Picked up Fisher Electra VIII console this weekend. Worth repairing?

Are these direct coupled outputs? On something this old I'd expect cap coupling. If you've got DC on the outputs with cap coupling, that speaks of a leaky output cap.
 
The DC offset is a function of the bias adjustments. Unlike later designs, there is a separate bias adjustment for the upper and lower transistors. This could be caused by your 17 mV bias if the other transistor is 26 mV. If you make adjustments, be careful because an intermittent open on one of the pots will cause problems.

There is a better way to wire the buck transformer. Connect the upper AC line to the upper end of the secondary and connect the upper end of the 110 V output to the top of the primary. The primary to secondary jumper will then be from the top of the primary to the bottom of the secondary. The input sees both windings in series with the output across only the primary.
 
The DC offset is a function of the bias adjustments. Unlike later designs, there is a separate bias adjustment for the upper and lower transistors. This could be caused by your 17 mV bias if the other transistor is 26 mV. If you make adjustments, be careful because an intermittent open on one of the pots will cause problems.

There is a better way to wire the buck transformer. Connect the upper AC line to the upper end of the secondary and connect the upper end of the 110 V output to the top of the primary. The primary to secondary jumper will then be from the top of the primary to the bottom of the secondary. The input sees both windings in series with the output across only the primary.

I realize that DC offset is caused by difference in individual transistor idle current. This is partly due to transistors themselves. I switched positive and negative sides around and bias reading is less apart, reflecting in lower DC offset (60mv) at 112vac input voltage. Close enough?
Bias pots have been disconnected and temp compensating diodes and emitter resistors added at some point of unit's life. Nothing to adjust there. By the way, after one very tragic attempt to adjust bias on Pioneer SX-929 I always clean pots before I touch them under voltage and I make first pass of adjustment under DBT while observing both, bias reading and lamp glow. I make sure wiper does not loose contact and then I set it low and go for adjustment from the wall.

Could you please post a drawing of the buck schematic you suggesting?

Thanks
 
If the bias has been changed to a non adjustable circuit, then you are subject to component tolerances. I had one that was good on one channel but quite a bit off on the other. That's when my bias modification was designed to return to an adjustable circuit. If the bias modification is similar to what they did for the 600-T, you might try adding a parallel resistor to slightly change the resistance value in the bias string. As an example, a 3300 Ω resistor added in parallel with a 270 Ω will bring the total down to 250 Ω. Since I don't have the exact schematic or the specifics of where the problem is, you would need to determine which resistor might need the change.

Do your transistors have a color dot on the flange? This indicates the beta group. They should be the same color or at least each pair the same color.

The transformer wiring is shown and explained at the following link. It is about halfway down the page.
http://sound.whsites.net/articles/buck-xfmr.htm
 
If the bias has been changed to a non adjustable circuit, then you are subject to component tolerances. I had one that was good on one channel but quite a bit off on the other. That's when my bias modification was designed to return to an adjustable circuit. If the bias modification is similar to what they did for the 600-T, you might try adding a parallel resistor to slightly change the resistance value in the bias string. As an example, a 3300 Ω resistor added in parallel with a 270 Ω will bring the total down to 250 Ω. Since I don't have the exact schematic or the specifics of where the problem is, you would need to determine which resistor might need the change.

Do your transistors have a color dot on the flange? This indicates the beta group. They should be the same color or at least each pair the same color.

The transformer wiring is shown and explained at the following link. It is about halfway down the page.
http://sound.whsites.net/articles/buck-xfmr.htm

Transistors have same color dot but still made the difference after the swap. I did clean sockets and transistor legs and re-greased them before taking first measurements.
Thanks for the link
 
If the bias modification is actually per the schematic posted by Larry, then my suggestion of a parallel resistor across the 330 Ω will not work because the diode will determine the voltage at the bottom of the transformer winding. It would be necessary to add a small resistance in series with the diode on the portion of the circuit with low bias.
 
Bias modification is per schematic:

upload_2019-1-4_23-31-46.png


Is 60-70 mv of DC offset significant enough to worry about? If so, what resistance range do you suggest? Could I use trimmer to dial on resistance value and then replace it with resistor?

One thing is puzzling me. I have both channels running into 8 ohm 50w resistors. Channel with uneven / low bias produces output equal to good channel but heatsink stays significantly cooler then on good channel. What am I missing? My dummy resistors are mounted on single heatsink so I can't be sure if they are getting different temperature. They seem to be same though.

I took a quick look at a frequency response and it's not that great. 60Hz - 10KHz flat. Below and above falls pretty quickly on the scope but I didn't take precise dB reading. Probably good enough for console sound?

I must add that tuner is impressive. I can pull 6-7 FM Stereo stations in the basement without antenna or even a wire connected. So far most sensitive I've seen.
 
The lower bias is resulting in slightly less heat but the difference should not be enough that you could notice it by touching.

60-70 mV offset is probably acceptable, though I'd prefer to see 10 mV or less. I'm presently fine tuning bias and offset of a new build and can dial the offset to ±2 mV but it does drift.

As for adding resistance, the 330 Ω bias resistors are passing approximately 72 mA. To increase the bias voltage from 17 to 24 mV, an increase of 7 mV, the added series resistance would probably be on the order of 0.08 Ω. Yes, such a resistor is available but this may not be worth the trouble. Don't add a trimpot because the residual resistance of the wiper is probably greater. If you have a pile of diodes, most certainly either germanium or Schottky, you could substitute. Another possibility is that the emitter resistor is out of spec. That would change the bias AND give a false reading.
 
I understand that lower bias would result in cooler heatsink, mostly at idle. Once signal is present and amplifier output is near full power for some time, both channels show same output level - heatsink temperature should be very close.
It is not. Any ideas?

I should have some 1N60 germanium diodes. Will they work? Aren't those Fisher Mod diodes temperature compensating units though?
 
1n60's won't work. The RCA diode is a Temperature compensating Diode. No longer in production but can be had online for about $20-30 each. Give me a couple days to look up the RCA # (it's something like 2628 but that's not it.) I've got a couple of 49A and 59A's (all working) with the Diodes in them. Just a matter of digging out the amps in the storage unit and locating the #'s. I think there is a thread by retrokid from back in 2014 or 15 detailing his quest for a few of these for his Ambassador. I'll look there 1st.
 
Well I had 2 of the numbers right. THE CORRECT # for the TEMP COMPENSATING DIODES is RCA 1N2326.

Like I said Earlier they are Not produced anymore and will likely have to be sourced from old stock stockpiles from Electronics Salvage houses.
 
Larry, I found a pair for $32 for sale. I took voltage drop across all 4 diodes yesterday and they all show 0.245 mV - 0.255 mV. Emitter resistors all show as 0.9 ohm on my meter which shows 0.2ohm with leads crossed. Close enough. I will move transistors from one channel to another and see if bias/DC and heat difference will follow. If not - will order pair of diodes. Does it sound like a good plan?

I found data sheet for those diodes if anyone needs it
 

Attachments

  • 1N2326.pdf
    32.8 KB · Views: 3
Concur on the diode part number. 1N3754 was also used by RCA in similar applications. This circuit is basically a design from an early 60s RCA manual. However, on the 600-T, the diodes were nowhere near the heat sinks, so they could only compensate for ambient temperature. Before purchasing diodes, you could also swap them and see if the problem follows a diode. Use an alligator clip on the leads as a heat sink while soldering as these are germanium. Get the iron hot and get on and off as quickly as possible.
 
I swapped transistors between channels.
Results are:

"Bad" channel with low bias / 60mV DC offset and cold heatsink now shows close 27mV bias across emitter resistors and DC offset is under 20mV
"Good" channel shows around 31- 32mV bias between emitters and DC offset is around 35mV.
Heatsink temperature is closer but same heatsink is still cooler. Warmer heatsink is closer to power resistors and power supply which I think may influence temperature difference between heatsinks on the top of bias being a little higher.

So, first I swapped transistors within the "bad" channel and DC changed from 100mV to 60mV while bias got closer but was still different.
Second, I swapped all transistors from channel to channel. Pos to pos and neg to neg. Bias on the "bad" channel is now equal but lower then on the "good" channel.
I think I have tired transistors more then I have a problem with temp compensators. As I noted in one of my posts, dropping resistor for remote speaker connection is open on the "bad" channel. I think, previous owners had something inappropriate or shorted hooked-up to that plug that possibly got the channel overloaded until resistor gave out.
I'll sleep on it but today, I'm thinking to leave it at that. 20mV and 35mV is much better then 60some mV of DC offset.
 
Concur on the diode part number. 1N3754 was also used by RCA in similar applications. This circuit is basically a design from an early 60s RCA manual. However, on the 600-T, the diodes were nowhere near the heat sinks, so they could only compensate for ambient temperature. Before purchasing diodes, you could also swap them and see if the problem follows a diode. Use an alligator clip on the leads as a heat sink while soldering as these are germanium. Get the iron hot and get on and off as quickly as possible.

Fred--

Your post came-in while I was typing my updates.
Temp compensating diodes in this chassis are also mounted inside of the chassis and sensing ambient temperature.
Please let me know your thoughts on my experiments and results above.
 
They may not have known much about temperature compensation in those days. I think the diodes were used to get a constant voltage without adjustments and any temperature compensation was coincidental.

Offsets of 20 and 35 mV and bias of 27 and 32 mV is probably as good as you can do without a bias control redesign. Those numbers should be good enough unless you are a perfectionist. Differences at this point would probably not be audible. Also, bias will change when everything gets warm and temperatures have stabilized.
 
without the diode mounted to the heat sink I doubt you're getting all that much temp compensation anyway. The stuff that I've owned that actually had compensating diodes mounted them to the heat sink directly next to the output device.
 
I agree about diodes hanging in the air. All I've seen until this one were mounted either in the top of the driver transistor or in the output heatsink.
I marked transistors so I can go back to last best configuration and moved them around trying to find optimal reading. In the end I settled on my last combination that Fred mentioned in his post.
I do not hear any distortion at low levels. Output is even. Heatsink temperature difference is most likely due to cold channel being further away from all hot resistors and tubes.
I'll be working on restoring cabinet appearance and will put it back together soon. Hopefully this week.
 
Back
Top Bottom