• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

Amplifier Distortion, DC-Offset, and You!

Haoleb said:
Well, i have this sansui au-517 sitting in my closet because it has distortion and its protection kicks in after like 5 mins. so i tested the dc and in one channel it was 80 something mv and the other was like 230mv Is this why i can hear distortion from the amp and why after a few mins the protection kics in?. what do you think i should investigate first to try and fix the amp??
Click to expand...

It would seem this is the dreaded VAS stage emitter resistor problem -common on AU517, 717, and the odd 710.

Look for 2 transisotrs on the power amp boards, 2SC1904 and 2SA899, they are black square types (TO126) and are mounted vertically on the board. In the immediate vicinity you will see two small greenish-grayish-whiteish resistors rised off the board surface via their legs being formed as spacers. These are metal film 150 ohm, though you will likely hardly be able to tell as they tend to overheat and change value, or go completely open. You need to substitute these with same value (150 ohm) about 1W, also mounted off the surface of the board for better cooling.

When these go open the amp will stay in protect mode due to sticking to one supply rail. When they drift too high up in value, offset will increase to the point where the protection may cut in and out intermittently, or cut out completely when they warm up. Distortion will increase drastically.

One of these resistors is often also covered with the famed Sansui corrosive glue, and tends to fail first - the one which is close to one of the 470uF local filter caps (on the power amp board).

I serviced a number of amps from this series (OK a large number) and would routinely change these resistors even before doing any other work, whatever it may have been.

ZN

PS, of course, after these are changed the amp will require a readjustment of bias and offset!

PPS: see attached schematic, sorry for the quality, this is one of my reverse-engineered hand-drawn diagrams...
 
Last edited:
i measured my Rotel RX-1603.
Left 3mv
Right 38mv :(
transistors are not the same left & right
transistors on the right side are new
aux,tape, phono stereo reverse ... makes no difference
so i think it must be the amp section
so it could be a resistor ?
 
Soran said:
i measured my Rotel RX-1603.
Left 3mv
Right 38mv :(
transistors are not the same left & right
transistors on the right side are new
Click to expand...

If transistors are changed, bias and offset need to be re-aligned as a matter of course. That being said, it is possible that a previous problem caused a deterioration of other components, I would look at the input pair and at the feedback DC blocking cap, usually an electrolytic.
 
The Rotel does have an adjustment for offset, and it can be nulled if you like. What you are reading at the present time isn't a terrible issue though...
 
I just fixed a Hitachi SR-304 receiver with around 80 or so mV offset on both channels (from memory, mV listed elsewhere here) - NO schematic to hand! Here's how...
1. Took a 1.5 volt dry cell, solder a wire to the negative and a 1 megohm resistor to the positive.
2. Connected the wire to the ground.
3. Connected both speakers (used my workshop test units, not my best PSB's!)
4. Connected a mV meter to each channel (or to one channel at a time if you only have one meter, but then you have to think more!)
5. Poked around the DC part of the driver and output circuits with the other end of the 1 meg resistor and watched the meters.
6. Found a point where the DC offset went down quite few mV, but not to zero (if it had gone up I would have reversed the battery polarity.)
7. Found the same current injection point on both channels (it also needed a positive voltage source, i.e. same direction.) BTW, clicks in the speakers were barely audible, no big thumps (used a large resistor compared to the typical circuit impedance and stayed away from the pre-amp circuits - no good there anyway as they are AC coupled, you ned to be in the DC part.)
8. Now know that I only need a positive DC source!
9. Used a forward-biased small S/C diode (per Echowars circuit), with a 15 K resistor from a convenient +13.5 volt regulated B+ point (going to the tuner section, IIRC.) Diode provided +0.65 VDC (measured.) (If negative would have gone from a B- source; if different polarity for the two channels would have used two diodes and both supply rails, etc.)
10. Bypassed the diode with a 470 uF 25 VDCWG capacitor, just to be sure.
11. Experimented with the resistor value that gave very near to zero on one channel. It was 470K - very convenient!
12. Repeated for the other channel. It was 610 K on the Fluke ohmmeter (happened to have this value to hand!)
13. Soldered all components under the PCB with carefully inserted cardboard over and under them to avoid shorts (to PCB and the foil on the base cover.)
14. Result: both channels now have a DC offset of less than 5 mV. No effect on the sound or balance; no hum, noise or instability.

I could have used a couple of 5K helipots to zero the DC offset exactly but I didn't want to use them when I could get close enough with simple preferred value resistors.
Cheers,
Roger
 
That'll work until they drift again. ;) (Ought to be good for a few years).

Hell, for that kind of effort, you could have build a servo.
 
Hi, EW.
Yes, I actually thought of doing that but I'd need to use an op.amp. as an integrator and set up a filter to ensure no feedback below, say, 15 Hz.
Quickly dropped the notion!
Cheers,
Roger
PS. Thanks for the original idea, also for providing the incentive to do something.
 
Last edited:
Marantz 4270, via Bay for $250 'checked over and in great condition', had outstanding photos, ....... DC offset -9.9. :)

-sf
 
First post

This thread was the actual reason for joining and registering at audiokarma.
I learned a lot reading all these pages.

Measured with an old digital RS (used to be called Tandy in Belgium) Micronta 22-198A on the 2-volt scale (lowest):

Kenwood KA-6100: L= 2mV R= 2mV (read as .002V)
Kenwood KA-550: L= 7mV R= 3mV
Kenwood KA-5040R L= 326mV :sigh: R= 19mV

The 5040 developed a well-known fault after some 12 years: noise and crackling in the left channel and finally the protection kicked in. Replacement of a leaked electrolytic and two transistors cured it. But as you see there is something else going on. I am going to order the service manual.
Funny thing is that when you breathe or blow on the board the value increases temporarily.

KR
Eduard
 
Sony STR-313L : 9V DC on right output!

Hi - I'm new here. Joined because you all seem to know your stuff pretty well and I need some help!

I got a Sony STR-313L (c. 1978) receiver off eBay recently and the left channel seems to work ok, with 55mV DC offset after 20mins. Unfortunately, the right channel has 9V (as in 9000mV) coming out of it. Any idea what could cause this? I've traced all the circuits up to the power amp and they seem fine (left circuit reads same as right), but somewhere between the power amp and the tone controls, the right channel picks up 9V. Reading some of the earlier posts, I guess I'm lucky I didn't blow my speakers!

Great forum, BTW.
 
Rtchoke...
Sounds like a differential pair needs to be replaced with a closely matched set. Most of the Kenwood's don't have an offset adjust. As far as blowing on it, much earlier in this thread I mention that thermal characteristics are a big factor in the gain of the transistors. Heating or cooling changes the gain, and will change the offset.

Zephyr...
9V? That's a problem. I'm not familiar with the internals of that unit, but seems as if something is amiss. Hard to believe that it stays at 9V with speakers connected, as that translates to over 10W!!! :yikes:

Unless someone more familiar with that unit chimes in, I'd just have to say that 'something, somewhere' in the amp section needs attention (leaky capacitor, burnt resistor, dying transistor....)
 
Perhaps a silly question (but not as bad, I hope, as 'what's the difference between current and voltage' a few pages back): how do you go about finding a leaky capacitor / dodgy transistor? And how on earth does 9V DC get into an amp output circuit just because of a failed capacitor?
(the answer might be because I paid $30 for it, but I'm hoping there more...)
 
That's the tough part...

The amp very likely has a few DC blocking caps in the feedback loop to reduce DC gain. If these are leaky, then they are allowing current to flow where it should not. How do you find 'em? Eh...amp driver boards generally only have 3 or 4 electrolytics anyway...if I suspect in any way that they might be a problem, I just rip 'em out and put in new ones.

If the amp is completely DC coupled (as a lot of amps are that use a FET differential pair, and Sony likes FET diff-pairs), then a bad transistor or burnt resistor is more likely.

Leaky transistors can generally be found by comparing the collector or emitter voltages to those you calculate by examining the circuit, or comparing to voltages given in the service manual or the other (working) channel.

It's not a rookie job, unfortunately, but any decent tech ought to be able to track it down fairly easily. I Googled the Sony, and I gotta admit it's a damn nice looking unit...ya oughtta get it fixed. :yes:
 
You've been a great help echo!
But more questions: I traced the inputs to the power amp chip on the board (I've got a schematic), and they are matched at 0V (=55mV) But the outputs from the chip show the same problem as the speaker outputs: 9V on the 2 right output pins (one goes through the tone control first, but both lead to the speakers), and the usual 55mV on the 2 left outs. Every transistor I could find matches the specificied voltages on the schematic, and there are no problems I can find upstream of the power amp. Does this mean the chip is to blame? It's an SI1125HD, if that means anything to anyone. If the chip is at fault, do you think I stand a chance of finding a new one?
 
It's a chip-amp...why didn't you say so? :D

Donberg in the Netherlands says they have them, as well as Dalbani in the UK. US sources seem to be only through large wholesalers dealing with old and excess inventory, and they ain't gonna sell you 1 or 2.

I'd like to see a scan of the amp section. If you can do that, start a thread in Vintage Audio or DIY.
 
I wonder if the Pioneer is a capacitor-coupled output. Have you been using it, and it 'seemed' to be OK?
 
I got it not to long ago and I can't put a left speaker on it or the speaker will buzz really loud and I think I found the problem theres a capacitor on the left channel on the amp board and it increases the power to 14V I tested the same kind of capacitor on the right side of the board and it increased the power from 13mv to 15mv.
 
You must log in or register to reply here.
Back
Top Bottom