Pioneer SX-727 Offset Issue

[ivandezande]

Got this guy with rail voltage on one of the channels. I replaced all transistors except for outputs on that channel (and now the bias is too high because of the higher gain parts but that's another problem), and now with matched KSA992's in the input stage of each channel, the bias won't drop below 200-150mV.
The hfe of the 992's was about 450 I believe. What gives?

 

[rcs16]

Probably it was these that were changed Q9-12? and with what? I am guessing you selected ksa1220/ksc2690
You can increase the value of R36(37) to correct this issue. is the pot (VR3,4) still 2K2? it can be increased as well to 5K to have the bias generator lower the voltage.

 

[ivandezande]

Probably it was these that were changed Q9-12? and with what? I am guessing you selected ksa1220/ksc2690
You can increase the value of R36(37) to correct this issue. is the pot (VR3,4) still 2K2? it can be increased as well to 5K to have the bias generator lower the voltage.

I did MJE15032/3 to be safe with thermal for the drivers, and then KSC3503 for 2SC1124, which might need to be KSC2690, I've heard varying reports.
Good to know, wasn't sure if increasing the resistance would just make it be able to bias more in the wrong direction. Glad I waited to order parts.

 

[ivandezande]

Wait R36/7 is for bias or offset?

 

[rcs16]

VR1,2 are offset
D4,5 are the temp sensors in the bias spreader. As they heat up they conduct more, developing a greater Vbe for Q7,8 which turns on more lowering its Vce, in turn lowering the Vbe for the push pull ( 4Vbe's)

 

[ivandezande]

VR1,2 are offset
D4,5 are the temp sensors in the bias spreader. As they heat up they conduct more, developing a greater Vbe for Q7,8 which turns on more lowering its Vce, in turn lowering the Vbe for the push pull ( 4Vbe's)

Yessir, got it. I'll probably just up the bias trimmers like you suggested and see where that gets me.
Don't mean to bother, but any ideas on the offset issue? As it goes +/- I really don't know what to change/omit to get it adjustable to zero.

 

[rcs16]

any ideas on the offset issue? As it goes +/- I really don't know what to change/omit to get it adjustable to zero.
What is your output offset range, from pot (VR1,2) CW to CCW extremes?
measure pins 5,6 to ground, pins 3,4
D3 clamps the -ve to -0.6V and the +ve is set by the voltage divider (R5,7), from the D1, 13V zener reference.

 

[ivandezande]

any ideas on the offset issue? As it goes +/- I really don't know what to change/omit to get it adjustable to zero.
What is your output offset range, from pot (VR1,2) CW to CCW extremes?
measure pins 5,6 to ground, pins 3,4
D3 clamps the -ve to -0.6V and the +ve is set by the voltage divider (R5,7), from the D1, 13V zener reference.

The closest to zero it can go is 150mV/200mV (left/right), measured at those pins. I replaced D3 with 2 1N4148's. Maybe I should replace the zeners.

 

[LesE]

Maybe I should replace the zeners.

If D1 cathode measures +13V and D2 anode voltage measures -13V, then the zeners should be good and not require replacement.

I replaced D3 with 2 1N4148's

D3 was replaced with two 1N4148's? Are they in series or parallel? One would have been more than sufficient. Two 1N4148's in series will result in a large offset with negative polarity. If D3 was shorted, it would result in a large offset with positive polarity.

The cathode of D3 must measure around -0.5V to -0.6V in order to obtain a usable range of offset voltage.

 

[ivandezande]

If D1 cathode measures +13V and D2 anode voltage measures -13V, then the zeners should be good and not require replacement.



D3 was replaced with two 1N4148's? Are they in series or parallel? One would have been more than sufficient. Two 1N4148's in series will result in a large offset with negative polarity. If D3 was shorted, it would result in a large offset with positive polarity.

The cathode of D3 must measure around -0.5V to -0.6V in order to obtain a usable range of offset voltage.

There it is then, thank you. I had 1S1212 = 2x 1N4148 in series.

 

[LesE]

In addition, using MJE15032/33 as drivers will often result in an inability to adjust the bias to 0 without circuit modifications due to their low Vbe. KSA1220/KSC1290 will likely result in lower bias.

 

[ivandezande]

The diode was indeed the issue. Dials in perfectly now.
I'll just be upping the trimmer value to deal with bias, I don't want to bother with heatsinks on the 1220/2690 combo.