DC-offset in my H/K PM665 Vxi starts at 500mV and takes about one hour to settle around 0-15mV. This can´t be right. Is there something to be done to speed up the DC-offset settle?
Thanks Kent
Thanks Kent
One channel only? Can you post a schematic?
You'll need to take some measurements while the amp is misbehaving.
You'll need to take some measurements while the amp is misbehaving.
Check/replace C405, 406 as well as the zeners for issues.
Likely replace the caps there to eliminate the problem.
Likely replace the caps there to eliminate the problem.
If this were my unit, I'd take a series of measurements, initially at power on and again after the unit has reached equilibrium.
This amp is bit unusual in that it has flat gain of about 60, extending down to DC. So it takes only a drift of about 8mV referred to the input to yield 500mV at the amp output. I assume that you have probably adjusted VR401 (DC BALANCE ADJ) to yield warmed-up amp output of 0V, but then see the objectionable initial error.
The 2SC2240s in the input have beta of 700, and I estimate about 24mV of bias voltage drop across R203, i.e. base Q403 should be roughly -24mV. Leakage through C401 could produce the offset you're observing, but there is a multitude of defects that might be responsible.
So I suggest measurement of voltages (re ground) at the following points:
base Q403
base Q405
amp output at R433
VR401
D403
Please report negative voltages as signed numbers. And again measurements at room temp power-on and again at warmed-up equilibrium.
This amp is bit unusual in that it has flat gain of about 60, extending down to DC. So it takes only a drift of about 8mV referred to the input to yield 500mV at the amp output. I assume that you have probably adjusted VR401 (DC BALANCE ADJ) to yield warmed-up amp output of 0V, but then see the objectionable initial error.
The 2SC2240s in the input have beta of 700, and I estimate about 24mV of bias voltage drop across R203, i.e. base Q403 should be roughly -24mV. Leakage through C401 could produce the offset you're observing, but there is a multitude of defects that might be responsible.
So I suggest measurement of voltages (re ground) at the following points:
base Q403
base Q405
amp output at R433
VR401
D403
Please report negative voltages as signed numbers. And again measurements at room temp power-on and again at warmed-up equilibrium.
I have replaced the C405 and C406 and there is no difference from before.
I guess i should read 0 mV on both positiv and negative side of the C401 and C402?
Now i´m reading -32 mV on the negative side of both C401 and C402 leading to base of the Q403 and Q404 resp.
I will replace these and see if it´s getting better.
Thanks.
I guess i should read 0 mV on both positiv and negative side of the C401 and C402?
Now i´m reading -32 mV on the negative side of both C401 and C402 leading to base of the Q403 and Q404 resp.
I will replace these and see if it´s getting better.
Thanks.
same issue with Radiotehnika 020. one channel -2V and other -0.15V . input stage transistor mismatch. replaced with original Onsemi KSA992 and now ~20mV
2SC2240 is obsolete and I doubt they will be matched or at least same batch
2SC2240 is obsolete and I doubt they will be matched or at least same batch
Other caps to suspect (due to the long timing):
C401 / C402 (input leakage), C411 / C412 (vbe multiplier), C417 / C418 (driver)
Replacement is easy. Then localised supply caps. Then small caps.
Inspect for pollution at both pcb sides too.
C401 / C402 (input leakage), C411 / C412 (vbe multiplier), C417 / C418 (driver)
Replacement is easy. Then localised supply caps. Then small caps.
Inspect for pollution at both pcb sides too.
Not necessarily. There is typically voltage across a cap--- otherwise no need for it to be there. 😉
You'd need to measure the voltage at power on and after settled drift to conclude anything. The notion being explored is possible change in leakage current though the cap.
I still advocate measurements above to get insight. Otherwise, you'll have to replace parts on guesswork.