• 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

High DC Ofset

QSilver

QSilver

Super Member
I've been working on an SA-710 recently re capping it etc and I wanted to finish with replacing the 2SA979's with new ones - I got the new ones, popped them in and left channel reads 16mV but right now read 50mV! The idea I had was the replacing these would get this reading down to a sensible level, not up even further. Have I done something wrong here?
 
Register to hide this ad
Where did you get the 2SA979's ?You may have leakage from one of the other transistors within that amplifier.
 
Got the transistors from LittleDiode.com cost the earth but I don't mind. Which transistors should I be looking at for leakage? Outputs or before that? Is the fact that both are not near 0mV showing that other transistors are leaking to varying degrees on either side of the amp?
 
If you are referring to Q13+Q14 those are the differential pairs. They need to be very carefully gain matched or they will cause your offset to be way off. Is this possibly the problem?
 
HVHiFI said:
If you are referring to Q13+Q14 those are the differential pairs. They need to be very carefully gain matched or they will cause your offset to be way off. Is this possibly the problem?
Click to expand...
The transistors in the front end are dual transistors on the same wafer. It's tough to get a better match than that :)

QSilver said:
I've been working on an SA-710 recently re capping it etc and I wanted to finish with replacing the 2SA979's with new ones - I got the new ones, popped them in and left channel reads 16mV but right now read 50mV! The idea I had was the replacing these would get this reading down to a sensible level, not up even further. Have I done something wrong here?
Click to expand...
What were the offset measurements prior to changing them? Are you able to measure transistor beta? A meter with hFE function?
 
:lurk: switch the dual transistors from one side to the other and see if the offset measurements follow. If not, you have other problems in the circuitry determining their operating points.
 
Last edited:
OverLoad said:
The transistors in the front end are dual transistors on the same wafer. It's tough to get a better match than that :)
Click to expand...

Just looked at the schematic closer...duh! mech986 has the right idea swap em and see if that's the issue, at least then you can narrow it down.
 
Ok swapping did nothing so I put the originals back in and then let the amp stabilize for 2 hours and then the dc dropped to 20-30mV - still no where near but better so it must be something else upsetting the whole system. This amp has been recapped and a lot of the low impedance resistors in the bias circuit have been replaced. When i say recapped - all electrolytics have been changed accept for those on the tone control sister board.
 
Which transistors could be causing this leakage? Any in particular? I'm hoping the outputs are OK.
 
Check q15, q17, and q19 for voltages. If you look at the schematic there are a few test points for your DMM to check. All three of those and the differential pair take down your voltages at each pin and post them. Also, compare sides.
 
I've been working on an SA-710 recently re capping it etc and I wanted to finish with replacing the 2SA979's with new ones - I got the new ones, popped them in and left channel reads 16mV but right now read 50mV! The idea I had was the replacing these would get this reading down to a sensible level, not up even further. Have I done something wrong here?
Click to expand...
I do not think so. I do not consider 50mV offset voltage an issue, that is if it is not causing another issue otherwise. On amps without dual bjts in the diff amp & a offset adj, 50mV is an acceptable number, especially using a single turn pot. IMHO, 50mV is nothing to concern one self about.
 
I've been working on an SA-710 recently re capping it etc and I wanted to finish with replacing the 2SA979's with new ones - I got the new ones, popped them in and left channel reads 16mV but right now read 50mV! The idea I had was the replacing these would get this reading down to a sensible level, not up even further. Have I done something wrong here?
Click to expand...
I do not think so. I do not consider 50mV offset voltage an issue, that is if it is not causing another issue otherwise. On amps without dual bjts in the diff amp & a offset adj, 50mV is an acceptable number, especially using a single turn pot. IMHO, 50mV is nothing to concern one self about.
 
If you wish to pursue this, swap q15 and q16 (2sc1775) between channels and see if the offset moves, if not post voltages of Q15 and Q16 e,c,b BUT the BLACK DMM LEAD GOES TO THE NEGATIVE SPEAKER POWER SUPPLY NOT GROUND!!!! (and not a regulated negative voltage either!! the RAW DC!!!) This is the -40 to -48v raw dc found on the huge filter caps.

All the other transistors in the power amp are under the control and balancing of the feedback system, the problem you are seeing is WITH the feedback system which is the 2sa979 and the 2sc1775. With new 2sa979's, that leaves the 2sc1775 and the components around it that form the current mirror circuit (r75, r77, d5)
 
Thanks for the reply mark, that really clears things up - I'll take a look at this as soon as I can and report back. Also, thanks rcs16 for your reply, this amp does not have any adjustment pots for idle current or dc offset and I thought that 50mV would start to cause some harmonic distortion - I'd like to get the amp running bang on spec if I can. Anyway, I will take a look and post results back when i get a spare few hours.
 
Ok, I swapped them over today and measured the dc offset after 30mins of the amp being on. Before I swapped: L 24mV R 20mV - seems to have dropped since last time.

I then swapped them and got L 24mV R 18mV

So then I measured the ECB of Q15 and Q16 with the ground connected to the negative Cap (-49.2V)

Q15
E 277mV C 1.22V B 0.88V
Q16
E 1mV C 1.22V B 0.89V
 
QSilver said:
Ok, I swapped them over today and measured the dc offset after 30mins of the amp being on. Before I swapped: L 24mV R 20mV - seems to have dropped since last time.

I then swapped them and got L 24mV R 18mV

So then I measured the ECB of Q15 and Q16 with the ground connected to the negative Cap (-49.2V)

Q15
E 277mV C 1.22V B 0.88V
Q16
E 1mV C 1.22V B 0.89V
Click to expand...

Ok, now swap q15 and q16 again and repeat the voltage measurements.

At first with the offsets barely moving, 24 to 24 and 20 to 18 I would say call it a day,
BUT
that 1 mV reading at the q16 emitter has me concerned.

This is a current mirror circuit, that is attempting to keep the emitter to base voltage (plus the offset required to turn it on) constant thus with 910 ohms and the diode drop voltage on the base leg controlling things (determined by the current flowing through them) the emitter and 1000 ohms will mirror that current.

So 1 MV across a 1k ohm resistor means the circuit is way out of balance internally.

IF the voltages remain the same after swapping the transistor - that means q16 and q15 are ok, and that something else is amiss in that circuit.
If so, read the emitter(black dmm lead) to base(red dmm lead) voltages on q18, q20, q22 and q24 - they should all be around 0.6 volts absolute (be it +(npn) or -(pnp) voltage, it should be 0.6v)

you may also want to look at the emitter to base voltages of the two halves of Q14 as well.

You see, a differential amp (what q13 and q14 are doing) can limp along (very poorly but perform) with half of itself dead...
either there will be two high value resistors from the q13/q14 collectors to the power supply voltage - OR the current mirror (q15/q16 with 2 resistors and an (optional)diode) that is used .

You could also try "turning around" q14 as well as swapping q14 and q13 and see what the offset and current mirror readings do.
 
Last edited:
I changed Q13 and Q14 round and the readings changed. I got DC offset of 13mV and 17mV.

I then measured the values on Q15 and Q16 again and these have changed too.

Q15 E 280mV C 1.2V B 890mV
Q16 E 280mV C 1.2V B 880mV
This seems better but I'm not sure how or why its suddenly decided to work.

I ran it for a short period and it seemed ok. However, I went to power it on later and it took a long time to go out of protection and once it came out, it went back in again and then out. So something looks to be unstable. I tried powering it on again and now the fluroscan meters jump around and the amp goes in and out of protection - DC offset is everywhere so what's breaking down?
 
I went back to this amp recently and found the issue with the rumbling in another post. Having replaced the transistors the regulate the voltage supply to these circuits, I went to check DC offset again. I replaced one of the 979's with a new one and comparing one channel to the other on powering on, the old 979 started at -70mV and slowly dropped to -35mV.

The new 979 started at -52mV and then dropped to -18mV.

I retook the measurements of BCE on Q15 and Q16 and got:

Q15
B: 863mV
C: 1.052V
E: 260mV

Q16
B: 867mV
C: 1.041V
E: 263mV

I' not sure if that's an improvement but I also checked the voltage readings from ground on Q13 and Q14 which seem to match spec - give or take 10mV.

Should I continue and measure B to E voltages on surrounding transistors?

I was under the impression that the offset on these amps should be nearer to 0mV but perhaps I've got the wrong idea? I've been trying to read the tuning fork guides about DC offset which have been helpful but I'm not sure I understand fully...
 
you will be hard pressed to get low dc offset. this amp does not have a ecap in the feedback path. its low freq ac gain is the same as the dc gain 16 or 24 dB.
you will need to add a offset control, dc servo, ecap in fb path to do much better than what you have now.
 
You must log in or register to reply here.
Back
Top Bottom