WEM PA100 DC Offset

[stratfordade]

I'm restoring a lovely WEM PA100. Just finished installing new filter caps (old ones measured fine but will future-proof it for a few more years) and the output is up to spec power but I still get a hum across the output with no signal. I measure 0.4V across the speaker when I use an 8 ohm load.

Is there any way to reduce this? The voltages on the schematic are not very helpful as they seem to show 0.2V offset 'officially', and confusingly a different voltage under load for the same point electrically, written just to the left on the power amp schematic!

There is a 100 ohm preset but I think this is for setting bias. I can turn it and the offset reduces to around 0.2V but noise level doesn't change. I can't see any preset for setting offset unless it's this one?

Would be nice to eliminate the offset/hum. Everything else in very good condition.

[g1]

The trimmer is not for bias, it looks to me like offset or symmetry adjustment but I'm not familiar with this amp.
There is a note about reducing quiescent (idle current) by adding 560 ohm resistor across one of the bias diodes.
You can check what quiescent for the power transistors is right now by measuring voltage across the 40411's emitter resistors (shown a 0R33 and 0R25) and calculating current.
It's possible with lower idle current the trimmer will do enough to get rid of the offset.  400mV or even 200mV that you have for offset is too much.  50mV is a better target, or maybe 100mV in worst case scenario.

[J M Fahey]

That 100 ohm potenttiometer i9s to fine adjust output DC offsetr.

It is a primitive version of famous RCA "Brute 70" amplifier, which was a revolution when it came out in 1969-70

I built many back in the day and is the base design for most modern SS amps.

But it uses a single input transistor (later ones used a differential pair there which self balances) so 40406 needs about -0.6V on its emitter (this is shown on schematic) which is "stolen" from driver transistor 40409 emitter.

Try to reduce offset, but 200mV is good enough, not sure your hum comes from there.

It may come from poorly filtered main rails (+/-42V) or grounding issues.

What happens if you short across input 18k resistor with a bit of wire?
Does hum change in any way?

[stratfordade]

Thanks for the insights on the role of the preset.

I'm getting closer to source of the hum. The nice thing about this WEM amp is that the preamp just plugs into the PA/power section.

If I disconnect the preamp/controls section hum disappears and I just hear very low level noise from PA. All the electrolytics in preamp section test good for value and ESR which is remarkable after 50-odd years. When I remove the board carrying the preamp and pots from metal front panel and reconnect to PA the hum is much lower and nearly imperceptible, but if I lay the board down again on the front panel so that one or more pots metal bushes touch the panel then the hum is present. The panel isn't grounded in any way. The solder joints on back of preamp board all look excellent (I had to reflow some on PA board which was nowhere near as nicely soldered). The input sockets are all switching correctly with the hot terminals grounded with nothing plugged in (they ground to mains ground via one of the 3 plug wires connecting the preamp and power amp sections - one  ground, one signal from preamp output, one carrying 18V supply from power supply back to preamp board).

But I'll try shorting out the PA input resistor as suggested but suspect hum will stop as it stops with no signal connected from preamp. It's clearly some grounding problem, but when sorted I'll get back to lowering the dc offset as suggested above.

[g1]

Not sure how this amp is set up, but are the jacks wired or board mount?
Sometimes jacks that have insulators get replaced with non-insulated, or the isolation washers get lost.
Then you can get weird ground loops, maybe because pot bodies are also grounded.
If you can dismount the jacks but leave the pots touching the chassis it's worth a try.  Then if it helps you just need to get insulating washers for the jacks.

[stratfordade]

The jacks are wired. I'll check how they're mounted and test them demounted too in case all are missing isolation washers. I did check all board mounted pots for ground continuity and they were all good (they ground via board track). Thanks for this suggestion.

Edit: just had a look and all washers present and correct on jacks. I'll double check grounds to pots and jacks. Put a scope on it next.

[stratfordade]

When I short the PA input with preamp plugged in it's beautifully quiet, which just confirms issue is preamp section.

Thought I was onto something when found one of the six pots had a metal washer instead of insulated, but made no difference when I changed it to insulated.

[Jazz P Bass]

here is the complete schematic.

[stratfordade]

I've gone through the preamp and all the DC levels are correct so believe it's functioning correctly.

The scope trace of the preamp output shows very high frequency hash. It may be external interference. If I orient the front panel a certain way I even hear a radio station! The layout of the amp with the preamp plugged into power amp with long wires may be making it vulnerable to interference.

Think I'll put the amp back in case and try it in different locations to see if theory about local interference holds up.

[stratfordade]

I do pick up RF interference but the main problem is mains hum from somewhere.

With a scope at the input to first preamp transistor I see a 50Hz sine albeit just the peaks. Amplitude is around 2mV but I think this is significant as the test readings for the amp ask for a 7.5mV input.

So I need to figure out how to ground out that mains signal. So far all the grounds I've checked seem OK so it's a puzzle!

[Tassieviking]

Is that the first time you pick up the 50Hz signal ?
What about on the input jacks, both sides of the 10k resistors and also both sides of the 100nF capacitor.
Also check both sides of that 470k resistor to Ch2, I like to measure on the components leg if possible since the signal could be picked up in a component or a track.

If you can't isolate it that way then probe with a wire to 0V ( ground ) on these areas and see what happens with the signal.

Components sometimes work like a great antenna for the magnetic field around that transformer.

[stratfordade]

Is that the first time you pick up the 50Hz signal ?
What about on the input jacks, both sides of the 10k resistors and also both sides of the 100nF capacitor.
Also check both sides of that 470k resistor to Ch2, I like to measure on the components leg if possible since the signal could be picked up in a component or a track.

If you can't isolate it that way then probe with a wire to 0V ( ground ) on these areas and see what happens with the signal.

Components sometimes work like a great antenna for the magnetic field around that transformer.

Thanks I will check through the path tomorrow for those points.

I have tried shorting several grounded parts of the amp back to the power connection ground but no change. Thought it may have just been old wiring/connections introducing resistance to ground but your suggestion of hum being picked up from transformer looks more likely now. Will track it down!

[J M Fahey]

Sorry Friend, but your amp is working like on its first day, nothing broken or to repair, "you have what you have".

It is a very old and dated amplifier, SS Technology was "just born", it works fine by 1970 standards.

What you see is normal.


The hash you see is just residual white noise. Unavoidable.

Some poor design choices were made, volume pot is straight at the front end, only after a unity gain buffer, so following stages are always at full gain, you see *their* noise.
Again, in 1970 most didn´t know any better.

I also see some supply noise.
Notice the alternating higher and lower valleys and peaks, that´s 50Hz (the larger waveform) mixed with 100Hz (the lower one), both always in perfect sync because they come from the same place, both are derivatives of Mains voltage and current.


This is textbook perfect ground supply hum.

Notice full wave rectifiers do NOT charge supply capacitors during all the 50 Hz cycle but only on narrow peaks:



IF load current is, say, 1A continuous, transformer is actually recharging filter caps only once every 1/100th of a Hertz, in a way larger narrow pulse, say 5A to 10A.
Which averages to 1A continuous, what load demands.

Now that STRONG current peak travels through chassis or some ground wiring.

Chassis is never ever true ZERO ohms, it is physically impossible, a few inches of aluminum or iron sheet (chassis) , a ground return wire, even if thick copper, will have *some* resistance.

Your meter will show zero (you already checked that), but it´s actually, say, 0.01 ohm  , "Nothing"

"Nothing" really?
It depends.  Remember those 10A charging pulses and compare them to 1A *average/DC* current demanded by load.

0.01 ohm times 1 A = 0.01V DC, so under load your supply will drop from 42VDC to 41.99Vdc .... nothing indeed.

Now for signal/noise/audible hum analysis

0.01 ohm times 10A =  100mV hum at 50 or 100 Hz!!!!!!

Straight at **ground**!!!!

Ground, your supposed friend, now is your enemy, because if you ground your preamp there, "everything" will have 50/100 Hz peaks mixed in!!!

That´s why layout, grounding, ground loops are SO important, and best results come from experience.
LONG experience.

Not much experience possible in the early days.

To boot, I see 2 problems:



1) preamp is on a separate chassis, connected to supply thnrough long wires, to boot using connectors.

2) I see main supply caps very far from each other, separated by the power transformer.

Modern practice is to place them side by side and join them by a large copper pad, to minimize resistance.
IF outside caps are used , those mounted to chassis with clamps, they often are joined by a solid copper *bar* , not kidding (or at least an aluminum one)

Bonus points for the 220V Mains switch mounted 1 inch away from 7.5mV sensitive input.

But again, not dissing that beautiful and very well made amplifier, "you have what you have", nothing to "repair" except a full redesign.
After which it would not be a WEM any more.

Hey!!! it was good enough for Pink Floyd!!!!     

[Loudthud]

What I find curious about the scope photo in post #9 is that the little peaks are 10mS apart but of alternating polarity. The only thing I can think of is that one or two of the diodes have failed open in the bridge rectifier effectively making it two half wave rectifiers. The plus peak of the 50Hz makes a little positive bump in the ground, then on the next half cycle the negative peak makes a little bump of opposite polarity.

Quick edit: Could be that one side of the transformer winding has failed open.

[J M Fahey]

What I find curious about the scope photo in post #9 is that the little peaks are 10mS apart but of alternating polarity. The only thing I can think of is that one or two of the diodes have failed open in the bridge rectifier effectively making it two half wave rectifiers. The plus peak of the 50Hz makes a little positive bump in the ground, then on the next half cycle the negative peak makes a little bump of opposite polarity.

Quick edit: Could be that one side of the transformer winding has failed open.

I may be wrong, but what I see there is:

1) waveforms are measured from one point until it repeats "x" time later, always the same..

So a 50 Hz sinewave has positive peaks separated by 1 second / 50 = 1000 milliseconds / 50=20 milliseconds.

Or alternatively, 20 milliseconds separating negative peaks, same thing.

We find both in the picture so I *guess* they are artifacts or parts-of a 50 Hz wave.

50Hz AC because I see no rectification there, both polarities are present, same amplitude,and in the proper sequence: + ... - ... + ... - ... and so on.

I can also guess *where* are they coming from.
Will try to find a drawing, too sleepy to draw anything from scratch now, 02:30 AM here, to show where that waveform is *guaranteed* 

I have found and solved this problem many times, but it plagued me in my early days ... some 40-50 years ago.

Grounding is walking a minefield.
You ground *here* and you have hum.
Screening is useless because ground itself, screen too, BOTH are Hummy     

Now you ground *there*, one inch away, same chassis, same ground bus bar, whatever, and it doe not hum   

You measure ZERO ohm between both points, of course