DC offset in Mission 777 amplifier

[tkwou]

I purchased a Mission 777 power amplifier on Ebay last fall. It's a great sounding amp that matches well with my Mission 776 preamp. It is a completely symmetrical dual mono design direct coupled with no input or output capacitors. It uses 2 pairs of Hitachi mosfets per channel. There is no output Zobel or output inductor. There are no potentiometers. Individual resistors are chosen in the input stage to balance the amp. I recently decided to check the DC offset at the outputs and was quite surprised what I measured. With no inputs connected the DC offset is 0.359 VDC on the left and -0.359 on the right. With the Mission 776 preamp connected, the offset rises to 1.296 VDC on the left and -1.296 on the right. With the amps inputs shorted the offset rises to 5.23 VDC on the left and -5.23 on the right. I'm not used to seeing offsets this high for DC amps. All my other amps measure very low. My Cyrus 2 measures 6mV and -12mV and my Outlaw measure 20mV. I wonder if I should make some adjustments to the input stage to see if I can reduce this offset?

[Mooly]

Hi,
Does it play O.K. no distortion etc. The offset's you are quoting are horrendous, it's not out of adjustment, there a serious fault, whether a design fault or component fault you can't say.
Have I read this right, 5 Volts across the speaker with the inputs shorted. You are putting your speakers at risk, the heating effect of half an amp plus, will overheat the voice coils.
You have come come up with a circuit I'm afraid, but it's weird both channels affected, perhaps there a common reference voltage somewhere thats misbehaving.
Regards Karl

[Tube_Dude]

Originally posted by tkwou
I purchased a Mission 777 power amplifier on Ebay last fall. It's a great sounding amp that matches well with my Mission 776 preamp. It is a completely symmetrical dual mono design direct coupled with no input or output capacitors.

Are you really sure that originaly the amp don't have capacitors, specially in the negative feedback side, and someone take them out to "improve sound"...

Or any servo that is not working properly, or also "missing"...

[tkwou]

I recently purchased a copy of the schematic on Ebay from another owner of this amp in Europe. The schematic is not a factory original. It was apparently drawn by a service technician and has an error in the feedback circuit which I have yet to clarify by openning up my own amp and trying to trace out the feedback loop myself. From the schematic that I have there are no capacitors in the signal path, even in the feedback loop. The design is completely mirror symmetrical but rather unconventional. Describing only the upper half, the input stage is an pnp emitter follower driven by half the negative rail voltage. This cascades into a npn common emitter stage which is biased in reference to half the positive rail through a resistor value which is apparently chosen at the factory. This stage then cascades into a pnp common emitter stage cascode which is loaded by a current mirror. The output of this cascode drives the voltage gain/driver stage which is three parallel npn transistors(2SD699) tied to the negative rail in common emitter configuration. The combined collector output of these parallel transistors then feeds a pair of p-type Hitachi 2SJ50 mosfets in parallel. The feedback loop is voltage divider which comes from the output back to the bases of the input transistors. It would seem to me that the DC offset is adjusted by changing the value of the resistors that bias the first common emitter stage. I've tried to put this circuit up on LTSpice but I can't get it to work because of the error in the feedback loop which I haven't had time to clarify. I think that I understand the circuit, unfortunately I don't have that much experience repairing or servicing amplifiers.

[Mooly]

Can you post the circuit.
Tube_Dude, I never thought of it being modded, that's a definite possibilty.
Regards Karl

[Tube_Dude]

Originally posted by Mooly

Tube_Dude, I never thought of it being modded, that's a definite possibilty.
Regards Karl

Mooly, from Ebay everything is possible...

[Mooly]

Oh my, you mean it wasn't as advertised people actually do that

[tkwou]

As far as I can tell it's all original. The amp looked mint inside. I didn't see any sign of modding but then I never took that close of a look. It looked identical to reference pictures that I have. The amp sounds great. Wide open presentation, dynamic, lots of detail. I drew the schematic up on LTSpice but there is a missing node in the feedback loop. I assume there must be a capacitor to ground missing otherwise there wouldn't be unity gain at DC. With the amp connected and running the DC offset sits steady at 1.296 VDC on the left and -1.296 on the right. These aren't particularly good figures so I'm worried about my speakers.

[tkwou]

Ok, I broke down and disassembled my amp this afternoon. I can't stand the thought of driving that much DC into my speakers. I pulled out the main board on one of the channels and was able to trace out the feedback circuit and check all the values on my schematic with the actual board. There were a couple differences in resistor values but relatively small. A 9.31K resistor comes off the source of the two N-power mosfets to the feedback loop. The sources of the N-power mosfet and the P-power mosfets don't meet until they come to the safety fuse. The output of the fuse goes directly to the positive speaker terminal and a second feedback loop with a 68.1 ohm resistor joins the 9.31K resistor. I guess this is the primary feedback loop with the second loop only coming into play if the fuse blows. I ran the circuit on LTSpice and it produced about 4V of positive DC offset. I threw an electrolytic capacitor onto the voltage divider resistor to ground and this corrected the DC offset. Unfortunately I don't know what this change will make to the stability of the amp. Also I don't know how accurate this model is to the real worldcircuit. I've attached a revised schematic.

[tkwou]

I found the resistors that are used to set the offset. On the LTSpice model I can completely zero out the offset. I will need to get some small pots that can replace these resistors. I won't need to touch the feedback loop.