• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

Improving the Fisher SA-100 with EFB II

@Pio1980,
Why are you discussing Williamson in stead of what @Coitboy2000 asked? Please have a look at his question as well as the title of this thread. Williamson topology is excellent and you points are well taken, but not the topic of the thread.
 
Last edited:
@Coitboy2000’s project is the Fisher SA-100 clone and his question was about suitable EL84 output transformers. @Pio1980, your post #116 took a tangent to Williamson and 6L6 amps. That was the reason for my question; perhaps I missed the relevancy.
 
My motivation is to try and use what I have…

A have several pairs old output transformers. Two were taken from an old Hammond H100 organ (6bq5 pp) and are push pull with 8k impedance into 8 ohms and probably 15-20W. They look really solid.
Given these circumstances IMHO I would build the Fisher SA-100 clone, using the Hammond H100 outputs, with the caveats in frequency response @nerdorama mentioned, and the need to tweak the feedback. If you like the potential well enough then acquire better outputs later, or just enjoy your Clone with Hammond outputs. Just build it with a future upgrade in mind in the component layout.

Good luck! And keep us informed.
 
@Coitboy2000’s project is the Fisher SA-100 clone and his question was about suitable EL84 output transformers. @Pio1980, your post #116 took a tangent to Williamson and 6L6 amps. That was the reason for my question; perhaps I missed the relevancy.
My mistake then, somehow thought he was cloning Heathkit WM-5.
 
Cold old Saturday here so I thought I would try to do a frequency spectrum analysis of both types of output transformers. Not sure if my testing methodology or equipment is suitable for testing transformers but the results look plausible.

I use the following setup
- signal source is my HP-8165a set to sweep from as low as 100hz up to 25kHz, 5V peak to peak terminated into 50ohms
- HP-35660a dynamic signal generator set for spectral analysis using both input channels. Measurement taken across the transformer outputs.
- outputs of transformers connected to 8ohm 50w wirewound resistor.

Results attached. I broke the sweep into 2 parts, 100hz or so to 6.4kHz and from 500hz or so up to 26kHz. Both the larger Hammond transformer and smaller Filmosound transformer were tested.

Photos should be self explanatory. Hammond seems to be 3db down at 20kHz, the Filmosound seems to be only 1db down at 20kHz.

How do these results look ? Decent transformers?

Richard
 

Attachments

  • hammond1.jpeg
    hammond1.jpeg
    120.9 KB · Views: 11
  • hammond2.jpeg
    hammond2.jpeg
    125.1 KB · Views: 10
  • hammond3.jpeg
    hammond3.jpeg
    124.3 KB · Views: 9
  • filmosound1.jpeg
    filmosound1.jpeg
    123.7 KB · Views: 8
  • filmosound2.jpeg
    filmosound2.jpeg
    133 KB · Views: 9
  • transformers2.jpeg
    transformers2.jpeg
    132.9 KB · Views: 9
  • transformers1.jpeg
    transformers1.jpeg
    160.5 KB · Views: 10
It is unclear how well your tests captured the actual capabilities of these transformer, and how that translates into audible goodness. The frequency responses do not appear to be extended, but perhaps they will sound fine when treated to some frequency extending GNF.

For comparison read about the frequency response curves of the clones with Heathkit/Stancor AA-100 outputs in this post #28 of @Dave451 and the frequency oscilloscope shot in the following post #29. They are flat from 20 Hz to 40 kHz.

 
Last edited:
It is unclear how well your tests captured the actual capabilities of these transformer, and how that translates into audible goodness. The frequency responses do not appear to be extended, but perhaps they will sound fine when treated to some frequency extending GNF.

For comparison read about the frequency response curves of the clones with Heathkit/Stancor AA-100 outputs in this post #28 of @Dave451 and the frequency oscilloscope shot in the following post #29. They are flat from 20 Hz to 40 kHz.

Ok. Will need to build an amp first to run similar tests :)

Curious whether the results are encouraging
 
Can you also get a phase plot with the response plot? Where you run into gremlins with output transformers in a feedback loop is the phase shift with frequency. If it shifts too far away from 180 degrees out of phase the feedback can become positive which turns it into a very efficient oscillator. Normally there is some bandwidth limiting in the amp to make it behave but if the transformer shifts phase too close to the audio band, that filter starts to affect within the audio band and thats not great.

I've used Hammond organ transformers before, not this specific one, but they just didn't do highs all that well. I put them on a budget console amp to get an 8 ohm output. I was able to get acceptable results out of it but they're not great by any stretch. Physically limited on space with that amp so there are only so many options. Mine were used on the AO-44 and AO-35 I think, 6bq5 or 6BM8 tubes.
 
Hi Dave, I have one additional question regarding R52 and R54 on your schematic, I assume they are the wires coming from the 4.7K resistors that are tied to the ends of the bias pots, correct? My schematic from HiFi engine has a different number for those two resistors, but the amp has the same white wire coming from that connection of those two resistors. Just wanted to get this right before applying voltage. Thanks!
 
Can you also get a phase plot with the response plot? Where you run into gremlins with output transformers in a feedback loop is the phase shift with frequency. If it shifts too far away from 180 degrees out of phase the feedback can become positive which turns it into a very efficient oscillator. Normally there is some bandwidth limiting in the amp to make it behave but if the transformer shifts phase too close to the audio band, that filter starts to affect within the audio band and thats not great.

I've used Hammond organ transformers before, not this specific one, but they just didn't do highs all that well. I put them on a budget console amp to get an 8 ohm output. I was able to get acceptable results out of it but they're not great by any stretch. Physically limited on space with that amp so there are only so many options. Mine were used on the AO-44 and AO-35 I think, 6bq5 or 6BM8 tubes.
I cannot get a phase plot with my DSA.
 
Quick question, do I have the 10 ohm resistors to each pin 3 and going to ground just that way? No switches or plugs, just the cathode (pin 3) going to ground via the 10 ohm sense resistors so that the cathodes are connected no directly to ground, but through the 10 ohm sense resistors, correct? Thanks!
 
Yes. I also removed the damping circuits which then frees up two terminals on each output barrier strip to use for monitoring the cathode current of each output tube in the relevant channel each strip serves. I ran a lead from each pin 3 to the appropriate (new) monitoring terminal on the output barrier strips, and then installed the 10 Ohm resistors from each monitoring terminal to a nearby chassis ground lug.

I hope this helps.

Dave
 
Yes. I also removed the damping circuits which then frees up two terminals on each output barrier strip to use for monitoring the cathode current of each output tube in the relevant channel each strip serves. I ran a lead from each pin 3 to the appropriate (new) monitoring terminal on the output barrier strips, and then installed the 10 Ohm resistors from each monitoring terminal to a nearby chassis ground lug.

I hope this helps.

Dave
Thanks Dave, going to finish this tonight! BTW, I am assuming I am to set each of the cathodes by measuring 0.3 volts across each of the 10 ohm resistors, correct?
 
Well geez, it's been nearly what, 13 years since I published the EFB™ modification for the Fisher SA-100 -- but like all things that change with time, so it is with this modification as well. As far as EFB action goes, the new version of the modification doesn't cause it to work any better than when first published, but there are a number of worthwhile improvements resulting from the new version none the less. Collectively they include:

1. Use of a better mosfet family with regards to mounting. The new mosfet is plastic, so it can simply be bolted to any convenient chassis hole for heat sinking purposes. With nothing more than a nut and bolt and a little dab of thermal grease then, you're good to go. The new mosfet also eliminates the need for the external Zener diode used in the original modification.

2. The mosfet still dissipates about 1 watt under quiescent conditions like before, but now only about half of what it did before under a full continuous boil in both channels (~ 2 watts).

3. In the spirit of bullet proofing the design -- similar to that which I did some time ago for the cathode regulator version of EFB -- there is now more protection built in for the mosfet as well as for the EFB bias supply regulator. There have been no known failures of these components with the original modification, but more protection is always good in case a tube decides to go sideways.

4. The new design also includes the use of a small Hammond 156R choke that can be mounted under the chassis on the end side panel near the output terminal strips. The choke can actually handle the full power current draw in both channels simultaneously, and makes itself known in two ways: At turn on, there is now no hint of 120 Hz hum as the output tubes warm up and balance out their current draw to cancel out the significant 120 Hz ripple voltage present at the OPT CT leads in the stock design. As well, less ripple in the B+ supply translates into less interaction with the audio signal passing through the output stage, and therefore reduces distortion.

5. Output stage plate voltage is elevated slightly as a result of the revised modification, which increases power output slightly -- on top of the increase already produced from the original EFB modification. EFB quiescent current draw remains the same, with the output tube plates dissipating little more than 9.5 watts each. This is just under 80% of the Design Center 12 watt Pd rating for the tubes.

Other changes on the revised schematic include:

A. The new schematic is more complete, now showing the driver stage B+ supply components as well.

B. The new schematic now specifies the CE C30X4-475 as the replacement can cap of choice. The original can cap was only rated for 450 volts, and therefore could also be rated for 85˚C operation. But this voltage rating is inadequate with today's higher line voltages and reduced current draw afforded by installation of the EFB modification. It is tempting to use a modern 525 volt can, but such cans are only rated for 55˚C operation, and therefore would not be advisable for use with the stock design since the stock amplifier runs quite warm. With the EFB modification however, the amplifier runs significantly cooler, allowing the 65˚C rating of the suggested can cap to operate with adequate temperature reserve.

C. The revised modification recommends the addition of a CL-80 inrush current limiter be installed between the AC line cord and the fuse post.

D. Finally, the revised modification also includes a heater bias circuit as the design is quite hard on the heater/cathode insulation in the phase inverter section of the driver tube. Even the official Fisher schematic of the SA-100 shows a heater/cathode voltage differential of 110 vdc, exceeding the Design Maximum rating for the 12DW7/7247 tube by 10%. Sams indicates it is also exceeded, although by only 5%, but both of these figures are based on operation from 117 vac. With the heater bias in place, the heater/cathode junction operates at only 72% of the rated allowable maximum voltage -- and this is based on operation from today's higher line voltages, and with the EFB modification in place.

For those lucky enough to own a Fisher SA-100, they well know that it is one of the best sounding 6BQ5/EL84 amplifiers ever produced. EFB doesn't change the character of the SA-100, but simply allows the amplifier to supply more of it, with longer output tube life, lower distortion, and cooler operation. And now, the modification itself is even better protected as well.

Happy listening!

Dave

Below -- Shows where the Hammond choke can be mounded under the chassis. The inner most mount screw passes through a cooling hole, while the outer most mounting screw requires a single small hole to be drilled in the side of the chassis to facilitate the screw:
SAM_2592.jpeg

The revised modification:

Fisher SA-100 EFB Modification 2025.jpg
 
Last edited:
Back
Top Bottom