Adcom GFA-565 again, fluctuating DC offset
I am new on this forum and want to thank everybody for the information I allready got, beeing no member.
I am just an amateur and sometimes repairing (mostly better) equipment for friends. In my spare time and not at all doing a business. Beeing a Dutch guy, please excuse me for maybe funny or bad English.
I have now an Adcom GFA 565, bought on a flee market internet site in Holland, beeing broken.
Well it was: five output transistors were blown, inclusive their emittor and base resistors, all at the positive side. All the "floating ground" fusible resistors were broken or vanished. And you guess allready: the tree "famous" capacitors leaking.
I think for the rest I was lucky. I could not find the schematic but I figured out the most important parts, on the boards, looking and measuring.
It appeared not easy in Holland but I found at last all the broken parts; the 2SD424 not original but a modern clone with the same type number. Also I found the OP97 (version F).
Looking on the threads on this subject (thanks to all!) I found out I had to clean the board, allso I had the fishy smell.
Not having an ultrasonic cleaner, I mounted out the board, desoldered all the electrolitic capacitors, the bias adjustable resistor and the heatsinked predriver transistors and also some bigger foil capacitors, and cleaned it with simple green with hot water and a tooth brush. Two times and in between some fine scratching with a screwdriver in the probably more sensitive areas (the darlingtons and later amplifiing transistors). Also I measured, just tipping on the board, with a digital Ohmmeter in the 200 Mohm range and were I measured a value I scratched as well.
Everything assembled again I had, without servo IC, about
-0,25 V offset, not bad I think. Everything seems rather normal:
2 mA tailcurrent and 20 mA in the predriver tree.
But now comes my problem: Also a small fluctuating voltage of maximal about 10 to 25 mV on the offset, very irregular, in "waves" of a few seconds.
With the OP97 mounted general offset is very small (a few mV) but the fluctuating seems the same, so reality is that there is an offset of about plus or minus 15 to 25 mV.
This is all without connected voltage to the power arrays. When I do so the results are about the same with a bias of about 29 mV but that seems of later concern.
So it seems for me more safe, just for this problem, of doing testing without all the current through the output arrays.
Allso their is high temperature sensitivity: when I touch some of the transistors of the first two stages their is a vast change and slowly stabilizing afterwards. But it seems rather logical: offset moving up or down depending witch transistor I touch, symmetrically about the same. First amplifing darlingtons are the most sensitive but also symmetrically about the same.
My idea is that the servo is working normal, but very slow, so it can't compensate for this fluctuations. Seems also logical for it must not compensate for audio signals.
Finally my questios are:
Is this kind of fluctuating offset normal? (I don't think so)
And if not, what can be the cause?
Can it be malfunctioning semiconductors, for instance the green
or are their still some kind of leaking currents of the spoiled
electrolite that cause the problem? And can that cause this
flúctuating problem and not just a steady offset?
All this questions to be a bit more sure about what to do next, if I have to look for ultrasonic cleaning or can avoid that.
Finally there is one other type of question: Have all these Adcoms this same problem of leaking caps or are theiir some series that have not?
I hope you can help me a bit and I will be very grateful for that.
With my kind regards,
I have some fluctuation with mine, but not much. The 565's, and the 585's were the ones plagued with the leaking Elna caps:(
Thank you for your reaction and well, now I am beginning to be pretty sure I as well have to walk on the glissy path of ultrasonic cleaning. Although I do not know really well the cause of my bigger fluctuations, and I am a bit afraid all this work of desoldering and resoldering again will not help at all.... But the owner of the amp will now try his contacts to get an ultrasonic cleaner.
Meanwhile I am still puzzling about a few things.
First the fluctuacion again, what can cause this? Can components be so noisy? Can leakage currents have a variation so much? In this state where the general offsett is allready pretty small? Depending on structure of the board, temperature, voltage differences,what else...? This without any signal, beeing the input open.
In the other threads I sometimes found some remarks about fluctuacion but never saw a reaction on this point. So it seems everybody accepts just a small fluctuation, like you as well I understood. But for me this fluctuation seems too much so I first will go on cleaning. But when someone knows a bit more about this fenomene and leakage currents I would like to know. For example I have a EC AW100 amp with abit other way of nulling, but it gives 0 mV and stable as a rock.. Allso I have a home made one (the first stages roughly a bit the same DC setup as the Adcom) but using normal zeners for the predriver tree and that one even has more fluctuations and with none leaking caps at all....
Secondly I found in your earlier thread (about the many adcoms with huges offsets) you intended to change all the power transistors beeing only one broken. Later you was saved of this problem finding the original ones of Toshiba. But why? Are these modern clones so bad? (You know from uphere I used 5 clones) And why looking to match them? Maybe I am thinking too simple and/or with not enough knowledge but I would think that using 10 of those in the ultimate current amplifing stage in a parallel way should give a fairly good generalisation of some individual differences. Maybe the bias will be a bit different for each individual transistor, but is that so important, will it hurt the quality of the sound? I know I am coming now on a pretty glissy path of thinking but I as well like making modifications or changing components in order to try to get "better/more musical" sound, but in this case I don't know.
Hoping to learn a bit more from your reactions, and thanks again,
I had some fluctuation on another amplifier. The problem was LF oscillation in the DC servo. The output moves between +/- 400mV.
The problem was faulty capacitor in the servo
Final DC offset as low as 25mV, even fluctuating, is not an issue. Anything below 100mV is considered acceptable, below 50mV is fine. I haven't worked on an Adcom though so i don't know if 250mV before the DC servo is good enough. Apparently it must be since the servo corrects it to 25mV...
Well progress is going on, but slowly.
I decided to look for original Toshiba's 2SD424 for the clones I bought in Holland do not match very well with the other ones that are still good.
In fact with the same bias (about 24 milivolt) the individual currents differ twice to four times... So that is not only a difference of the beta (I measured them) but also a question of different voltage-current behaviour, I think. And that will get problems later when asking some power, I am afraid.
Also the fluctuacion puzzles me but I think I have to accept your advice, maybe do onely one more cleaning.., but furst the original transistors.
Any remarks or suggestions? I would like so.
|All times are GMT. The time now is 01:11 PM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright ©1999-2017 diyAudio