official GDIY51PSU Help & Support Thread

[kante1603]

Hi,

the lm350 is speced  at "Operating Junction Temperature Range TJ −40 to +125 °C",that's 257 °F before thermal shut down.I'd recommend the same as silent arts and bruce0, swap it out,it's just a few cents.

Best regards,

Udo.

 

[Tajmoben]

I think you have a bad regulator (or a bad bridge rectifier).  Given that the regulator gets hotter than the good one... it is probably the culprit.

If you don't have a spare... you could always swap the +16 and -16... and see if the problem moves.

Hey Bruce, I swapped the regulators on the +16v and -16v rails and the issue stayed on the +16v rail. I'm guessing I should change the bridge rectifier next? The crazy thing is that the bridge rectifier was one of the first things I replaced when the issue started. I actually put new rectifiers on both the + and - 16v rails. How likely is it that the new rectifier would have the same problem that the other one did?

 

[bruce0]

I swapped the regulators on the +16v and -16v rails and the issue stayed on the +16v rail. I'm guessing I should change the bridge rectifier next? The crazy thing is that the bridge rectifier was one of the first things I replaced when the issue started. I actually put new rectifiers on both the + and - 16v rails. How likely is it that the new rectifier would have the same problem that the other one did?

If you have changed that rectifier then I have to figure that something else is wrong. 

Summarizing where you are (for my own benefit... and to check with you):
1) You are pretty sure that the load is not the problem, because you are running your rack through an ammeter, so you know it is fading when the amperes are well within spec.
2) You moved the regulator and the voltage drop still happens on the new +16V regulator.
3) You had already swapped the bridge rectifier, and your voltmeter test shows it is fine too.
4) You have tested the transformer (more on this later).

So it is probably not the regulator or the rectifier... what else is there:

Looking at the schematic, there is not much else: http://www.silentarts.de/DIY/GDIY51X/GDIY51PSU_rev02-Schematic.pdf
a) Unlikely... It could be some sort of sneak circuit through LED31 (hard to imagine that it could carry the amp! it certainly wouldn't be lit!).
b) Unlikely... There could be some other sneak circuit, drawing power off that regulator (you could put an ohmeter on the output pad to ground, and see if there is low resistance there (you might have to pull the LED or it's resistor out , or the regulator... I don't know how that output works)  Basically you are looking for some sort of circuit to generate load that doesn't flow over to the rack.  Leave the cables plugged in all the way to your ammeter.
c) It could be that there is something flaky about D32, if it leaked it could cause the voltage to fall (1N400x parts don't fail in my experience, but I am sure they can)
d) It could be that somehow one of the "~" connections on that rectifier is grounded somehow (flaw in board, solder bridge) which would simulate the condition I first described.  This could be checked easily because if you put an Ohmeter from each of those connections to ground, neither should be grounded).  There is a lot of ways to ground them, it could be at the screw terminals or a solder bridge on the board or something in the wiring between the transformer and the board.  Note that even if one side is grounded, you would still see the appropriate winding resistance and ac voltage off the output of the transformer might look right depending on your meter.  But the result would be the PSU would be running off a half wave rectifier, and you might see what you are seeing.

If I were to guess, I would guess the problem is d... followed by c, b, and a in that order.
d is easy to check.  Turn it off, and put an ohmeter between the 16v winding and ground, then the other 16v winding and ground.

Lastly... this has gotten "curiouser and curiouser" and I would love to see a picture of the board.  I would be looking for components oriented incorrectly and solder bridges.

bb

ps... I suppose there could be something flaky in the regulator control potentiometer... but seems unlikely.

 

[bruce0]

The voltage on that rail starts to sag under any load, no matter how small. At 25mA it sags by .02v, at 50mA it sags by .04v. When the total load reaches 250mA the voltage starts a slow creep down until it reaches +7v and then it crashes to +2v and stays there until I lighten the load. At 350mA it crashes fast after about 5 seconds.......


... We've ruled out the Toroidal Transformer, ....

Ben:
I went back and looked at your original post.  And I have changed my theory a little:

The fact that the voltage starts to sag ONLY under load probably means that the sneak circuit suggestions (a and b) are probably wrong. 

I still like theory d.  I like it because it is EASY to check, and there are too many ways to cause that short (those little caps can be solder bridged).  But I admit that the fact that the voltage starts to sag under EVEN LIGHT (.25MA) light load makes d less likely.

Theory c however might explain it, the diode could be leaking just a little, and you adjusted the pot to give you 16v and corrected it (making the pot carry more current than it should), and now you are in a temperature and voltage sensitive balancing act that is unstable.  The diode leakage could be temperature sensitive and voltage sensitive, and it might leak a lot less at lower voltages (in a normal diode leakage is logarithmic and you can build a log amp based upon this fact) and very temperature sensitive.  If it leaked less the voltage at the adjustment terminal would fall, and the output would fall and the diode might leak less... in a sort of death spiral.

The pot could be damaged, and could be getting warm and changing resistance.

If this is the case, changing the diode D32 will fix the problem, you won't have the sagging any more...

If you don't have another 1n400x... (you said you were new to this, most of us have these lying around) then you could just pull it.  D32 is a protection diode so that if there is an output short circuit the circuit doesn't discharge through the regulator and fry it.  But as long as you don't short circuit the output you could do this for a test.  (Note you might be risking that LM350T but you know that).


Anyway... if you did have a leaky diode you may find that you have heat damaged the pot (RV31) and have a hard time getting precise voltage adjustment, but whatever adjustment level you do achieve, it should be stable.



Lastly... if none of this helps... then tell me... what do you mean you "ruled out" the toroidal transformer.  How did you do that?

 

[mulletchuck]

I was looking thru this thread again, and i'm wondering where everyone who built a dual PSU got that faceplate that looks like the official faceplate, but with 2 rows for LEDs and Fuses.   

 

[sahib]

Hey Chuck, I would imagine from me. 

On the later versions I placed the fuses internally and only had the LEDs on the front panel. I now have a couple of Gold facias with two rows of LEDs (and the power switch).

For the fuses I have what I called a fuse board (thanks to Jeff for doing the pcb) with five on board fuse holder and LED connections through 2 way molex headers.

 

[mulletchuck]

soooooooooo, what's a guy got to do to get one of those dual-row gold faces and maybe the fuse board?

 

[sahib]

No probs. I'll check out the shipping cost and pm you the total.

 

[rob_gould]

So I've just made a total beginner error powering on the PSU for the first time.

I had left the 240V / 120V selector in the 120V position after testing it to see how it should be soldered, and connected 240V mains to the PSU.

The main fuse blew, but now the neon light in the power switch lights in the '0' position and turns off in the '1' position.  I can't have wired this wrong surely.  It's just a DPDT switch with live on 1 & 1 and neutral on 2 & 2 right?

How likely is it that the transformer has been cooked to death?  Is there any way I can measure it to check?

Bugger!

 

[zayance]

How likely is it that the transformer has been cooked to death?  Is there any way I can measure it to check?

usually they don't like that, for testing just use a continuity test between windings of the primary.
No beep means ------> I'm sorry.

 

[rob_gould]

How likely is it that the transformer has been cooked to death?  Is there any way I can measure it to check?

usually they don't like that, for testing just use a continuity test between windings of the primary.
No beep means ------> I'm sorry.

Thanks Zayance.

I pulled the transformer out and tested the primaries and there's continuity between yellow and white on both primaries.

But, I rewired the primaries into the 240 / 120 switch, double checking that the wiring was correct and the problem is the same.  When the power switch is at '0', it lights dimly.  When the power switch is at '1', the power switch neon turns off.  Weird.

Anyway, I disconnected the primaries and tested again just to ensure that there was no problem internally with the 240 / 120 switch, and the power switch does what it is meant to, so it seems that the issue must be with the transformer.

Oh well...

 

[sahib]

Rob,

Try the switch without the transformer connected to its output.

Also try the transfromer without the voltage select switch. Just wire in series and check it.

 

[[silent:arts]]

Rob,

disconnect the PSU transformer and measure DC resistance between the windings.
forget about the power switch first, might be another problem.

the Rondo-Mueller transformer is rock solid, and in my observation only the 48V secondary dies if it is connected to 240V in 110V configuration.

important note: this is just my observation from customers! never tried this at home!

 

[rob_gould]

Rob,

Try the switch without the transformer connected to its output.

Also try the transfromer without the voltage select switch. Just wire in series and check it.

#1 : I've tried this and the switch works fine.

#2 : I've disconnected the transformer now so I'll try some other stuff before reconnecting like...... this!

Rob,

disconnect the PSU transformer and measure DC resistance between the windings.
forget about the power switch first, might be another problem.

the Rondo-Mueller transformer is rock solid, and in my observation only the 48V secondary dies if it is connected to 240V in 110V configuration.

important note: this is just my observation from customers! never tried this at home!

Both the primaries measure 4.4 Ω

48V black > black = 6.8 Ω

24V red > green = 0.7 Ω

24V blue > white = 0.8 Ω

16V orange > grey = 0.6 Ω

16V brown > purple = 0.6 Ω

 

[[silent:arts]]

looks like a healthy transformer  8)

 

[rob_gould]

Hmm, so the power switch and 240 / 120 switch seem to check out with the transformer disconnected, and the transformer seems to check out when disconnected from the rest of the circuit.

Must be something to do with how I'm wiring the primaries to the 240 / 120 switch, though I have checked and double checked this.

I think I'll give Cemal's advise a try and try powering the transformer directly, bypassing the 240 / 120 switch, just to double check...

Thanks for the input!

Rob

 

[rob_gould]

Hi 51X gurus,

What are the values of the fuses used in the PSU?  Both AC fuse and DC rail fuses?

I want to put in a Rapid Electronics order while back in the UK for a business trip, and didn't note the values down before leaving home.

They're 20mm fuses aren't they?

Thanks,

Rob

 

[sahib]

51x guru? Who? Where?

They are 20mm.

Keep the + - rails to 2A and the + 48 to 150mA to 200mA.

For the mains 3A would be fine.

 

[rob_gould]

OK, so I've finally had a chance to come back to this.

I first tested with both 120V / 240V selector and the power switch out of circuit.  After tweaking the trimmers, all voltages were perfect.  Phew - the transformer is alive.

Now I've just added the power switch back into the circuit.  Is it normal for the power switch to be faintly illuminated in the 'off' position?! My guess is not!

When I turn the power switch on, I see the correct DC voltages, but when I turn the power switch off, I still see small  (6 - 8V) readings on the DC terminals on the PSU board.

I can barely believe I'm having to ask this as I've verified it using the continuity setting on my multimeter about ten times, but this (attached pic) is how the power switch should be wired isn't it?  Looking from the rear of course...

 

[rob_gould]

Anyone?