too low at -0.7V or maybe it is very wrong at +0.7V. Maybe because D5, D6, D7, C2 or C3 is connected backwards, so check them.

Thanksk for the feedback Audioguru. I will replace opamps with 44 V. I have ordered MC3 4071 and sockets.

In the meanwhile, it seems that there is another issue with negative rail as you  also pointed out. I cant have proper negative voltages there. I have measured it all over again just now. All diodes and capacitors are well connected, polarities are ok. The only negative values I get are :  -0,001 at D3,D4 anodes and   -5,5 V at negative pin of C3 but not at R3 or elsewhere, meaning no opamps get negative voltage at pin 4 ( they actually get +0,8V), same goes for R14 obviously. It seems that voltage pumping arrangement is not working right.

Any thoughts guys ?

BTW, I have also finished this DIY little function generator. It works all right!
https://photos.app.goo.gl/ya7ZzxYdlKS7N2Zf2

Thanks

[Audioguru]

The MC34071 is not made in a DIP package anymore. The tiny surface mount ones might overheat. The TLE2141 is also 44V and available in a DIP package.
The MC34071 and TLE2141 do not need and do not want the -5.6V negative voltage needed by the TL081, their negative voltage is only -1.3v because zener diode D7 is replaced with two 1N4148 ordinary diodes in series. You should use the latest schematic and parts list.

Your -5.5V at C3 is almost correct but since your negative voltage is positive then maybe your R3 is 220meg ohms (too small and is burned out?) instead of 220 ohms 1/2W.
Measure voltages with the black lead of your meter on the negative wire of C1 because the output 0V goes positive when the load causes current and a voltage drop in R7.

[cspwcspw]

I built the kit with the MCU control, superficially it seems working just fine.   But when I put a scope on the power I had doubts.  I found no specs about levels of ripple current or stability, so a few questions...

a) What do we guess "sensible" specs would be for ripple at quite modest voltages and load (6v, 500ma) say?

b) Has anyone looked at their supplies (either the original one without the MCU, or the computer-controlled one) with a scope?  Can you confirm your observed levels of ripple / noise?

c) My picture is below, 50mv per division, 5ms per division in the time axis.  Clearly, with no changes in load or voltage, and nowhere near current limit kicking in, I have something turning on then off.   Even with zero load, this artifact is present.    Can anyone else with an MCU-based supply let me know if they see similar artifacts? 

Thanks
Peter

[Kleinstein]

IN voltage regulation more the ripple should be relatively low. It depends a low on the quality of the layout. So I would expect ripple similar to cheap small supplies. So a few mV of ripple at most.

However the layout is bad when it comes to ripple in the current regulation case. This is due to a poor layout on how to connect to the shunt. So current ripply is rather high, but could be improved a lot with some small modification on the layout.

The shown curve does not look really good.

[soligen]

Peter,

The scope trace I get is nowhere near that bad, but did a major upgrade on the main filter capacitors and added another output capacitor to clean up some oscillations.

Can you link the exact kit you are using?

[Argiros]

Hallo to the group. I build one too works fine excep the led does not lights up. The Volt and Curent potentiometer are working and the volt and ampere are adjusting but the led does not light. I test it with load. The catch I found is that the resistor in series with the amp potentiometer above the TL081 the R18 was 46K I test it also with 75K and the same results.

** finally it was a stupid mistake: I put the red led in wrong direction. Everything works fine. As for the R18 value from 56K or 46K or 75K is only for limiting the curent of the board.

[Fabius_89]

Hi there.

I bought this psu kit from aliexpress. And started reading forum after forum

I made some of the modifications noted by Paul´s DIY Electronic Blog. (http://www.paulvdiyblogs.net/2015/05/tuning-030v-dc-with-03a-psu-diy-kit.html?m=1)

And wanted to share some of my insight.


First, modifications.

I replaced R3 (220 ohm) with an LM337. First configured to -3.3V output, then -1.8V.
This caused the maximum voltage output to fall from around 32V (with -3.3 Vref) to 29.1V (with -1.8 Vref). I´m thinking in replace it with the 220 resistor again and putting 2 1N4148 in series in place of D7 as Audioguru´s approach to leave -1.3V (needs testing)

I'm using a 24V transformer with 4A DC rated output.

I also put a 200K trimpot in series with R18 and reduced the output current (which originally was 3.81A to 2.5A) Mainly to reduce power dissipation in Q4 and the bridge rectifiers, which were pulling 4.95A from the transformer and this also reduced the input raw voltage from 35V to 26V. Also past 3A high ripple was present at the output at 100 hz rate (I believe the C1 3300 uF capacity was overloaded and the output was "sucking" the unfiltered voltage. I have plans to add a 2200uF (the one I have around) in parallel to C1).

Aside the Lm337, and the 200K trimpot I also added a 470uF cap in parallel with C7 which was 1uF, at the output (this caused that the CC mode is present in the milliamps spectrum so further testing is needed to see at which current setting the CC is disabled) the rest of the PSU is original. I'm also planning to add a cutoff switch to ground the non-inverting input of U1 in order to kill the output voltage of Q2. and some more stuff.

So it seems until now that reducing the negative voltage to -1.8V using the LM337 AND reducing the max output current to around 2.5A leaves this kit pretty much usable.

I measure the ripple with a DSO138 oscilloscope with 50mv and 5us divisions and the noise was 18mV peak to peak.

Hope this information benefits beginners (like me) who just bought this kit and are overwhelmed by its flaws and how to fix them. And also I'm not saying that all that was discussed (like changing the op-amps to 44V ones is not good)

It's just that in some countries is more difficult to come by with supplies (I live in Uruguay for example and my best chance to get good op amps are buying from Ebay. I cannot even get a D1047 or a SC5200 here)

Still my psu is under test and construction my next step is to get a K-Type thermocouple and start measuring the heat sink temperatura of the D1047.

Last but not least, I putted a 5A fast blow fuse and a NTC thermistor rated for 5A between the output of the transformer and the input of the PCB. And I can still hear a "PLOC" in my Pc speakers when I turn on and off the PSU. Even with the 3300uF cap (still not upgraded to 5500uF)

[Kleinstein]

Hi there.

I bought this psu kit from aliexpress. And started reading forum after forum

I made some of the modifications noted by Paul´s DIY Electronic Blog. (http://www.paulvdiyblogs.net/2015/05/tuning-030v-dc-with-03a-psu-diy-kit.html?m=1)

And wanted to share some of my insight.


First, modifications.

I replaced R3 (220 ohm) with an LM337. First configured to -3.3V output, then -1.8V.
This caused the maximum voltage output to fall from around 32V (with -3.3 Vref) to 29.1V (with -1.8 Vref). I´m thinking in replace it with the 220 resistor again and putting 2 1N4148 in series in place of D7 as Audioguru´s approach to leave -1.3V (needs testing)

I'm using a 24V transformer with 4A DC rated output.
....

Reducing the negative supply does not make much sense with the original TL081 OPs. They may need the negative supply to stay inside the common mode range, that may start at some 3 V from the negative supply. So some reduction to -3.3 V may be OK, but not below. Outside the comon mode range the TL081 can show phase reversal and thus possibly nasty spikes.

I see no reason why the maximum voltage should drop with less negative supply - maybe a variation in mains voltage happened.
With a 24 V transformer the TL081 are already used to beyond there rated maximum voltage. One may be lucky for some time, but there is a risk to this. The reduced negative supply help a little.

Reducing the current is a good idea. Even 2.5 A is a hefty load for the SD1047.

Increasing C7 to 470 µF may be too much of a reasonable idea. A relatively small capacitance at the output is one of the few good point's of this supply. A little more (e.g. 10 µF) may be good though.

[Fabius_89]

Thanks for the fast response.

The max voltage of 29.1V is with no load attached (is being measured by the amp/voltmeter) Maybe it would be a good idea to check it with a DMM to be sure. Maybe the difference is there, sorry my bad I just realized now.

The idea of lowering the negative voltage was to bring the infamous TL081 closet to 36V. In my current setup they are at 37.3V with no load and around 35V with load. I checked for transients at the output and saw a difference with the added 470uF capacitor (I can see it´s discharge curve in the oscilloscope once I cut mains power, this did not happened before adding it) also with your input I will consider removing it again.

As for the rest I´m still tinkering with the supply but if anyone is interested I can continue to post my findings.

For now I will stick to the TL081 as per said in my country is not easy to find electronic spares (the electronics store here charged me around 7 dollars each LM337, so you get the picture).

Still I assembled the op amps with socket so my idea in the future is to remove the +VCC pin of the three of them from the socket and put 10V zener diodes in series or feeding them with a Lm338 configured to 33V and leaving the negative rail at -3.3V as pointed out. The irony is that a LM337 cost me the same as the whole PSU kit (with no heatsink nor transformer) haha so, it sucks but still learning is fun

[Fabius_89]

Reducing the negative supply does not make much sense with the original TL081 OPs. They may need the negative supply to stay inside the common mode range, that may start at some 3 V from the negative supply. So some reduction to -3.3 V may be OK, but not below. Outside the comon mode range the TL081 can show phase reversal and thus possibly nasty spikes.

Thanks for the valuable info, this behavior can occur at anytime during operation, or in a current/voltage range? Sorry for the ignorance I´m trying to grasp all the concepts and learn some more

[Kleinstein]

The U3 Op does not need the full positive supply - so for this OP the 10 V zener in series to the positive supply could be OK.

There are simpler ways to reduce the negative supply (e.g. just change D7 to a 3.3V or 3.9 V zener). A small BJT wired as a VBE multiplier could also work.

With to little negative supply, to low a common mode voltage could be reached at output voltage zero or close, e.g. during a short or turned down to zero.

To be on the safe side for the supply to the TL081 one could use a series resistance and a zener to limit the voltage to some 30-33 V.  If one wants one could compensate for the voltage lost by using 2 extra diodes and an extra filter cap for the supply of U1. This way U1 would not see the ripply and might even give a little higher output under high load.

[Fabius_89]

Thanks for your valuable insight and time Kleinstein. Much appreciated.

I will continue to tinker with the supply and if wanted will post here the workarounds. If this post was closed then please just let me know, so I do not restart a closed discussion.

Best regards to all!