Maker Pro
Maker Pro

audio amplifier with unstable output offset voltage

yulyph

Dec 6, 2012
15
Joined
Dec 6, 2012
Messages
15
hi,
i have on old amp (amfiton 50Y-202C) that has one of the chanels oscilating after 30 min. i measured the output and the offset goes down from 30mv to -50mv in about 2 min. i replaced a few capacitors and resoldered most connections to no avail.
i've attached a schematic of the amp.
can someone please help ???
 

Attachments

  • sxamfiton.jpg
    sxamfiton.jpg
    76 KB · Views: 807

Harald Kapp

Moderator
Moderator
Nov 17, 2011
13,700
Joined
Nov 17, 2011
Messages
13,700
This sounds like a thermal problem. You could:
- check the fuses. I only recently experienced fuses that were not blown but became high-resistance.
- check the offset working from the input forward (or from the output backward) in the signal chain. Find the point where the amplifier is still stable and becomes unstable after the next stage. Check all components of that particular stage, escpecially capacitors and transistors. See if any component becomes unusually warm. If you have cooling spray, cool this component. Does the effect change? If so, replace that component.

Harald
 

yulyph

Dec 6, 2012
15
Joined
Dec 6, 2012
Messages
15
fuses are ok, even changed between them to make sure, still the same.
offset is going bad after (forward) c6 (100pF) and r12 (18k). i've replaced c5(50mF) earlier (was bad) still no effect. i've noticed VT6 a few degrees above room temp (haven't cooled to see difference - have to get a cooling fan to see) and checked comp around it and vt6 itself, to no avail. still checking...

edit:
replaced transitors, checked capacitors but still no effect. i've noticed that voltages on negative side increase in value and the positive side decrease...
thanks for your help.
 
Last edited:

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
As Harald suggested, the problem is probably thermal and you can use rapid freeze (in an aerosol can) to find which component is causing it.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
"canned air" works fine. Invert the can and you get cold propellant -- instant freeze!
 

yulyph

Dec 6, 2012
15
Joined
Dec 6, 2012
Messages
15
thank for suggestions.
the only transistor to overheat is VT6 (kt814r) witch i replaced with bd140 with the same result. replaced some other nonpolar caps that changed value just from the heat off my hand :confused:.
so the only conclusion so far is that the problem is around VT6 (it's the third transistor from top left, the one with two diodes in the base of the transistor) , because i was blowing a bit on it and the offset jumped...
 
Last edited:

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
That schematic is hard to read. Line drawings are best in GIF or PNG format, not JPG (JPG is designed for continuous-tone images like photographs). I assume you scanned it at a higher resolution than the image you uploaded? I would convert it to greyscale, increase the contrast so ALL the background is 100% white, then reduce to 16 colours, or even 2 colours if it looks good. You can do this at a relatively high resolution. When you convert to GIF, the large white areas will compress well, and you should be able to get an image smaller than the 100k limit with plenty of resolution and high readability. This would be worthwhile because we are interested in the type numbers of the semiconductors.

With an old amplifier like that, many of the semiconductors may be going leaky. If you replace one or two, it may just be a matter of time before you need to replace more.

A DC offset of 50 mV is not a major problem. The fact that the DC offset varies as it's warming up isn't necessarily a problem either. Have you checked the other channel? It may be the same.

What's the condition of the amplifier in general? Is it possible there's dampness or some other kind of leakage paths on the PCB? Or off-board? Can you upload some photos?

The two diodes next to VT6 could also be faulty. You probably can't blow on VT6 without blowing on them as well, right?

Just because VT6 was the only transistor that overheated, doesn't mean that there aren't any other components whose characteristics could be changing as they warm up. That's why I suggested rapid freeze, or any aerosol I guess. If possible use a thin tube to direct the spray directly at the component, and use some curved cardboard or similar to prevent the spray from going on components behind it. Try each semiconductor in turn, after the amplifier has warmed up.
 

yulyph

Dec 6, 2012
15
Joined
Dec 6, 2012
Messages
15
this is the best i could find on the net (a few transistors have different markings). i don't have the original papers...

edit1:
if the offset would vary a few mv, it wouldn't be a problem but from +30mv to -50mv in 2 min (and still going down till oscillation) or less is odd. the other channel is fine and it stabilizes it's self at +20mv.
looking at the pcb there isn't any sign of trouble (leaking caps, signs of burned components,....).
as for the diodes of vt6 there are measured and fine, but could try replacement if i knew the equivalent as i can't find any info on the net....
as for the cooling spray i can't say that i have access to any cans (i have to order one).

edit2:
after replacing the mirror transistors and the offset pot it seems that the hole thing offsets (maybe i didn't measure right the first time) right after c1(5uF). so could it be the supply ? i have replaced the filtering caps as they were bad, so could the rectifying bridge be faulty or something simple like that?
 

Attachments

  • 1.png
    1.png
    90.9 KB · Views: 869
Last edited:

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
this is the best i could find on the net (a few transistors have different markings). i don't have the original papers...
Thanks for that. It's an improvement... I think!
if the offset would vary a few mv, it wouldn't be a problem but from +30mv to -50mv in 2 min (and still going down till oscillation) or less is odd. the other channel is fine and it stabilizes it's self at +20mv.
OK.
looking at the pcb there isn't any sign of trouble (leaking caps, signs of burned components,....).
There can be subtle clues. Also this is an unusual amplifier. I think it could be quite helpful to see some photos of the amplifier. Both sides of the PCB could be relevant.
as for the diodes of vt6 there are measured and fine, but could try replacement if i knew the equivalent as i can't find any info on the net....
When semiconductors get old, the faults they exhibit can be quite subtle. It's quite possible for a diode to measure OK, but still cause DC shifts as it warms up. They can also become sensitive to being moved or tapped, as well as temperature of course. It MAY be possible to locate these components by measuring voltages and currents, but this may be impractical and it's often easiest to locate them using freeze spray and/or a heat gun, or just replace any suspect components.

What is the part number for the diodes? I can't read it on that schematic.

as for the cooling spray i can't say that i have access to any cans (i have to order one).
I think you should do that.

after replacing the mirror transistors and the offset pot it seems that the hole thing offsets (maybe i didn't measure right the first time) right after c1(5uF). so could it be the supply ? i have replaced the filtering caps as they were bad, so could the rectifying bridge be faulty or something simple like that?
I doubt it would be the power supply. But this raises another issue. What is the history of this amplifier? When was it manufactured? How did it fail? What else have you done to it? Tell me everything you know about it :) There may be a clue in there somewhere.

Having a look at the schematic, I notice that VT1 (type KT315A) is used as a zener diode to regulate the current in the differential amplifier (VT2 and VT4). This probably isn't a good idea. You could measure the collector-emitter voltage of VT1 and see whether it varies noticeably as the unit warms up.

I searched for KT315 (several transistors of this type in this amplifier) and discovered that it's a Russian device. Actually, it's a series of devices with significantly different specifications - VCEmax ranges from 15 to 40 volts! This might indicate poor control of the manufacturing process.

The data is at http://www.5v.ru/ds/trnz/kt315.htm and there are photos at http://www.freewebs.com/projektai/fm/tranzistoriai.htm and http://www.155la3.ru/kt315_2.htm.

They seem to be packaged in some kind of potting resin, rather than being hermetically sealed. If that's what you have in that amplifier, I would be suspicious of all of them.

BC547 is a suitable replacement for these KT315 transistors, although the pin order is different and you would have to choose a gain suffix according to the specification of the KT315 variant you're replacing. The 2SC945 would be better, but neither Digikey nor Mouser carries it :-(

If the main smoothing capacitors had failed, I would replace (or at least measure) any other electrolytics. There's one in the feedback path back to the input stage, C5 I think.

Beyond that, I would suspect VT2 and VT4.
 

yulyph

Dec 6, 2012
15
Joined
Dec 6, 2012
Messages
15
the hole god damn thing is a russian crap of an amp.
the history is fussy at best, as it was a friends amp and i just got possession of. it was repaired before so....
as for vt1 i'l try to replace with bc546 (i have only these ones at the moment) but i don't expect much as i replaced vt5 (KT315B) with bc546 and the amp got the negative rail at the output...:confused:
c5 was replaced and as for vt2 and vt4 were replaced with bc556, also no effect.
all replacement transistors are new so no joy there.
could you specify a replacement for the diodes? they are king of flimsy by looking at there legs (KA103A - the A after the K is a russian A and the last one is a normal A) :D
as for pictures of the amp tomorrow.
 
Last edited:

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
A quick Google search doesn't show any specifications or equivalents for KA103A, only a Russian supplier who will sell them to you. In this case any small silicon diode will work; I'd use a 1N914 or 1N4148. But I doubt that the diodes on VT6's base are at fault. I was just making a point.

Why don't you tell me everything you've done to it, and everything you've noticed.

I would like to see photos of both sides of the board.
 

yulyph

Dec 6, 2012
15
Joined
Dec 6, 2012
Messages
15
ok, here are the photos and it's the best i could do with the phones camera:
http://imageshack.us/photo/my-images/853/78280683.jpg/
http://imageshack.us/photo/my-images/254/35400100.jpg/
http://imageshack.us/photo/my-images/443/32336510.jpg/

i kept everything stock as new transistors didn't help, only the dark red non-polar caps were replaced as the originals kept changing value with temperature change.
the board is a bit different from the schematic but it is largely the same.
vt1 (kt315A) maintained a stable collector-emitter voltage as the offset kept falling.
 
Last edited:

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
OK. Well, it's difficult to be sure what to do if the schematic doesn't match the unit.

Rapid freeze could tell you which component(s) are failing. My guess is that a transistor may be becoming leaky as it warms up. If VT2, 4, 5 and 7 match the unit, they seem to me to be the most likely candidates to cause DC drift. You could try replacing VT5 and VT7.

But I think it's likely that all of the semiconductors are in the process of drifting out of specification. Quite possibly the resistors too, and the capacitors. And the trimpots! Did I leave anything out? The wires are probably OK :-/

Another approach would be to replace all the semiconductors, measure all the resistors, and so on, but even if you fix it, it's not going to give very good performance. Personally, I would scrap it, unless I had a boat that needed an anchor. If you like the look of the unit, you could replace the amplifier boards with nice compact IC-based power amplifiers; either kitset or preassembled.
 

yulyph

Dec 6, 2012
15
Joined
Dec 6, 2012
Messages
15
so i replaced vt2,3,4 with bc556 no change, but as soon as i replaced vt7 with bd139 and r16,r18 the offset started to jump around from + to - still going down... so i guess here lies the problem...

as for the hole thing i know it's a piece of ... russian :D, but i would like to have it fixed.
thanks
 

yulyph

Dec 6, 2012
15
Joined
Dec 6, 2012
Messages
15
replaced vt1 (kt315) with bc546 and it didn't work so well, it's got almost full rail between emitter and collector.
did a bit of replacing like vt2,3,4 and everything around it and it seams a lot more stable. the offset stabilizes itself for a while after going down slowly.
r11 and r13 seam like odd values so i have to figure out a replacement combination of resistors.
any other suggestion would be appreciated.
thanks
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Are you sure you have the pinout right? It looks like the two versions of the KT315 have different pinouts - the flat one is b-c-e and the TO-92 one (cylindrical with a flat section) is e-b-c, according to the diagram I linked to in post #11.

I wouldn't bother replacing VT1 anyway. You already found that the voltage across it doesn't change as the amp warms up and the output voltage drifts.

I would suspect VT2, VT4, VT5, VT6 and VT7. You've already replaced VT2, VT4 and VT6 earlier, but I think you put the originals back for VT2 and VT4. I would replace ALL of those transistors. Keep the originals if you want, but don't start putting them back until you've fixed the drifting and oscillation problem.

I doubt R11 or R13 could cause the problem. They only affect the AC gain of the amplifier, not the DC conditions. If you want to go replacing resistors, I'd suggest R3, R8, R5 (trimpot), R4, R9, R16 and R18.

I really think you could save yourself a lot of time by using rapid freeze, or any aerosol really, even an air freshener, to find which component is drifting with temperature.

I would also check for possible leakage paths around the transistor shown hanging out from the bottom of the diagram. It looks like it's marked VT1, but there's a VT1 at the top left corner. This transistor will be mounted against the heatsink. Check there are no possible leakage paths from any of its terminals - maybe there's some thermal grease that's gone bad.
 

yulyph

Dec 6, 2012
15
Joined
Dec 6, 2012
Messages
15
thanks man for pointing out to me that i misconfigured the bc546 connections. don't you just love chinese parts and russian engineering ??? only after looking on the net and seeing two different configs for bc546, i measured the hfe and... the first configs you get on the net are wrong...
replaced vt5 with bc546 and now it's working well.... for a russian amp.
many thanks
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
You're welcome.

I hope it gives you many more years of good quality sound and reliable operation. LOL!
 
Top