Quad 2805 power supply malfunction?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Sheldon, much kudos for you too. Many professionals wouldn’t dream of helping a newb like me.

I'm not really a professional though. :)

I am a mechanical engineer and designer primarily (that pays the bills), and quad repairs and audio are a hobby. So I charge less and am very willing to help folks out. I do it because I really like the speakers and what they do.


Sheldon
quadesl.com
 
Meeeeehhh, distortion again

I got a lot of help in this thread and was so glad that I was able to repair my Quad 2805.
To recall what happened: one of my 2805 ESL's produced a terrible distortion
on some types of music. With the help I've got in this thread I replaced c28 (1000uF)
in the power supply and my beloved ESL was like new.

Last week I noticed the problem was back. The distortion is not a constant thing.
It's coming and going in louder passages. I thought how is that possible ?
Is the same cap bad again ? A faulty repair job ?

So I put the ESL back on the "test bench". Opened the electronics compartment
at the bottom and did some measurements. At first without music playing.
I measured Vdc over c28 (the new 1000uF cap in the power supply): 15.5 V.
Then Vac over c28: 70mV. Both figures seem okay to me.

Subsequently I connected the ESL to my amplifier and while I was playing some tracks with
bursts of serious distortion repeated the aforementioned measurements.
When there is no distortion Vac over c28 (the ripple) is about 70mV, then suddenly
when distortion is setting in the ripple doubles to about 130mV. At the same time
Vdc over c28 is dropping from 15.5 V to 14.8 V. A moment later distortion is gone and
both figures are back to normal.

The new cap is sold to me as a Panasonic FR 1000uF 25V. I bought it from tme.eu.
Could it be a fake ? I really can't believe it is although there is no manufacturer indication
printed on the cap. Did I ruin the cap by applying too much heat when soldering ?

And why is the ripple twice as high during distorted passages and the
supply voltage somewhat sagging ? Does someone has a clue for me?
 
I am pretty sure that the higher VAC readings you are getting is because of the fact that the 555 timer circuit has fired and is clamping the audio signal with the Triac (that's the distortion you are hearing is the forced clipping from the triac). So that is normal. Now the question is why is the clamp circuit firing?

Things to check:

- Neon bulb blink rate
- watch neon bulb when the circuit fires on the bench, it may flash brightly or quickly at the times the clamp circuit fires. That would indicate something wrong with a panel.
- I've had bad traces on the boards and even bad delay line inductors that cause symptoms like you are hearing.


I know it's a lot of work, but you could swap boards across speakers, and even transformers and such from speaker to speaker to determine if the problem is panels or electronics.

My money is on you having a panel going bad. A stator is breaking free of the support structure and touching the diaphragm during loud or bass heavy passages.

Sheldon
quadesl.com
 
Sheldon, thank you (again) for your comprehensive answer.
One of the tips you gave me was to watch the neon bulb when the ESL was misbehaving.

Well, it looks like the devil has decided to play games with me today.
Because not one false note, not one snippet of distortion in more than an hour of listening.

You've said you are pretty sure the timer circuit has fired and the clamping of the triac
is producing the distortion I am hearing. To confirm this is what's happening I would like
to know if it's possible to measure the actual clamping action. This would give me some extra certainty.
 
…To confirm this is what's happening I would like
to know if it's possible to measure the actual clamping action. This would give me some extra certainty.

The NE555 timer is operating in monostable mode. 555 timer IC - Wikipedia
When the trigger pin(2) is pulled low(ie to ground) the output pin(3) goes high, activating the TRIAC clamp, and stays high for about 3 seconds after the trigger is removed. You can monitor pin(3) , or it’s connection to R17. It should be close to ground potential normally, and close to 15VDC when activated.

For testing purposes, you can activate the clamp manually by shorting pins 1 & 2 together.
 

Attachments

  • Clamp_test_01.png
    Clamp_test_01.png
    445 KB · Views: 224
  • Clamp_test_02.png
    Clamp_test_02.png
    444 KB · Views: 229
I have the exact same thing going on with my 2905s. One speaker distorts. It was once in a while at first now it’s all the time with the speaker occasionally sounding good for a couple of seconds. Here is the story so far
All panels removed and membranes visually checked to be fine
Main board swapped and distoration continues with the same speaker
Bulb does NOT fire when it’s distorting but keeps blinking at a steady rate of every few seconds.

That leaves me with the transformers and the delay lines. Does that make sense?
 
@Bolserst, thank you for your very useful reply. After the post by sheldon I was pondering how to proceed. As he suggested I could swap boards between both ESL's but I am hesitant to do so because I would have to "dismantle" an otherwise good functioning speaker.

So my first thought was: is the triac really clamping ? At the same time I realized that as I have no background in electronics I'm really handicapped. I'm aware of the basics, build a few amps and did a lot of reading but nevertheless a newbee.

The last few days I was reading on the net about the ne555 timer and the triac and there It was - your post - exactly what I was looking for. Now I know how to check if the triac is clamping and how to activate the triac manually.
Now my assumption is, that if the distortion that I am hearing is indeed caused by clamping activating the clamp manually would create the same distortion?
Or am I talking nonsense here ??
 
I recommend switching the main boards. It’s a bit daunting but actually not that hard and it Might clarify whether you have a panel issue. Also I highly recommend inspecting all the panels. I found two panels that were severely compromised that showed only intermittent problems. It’s a lot of work but actually straight forward. With those two checks done you’ll have a lot more information. If you don’t want to completely disassemble you can easily check if the stator glue has failed just by touching the stators.
 
That's a great write-up by bolserst. The clamp testing is spot on.

In my limited experience, these intermittent problems with clamping are often a panel problem more than an electronics problem. And it either a stator glue problem, just starting where the de-lamination is very local. But the bigger issue these days is diaphragm rot, where the diaphragm starts to split. It's really hard to spot without removing and splitting open each panel. The split often forms a corona and causes a small leakage, and forms an area that is easy for a corona to turn into an arc.

Here's a picture of diaphragm splitting ( this one is pretty advanced), but it can be somewhat subtle to spot if it's just starting.

http://quadesl.com/diyaudio/IMG_2887.jpg


Sheldon
 
I asked myself if it was possible to activate the clamp circuit on purpose so that I could check if the perceived distortion was like the distortion the ESL itself was producing on loud music (sometimes). To make matters more complicated the recent days the culprit was behaving impeccably. No distortion at all. Grrrrrr !!

Bolserst showed me how I could activate the clamp. At the same time I measured Vac (the ripple) over C28. The latter because I noticed before that the ripple doubled during periods of distortion. The outcome is that the distortion caused by forced clamping more or less resembles the original problem and that it is indeed the clamp circuit that is causing the increased ripple. Of course Sheldon said all this before :worship:

Unfortunately why the ESL is misbehaving is still unclear.

What are my options ?
Sheldon and Malthuse suggested swapping the electronics between both ESL's. I must admit I am still scary to tear down a perfectly good ESL. Furthermore Sheldon wrote
My money is on you having a panel going bad. A stator is breaking free of the support structure and touching the diaphragm during loud or bass heavy passages.
But why then is the distortion away for hours and then suddenly reappearing without apparent reason. From Sheldon I have got the impression that panel problems with newer ESL's are relatively widespread. And my ESL's are from 2007 !!
Conclusion: I am about to dismantle the troublesome speaker. Wise decision or not ????
Other things to check first ????
 
I hate to keep repeating, but the clamp circuit is clamping for some reason. If that 1000uF cap was bad, the clamp circuit gets overly sensitive. It's possible that you have an electronics reason for the clamp circuit firing. For instance a cold solder joint or a cracked trace on the PCB (of which quad sometimes provides straight from the factory) can spark and trigger the clamp. A bad coil in the delay line can also make the clamp trigger. But the most frequent, and most likely problem is that you have at least one bad panel. It could be bad stator glue, or it could be a diaphragm going bad (that is a feature on the newer quads that rarely happened on the older ones).

The reason it only happens sometimes (which can be maddening for the owner and myself as well) is that you problem is right on the edge of boing a problem. Let's assume for a moment that the stator glue is failing on one panel. The stator will be pulled toward the diaphragm by static attraction, so the gap is smaller. Depending on frequency content and amplitude, the diaphragm excursion will get close enough to arc or not. If you have a shutdown, you may be able to rewind your music and repeat it at the same passage again and again (giving time for the diaphragm to recharge).

If it's a diaphragm problem, where there is a split or tear, you will probably develop a corona at the site of the tear. Turn off your amp and put your ear right on the speaker and move around, you will hear a very faint hissing or buzzing sound. The corona is already locally ionizing the air (and producing ozone which is destroying the glue on your other panels), and this corona makes it easier for an arc to occur there if the diaphragm gets close enough to the stator. But the heat of the arc is often enough to locally melt the diaphragm and then you have a slightly different geometry that may be easier or harder to arc than the last time.

It's an insidious problem, and one that is affected by humidity, wind conditions around the speaker, and music volume and spectral content. But one thing that doesn't happen is that it never gets better on its own.

Sadly, it isn't easy to diagnose exactly what is wrong via internet forum. I'll be glad to help you figure it out, you just have to bring your speaker to New Hampshire. :)


Hope that is helpful,
Sheldon
quadesl.com
 
My experience with Quad 988's supports what Sheldon is saying. I had buzzing and distortion that was horrible one day and gone the next. I could not find any correlation to temperature, humidity or music content. My confusion ended when I started hearing the dreaded panel snap sound.
I have completely rebuilt the speakers. Now they sound fantastic. It is clear to me that they were going for a long time, probably for over a year. I did not make any modifications when I did the initial rebuild.
I ended up doing the usual electronic mods and increased the stiffness of the panel support frames. Both made a good improvement in sound.
It looks like a board swap will not be too difficult from the pictures I have seen on the net for the 2805's.
 
At the moment the quad is on my "workbench". That is the dinner table in the living:) My wife is not at all amused to have this thing on the dinner table much longer. So this weekend I will move it up one floor to a free room where I can tear it apart. Yes I am convinced now that checking the panels is probably the best approach. First thing I did was buying the Vessel JIS screwdriver as Sheldon warned in another thread for the crappy flat-head screws in the grilles. As the job is proceeding I will take a lot of pictures and upload them here. Perhaps other forum members will find them useful. And I am sure I will have more questions. To be continued !!
 
At the moment the quad is on my "workbench". That is the dinner table in the living:) My wife is not at all amused to have this thing on the dinner table much longer. So this weekend I will move it up one floor to a free room where I can tear it apart. Yes I am convinced now that checking the panels is probably the best approach. First thing I did was buying the Vessel JIS screwdriver as Sheldon warned in another thread for the crappy flat-head screws in the grilles. As the job is proceeding I will take a lot of pictures and upload them here. Perhaps other forum members will find them useful. And I am sure I will have more questions. To be continued !!

Don't hesitate to email me if you have questions or run into issues. I'm much more responsive to my direct email than the PMing on this forum.

Sheldon
stokes@quadesl.com
 
To summarize this thread, one of my Quad 2805 ESL's has an intermittent distortion problem. It's playing fine for hours even days and then at once the sound is distorting heavily. Conclusion is that for some reason the clamp circuit of the Quad is activated. To check the stators and the diaphragms I dismantled the speaker completely. A very time consuming job and you have to operate very carefully not to damage something. Now to the findings. The stators are properly glued to the frames, no problems here. At first sight the membranes are looking fine too, a few stains here and there which look like coating spatter. On a closer look one of the panels has a very very tiny what looks like a small tear at the edge. It's just two or three millimeters. See pictures. Could it be the beginning of what Sheldon calls "panel rot". On the third picture at the left there are two stains that look like coating spatters and in the middle the small tear.
Now my questions: could this be the reason for the clamping circuit to activate and second is it necessary to renew the membrane ??
 

Attachments

  • Screenshot from 2019-05-16 17-56-45.png
    Screenshot from 2019-05-16 17-56-45.png
    550.2 KB · Views: 239
  • Screenshot from 2019-05-16 17-59-37.png
    Screenshot from 2019-05-16 17-59-37.png
    626.2 KB · Views: 241
  • Screenshot from 2019-05-16 18-01-54.png
    Screenshot from 2019-05-16 18-01-54.png
    572.4 KB · Views: 229
Aha! you have a bad diaphragm; I warned that it can be hard to find without completely taking the speaker apart. In my limited experience, that diaphragm rot will continue to get worse on the effected panels until it fails completely (see my previous pictures). I've never tried to repair one, because it's so much less effort for me to replace it, and then I know it's good. I've seen quad panels from the factory with tape repairs, you could try a small piece of tape to cover those areas. The problem is that the ragged edges that stick down may still arc and trigger the clamp circuit.

To answer your questions, yes, that is a reason the clamp circuit could be activating. In my opinion, that is THE reason it is activating. And to answer your second question, in my expeerience that panel will continue to get worse.

Sheldon
quadesl.com
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.