Help with non-charging capacitor off of a bridge rectifier

[Ian.M]

1N540x diodes are *CHEAP* compared to the time and effort you've put in so far and the value of the boards and risk of damage if you have to keep removing/refitting parts to test them.   I'd 'shotgun' it by replacing *ALL* the diodes in the bridge rectifier, checking each track for pad to pad continuity while they are out.  When bending the leads of the new diodes, support them next to the body with needle nose pliers so you don't micro-crack the epoxy encapsulation which can lead to immediate or early failure.

Its likely that will cure it, but if it doesn't things are going to get *INTERESTING*   

[mapleLC]

I've rebuilt the rectifier with new diodes, and soldered new traces to the little caps under the diodes that seemed suspect.

I get continuity everywhere, a capacitor that charges, but only partial voltage (about -17) on one side of the aisle, and 50+ volts on the other.

At this point, I am wondering if I should rebuild the rectifier off circuit and patch it in, or perhaps introduce a single component rectifier that would do what I need here instead?

The more I mess with the board, the more it deteriorates, so its a risk to try and patch this rectifier again unless there is a component we have not discussed that could be influencing the voltage difference?

I guess it gets interesting?   is not the emo I had hoped for   

[Ian.M]

We are all going to be kicking ourselves for not thinking of *whatever* when you finally find the cause.

Lashing up a clone of the rectifier + reservoir capacitors circuit on stripboard or tag strip then measuring voltages round the clone would be a good sanity check, though I wouldn't actually use the clone to replace the original as a repair.

Its an extreme longshot, but it is remotely possible the fault may be due to poor contact somewhere in the circuit on the AC side of the bridge:  A rectifying junction can form between oxidized copper surfaces. Its known that a copper oxide rectifier can have a Zener-like characteristic with a fairly low reverse breakdown voltage ([ref]), so its remotely possible if *BOTH* secondary connections have developed such a fault that you could get an asymmetry between the supposedly equal positive and negative bridge rectifier output voltages.  I would normally regard this as vanishingly unlikely, but the presence of crimped spade connectors on the transformer wires could provide the conditions required if insufficient crimping pressure to obtain gas-tight contact was used.  Continuity checks should reveal any such faults, which will show as the resistance changing when you reverse the test leads.   If it is a rectifying contact issue, go out and buy a lottery ticket as inadvertent ones are *RARE* and stable one even more so, and you are lucky enough to have two sitting in front of you! (A single secondary properly connected would still produce the correct output voltages by half-wave rectification, at half the available current and twice the ripple.)

[StillTrying]

Have you checked the via in the center of the star point.
Is the transformer 35-0-35 VAC.
Are you sure it's just this circuit.
Does that -Ve track go directly to the -Ve cap.
I'm running out of ideas.

[mapleLC]

Here is the scope trace.  Let me know if there is another view you wish to see.

[mapleLC]

Have you checked the via in the center of the star point.
Is the transformer 35-0-35 VAC.
Are you sure it's just this circuit.
Does that -Ve track go directly to the -Ve cap.
I'm running out of ideas.

Yes, the circuit is mostly as you describe, with the small caps under the diodes.  I have not added the second cap to the breadboard, I wanted to see this outcome first I described above.

I think I damaged the 2 sided-ness of the board somehow, that is my suspicion.  Some of the diodes trace on one side, the others, the reverse.

I can add the second cap if you wish?  I might as well rebuild it at this point 

[mapleLC]

My posts are out of order... please treat this as first, then the scope trace


I managed to rebuild the rectifier with the diodes I had in place, and the problematic cap that wouldn't charge.  I only did this with a single cap.

At 80v thru the existing transformer, I am seeing the following output on the neg end of the cap on the scope.

This seems to look ok to me as output verifying these parts work, so I am left with something bad on the board, I suppose.

I have to attach the scope trace to another post bc of file size.

[David_AVD]

It's not uncommon for large electros to leak, corroding one leg between the cap case and the PCB. From underneath, the board looks fine but the cap is pretty much O/C.

Have you removed those caps to confirm that the leg or a PCB trace under the cap is not corroded?

[mapleLC]

It's not uncommon for large electros to leak, corroding one leg between the cap case and the PCB. From underneath, the board looks fine but the cap is pretty much O/C.

Have you removed those caps to confirm that the leg or a PCB trace under the cap is not corroded?

Yes, I did so early on.  I have new caps in place at the moment.

[mapleLC]

I have added the 2nd cap and the rectifier is behaving exactly as expected on the breadboard.

I am getting 40v AC in at the v+ and v- and a +20v and -20v at each of the caps.

Does anyone know what the value of the small caps is/typically is?  I went in with the microscope and saw no writing on it.

[StillTrying]

Don't go much over 80V on that cap.

What would be more useful is to see the -17 VDC trace, gnd on the star point, signal on the -17VDC, leave the -Ve cap off if it's still off. That should show if there's the full wave -Ve rectification there, with peaks a lot higher than the average of -17V.

"Does anyone know what the value of the small caps is/typically is?"

I think they're usually 1nF to 10nF, they just stop a bit of noise, they're not very important, unless one has gone short.

[mapleLC]

Don't go much over 80V on that cap.

What would be more useful is to see the -17 VDC trace, gnd on the star point, signal on the -17VDC, leave the -Ve cap off if it's still off. That should show if there's the full wave -Ve rectification there, with peaks a lot higher than the average of -17V.

"Does anyone know what the value of the small caps is/typically is?"

I think they're usually 1nF to 10nF, they just stop a bit of noise, they're not very important, unless one has gone short.

This is what I tried numerous times before and was never able to get it to go off of the 35v vs -17v thing in my readings.

What I am seeing now with the same parts is that they work, so I wonder how I can merge my known working piece back over to the board to test it out?

I come from the software world, where this kind of hodge podge is easy, but here, I am not sure if its possible.

Would tracing a new lead into the old board at the correct points even work?

[StillTrying]

You need to find out where/why the -17 VDC was there. I don't think it could be a large load on the -Ve supply because something would be getting quite hot.

Smoothed peaks of 53 VDC just happen to be -17 V average half wave unsmoothed, and ~32 VDC full wave unsmoothed, which seems quite a coincidence that you get those values on the -Ve side. I think the -Ve cap isn't charging on the board because it's not connected to the -Ve diode(s) somehow.

Can't you scope the -17V/-30V, at the cap. the trace shape should/might give some idea of what's happening.

https://en.wikipedia.org/wiki/Rectifier#Half-wave_rectification

[StillTrying]

You must have found out what's wrong with the -Ve side by now.

[mapleLC]

You must have found out what's wrong with the -Ve side by now.

Ha ha not yet.  Some work has been going on in the house so my lab has been shut down since power is constantly going on and off in the house.

I am completely committed to fixing this whether its  or  or even a 

[Ian.M]

I say again:

We are all going to be kicking ourselves for not thinking of *whatever* when you finally find the cause.

*DON'T* give up - electronics repair is approx. 5% technical knowledge, 5% skill and 90% sheer bloody-minded perseverance!