Fisher 500-C Bias - A Case History

tcdriver

AK Subscriber
Subscriber
Fisher 500-C Bias

The Story: Recently, I pulled my Fisher 500-C out of the barn, where it had been stored since 1985. Cosmetically, it still looked to be in pretty good condition. I knew the challenge would be in getting it operating again, after 33 years sitting in the barn.

The first task was to test the B+ electrolytic capacitors. This was accomplished by disconnecting them from the circuit and slowly applying voltage while monitoring leakage current, using my trusty Sprague TO-6. As it turned out, I was able to reform all the original capacitors to where they showed low leakage up to their full rated voltage. Seeing has how they were still all good, I left them in place.

The second step was to carry out some modifications based on posts I had read here on AudioKarma. I added screen grid-stopper resistors and 10Ω cathode resistors for the output tubes. I also modified the bias supply, changing out the original capacitor and selenium rectifier and adding a bias adjust potentiometer, reducing the output tube control grid resistors from 330kΩ to 220kΩ and changing the four coupling capacitors feeding the output tube grids from 0.047µF to 0.068µF. A CL-80, current limit resistor, was added to the line input.

The third step was to slowly power up the receiver using my trusty Variac™. The good news is that as more voltage was applied, the Fisher awoke from its 33 year slumber and produced sound. Although, the FM tuner was a bit off, I was able to pull in radio stations in stereo. The less good news is that I discovered the output tube bias voltages and currents were not particularly well matched. The four output tubes were GE brand 7591A purchased sometime in the early eighties. The tubes were not a matched set when purchased and they were not well matched in terms of bias current for a given grid bias voltage. If I set the bias voltage to correctly bias the highest current tube, the lowest current tubes would be running about half current.

The fourth step was to modify my bias circuit to add individual bias voltage adjustments for each of the four output tubes. Once done, I now had a very good ability to adjust the individual bias current on each of the output tubes. Of course, once one can measure the bias current, one finds that it is not necessarily a set once and forget proposition. Over time there is some drift as the tube warms up and settles in and as the input AC line fluctuates the bias voltage and currents will also vary. At first, I recorded some readings manually, however; it was time consuming and a bit awkward.

Four Channel Bias Board:
Bias 4-ch (800x700).jpg

The fifth step was to automate the recording of the bias current VS time. For this task, I used a DATAQ DL-145 data logger. For useful readings I had to build a buffer / gain interface between the 10Ω current sense resistors and the DL-145 data logger. The end result is that I now have the ability to monitor the bias current over time.

Here is an example of a Bias VS Time chart:
Fisher 500C Bias VS Time.jpg
 
Last edited:
Great thread TC, lots of good help in the Fisher Forum if you need it, nice job on that 500C!
Thank you for the compliment and thank you for the recommendation for the Fisher Forum. I will see if the moderators can move this thread over there.

Is that board a manufactured item available for purchase? That's very neat work.
I built that board myself, so, to answer your question: no, it is not a manufactured item. I have slightly modified the original Fisher circuit to give more bias voltage range.

Here is a schematic:
Bias Circuit.png
 
Last edited:
Looks great Tim!

Couldn't you take it a step further and put an Arduino and some digitally controlled resistors in there and have it maintain your bias setpoint without ever needing adjustment?
 
An Arduino-based automatic bias control would be very interesting. Might have to be outboard due to size and heat......
 
Couldn't you take it a step further and put an Arduino and some digitally controlled resistors in there and have it maintain your bias setpoint without ever needing adjustment?
That is an interesting thought. If properly implemented, there would be a couple of advantages. The idle current would always be optimized, tube matching would be less critical and the chance of tube thermal run away could be eliminated. Of course, once developed, such an auto-bias circuit could be applied to many other amplifiers.
 
An Arduino can do a lot. If you got motivated you could probably predict the need to change tubes, add remote volume and power on/off control and support streaming in various formats. I started looking into it, but don't know enough about the coding to do it.

Other than streaming it would be a sort of parasitic add on that doesn't get into the analog signal path.

Yes, could be applied to many tube receivers.

Sorry to hijack your thread.
 
Last edited:
... I started looking into it, but don't know enough about the coding to do it.....
I do not really know what Arduino is. I was counting on you to do the coding. Is there an AK volunteer interested in such a project? Let us know.
 
Schematic showing some of the changes to the output stage:

Added 10Ω 1/4W cathode resistors between pin 5 of the 7591 output tubes and chassis ground to enable measuring cathode current.

Added 82Ω 1/4W resistors in the screen leads of each output tube to prevent oscillation and protect against shorts.

Added an individual output tube bias adjustments.

Changed R121, R122, R123 and R124 grid resistors from original value of 330kΩ to 220kΩ

Replaced coupler caps C77, C78, C79 and C80 between driver tubes and output tubes. The value was changed from 0.047µF to 0.068µF to keep the time consistent similar, now that the grid resistors were decreased in value.

Fisher 500-C Output Stage Schematic.jpg
 
Last edited:
So how often would you want the arduino to adjust the bias? If you do it too often I imagine you would hear it. I mean we manually adjust the bias and it can sit that way for months or years, right? With an arduino would it be necessary to check more than once ever 15 minutes or maybe once per hour? Would you want it to jump to the new value or would you want it to step toward the new value in small increments each cycle?

The arduino could check the bias voltage vs the cathode current too and you could set flashing LED's if the two values get too out of whack.

BTW, is "bias current" a real term w.r.t. tubes? I'm not an expert. Just trying to learn.
 
So how often would you want the arduino to adjust the bias?
Good question. I do not think that it would be too critical. Any time there is no audio signal the bias can be adjusted.

If you do it too often I imagine you would hear it. I mean we manually adjust the bias and it can sit that way for months or years, right?
Yes, if the bias is continually being adjusted, while music is playing, it will screw up the bias and have a negative effect on the sound. You are right in that in most cases, on many or most amplifiers, the bias is either never adjusted or only adjusted on a monthly, yearly or longer time interval. In the case of the Fisher 500-C, the bias voltage is preset and there is no adjustment. This can be OK if you have very well matched tubes that have a bias voltage requirement that matches the bias voltage provided by the original circuit. Many of the modern tubes sold as 7591 require a different, usually higher, bias voltage than the original circuit provides. This can cause the output tubes to run too hot which can result in early tube failure. Even in my case, where I was/am using original production GE 7591A tubes, the original bias voltage was insufficient to properly bias all four of the output tubes.

With an arduino would it be necessary to check more than once ever 15 minutes or maybe once per hour? Would you want it to jump to the new value or would you want it to step toward the new value in small increments each cycle?
A jump to the new value would probably be best if the adjustment is only being made every fifteen minutes or hourly. As long as the amplifier is in a quiescent state, no audio signal passing through, the adjustments could be made continuously.

BTW, is "bias current" a real term w.r.t. tubes? I'm not an expert. Just trying to learn.
Good question. I had not thought about it before. I have been using the term bias current for many years. In the context of an audio amplifier, I define bias current as the measured steady state current when no signal is applied. In the case of output tubes in a class AB design, like the Fisher 500-C, the current measured at the 10Ω cathode resistors will increase as an audio signal is run through the tubes. In a class A amplifier there is little or no change in operating current, audio signal present or no.
 
Last edited:
I did look into Arduino on the official website. I really requires a lot of power. I cannot imagine using any of the internal voltages in the Fisher 500-C to power it. There must be lower power similar devices that may have a chance of being used in an autobias scheme with the Fisher. Can anyone point me in the right direction?
 
Last edited:
The arduino itself doesn't require a lot. Arduino runs on: 9 to 12V DC, 250mA or more. I.e. Just a few watts.

Do you mean the bias network it would control needs more power?
 
The arduino itself doesn't require a lot. Arduino runs on: 9 to 12V DC, 250mA or more. I.e. Just a few watts.

Do you mean the bias network it would control needs more power?
No, the bias network should need very little power. Where will one get the 9 to 12V @ 0.25A in the Fisher 500-C? When I went to the Arduino website they specify a 9 to 12V @ 2A power supply. I could not find what the actual current draw of the module is. You are saying that it could only draw as little as 0.25A? To me that is still a lot of extra current. The only possible place I see to draw that kind of current, in the Fisher 500-C, is from the bias/phono filament supply.

I think a micro-controller with a 1 to 2mA current draw is what is needed. High speed operation is not a consideration for the bias control.
 
No, the bias network should need very little power. Where will one get the 9 to 12V @ 0.25A in the Fisher 500-C? When I went to the Arduino website they specify a 9 to 12V @ 2A power supply. I could not find what the actual current draw of the module is. You are saying that it could only draw as little as 0.25A? To me that is still a lot of extra current. The only possible place I see to draw that kind of current, in the Fisher 500-C, is from the bias/phono filament supply.

I think a micro-controller with a 1 to 2mA current draw is what is needed. High speed operation is not a consideration for the bias control.

You need to add a little power supply. A 120/12.6VAC filament transformer of the right current capacity will work just fine. Don't want to add load to the current filament transformer. You'll then need to build up the power supply rectification and filter; and a dropping resistor or voltage regulator. Shouldn't be hard to find a circuit on line to build from. The real challenge is to control the bias voltages - not to power the arduino. You can start with a simple wall wart until you build something more permanent. The wall warts are ubiquitous for arduinos.
 
Good to see the Arduino thread coming to life. There are other options such as Raspberry PI that can do the same job, but I think either Arduino or RP may be overkill for current regulation - I think there are dedicated voltage/current regulator circuits out there. I will ask around among the EE's at my work. Both Arduino and RP have enough IO that you could monitor performance of every tube in the receiver, add remote power on, send you a text when something is wrong and more.

I don't think the bias adjustment on the fly would affect sound since we are talking about a system that is pretty slow to react - think about warm up time. It's also fully programmable, so you would only have it try adjustment if it creeps out of a defined envelope, if it gets to a point where it is told a tube is bad it could shut down and send a message. Resistance steps could be very small too.

There are digital potentiometers that could be controlled by the Arduino to set the bias.

We need an electrical engineer and a software engineer...
 
Last edited:
Software engineer for 35+ years here. If you can tell me what you want it to do, I can program it.
 
Back
Top Bottom