shitty chinese psu kit

[m3vuv]

Hi all,i have one of these:https://www.aliexpress.com/item/32986139693.html?spm=a2g0o.productlist.0.0.627b1d98IrSbAW&algo_pvid=9607a120-ddef-4f47-bfbb-cc527331571d&algo_exp_id=9607a120-ddef-4f47-bfbb-cc527331571d-6&pdp_ext_f=%7B%22sku_id%22%3A%2266875672074%22%7D&pdp_pi=-1%3B12.8%3B-1%3B-1%40salePrice%3BGBP%3Bsearch-mainSearch
Firstly it came without the 15v and 6.2v zeners as marked on the schematic that arived with it,ive just built it but have no place/holes/silkscreen markings for the 6.2v or 15v zener diodes,has anyone else built the same if so whats the score,do i have a reject board,i havent put power to it yet ,i would rather know what is what with it before doing that,i have mailed the seller but no responce as of yet,maybe due to time zones as im in uk,TIA m3vuv.

[Algoma]

Simplified URL: Links can exclude everything past that first: ?

https://www.aliexpress.com/item/32986139693.html

Oh, those kits often copy and paste other listings and then ship the cheapest knockoff version they can get from a local market.  Can you post a few pictures of the board you did receive?

[magic]

Why did you start another thread on this?

Anyway, if there is no zeners in the kit and no place for zeners on the board, it goes without saying that you must have received wrong schematic. At best, they use open circuit output voltage of the transformer without any regulation, at worst it's a completely different circuit and you haven't noticed during assembly

[m3vuv]

not a lot of point sending pics of the board as its now fully populated with components,it looks %100 identical too the one in the listing tho,even the bom checks out against the schematic,but lists d10+d9 that it doesent appear to have.

[magic]

If lack of zeners is the only difference, the CC/CV circuit supply voltage will be unregulated and higher (how much? no idea, disconnect the CC/CV transformer and measure its output). And you may have trouble with adjusting CC mode regulation, perhaps. Maybe they gave you different resistors in this area to compensate for different supply voltage.

Furthermore, the problem with pot connections still stands. It should be done like below (the diode can be anything, like 4148 or 4007, it's reverse biased under all sane conditions).

[MK14]

This kit seems the same:
https://www.aliexpress.com/i/32946735505.html

But it seems to have clearer pictures, including of the PCB.

Although I can't see the markings for D9/D10, I can probably see where they go, in the bottom left hand corner, between the pair of 680 Ohm 1 watt resistors.

If you have the same PCB layout as the picture in my link (click view details and buy, then please scroll down, to see big picture of PCB).

[magic]

This kit seems the same:
https://www.aliexpress.com/i/32946735505.html

Finally a legible schematic

I see that negative rail voltage has little effect on CC setpoint, just 15~20mV to compensate for LM358 offset voltage. It's still better to have it regulated, though. I don't think there would be problems using 5.1V instead of 6.2V.

The positive supply looks like it could be reduced too, if not to 5V then maybe to 10V at least, but some resistor would need to be recalculated.

If you feel like it, those stupid shunt regulators could be replaced with a 78L/79L pair. (With some care).

[Electro Fan]

Maybe it's just the particular supplier's parts, or maybe trying to squeeze out a bit more performance (35V, 5A) than can be expected for the price.

These kits (30V, 3A) from various sources sometimes including Amazon perform pretty much as advertised:
https://www.ebay.com/p/1474757007
Some versions come with a heat sink and small fan (see photo below).

Some learnings from building the kit:
https://www.eevblog.com/forum/beginners/question-on-power-supply-diodes/

Also, might try a slightly less derogatory subject title too.  Jus' sayin'

[ozcar]

. . .
If you feel like it, those stupid shunt regulators could be replaced with a 78L/79L pair. (With some care).

Take a look at this other listing.

https://www.aliexpress.com/item/32946735505.html

From a quick look the schematic looks similar. However, this other listing has a pretty good PCB picture, and it is easy to see where the two zener diodes and the associated 1W resistors are placed.

In contrast, the PCB photos in the listing given in this thread appear to show fixed-voltage regulators, 7912 and maybe 7812 in the equivalent place on the PCB. So, may not be necessary to make any changes after all.

I’d say the schematic and PCB in this thread are only distantly related to each other, so maybe follow the advice given, and stand well clear of it when turning it on!

Ah MK14 beat me to finding that...

[magic]

Those big regulators are for other stuff. We are talking the CC/CV regulator (top of the schematic) running from a small floating PSU.

This design seems reasonably foolproof, because current from the output can't pass through the CC/CV circuits to ground, except through that dangerously connect CV pot.

I would rewire the CV pot, start the CC/CV circuit without power on the main transformer and verify if everything works, then power the rest up. Also, I would mod CC/CV to run on ±5V because I have a suitable crap transformer with no better use for it.

Try to stay away from the motherboard when powering on for the first time! Prevent electric gun incidents caused by various welding errors! ! !

If the potentiometer is disconnected during use, there will be 30V voltage and 5A current output! ! !

Exactly what I said. And moreover, zeroing / shorting out the CV pot with high voltage present on the output could blow up the CV opamp. And the circuit can be fixed easily, they are idiots

[xavier60]

The control transformer doesn't need a center taped secondary if it's being half wave rectified as shown. A single 12VAC secondary could be wired to work just as well.

[xavier60]

The CV Pot could be replaced with a fixed divider and the Pot used to adjust the reference in place of R9 and R10.
The stability will need to be checked and possible compensation tweaks.

[m3vuv]

I think its possible they use the linear regs as mentioned,im just finishing the case then will power it up and try it,will be tomorow as 20 to midnight here,if i wanted to hook up a cv indicator where would i hook it up too?,cheers pAUL M3VUV

[magic]

The control transformer doesn't need a center taped secondary if it's being half wave rectified as shown. A single 12VAC secondary could be wired to work just as well.

Only if the capacitors are sufficient. A dual secondary recharges them twice as often. But you are right, it's another missed opportunity for improvement.
LOL, you are right. Those guys really were on some hard drugs.

The CV Pot could be replaced with a fixed divider and the Pot used to adjust the reference in place of R9 and R10.
The stability will need to be checked and possible compensation tweaks.

IMO not worth it, just fix the damn pot. It's nice as it is, the opamp runs consistently at unity gain regardless of output voltage thanks to C7. It's like a discrete build of LT3042 with a slightly noisier buffer

if i wanted to hook up a cv indicator where would i hook it up too?

Same way CC is done

[m3vuv]

ok,the cc led gets its from ua output so saying the same as cc dont help!,so what pin should i use for a feed for a cv led indicator,or are you taking the piss? lol.

[xavier60]

ok,the cc led gets its from ua output so saying the same as cc dont help!,so what pin should i use for a feed for a cv led indicator,or are you taking the piss? lol.

D17 could be replaced with an LED also but LEDs can and do go open circuit.
A safer option is an LED with series dropper resistor from the +15V control rail to the output of the CV opamp, pin 7.

[xavier60]

If D17 is replaced with an LED, a Zener diode or even a resistor placed across it would make the LED failure safe.

 Extra: The LED might dim noticeably with increased PSU loading. 

[xavier60]

Actually the position of VR2 is rather clever. It makes an open wiper contact fail safe, the PSU's output voltage will go low.
But its position would also normally prevent the PSU's voltage from being able to be adjusted to 0V. The R9/R10 divider fixes this.

[magic]

VR2 is
It's J2 where you plug the CV pot. OP said so in the previous thread.

[xavier60]

VR2 is
It's J2 where you plug the CV pot. OP said so in the previous thread.

So it's not fail safe at all.

[magic]

Obviously.
And it gets better. Consider such cases:
1. the output cap is charged, CV pot is rapidly turned to 0Ω
2. CV pot is set to 0Ω, 12V is backfed from outside

[m3vuv]

Thought i would just mention it,this thing has what looks identical to a zener diode,but aparently its a thermistor,ive measured it and it is in fact a thermistor,has anyone else seen one like that?,all the ones ive seen are totaly different to this either a bead type or what looks almost like a disc ceramic cap.

[Terry Bites]

You cannot gild a turd!

[MK14]

Thought i would just mention it,this thing has what looks identical to a zener diode,but aparently its a thermistor,ive measured it and it is in fact a thermistor,has anyone else seen one like that?,all the ones ive seen are totaly different to this either a bead type or what looks almost like a disc ceramic cap.

Are you sure it's not just a faulty zener diode ?

[magic]

DO35 thermistors are a thing.

Sometimes even a good zener could be confused for something else if tested at insufficient current. Low voltage ones need a few mA.

[m3vuv]

take a look at the pic,its marked thermister but looks like a zener.

[MK14]

You were right about the title of this thread. It does seem to have quality control issues.

They should call it, power supply kit with built in, set of 20 complicated technical puzzles to solve. We have randomly changed component names, values and moved things around to create a wonderful electronics problem solving arena.

Anyway, I can see why people don't think too highly of these Chinese 'bargain' kits.

[MK14]

Anyway, it looks like a small signal diode (1N4148) or zener, not a thermistor. As someone else mentioned, they can be cased like that, but that is extremely unlikely, in this case.

Rarely diodes can be used as temperature sensors (because their voltage drop, changes somewhat consistently), so it could be temperature related. Another great puzzle to solve ?

[magic]

It looks like a thermistor. If it measures equal resistance regardless of polarity (near 10kΩ is probably the most common value) and decreases fairly fast with temperature, it's a thermistor.

[m3vuv]

well it measures about 6k both directions at room temp and drops to 100 ohms about if heated in a flame,its a lot bigger/thicker than a regular 1n4148!

[magic]

Not sure if I would want put one in flame, but those things can indeed take more heat than silicon.
Similar ones are found in some kitchen thermometers going up to 300°C, for example.

Schematic and board:
https://ae01.alicdn.com/kf/HTB1zsMyXPDuK1Rjy1zjq6zraFXao.jpg_Q90.jpg_.webp
https://ae01.alicdn.com/kf/HTB1atZ0XhrvK1RjSszeq6yObFXag.jpg_Q90.jpg_.webp

[MK14]

well it measures about 6k both directions at room temp and drops to 100 ohms about if heated in a flame,its a lot bigger/thicker than a regular 1n4148!

Thermistor then.

[MK14]

Labeled NTC (Negative Temperature coefficient), for the fan control, and VR5 adjusts the setting. Makes sense.
NTC = Resistance drops, somewhat dramatically as the temperature increases. So useful for fan control.

[m3vuv]

what a pile of shite!,well i hooked this up,firstly one of the caps exploded!,replaced all with higher voltage items,2nd try,it seemed to work,next thing i heard a crack and lost the display but still had output,the lm7812 smoked and got red hot,i just cut the leads too it and powered up the psu to find the input to the lm7812 has 50 odd volts on it,can anyone shed light on this?,im sure the lm 78 series regs spec a max input of 37v?,Tia.

[MK14]

what a pile of shite!,well i hooked this up,firstly one of the caps exploded!,replaced all with higher voltage items,2nd try,it seemed to work,next thing i heard a crack and lost the display but still had output,the lm7812 smoked and got red hot,i just cut the leads too it and powered up the psu to find the input to the lm7812 has 50 odd volts on it,can anyone shed light on this?,im sure the lm 78 series regs spec a max input of 37v?,Tia.

That's hilarious (sorry  ).

Would you ever trust it to be left on (unattended), and not attempt to set the place on fire ?

Or the output, when set to perhaps 5V, for it to suddenly go to 50V, and blow up the expensive/rare thing(s), connected to the output terminals.

On the other hand, I suppose you can eventually fix all the problems/issues. But even then, the overvoltaging and overloading, may have already wreaked havoc, on the remaining components.

[MK14]

the lm7812 has 50 odd volts on it,can anyone shed light on this?,im sure the lm 78 series regs spec a max input of 37v?,Tia.

If I've understood what you have said, it seems it should be coming from a rectified and smoothed, 15V AC tap, from the transformer. So, presumably, something has gone wrong with the transformer wiring, associated relays, or something around that area. Which may well explain why all those components, exploded, overheated, went pop and fizzled, etc.
Since it is a PCB, it shouldn't be able to be badly wired up as such (Correction: Assuming the external transformer wires, are connected correctly to the screw block connector). On the other hand, possibly poor instructions, and other issues with the kit, could explain it.

[m3vuv]

Well according to the paperwork,i need a taps of relative to 0v of 15v then 25v then 35v,i was just using 13v instead of 15v and a 36v tap,am i right in thinking to get the rectified smoothed voltage from the transformer secondary raw ac, i need to times it by 1.414?

[m3vuv]

if thats so it comes out at 49v using a 35v tap,that seems over the top for 78 series regulators!,am i correct?

[MK14]

am i right in thinking to get the rectified smoothed voltage from the transformer secondary raw ac, i need to times it by 1.414?

Yes, it is. There is a slight increase (perhaps 5% or 10% or so, depending on transformer specifications, usually bigger transformer = smaller percentage, called transformer voltage regulation) if the transformer is very lightly loaded on that secondary. You will lose around 1V or 2V, when a diode (or bridge rectifier), has lots of current. There are other more minor effects, if you really read up on transformers.

So basically, yes, multiply by root 2 (1.414 ..).

[MK14]

if thats so it comes out at 49v using a 35v tap,that seems over the top for 78 series regulators!,am i correct?

Well a quick look at the schematic:
https://ae01.alicdn.com/kf/HTB1atZ0XhrvK1RjSszeq6yObFXag.jpg_Q90.jpg_.webp

Seems to show, it should be the 15V tap of L2, Not the 35V one, at that point. But, I'm still not anywhere near 100% sure how you have wired it up.

The voltages (in AC mode on a multimeter), should read the approximate voltages on the transformer tap PCB screw connector block. If not, that could explain why the 15V AC line to the 12V regulator, is actually the 35V AC one.

[MK14]

if thats so it comes out at 49v using a 35v tap,that seems over the top for 78 series regulators!,am i correct?

Sorry for the triple post, but no one else seems to be responding at the moment.

So, you seem to be using 2 transformer taps, when the circuit is expecting 3 transformer taps ?

If so, how come the 35V AC tap, is getting into the 15 V AC tap connector ?
Hence creating around 50V, instead of around 18V to 20V.

[m3vuv]

Tbh in not sure,i replaced the 7812,also the tip42,( i used a tip122 instead as it was all i had),also using a 17v tap and a 23v tap,it now seems to be working fine only goes 0-28v but i can live with that,cheers all,may post a pic when its finished.

[m3vuv]

may fitanother meter and a 1l resistor in it with some sprung loaded jacks for testing zener voltages,also a usb port running off of a 7805 for phone charging etc,seems a handy tool,just using it to charge my prc-320 battery,good to see the current limit is working ok,currently set to 28v 400ma charge,battery has come up 1v so far,4ah battery pack nicad so 1/10c 400ma seems ok ?.

[m3vuv]

a pic of it charging my clansman battery,its just loosly put together,quite happy how its turning out,the case is an old pc sub woofer

[m3vuv]

another pic,this showing the current thru my fluke,the psu readout is a tad off,need to adjust it.

[magic]

Shitty build of a shit PSU kit

Sorry, I don't think there is any way to end up with 50V on the 7812 without screwing up transformer wiring somehow

[Gyro]

another pic,this showing the current thru my fluke,the psu readout is a tad off,need to adjust it.

That has to be in the running for shitty eevblog photo of the year. 

[m3vuv]

yes truly crappy pics,i was using my phone,will take some proper ones with my dslr when its finished,sort of been beefing it up,made all the tracks that carry power a lot bigger ,fitted a 22.000uf phillips 63v smoothing cap+ a 50 amp bridge rectifier,and 5x2sd1047 pass transistors with 0.22 ohm 5 watt emitter resistors,these are bolted with allen bolts to a pc copper heatsink with heat pipes and fan,hope to get about 30+ amps from it,the transformer alone weight 6kg,fingers crossed it works out,if it does it should be a handy piece of kit.,i hope to have it finished in 2 or 3 days all beeing well.

[m3vuv]

forgot to say i may use 12v car glow plug 50a relays if needed.but i will see how the stock ones fair.

[xavier60]

Even with 5x2sd1047, you will need to avoid short circuiting the output at 30A.
What is the expected voltage drop across the output BJTs?
https://www.farnell.com/datasheets/2124326.pdf

[m3vuv]

Well the bjt 2sd1047 is rated at 12 amps max,so less than half there rated current at 30a,i suppose the voltage drop will depend as to where the output voltage is set to, and which transformer tap is switched in,,hopefully the opamps are fast enough to protect it,we shall see.

[magic]

It's all about SOA of the transistors, check the datasheet. Opamps are quite fast, but relays aren't - I suppose it's some milliseconds to switch them.

Did you fix the potentiometers?
DId you check what happens when you set CV to zero with a 12V battery on the output? No, definitely don't do it

[m3vuv]

well for the pots i fitted 10 turn 10k jobs,when i charged a car battery using it even on full current output the cc led was on for about 3hrs until the battery got to 13v then went off,at this point i took it as charged and cisconnected it,one thing that did happen tho was the output shot up to 32v,this was charging at 1.5a,the heatsink was burning hot,one pass transistor had gone sc,the wire had melted off the other but it was still ok,thats when i decided on the big copper pc heatsink and more x pass transistors.

[m3vuv]

getting there slowly.

[m3vuv]

still have 1 more 2sd 1057 to fit+ the current sharring resistors etc.

[m3vuv]

Well i did all that,heres the issue,it outputs max 10ma,as soon as the load draws 10ma the relay/relays click and the output volts drop to zero,same goes if i try and draw more than 10ma from the pass transistor driver too,ive looked all over and cant see any faults,ive renewed the tip42 driver and the op amps,diodes test ok(havent tested zeners tho,just basic diode test on those,but measure 0.7v drop foward and ol reverse) at my wits end,any ideas what to check,i was going to look and see what fed the relay coils but the schematic is missing big parts,emailed the supplier,they say they cant get the rite schematic,what a shower of brown stuff!.

[m3vuv]

Also when it all drops out,relays etc,i still have the volts/amps display so that is ok supply wise,also i can pull 40amps from the bridge rectifier ok into a load so the input voltage isnt sagging.

[xavier60]

Check to see if it's going into CC mode regardless of what the LED shows. Measure the output voltage of the CC opamp WRT the positive output terminal before and after.
In other words, measure it with the PSU unloaded then loaded.
Measure the output of the CV opamp also.

[m3vuv]

What would be what then as regards op amps and outputs,would i be right in thinking ul1A iout is high for cc and low for cvor is it u1B,i assume u2A+B are to do with the relay tap switching,the schematic is very poor.TiA

[m3vuv]

anyone? TIA

[MK14]

anyone? TIA

I can't speak for everyone else. But I will speak for myself, at least.

If you want help, it would be much better, if you supplied a complete and clear schematic, which includes all the modifications you have made. Along with a specification of the PSU, you are trying to achieve.

Otherwise, it would be very difficult, at least for me, to actually give you much useful help.

You seem to be saying things like, added a transistor or so. Not sure how transformer is wired. This or that component(s) burnt out, exploded or smoked. Changed this and that, etc.

Also, you are supplying photographs, but they seem to not really help that much.

In another thread, you seem to have said your PC, can't handle schematics (actually you were referring to PCB design software, which often/usually includes schematic creation capabilities).

Without a schematic (small/simple ones, can be done by drawing them on paper, and photographing them. But, I don't think that would help here, as it is too big/complicated a schematic, and you probably would supply a sub-par photograph, which is too unclear to be especially useful).

As it stands, I can't make much sense of what exactly is going on, what has been damaged, what has been replaced, how the thing is wired up, what you intend to make (specification wise). Even the original schematic seems to be way too blurry, to make much/easy sense of.

You don't seem to pay much attention to peoples comments, either. Which further muddies the water.

Don't get me started about easily readable forum posts (grammar etc).
E.g. What is a TIA ? (Probably Thanks In Advance, I suppose).

[m3vuv]

This is the schematic that arived with it,but not complete or !00% correct,the mods i have done is fit 2 more pass transistors and emitter resistors,a 50a bridge rectified off board and a 1000000uf 63v filter/smoothing cap,all that seems ok,i havent checked the aux supply yet tho,maybe its that,before it maxed out at about 33v now its about 23v max,no current tho,it only outputs 10ma before it drops the relay out and goes to zero.TIA

[MK14]

What power supply specifications are you aiming for, when it is working correctly. I.e. Maximum output voltage, and maximum output currents ?
I'm presuming the maximum voltage and currents are available at the same time (which is not always the case, some limit the maximum output power, usually for switch mode supplies, rather than linear types, like yours).

Also (if you already stated it, I missed it). What the specifications of the mains transformer is. I.e. Type, AC output voltage (per tap if applicable), Max AC output currents, number of taps.

[magic]

I see nothing obviously missing.

J1,J2 are the relay coils and T1,T2 drive them under control of that LM358.
The other LM358 controls CC/CV by stealing R4 current through diodes away from the pass transistors.

It couldn't be simpler.

[m3vuv]

dont know the exact specs on the transformer but it weigs about 6kg,its a big ex milatary job,it has taps at 36v-25v-17v+ others,ive powered 4 diesel glow plugs from it no problem,guess 50+ amps,the original kit was 5A at 30v,but it lists it as 50v as well,its not clear bit like the schematic,magic can you say what opamps do what,ie u1a u1b + u2 a+b?.Tia,i wonder if its trying to get its power from the small auxillary transformer for some reason,ive not looked into that yet.,the paperwork that came with the kit is in chinglish and the seller is as much use as tits on a bull!.

[m3vuv]

Also magic,you say the relays are driven by t1+t2,i see t1 but no t2,am i going blind?.

[MK14]

Also magic,you say the relays are driven by t1+t2,i see t1 but no t2,am i going blind?.

Yes.
Take the following image, twice a day, after meals with water, it should fix it.

[MK14]

Also magic,you say the relays are driven by t1+t2,i see t1 but no t2,am i going blind?.

If you mean the transistors themselves. T1 is labeled as T1. But T2, seems to have had the 2 shaved off it. It not only is blurry (entire schematic), but seems to have had the right hand section, removed almost entirely.
It's right by T1, just look an inch or so, to the right of it, for a T, missing the 2 bit (T2).

[magic]

Good catch

I didn't even notice that they cut the number, just guessed it's T2

Anyway, obviously, I mean the transistor which pulls down on J2. Look for J1/J2 labels on the schematic and draw connections between them - that's two transistors driving those two relays.

It looks like divided output voltage goes to IN+ of the opamps and you are supposed to set switching thresholds with pots going to IN-. If you simply insert pots without adjustment, it will change transformer windings at some random output levels.

[m3vuv]

So magic are you not going to tell me what section of what opamp does what?

[ledtester]

The youtuber Steve Wagner is current building the same supply and he's redrawn the board layout and schematics which you can find on his web site:

https://gswagner.com/35%20V%205%20A%20Supply/

Look at the files named "35 V 5 A Power Suppl..>".

The schematic that's been posted here is also present in that directory under the file name "Wrong but similar Schematic".

His build videos (so far):

Part 1: https://youtu.be/P2_CoiQThQ8

Part 2: https://youtu.be/3yn66h7hS5A

Part 3: https://youtu.be/gbecNRaFl8g

[m3vuv]

Thanks for that,thats brilliant,should be able to sort it out with that propper schematic etc,cheers.

[xavier60]

Well i did all that,heres the issue,it outputs max 10ma,as soon as the load draws 10ma the relay/relays click and the output volts drop to zero,same goes if i try and draw more than 10ma from the pass transistor driver too,ive looked all over and cant see any faults,ive renewed the tip42 driver and the op amps,diodes test ok(havent tested zeners tho,just basic diode test on those,but measure 0.7v drop foward and ol reverse) at my wits end,any ideas what to check,i was going to look and see what fed the relay coils but the schematic is missing big parts,emailed the supplier,they say they cant get the rite schematic,what a shower of brown stuff!.

What are you using to load the PSU? Does the output voltage recover when the load is reduced?
If there is always voltage present on the large filter cap, there is no reason to look for problems in the tap changing area.

[magic]

So magic are you not going to tell me what section of what opamp does what?

U2 is doing nothing clever, it's a comparator. IN+ is half of output voltage, when it goes above the level set by the pot at IN- the output goes up and T1/T2 collector goes down. This energizes the relay and also decrease the IN- threshold voltage by means of D7,R26 so that the opamp won't switch back until the output voltage decreases a little. This prevents relay chatter.

I suppose you can figure out which opamp goes to which relay from the schematic. Or, better, from the board

[m3vuv]

Well so far i have found the voltage on the filter caps goes up in steps as each relay pulls in,if i i wind the current pot fully ccw the led comes on+ output drops to zero,seems fine without a load on the output,any load makes the output drop to zero and the relays drop out, both tip 42 trannys have full rail power on there collectors and the main pass transistors too,they get enough current to light a 12v 21 watt bulb .any ideas? tia.

[magic]

The relays are supposed to turn off when output voltage is low to reduce heat on TIP42. So this seems to be working as intended.

Probably CC or CV acts up and regulates the PSU down to zero. So it's gonna be U1 today.

For CV adjustment, the output should simply follow CV pot voltage delivered to U1 pin 5 (output should be ~80mV less because it's fed to pin 6 through divider R9/R10). If this works under no load than the problem is in CC.

On CC side, R6,R14 bias the low side of the pot to ~18mV less than the output - check if this works right. R3,VR1 limit how far the high side of the CC pot can go - this may need adjustment. CC kicks in when current sense voltage coming from those 0.22 resistors through R11 exceeds the threshold set by the CC pot. Because 5A·0.11Ω is 0.55V, you should adjust VR1 so that maximum CC pot setting gives 0.55V above the output voltage going to U1 pin 3.

Once all load is removed, CC should disengage and allow the output to rise up again.

[m3vuv]

is that u1 a or b?

[xavier60]

is that u1 a or b?

U1A is the CC opamp
U1B is the CV opamp.
You have not responded to my earler questions.

[m3vuv]

im using a 5w 12v bulb to load it,as soon as i remove the load voltage on the output recovers,havent measured u1a or b yet,its late,will look tomorow.

[magic]

So it's working as intended, but the CC setpoint is too low. Likely a matter of VR1 tweak or some resistors are wrong.

U1 is one chip and pin numbers on the schematic correspond to those in LM358 datasheet .

[m3vuv]

one thing i do remember is when i was messing with it was the tab of the 7912 shorted to the tab of a 7812 as there back to back,that popped but i renewed it,the 7812 seems ok tho.

[m3vuv]

Well not tested yet as have to put it back together,i checked resistors around the lm7912 that shorted,i found R5 33k to be open,looks like this is one of the feeds too pin 2 of u1A,fingers crossed.

[m3vuv]

i find these new metal film resistors just go open at any sign of abuse,not like the carbon resistors that could glow red hot and still sort of work!.

[m3vuv]

Well i dont know if i mentioned it but i found r5 open,replaced that also added more  pass transistors and current sharing resistors x2 so now 6x pass transistors,and it worked,so started testing into a resistive heater element as a load,had ,i also was running the heatsink fan from the output of the lm7812 vreg,it was going well for as i wound up the  output up from zero on max amps ,it got to 10 amps think it was reading 25v,then the fan slowed down and output vanished!,i found r5 open again so replaced it with a 2 watt item,now all transistors ie pass transistors and the tip41 driver test good but still no output,fan is normal tho and voltmeter works,any ideas?,seems i am chasing my tail with this! lol TIA.

[m3vuv]

forgot to say i renewed all of the opamps too.

[xavier60]

It is impossible to electrically damage  R5 33k with the available voltage.
Is the +12V control rail ok?
If so, measure the CC and CV opamp outputs WRT the + output terminal, unloaded then loaded.
Also say what the output voltage is for each test condition.

[magic]

Check if C4 isn't shorted and if the resistor you are replacing isn't R6 or some other one.

[m3vuv]

Well,the 12v power rail from the 7812 is ok,c4 is not shorted and it was r5 i renewed to get it working after the first blow up,to test the output ot the cc/cv opamp,against the + output,do i need my red+ve meter lead in the +ve output jack or black-ve?,all diodes test ok as well in zeners ,but havent tested zener voltage,just normal voltage drop in foward direction,they drop .7v others 0.6..

[xavier60]

Have the black meter probe on the + output terminal which is the same as the 0V for the control rails.

[m3vuv]

Well,dont know if this is good or bad news! lol,i was removing the 7912 with a hotair smd heat gun and when i pulled it,one leg stayed in the pcb,i had been bending it a while ago when fitting the added pass transistors,i assume this will cause the zero output?,i have a lot of resistors removed from the board at the moment that i removed to test as in cuircit they gave strange readings,i was just gonna lift one leg then thought no,i am going to change them for carbon types as they seem more resiliant to abuse!( is that a good idea?),i have checked all resistors associated with the cc/cv opamps and they are all good,also originaly had q7 blowup when i first connected the transfo taps wrong on the first powerup after building it!,i replaced that with a tip122,now thatd testing as a double diode!,may test the 8050 driver too,also is there a simple test for the tl431 out od circuit,i have spres so may just replace it.TIA.

[xavier60]

Well,dont know if this is good or bad news! lol,i was removing the 7912 with a hotair smd heat gun and when i pulled it,one leg stayed in the pcb,i had been bending it a while ago when fitting the added pass transistors,i assume this will cause the zero output?,i have a lot of resistors removed from the board at the moment that i removed to test as in cuircit they gave strange readings,i was just gonna lift one leg then thought no,i am going to change them for carbon types as they seem more resiliant to abuse!( is that a good idea?),i have checked all resistors associated with the cc/cv opamps and they are all good,also originaly had q7 blowup when i first connected the transfo taps wrong on the first powerup after building it!,i replaced that with a tip122,now thatd testing as a double diode!,may test the 8050 driver too,also is there a simple test for the tl431 out od circuit,i have spres so may just replace it.TIA.

The broken pin on the 7912 could have been causing a problem, only a voltage measurement would have made it more certain and might explain the intermittent nature of the fault.
Replacing the resistors isn't necessary at all.
I'm having difficulty following circuit references. Which schematic are you using?
Like a lot of things, the proper function of the TL431 is best done with voltage measurements. Just check that it's shunt regulating 2.5V first. If it is, then move on. If it isn't, then investigate further.

[m3vuv]

Hi the schematic and info i am using is on the link here :https://gswagner.com/35%20V%205%20A%20Supply/35%20V%205%20A%20Power%20Supply%20Schematic%201%20of%203.pdf
too late on the voltage ref tl 432?,ive replaced it already,will post when ist all back together.

[xavier60]

Hi the schematic and info i am using is on the link here :https://gswagner.com/35%20V%205%20A%20Supply/35%20V%205%20A%20Power%20Supply%20Schematic%201%20of%203.pdf
too late on the voltage ref tl 432?,ive replaced it already,will post when ist all back together.

I couldn't get that link to work, I assume it's this one, "ledtester" posted earlier.
https://gswagner.com/35%20V%205%20A%20Supply/
"Vpwr Control" and "Apwr Control" are interchanged between sheets 1 and 2.
R5 circuit position is also wrong. It should connect to the RH end of R14. R5 is not important to PSU operation.
R6 and C4 as shown, do nothing useful.

[magic]

So halfway in the thread you started using a different schematic with different component numbers without telling anyone anything?

Isn't that guy's schematic wrong? You said on the first page that your board agrees 100% with the schematic and BOM from the auction.

[m3vuv]

magic,my board matches the schematic i got from the seller(sort of - the lm7912+7812 vregs,same with the bom,i am now using the one on wagners site as that seems more complete,hope that clears it up.

[m3vuv]

magic,my board matches the schematic i got from the seller(sort of - the lm7912+7812 vregs,same with the bom,i am now using the one on wagners site as that seems more complete,hope that clears it up.

[magic]

Wagner shows a pair of LM7812/LM7912 being used to power U1 and the AliExpress schematic shows 680/1W resistors and 15V/6.2V zeners instead. The AliExpress board photo somebody found also has space for those resistors and zeners, but not for regulators (except those 7812/7805 on page 3 of Wagner).

And there are those weird differences in resistors that xavier60 pointed out.

Obviously, both schematics can't be true at the same time

Given that you talk about 7912 regulators and I can't see any of that on the AE schematic, I suppose that your board version is the same / similar to Wagner's?
Are the CC/CV pots also connected to the board with three wires instead of two?
Are those resistors R5, R6 and cap C4 exactly as drawn by Wagner?

[m3vuv]

my pots connect with 3 wires not 2,r5+6 also c4 are as per wagners scematic om my board.

[m3vuv]

are there higher power pass transistors with the same tab mountings i could use in this that could be driven to 30+amps,without killing the opamps or driver ?,several of them in parralell of course.

[magic]

Okay, so what's the current state of this PSU?

There are several regulators (I use Wagner's numbers now).
U6 7812, U7 7805 that power some opamps, relays, fan etc. AFAIK there is absolutely no problems with this part of the circuit, no magic smoke, no nothing, correct?
U3 7812 which powers the U1 CC/CV opamp. This regulator is still OK too?
U4 7912 which powers U1 - this blew up once or twice already?

And the failure is always the same: blown 7912, blown U1 and blown R5 between these two ICs? I'm not sure what could be causing it, but at least it makes some sense because these components are connected together.

What sort of power supply are you using for U3,U4? It's supposed to be some low voltage transformer, separate from the main transformer.
What's the voltage on Vcont+ and Vcont- and on the inputs of U3,U4? (Black probe on Vcont each time).


There are two things I would change right off the bat:
1. Most negative regulators require output capacitors for correct operation. The big LM79 regulators in TO220 are supposed to work with at least 25µF aluminium electrolytic per National datasheet. So I would bodge some 33µF or more into the circuit.

2. LM358 really doesn't like its inputs pulled below the negative supply. I would add a diode from Vcont to U1 pin 5. I posted a pic with this mod on the first page. It's nice that they corrected pot wiring on this new board revision, but protection diode is still missing.

In general, I can't see any sane reason for U1 to blow up here, except for these two problems above: misbehaving supply regulator or input voltage out of range.

Output base current is provided by R4 from U3 (a few mA) and U1 only removes some of that current before it reaches Q1 in order to throttle down the output. It really has no good reason to blow up.

[m3vuv]

would 2x buw49 transitors work ok as the pass elements in this?,s look beefy items 30a a pop.?

[m3vuv]

would r4 deliver enough current to drive the driver tip 42c enough to drive all the current pass transistors (6x 2sd1047) to full power then or need to be a lower value or higher wattage?

[m3vuv]

So anyone should r4 be changed for a bigger wattage to drive the extra pass transistors,be good to know before i put it back together.Tia

[magic]

Dunno, it all depends on output current and β of the transistors. Q1 has to provide base current to Q2,Q3 and R4 needs to provide enough base current to Q1. It should be simple calculation using the datasheets.

[m3vuv]

Well having another go at this,found out i had the 7812 and 7912 regs mixed up,now output goes 21v max but sinks to about 8v halfway tru the volts pot halfway point then rises againe,cc light is lit all the time,getting volts out but no current,u1a output is vpot fully ccw=10.5v,turns negative halfway thru its travel then rises up agine,u1b  sits at -10.9 and doesent change,any help welcome! TIA.

[m3vuv]

forgot to mention i replaced both regs with new ones.

[xavier60]

Well having another go at this,found out i had the 7812 and 7912 regs mixed up,now output goes 21v max but sinks to about 8v halfway tru the volts pot halfway point then rises againe,cc light is lit all the time,getting volts out but no current,u1a output is vpot fully ccw=10.5v,turns negative halfway thru its travel then rises up agine,u1b  sits at -10.9 and doesent change,any help welcome! TIA.

Just considering the CC LED being always on, it shouldn't when the CC pot is at a high setting with no PSU load.
The input difference should such that the opamp's output is fully high, close to 12V.
I assume that the control rail voltages have been  checked and are correct?

Edit: Im referring to the CC opamp's inputs and output.

[m3vuv]

well just checked voltage over pins 1 and 8 on both opamps,u1=24v u2=11.9v.

[xavier60]

well just checked voltage over pins 1 and 8 on both opamps,u1=24v u2=11.9v.

Best to refer to them as CV and CC opamps. Are you measuring WRT to the PSU's positive output terminal?

[m3vuv]

was measureing over pins 4+ 8 on both
Heres some more readings both pots fully cw ie full volts+ current opamp power rail=24v,u1 a out=1.1v u1b out= -22.8v
u2 power rail=11.9v out a=0v out b=0v
this is mesuered with ref to both opamp pin 4 ground pins.

[xavier60]

That's difficult to follow. Take all measurement With Respect To the PSU's positive terminal.
But there seems to be something wrong with the CC opamp. Its output should be close to 12V.
Has the opamp been replaced recently? What voltages are on the CC opamp inputs.

[m3vuv]

ok so are you saying +meter lead to + psu output and neg lead on opamp outputs?,also where do the pots need to be?,i swapped one opamp as it was getting hot,cant recal what one tho i think it was u2,cant be sure the replacement is ok tho.

[xavier60]

ok so are you saying +meter lead to + psu output and neg lead on opamp outputs?,also where do the pots need to be?,i swapped one opamp as it was getting hot,cant recal what one tho i think it was u2,cant be sure the replacement is ok tho.

Black meter lead to the PSU's positive terminal. Take measurements with the red, noting the sign if minus. Both pots at half way.
The CC loop can be disabled by removing the LED to check if the CV loop is ok, but you must be careful to not short circuit the output.

[m3vuv]

will do it later as its 3.50am here.

[m3vuv]

Ok this is what i get(black meter lead to psu +ve terminal),opamp 1A ie pin1 measures +5.93volts
1B ie pin 7 measures 10.9v ,that is with both pots at halfway point.

[xavier60]

Ok this is what i get(black meter lead to psu +ve terminal),opamp 1A ie pin1 measures +5.93volts
1B ie pin 7 measures 10.9v ,that is with both pots at halfway point.

That indicates multiple faults. The opamp with the lowest output voltage has control of the output stage which is opamp 1A which should be the CC opamp. And the +5.93V is abnormally high.
Tell us what happens to the output voltage with a small load, 100mA or so,
Then measure voltages starting at the anode of D17 and every point along the drive path.
 

[m3vuv]

is that with meter -ve lead on the psu+ve output againe ?

[xavier60]

is that with meter -ve lead on the psu+ve output againe ?

Yes, always. You can measure the PSU output voltage with the red lead probing the negative output terminal. If you don't have a 2nd meter. 

[m3vuv]

I have a few avo mk8 meters but they take up too much bench space so use my fluke 83,heres some readings D17 anode =11.36v,D17 cathode=11.83v r4=12.14v, one leg,other leg is 11.83v,7812 out is 11.76v.,thats with a load on the psu output,black meter lead to psu +ve out,one thing that i notice is with the load connected the led lights,goes out with no load.

[xavier60]

I have a few avo mk8 meters but they take up too much bench space so use my fluke 83,heres some readings D17 anode =11.36v,D17 cathode=11.83v r4=12.14v, one leg,other leg is 11.83v,7812 out is 11.76v.,thats with a load on the psu output,black meter lead to psu +ve out,one thing that i notice is with the load connected the led lights,goes out with no load.

There is too much information missing.

I need the voltages at every point along the drive path on the way to the output.

[m3vuv]

what info do you need?,thats all points on the path from d17 to the 7812 output,what else do i measure?

[m3vuv]

just tried measuring the psu output with the load etc,no load =15v,loaded 0v,same with led disconected is the same.ie goes to zero with a load,thats the psu output.

[xavier60]

Starting at the Base of Q1, work your way out to the output, measuring every point.
After the Base of Q1, the next point will be the Emitter of Q1.

While it's loaded.

[m3vuv]

did i say the led goes off with a load,its on without,sort of working backwards.

[magic]

https://www.eevblog.com/2014/04/06/eevblog-600-opamps-explained/

[m3vuv]

base of q1 is 0v,emitter is-11.93v

[m3vuv]

ive tested d5 thats ok,does the output of u1a feed q1 via c3?,i see 7.7v on one side of c3,and zero on the other leg,could it be open?,but cant see how that can turn on q1 unless its a sort of pulsed dc signal,ie pwm?

[xavier60]

base of q1 is 0v,emitter is-11.93v

The Base of Q1 should be much higher based on the previous reading at the anode of D17.
And I have no idea yet as to how there can be -11.9V at the Emitter of Q1, the TIP41.
But for what ever reason you are reading those voltages, Q1's B-E junction must be open circuited.

[m3vuv]

it looks like the led is wired in reverse,would that cause the issues?

[xavier60]

it looks like the led is wired in reverse,would that cause the issues?

Yes, that would explain a lot of the problems.

[xavier60]

When the LED was removed earlier, there should have been normal voltage regulation if there were no other problems.

[m3vuv]

could the regulators that i fitted wrong killed q1?,i now get volts on the base that tracks with the volts pot but no amps out.

[m3vuv]

would the fact,i amm trying to drive 6x pass transistors be the issue?

[xavier60]

would the fact,i amm trying to drive 6x pass transistors be the issue?

That would be just guessing. The recent readings dont make much sense at all.
With the LED removed and a small load, take voltage readings starting at the anode of D17 and at each point along the drive path ending at the output.

[m3vuv]

Well i have just ordered some new opamps,so this can go on the back burner untill they arrive,thanks for the help so far.

[xavier60]

Those voltage readings would have helped me to make sense of it.

[xavier60]

There is a test for the output stage that can be done while the opamp is removed.
The output should go up to the  full unregulated voltage, then test to see how it holds up with a load.
But of coarse, dont short the output.

[m3vuv]

Well this is on the tuit pile,I have a new kit the same that i am just building,i have a few mrb  5200 diodes,would i be better using one of these in place of D5,that is a 1N5408 ? ,its a  sschotty diode so .3v drop the original is  about .6v TIA

[daisizhou]

Generally we do not use this power supply.Sellers only sell open source solutions.Many electronic components are cheap.There are important parts such as imitation or inferior mos tubes.
Most items sold on AliExpress are imperfect, or come from informal manufacturers. If there is a need for quality, it is recommended not to buy.

[m3vuv]

Generally we do not use this power supply.Sellers only sell open source solutions.Many electronic components are cheap.There are important parts such as imitation or inferior mos tubes.
Most items sold on AliExpress are imperfect, or come from informal manufacturers. If there is a need for quality, it is recommended not to buy.

and what has that got to do what diode to select what i asked?,maybe there is a hidden answer in there somewhere,do i need a crytstal ball to view it??.

[MK14]

Generally we do not use this power supply.Sellers only sell open source solutions.Many electronic components are cheap.There are important parts such as imitation or inferior mos tubes.
Most items sold on AliExpress are imperfect, or come from informal manufacturers. If there is a need for quality, it is recommended not to buy.

That makes a lot of sense. The thing is, some people see something that looks shiny and cheap, and they just can't resist going for it, time and time again (it seems). Despite good advice from people like you.

I suspect there are a number of proper/decent power supply schematics, which would be considerably better. But it is not necessarily obvious, which those are. As there are plenty of bad/problematic ones, available on the internet.

Anyway, your advise has helped me, be even more sure, that these cheap Chinese (not so cheap these days) kits, are not as wonderful, as they might first appear.

[m3vuv]

So for the 3rd time do change the silicon diode for a Schottky Diode ?

[m3vuv]

with most aliexpress kits i normaly replace the components with what i have in my junkbox and use the pcb ,knobs etc

[xavier60]

There is no advantage in changing D5 from an ordinary silicon type.

[magic]

It's specified as 1N4007, not 1N5408, and there is no need for a high current diode there. And the substitute part number that you posted doesn't exist.

The circuit doesn't look too bad. I like the floating power supply to the regulator - it makes the regulator operate at constant (unity) gain, breaks any cap on output voltage due to opamp supply limits and renders some types of faults completely impossible as there is no DC connection between the two power supplies. The few missing protection elements could be hacked into it.

[ledtester]

fwiw, Steve has a new video on his PSU build...

This one focuses on the internal placement of components within the enclosure he's using.

https://youtu.be/nkn7lQgIrfM

[m3vuv]

It's specified as 1N4007, not 1N5408, and there is no need for a high current diode there. And the substitute part number that you posted doesn't exist.

The circuit doesn't look too bad. I like the floating power supply to the regulator - it makes the regulator operate at constant (unity) gain, breaks any cap on output voltage due to opamp supply limits and renders some types of faults completely impossible as there is no DC connection between the two power supplies. The few missing protection elements could be hacked into it.

you dont seem to know what i am talking about,i have the pcb and parts infront of me,you are 100% WRONG!!!!
what are these i brought then? https://www.ebay.co.uk/itm/134027369085?_trkparms=ispr%3D1&hash=item1f34a75a7d:g:hv8AAOSwPzhaITBU&amdata=enc%3AAQAGAAACkPYe5NmHp%252B2JMhMi7yxGiTJkPrKr5t53CooMSQt2orsSHYXPhGXR5uguexJBeHwfgdfqzbAWa0SaMMAeG5vPfVxF5kfWaZs%252FgMGAq%252Fw94q8f0M1f0UBg%252B0ytSjPJWOEgbfo66W05WVyF7M0W50luZiVXNH4bmoXr8m97sYs0HKWcPLwm5lzevw6GOG8FoidqaO3iBFI7dxa3LaOIcGmfgbsPZgMPXk7Qy4cDhTqYds92%252Bz60PptwjDfTsReTVQzaWe9LvjLLtGM78LX2plySdU%252BbWpitX8t7dqjlnYk459WGJf0DR3z6ZeiPdn8T1xrm6AIyxbZW5ra7YxxzE%252BILU5pHPeJHeBfGOWFRDXIuHwG4TduY64UGM1vBm9N57HONvQe4MyaiqboUXn%252B5n9BonJqbQmIxeNRkZMkx%252F%252BDri2IcdcNh1dZaRt1wolglYVSsgkP%252BqKhzPoW4mpkgmLZyL8PkBtCZojWBi6J7CMScIck8MnSU7bs8WEJpwO%252FbQlgoMecHr270hU24Qk2y3R48rD4AnsE2IDCNaXkKXKYxF1%252F2nRES9GG6lxfesY0HHl0ROAmYsvZB7dsVNnUcWcubOX%252FCE5fv%252FcxOfm%252BSiK8QVy8cLRUP7n40kpB9wmrEEDYLD8c4Hahplju9AmP9mjSOS3rS0P%252B83qlsHz%252FmnEWNNax6EK8O%252BNbUoyGryFdpVF6zhLK%252BihvJ2tMN9iJrBJCv68dggWvvd%252BeTNGbjOP1NY%252FZYzuJFkuA1J3q09ODiQUGx5e2YyOhEc87x%252FQGMBDLUruBBzlkNO7w7YSBXVhQn6QN5cDKnYx88urwlVH60YRHke8HcATOb7nWur0nF8WBoVrlXkJJ9qABcKF6xKd2Wv%252Fh2%7Cclp%3A2334524%7Ctkp%3ABFBM2oyZ2PFf

[magic]

Then nobody here knows what D5 is because we don't have the same PCB in front of us and you have used component numbers from at least two different schematics so far.

The diodes you linked are of course either SR5200, SB5200 or MBR5200 and maybe it's all the same, but they aren't MRB5200. I supose that was a typo, fair enough.

[m3vuv]

sorry it was d6 i made a typo.

[m3vuv]

Those diodes were ordered a few months ago to repair an smps in a 22" phillips flat screen monitor.

[magic]

I suppose it could work as D6, that diode is just sitting there under reverse bias.
The PSU would even work without it, but it protects it from externally applied negative voltage which could damage things.

Beware that schottkys leak more than silicon diodes, check the datasheet "reverse characteristic". Leakage may even be a few mA for some large diodes close to their peak reverse rating.

[m3vuv]

I have 5 of these,would 5 of then be ok for 20a at say 5v?,or do i keep the 2s1047 trannys fitted at the moment( originals were tip 3055 types.

[m3vuv]

LINK DIDNT WORK,heres the data sheet,am i correct in thinking they are no good in non switching applications?

[magic]

ISC / InChange Semiconductor and SavantIC transistors are made of 100% pure Chinesium semiconductor material.
Rumor has it that even their replacements of "linear" transistors have SOA inferior to the originals (and SOA is rarely specified in their datasheets even if it should).

It's probably crap. It may work with low input-output differential.

[xavier60]

The TIP35C BJTs I used in my bench PSU came from this Aliexpress seller, https://www.aliexpress.com/item/32539913063.html?spm=a2g0o.order_list.0.0.21ef18022M3Ta4

[m3vuv]

Morning all,i have just got this working,with one issue,it puts out 6a max,i assume thats down to the current sense resistor value,would that be R19? TIA.

[xavier60]

Morning all,i have just got this working,with one issue,it puts out 6a max,i assume thats down to the current sense resistor value,would that be R19? TIA.

The Emitter resistors summed together are used foe current sense.
It it's going into CC mode, the LED should light up.

[m3vuv]

So should i double r15+r 16 then for double the output?

[xavier60]

So should i double r15+r 16 then for double the output?

Altering R15 and R16 wont help. Is it going into CC mode? Or is it a drive problem?

[m3vuv]

seems a drive issue as the cc led doesent illuminate.

[xavier60]

seems a drive issue as the cc led doesent illuminate.

With it still loaded to 6A, confirm that the  opamp outputs are both close to 12v.
Starting at the anode of D5, measure the voltages at every point along the drive path.

[xavier60]

If the voltage measurements dont show any obvious fault, try reducing R4 by about half.
Q1's current gain might be too low.

[m3vuv]

Well at the moment,it has 4x buw49 pass transistors ,the 2sd1047 trannys(aliexpress junk!) went sc at 21v 5a output,at the moment it will output 36v max at 3a,i havent reduced r4 yet for more drive,i notice the 7812 must be about 80 deg c,it gets to hot to hold your finger on!,i guess if i reduce r4 that will put it under more load and kill it?.

[xavier60]

Well at the moment,it has 4x buw49 pass transistors ,the 2sd1047 trannys(aliexpress junk!) went sc at 21v 5a output,at the moment it will output 36v max at 3a,i havent reduced r4 yet for more drive,i notice the 7812 must be about 80 deg c,it gets to hot to hold your finger on!,i guess if i reduce r4 that will put it under more load and kill it?.

The few milliamps through R4 isn't a great load for the 7812. What is the input voltage and is it powering anything other than shown in the schematic?

[m3vuv]

The input to the 7812 is 22v output is 12.1v,there is nothing connected on it apart from whats on the schematic.

[xavier60]

There should be about 10mA load, very little. Is the opamp getting hot?

[m3vuv]

No both opamps are cool and seem to be working ok,everything seems fine apart from the lack of output current,seems to max out at about 2a, i am using a 1 ohm 100 watt resistor,am i correct in thinking 5v into that should draw 5a?.,as i said i get about 2 max,no mater where the current pot is.

[m3vuv]

I just unplugged the led,if i wind the voltage up i get about 5a at about 11.5v,couldnt go any more as wires got hot on my leads and grabbers!

[xavier60]

No both opamps are cool and seem to be working ok,everything seems fine apart from the lack of output current,seems to max out at about 2a, i am using a 1 ohm 100 watt resistor,am i correct in thinking 5v into that should draw 5a?.,as i said i get about 2 max,no mater where the current pot is.

Yes, 5V across 1 ohm will cause 5A of current.

[m3vuv]

I have some 1 watt 1k2 resistors,would it be safe to try one of those in place of r4,or is that pushing it?

[xavier60]

I just unplugged the led,if i wind the voltage up i get about 5a at about 11.5v,couldnt go any more as wires got hot on my leads and grabbers!

If disconnecting the LED causes increased current output, it indicates that there is a problem with the CC control side.
Is the LED faulty in some way? It's important that the LED functions a a good diode for the PSU to work properly.

[m3vuv]

well it ligts up! lol,has a 2.122v drop using the diode test on my fluke,its a red one from my junkbox.

[xavier60]

I have some 1 watt 1k2 resistors,would it be safe to try one of those in place of r4,or is that pushing it?

Yes, try it. Although it's the wrong fix. You either still have a fault in the control circuit or have transistors with unusually low current gain or maybe a faulty one hogging drive current. Or a wiring mistake. For example, a bad collector connection to one of the output transistors.
Test carefully  with the CC disabled by leaving out the LED.

[m3vuv]

Well,i just tested the led that came with the kit,its way brighter and has a vf of 1.97v,could that be the issue?,everything else looks good.

[xavier60]

Also, Oscillations can be the reason for odd PSU behavior. Best checked with an oscilloscope.

[xavier60]

Well,i just tested the led that came with the kit,its way brighter and has a vf of 1.97v,could that be the issue?,everything else looks good.

That wont cause a problem. Carefully test the PSU's output with the LED removed for now.

[m3vuv]

Hi,your answer isnt clear,how exactly do you want me to test it?

[xavier60]

confirm that the voltage regulation is ok then see how much current it can safely supply before the voltage drops.

[m3vuv]

how can the voltage drop without the led connected?,i dont want to try it without,the transfo is huge,testing max without any regulation will destroy things,the transfo can deliver 100 amps +.

[xavier60]

With the LED missing, there will be no current limiting. It will be safe to test the PSU so long as it's not overloaded.

[m3vuv]

thats my pont,sounds iffy to load it untill the v out sags,that will be a lot of amps without the led,what if i stop at say 10 amps?

[xavier60]

thats my pont,sounds iffy to load it untill the v out sags,that will be a lot of amps without the led,what if i stop at say 10 amps?

If it has been designed to supply up tp 10A then it should be safe regardless of the current limiting not working.

[m3vuv]

well it will supply 20a ok before the voltage drops,its od tho if i connect the load of 1 ohm then wind the  volts up it gets to 13v and tops out,if i wind the voltage upto 20v first then connect the load it sits there happily untilthings get hot then fan starts,10a is the most i can measure on my fluke or avo 8,thats without the led.

[m3vuv]

Well just an update,i changed the led for a decent one,adjussting vr1 gives me 12amp before current limiting,i will leave it at that,seems better than the stock 5a,goes up to 36v too,i class it as a good result,thanks for everyones help.

[xavier60]

That sort of odd behavior can be caused by loop instability resulting in oscillations.

[xavier60]

Well just an update,i changed the led for a decent one,adjussting vr1 gives me 12amp before current limiting,i will leave it at that,seems better than the stock 5a,goes up to 36v too,i class it as a good result,thanks for everyones help.

What actually fixed it though? Is the 7812 over heating? Is there a separate 7812 for the tap changer?

[m3vuv]

i'm not sure,the vreg seems not to be getting too hot,the tap switching is stock.

[m3vuv]

in a nutshell i used four BUW49 transistors as the pass elements,fitted a 50amp bridge offboard,a 22000 uf offboard filter cap and changed r4 too a 1 watt item.the pass transistors are on an old amd pc heatsink with an old ps2 fan.Just need to fit some vents on the case now,its an old pc speaker sub case.

[xavier60]

I wouldn't be surprised if it has some instability.  R6 and C4 are positioned in a way that does nothing useful. Series RC are usually for frequency compensation, so should connect back to the CV opamp's input.
The position of R5 makes R33 totally redundant as a bleeder for the CC pot's wiper. The top of R5 should connect to the bottom of the CC pot.

[m3vuv]

one thing i notice is when the led illuminates i can stil increase the output current using the volts pot!,is that normal?

[xavier60]

one thing i notice is when the led illuminates i can stil increase the output current using the volts pot!,is that normal?

Not normal. Check the stability. Or there could be some sort of interaction between the 2 pots. Check the 2.5V reference.

[m3vuv]

how do i check that then?

[m3vuv]

I assume thats the tl-431?

[xavier60]

I assume thats the tl-431?

That's it.

[m3vuv]

When two of the 2sd 1047 trannys going sc killed the tl430 as well?

[xavier60]

When two of the 2sd 1047 trannys going sc killed the tl430 as well?

No. What does the reference  voltage measure?

[m3vuv]

I dont know yet as i am getting ready for a hospital visit,at least as it is i can set the current output using the led,i dont see that as an issue,ie set the voltage then adjust the current pot to whats needed?.

[xavier60]

When two of the 2sd 1047 trannys going sc killed the tl430 as well?

On second thoughts. Any fault that causes the PSU's output to go to max voltage can cause a strong reverse bias to be applied to the TL431 when the voltage pot is set to minimum. The TL431 needs a protection diode across it.

[m3vuv]

Well,i have this fitted into the case finaly,i am just charging my clansman battery ,feeding it 28v at 400ma,its working perfect,i must of been having a brain fart before! lol.

[m3vuv]

would it be possible to fit a cv led to this,if so where to connect,is it same place as the otherone just polarity reversed?

[xavier60]

D17 could be replaced with an LED but the current might be too high for it. A better way is to put an LED and resistor in series between the 12V control rail and the output of the CV opamp. Start with 10K or a bit lower to see how the LED brightness is.

[m3vuv]

is that u1a pin1 or 7? and the amp control rail or volts control rail?,thanks.

[xavier60]

is that u1a pin1 or 7? and the amp control rail or volts control rail?,thanks.

If you want the LED to indicate CV status, it needs to connect to the CV opamp U1B output pin 7.
The power source needs to be 12V from the 7812.
What did the reference voltage measure??

[m3vuv]

what ref voltage are you talking about?,i havent checked any,is that not how the led is alread conected?

[xavier60]

what ref voltage are you talking about?,i havent checked any,is that not how the led is alread conected?

Im still curious about the 2.5V reference I asked you to measure yesterday. In case the TL431 really has been damaged.

[m3vuv]

didnt check it as the psu was working ok,well up to 10a anyway,i am more concerned over the cv connections,although thats icing on the cake imho.

[xavier60]

didnt check it as the psu was working ok,well up to 10a anyway,i am more concerned over the cv connections,although thats icing on the cake imho.

I'll standby until you get a chance to check it.

[m3vuv]

whats the point in risking shorting pins probing to check the ref voltage when the psu is working fine,i just want to know where to connect a cv led.

[ebastler]

Wow, five weeks of tinkering and nine pages of forum dialog on a $20 power supply kit. I guess we can't complain about the value for money one gets out of this kit... 

[m3vuv]

at least it gave you some reading!

[MK14]

Wow, five weeks of tinkering and nine pages of forum dialog on a $20 power supply kit. I guess we can't complain about the value for money one gets out of this kit... 

I suppose, it shows that there are many different types of people, in this world.

A reasonably decent kit, with PCB and decent enough supplied components, should be able to be constructed, reasonably quickly and tend to work, first time.

They (presumably) are either buying sub-standard quality kits and/or need to improve their construction techniques, so it doesn't short out and blow components up, every 5 minutes and/or they just can't listen to good advice, and so end up with something so badly bodged together, that it keeps on failing.

[MK14]

at least it gave you some reading!

You have got to be joking. The gist of the way it looks to me, is that you keep on saying you swapped this component for another component. This kit for another kit. This component or that component smoked, burnt out, got too hot, etc.
Making it extremely difficult or not possible, to follow/help or get anything sorted out.

It really needs you to supply a schematic, and keep that schematic updated with any changes you make.

[m3vuv]

well mines been a bit of a battle mostly due to doubling the output,was a good learning curve tho,at the end of the day i have a 0-36v 10amp+psu with cc control for about 30 quid!,whats not to like?.

[m3vuv]

at least it gave you some reading!

You have got to be joking. The gist of the way it looks to me, is that you keep on saying you swapped this component for another component. This kit for another kit. This component or that component smoked, burnt out, got too hot, etc.
Making it extremely difficult or not possible, to follow/help or get anything sorted out.

It really needs you to supply a schematic, and keep that schematic updated with any changes you make.

if you look the schematic + links have been posted.

[MK14]

if you look the schematic + links have been posted.

But the problem is, I keep on seeing posts, such as this one:

in a nutshell i used four BUW49 transistors as the pass elements,fitted a 50amp bridge offboard,a 22000 uf offboard filter cap and changed r4 too a 1 watt item.the pass transistors are on an old amd pc heatsink with an old ps2 fan.Just need to fit some vents on the case now,its an old pc speaker sub case.

Which keep on changing components, component values and possibly circuit configuration.

Leading to confusion, misunderstandings and difficulty with following what exactly is going on.

[m3vuv]

So am i correct in thinking the cv led goes to the lm7812 output via a dropper resistor and u1 pin7?

[m3vuv]

well i left this running overnight charging a car battery at 14v 1 amp,woke up this morning to find zero output,left it disconnected for an hour and its working againe,could that be the 7812 shutting down thru heat?,if so i plan on building this and fit offboard,is that a good plan?or just fit the existing reg offboard on a bigger heatsink?

[m3vuv]

Seems mr wagner made a ballsup with u3 +u7,he has 2xu3 but no u7. on the board layout etc.

[m3vuv]

seems we have a u6 on the schematic but not on the pcb!

[not1xor1]

well i left this running overnight charging a car battery at 14v 1 amp,woke up this morning to find zero output,left it disconnected for an hour and its working againe,could that be the 7812 shutting down thru heat?,if so i plan on building this and fit offboard,is that a good plan?or just fit the existing reg offboard on a bigger heatsink?

Unless you attached the wrong schematic or are just trolling that circuit will never work because you will never get 14V and 1A out of 2 78L12 ICs.

[xavier60]

well i left this running overnight charging a car battery at 14v 1 amp,woke up this morning to find zero output,left it disconnected for an hour and its working againe,could that be the 7812 shutting down thru heat?,if so i plan on building this and fit offboard,is that a good plan?or just fit the existing reg offboard on a bigger heatsink?

Unless you attached the wrong schematic or are just trolling that circuit will never work because you will never get 14V and 1A out of 2 78L12 ICs.

"m3vuv" is suspecting that the LM7812 supplying the 12V control rail is shutting down.

[m3vuv]

well i fitted a bigger heatsink to the 12v reg but after steve wagner corrected his pdf docs i realised i had fitted it to the wrong 12v reg,gonna redo it now,may fit a usb charger port hung off of the 5v reg too.

[m3vuv]

i see the shutdown as a heat issue as the output reterns when its cooled,my board is different to the last as in u3 is hard up against the tl431 voltage ref so dare say heat could cause issues there,going to mount the 12v u3 offboard on a heat sink with caps on the in+ out and try that.

[m3vuv]

hope the corrected pdf files help if anyone is planning on building one of these.

[Terry Bites]

Will this saga ever end?

[Kleinstein]

hope the corrected pdf files help if anyone is planning on building one of these.

With the circuit divided in 3 files the plan is a bit hard to read.
The last version still has the voltage regulation messed up somewhat - the divider for the main feedback is somehow missing. The lables between the files also don't match.

[m3vuv]

Hopefully this should fix it,as it is it works flawless upto 36v and 10 amp output,it will do more but melts my test leads and my avo8 and flule 83v breakout into a sweat over 10 amps,its just the no output when its been on a few hours and is hot,i hope the big heatsink on u3 fixes it,at least it wont be cooking the tl431 voltage ref ic,fingers crossed,getting short of room in the enclosure now tho with 2x transformers a 22000 uf cap + 50a bridge heat sink!

[m3vuv]

hope the corrected pdf files help if anyone is planning on building one of these.

With the circuit divided in 3 files the plan is a bit hard to read.
The last version still has the voltage regulation messed up somewhat - the divider for the main feedback is somehow missing. The lables between the files also don't match.

Have a look yourself :https://gswagner.com/35%20V%205%20A%20Supply/

[m3vuv]

well 1hr and its looking good at 4v+ 4a so far!

[m3vuv]

neres a couple of pics of it.loaded thru my avo into a 1 ohm resistor.

[m3vuv]

and another pic.

[m3vuv]

a few more to come,hassle with the filesize limit tho as cant be assed to resize them.

[m3vuv]

and another,sideview of the guts of it - the main transfo + filter cap, bridge + 7812 heatsink,they are all in the main case,the aux tranfo+ pcb,pass transistors+ pc heatsink and fan are all on the front panel,it started life as an old pc speaker sub woofer.

[m3vuv]

last one so far untill the next blowup lol.

[m3vuv]

I lied lol,heres a summary of changes to the original pcb etc,firstly the original bridge rectifier removed and replaced with screw terminals to mount a 50a bridge offboard,secondly r4 replaced with a 1 watt resistor,one of the filter caps has had terminals fitted to its legs to allow connection to a 22000uf 68v cap offboard,a bigger heatsink added to u6,the original tip 3055 transiistors replaced with 4x BUW49 transistors with a 30amp max collector current each,along with .22 ohm 5 watt emitter resistors,u3 mounted offboard an a 2x2 inch heatsink,i think thats it,im happy the constant current mode works well and max vout is 36v and tested so far at 10a max as thats the limit of my test gear.Hope this helps others who want to build this/uprate this. 73 de m3vuv.

[m3vuv]

this is the link for the kit i got: https://www.aliexpress.com/item/32986139693.html?spm=a2g0o.cart.0.0.3e143c00nN48NA&mp=1

[not1xor1]

well i left this running overnight charging a car battery at 14v 1 amp,woke up this morning to find zero output,left it disconnected for an hour and its working againe,could that be the 7812 shutting down thru heat?,if so i plan on building this and fit offboard,is that a good plan?or just fit the existing reg offboard on a bigger heatsink?

Unless you attached the wrong schematic or are just trolling that circuit will never work because you will never get 14V and 1A out of 2 78L12 ICs.

"m3vuv" is suspecting that the LM7812 supplying the 12V control rail is shutting down.

and so ... ? Probably those are 2 LM7812 unlike the L type ones of the schematic he posted.
In any case I can't see how he can get 14V with just a single diode in series with the common.
Besides that I do not see any current regulation so where does that 1A come from?

This is the beginner section. Do you want to make beginners waste time and components with such an outrageously wrong schematic     .

[xavier60]

well i left this running overnight charging a car battery at 14v 1 amp,woke up this morning to find zero output,left it disconnected for an hour and its working againe,could that be the 7812 shutting down thru heat?,if so i plan on building this and fit offboard,is that a good plan?or just fit the existing reg offboard on a bigger heatsink?

Unless you attached the wrong schematic or are just trolling that circuit will never work because you will never get 14V and 1A out of 2 78L12 ICs.

"m3vuv" is suspecting that the LM7812 supplying the 12V control rail is shutting down.

and so ... ? Probably those are 2 LM7812 unlike the L type ones of the schematic he posted.
In any case I can't see how he can get 14V with just a single diode in series with the common.
Besides that I do not see any current regulation so where does that 1A come from?

This is the beginner section. Do you want to make beginners waste time and components with such an outrageously wrong schematic     .

The "battery at 14v 1 amp" refers to what the PSU's output was set to, to charge a battery. The circuit with the two LM7812 regulators was intended by "m3vuv" to power the PSU's control circuitry with 12V, nothing to do with 14V.

Also, through the thread I have pointed out the mistakes in the schematic. Which schematic?

[m3vuv]

my good god even i am getting lost now! lol.

[m3vuv]

well i left this running overnight charging a car battery at 14v 1 amp,woke up this morning to find zero output,left it disconnected for an hour and its working againe,could that be the 7812 shutting down thru heat?,if so i plan on building this and fit offboard,is that a good plan?or just fit the existing reg offboard on a bigger heatsink?

Unless you attached the wrong schematic or are just trolling that circuit will never work because you will never get 14V and 1A out of 2 78L12 ICs.

"m3vuv" is suspecting that the LM7812 supplying the 12V control rail is shutting down.

and so ... ? Probably those are 2 LM7812 unlike the L type ones of the schematic he posted.
In any case I can't see how he can get 14V with just a single diode in series with the common.
Besides that I do not see any current regulation so where does that 1A come from?

This is the beginner section. Do you want to make beginners waste time and components with such an outrageously wrong schematic     .

The "battery at 14v 1 amp" refers to the what the PSU's output was set to, to charge a battery. The circuit with the two LM7812 regulators was intended by "m3vuv" to power the PSU's control circuitry with 12V, nothing to do with 14V.

Also, through the thread I have pointed out the mistakes in the schematic. Which schematic?

maybe should understand what i am talking about and not troll!!!!,looks like norx1 whatever his name is, is not quite switched on with this??

[xavier60]

my good god even i am getting lost now! lol.

To clear this up, what was the intention of the circuit that you attached in  Reply #216?

[m3vuv]

was that the one with two 7812 regs?

[xavier60]

was that the one with two 7812 regs?

Yes, that one.

[m3vuv]

That schematic was thoughts on a possible solution to the 7812 shutting down,ie share the load and cut down the heat of the probblem 7812,mounting the original on a heatsink offboard seems to of fixed it tho,unless it was the heat soaking to the 431? ic tho as how they are on the pcb, means the 431 is hard against the 7812 heatsing,just a shitty layout!.

[not1xor1]

That schematic was thoughts on a possible solution to the 7812 shutting down,ie share the load and cut down the heat of the probblem 7812,mounting the original on a heatsink offboard seems to of fixed it tho,unless it was the heat soaking to the 431? ic tho as how they are on the pcb, means the 431 is hard against the 7812 heatsing,just a shitty layout!.

OK... so it is just a matter of misunderstanding... I haven't followed the whole thread and had understood you were charging the battery with that L78L12 circuit. 
(I link the image below)
I apologize.

[m3vuv]

its ok,we are only human,i didnt build the circuit as moving the 7812 offboard fitted on a 2"x2" heatsink seems to of cured it shutting down.time will tell.

[m3vuv]

I have just noticed,if i switch this on with any load connected i get zero output,if i disconnect the load and turn it on then reconnect the load its fine,any ideas whats causing this? TIA.

[xavier60]

Measure the voltage on all pins of U1 with respect to the + output terminal.

[m3vuv]

when do i do that,when it wont output with a load connected or what ?.

[xavier60]

when do i do that,when it wont output with a load connected or what ?.

While it's in the fault state, Watch out for the fault clearing while the pins are being probed.

[m3vuv]

Would it be ok to use the 5v meter supply to power a usb socket for charging phones etc?.,or would it cook?

[m3vuv]

well i fed a 7805 from the onboard 7805 feed,fitted it to a heatsink,working fine charging my vape,happy days.

[m3vuv]

Well,i had a disaster with this,i connected it backwards to a charged 12v car battery( polarity wrong),now no output,any idears what to check in what order? TIA.p.s the 5v usb still works and meter reads ov.

[Gyro]

If you did that, buy another one. 

[MK14]

Well,i had a disaster with this,i connected it backwards to a charged 12v car battery( polarity wrong),now no output,any idears what to check in what order? TIA.p.s the 5v usb still works and meter reads ov.

That's why it is a very good idea to use quality bench power supplies and/or decent quality and/or professional, schematics or better, if you're making your own.
Typically, they will survive such mishaps. As it will have taken such issues into account, when designing the power supply.

EDIT:
I'm assuming the power supply is built to a decent standard, without lots of shortcuts, in its construction, with reasonably decent quality components. Without significant over-loading of those components, etc. Otherwise, the power supply will tend to be something which is very unreliable and may break, even if you only look at it in a funny way.

[m3vuv]

Firstly thanks to mk14 and gyro,that isnt what i am asking tho is it?,buying a new one is the easy option,and i live mostly in a world of skint as off work waiting for an operation!,cheers anyway Paul M3VUV

[MK14]

Firstly thanks to mk14 and gyro,that isnt what i am asking tho is it?,buying a new one is the easy option,and i live mostly in a world of skint as off work waiting for an operation!,cheers anyway Paul M3VUV

If it has got a (normally) reverse biased (power) diode, across (internally) the output terminals. Then that diode, may have gone short circuit, because of the reverse polarity.
If it hasn't, then you might have blown your PCB tracks (Maybe!!!!, lead-acid batteries, can give out huge currents) and/or burnt out a current sensing resistor, along the output section. Which also might mean the output transistors, have (probably) gone short circuit.
But I'm guessing, there are many possibilities.

I'd normally check the schematics, but there are so many different ones in this thread, I've given up on trying to identify which one, is the correct one.

[m3vuv]

So far all the emitter resistors were open,the diode mk14 is talking of was faulty as well,i have replaced those but still no output,all tracks look good but i beefed those up when upping the power output from it!,cheers m3vuv.

[Gyro]

Well if the emitter resistors are fried, then the pass transistors probably are too, and their driver transistor etc (cascade failure). It's basically a process of working backwards from the output until you weed out all the components that were damaged by all that reverse biassed current (hence the comment about buying a new one, it's the more expensive parts that suffer in such situations).

It would be best to check all components without powering up in between to prevent repeat damage, eg. Shorted pass transistors will take out a new driver.


P.S. Don't assume that, if the pass transistors aren't shorted, they are ok. Several hundred amps from a car battery can easily blow the bond wires open. At least do diode tests on their junctions.