Newly-Acquired Fisher 400

Thanks, Dave! The latest set of readings, including CGR output: B+ 416 VDC, SG Out: 302 VDC, IBAM Out: -17.7 VDC, CGR Out from EFB: -22.3 VDC. So I've driven down the CGR fairly negative, with a net 4.6 volt "drop" across the IBAM. It has the 22K resistors installed instead of 10K. Dave
 
BRIEF TIMEOUT to show the next project (cold weather is coming on!). Picked up a Fisher 500B today from CL in working condition and not in bad shape (I have brights for the knobs, at least!). Was playing into a pair of large Advent speakers and sounding quite good when I picked it up. Good stereo FM, eye closing totally on close by strong FM signals, excellent sound on CD input. Dial cord has slipped. Of course, noticed an odd tube right away--turns out to be a Fender 6V6 stuck in among what appear to be Westinghouse 7591's! I didn't even know that would work! It was playing, though, and sounding good. I haven't been underneath yet to see if any modifications were made to use the 6V6GTA. The owner said he 'had a tech go over it' and that it had played reliably and well since. One tone control 12AX7's replaced with Russian Tung-Sol reissue. Phono tubes are "Made In Holland" Fisher labels. Other 12AX7's are Telefunken ribbed plates, except for a GE 6201 (12AT7) and TF 12AT7 in the multiplexer. I'll let you know what I find, but first to continue with the 400! Dave
 

Attachments

  • IMG_2686.JPG
    IMG_2686.JPG
    73.1 KB · Views: 49
  • IMG_2676.JPG
    IMG_2676.JPG
    87 KB · Views: 47
  • IMG_2678.JPG
    IMG_2678.JPG
    77 KB · Views: 44
  • IMG_2679.JPG
    IMG_2679.JPG
    63.6 KB · Views: 43
The 6V6 doesn't work -- and more importantly, can't work. Plugging a 6V6 in probably won't hurt anything in the set except the heater balancing resistors if the heater/cathode insulation within the 6V6 breaks down -- which it could since the cathode has over 350 volts applied to it in this instance. But even if the socket was rewired to accept the altered pinout of the 6V6, there's not enough bias voltage in the set to keep the tube from burning itself up rather quickly. Most likely, you were just hearing the other 7591 of that channel while it was still operating in class A -- which it will for the first couple of watts or so, while the 6V6 was just looking good with its heater all lit up, but doing nothing.

Dave
 
Yep, after a look at the 6V6 data sheet I sort of thought so too. Hopefully the unsupported 7591 wasn't sacrificed. I've got a good set of 7591a's I can use in due course as needed.
 
I installed the EFB/IBAM and, after triple checking the wiring, I fired the 400 up without the output tubes. Nothing amiss, B+ and screen a little high without the output tube load, but OK. So, put the tubes in and crossed my fingers and fired it up. Did fine, but apparently I'm not done yet. Bottom line: the tubes are so cold running at the lower screen voltage, I couldn't get up to 210 mA quiescent current despite turning "down" the bias (making it more positive, i.e.fewer minus volts) all the way. Here is the data:

Line Voltage: 121 VAC (before CL-80), 119 VAC after CL-80. AC to EFB supply (from DC heater/bias supply): 25.1 VAC
B+: 415 VDC EFB Screen: 304 VDC EFB CGR out: -21 VDC
Roughly balanced at the tubes: Bias voltage R: -15.5 VDC each Cathode voltage: 185 mV each Plate current: 15.5 mA
Bias Voltage L: -15.7 VDC each Cathode voltage: 176 mV each Plate current: 14.6 mA

So, I think I'm going to have to go back and either 1) reduce the 68K NFB resistor to 47 K again and/or 2) reduce the two 22K in the IBAM back to 10K. I guess you had a reason for putting it at 47K, Dave. Where I went wrong was centering the EFB/IBAM bias to the tubes at the current normal level of around -17.5 VDC. Apparently with the screen voltages so much lower, I need a more positive grid bias to bias the tubes to the recommended 21 mA per tube.

Sound right?
Dave
 
Dave G: just re-read your prior post about needing -15 VDC to get 21mA on your tubes. I expect I'll be close to that, maybe a little more positive given the lower B+. some of us (read ME) are just slow learners. Dave
 
You're doing great! But you surmised correctly. The screen grid is in fact a grid -- just like the control grid is. Therefore, it is the lower screen grid voltage coupled with the lower control grid voltage that reduces the quiescent current draw down to the 21 ma level. You're very close!

Dave
 
Installed the EFB/IBAM module in the Fisher 400 today after getting the NFB resistor finally right. It was like the Three Bears: 47K was too small, 68K was too big, but 56K ohms was 'just right.' Looks like -16.0 VDC to -13.5 VDC with the IBAM pots centered with the tubes in. Was able to bias the tubes now to 21 mA quiescent cathode current for each tube at grid bias of -15.1 to -15.3 VDC, depending on the tube. EFB CGR was -19.7 VDC, B+ = 418 VDC, screen = 304 VDC, Plates = 415 VDC.

That gives 18 mA idle plate current per tube (assuming 3 mA for the screen grid), and that translates at 415 V plate voltage to 7.5 W quiescent dissipation per tube; about 39% of 19W max. Pretty cool!

The heat sink on the MOSFET got only mildly warm after a fair amount of testing and operation.

I adjusted the phase inverter using an old Scott method: at 3/4 output into an 8 ohm load, adjusted the PI for the minimum (null) of 2nd order harmonics. Then checked the power output at clipping (1 kHz-single channel driven) as follows:
-Left Channel--13.7 VRMS gives 23.5 WRMS with about 1.5% THD (below)
-Right Channel--13.5 VRMS gives 22.7 WRMS with about 1.2% THD.
Plates were 416 VDC, screens at 306 VDC no signal and 297 VDC at full steady output.

I did the 'rough and ready' THD calculation by measuring the peaks at the different harmonics using FFT on my scope (no distortion meter) and used the accepted formula for the calculation. This is, of course, higher than what Dave G got, but I'm can't recall what power level he measured. I'll show the scope output in the next post. I wouldn't bet the ranch on these numbers, but the difference is probably real based on the scope pictures.

This is basically the same output as the pre-EFB measurement I made (prior post).

Sound check was good, so I decided to remove the 150K 'noose' from the phase inverters (R115 and R116). Clipped one end and covered it, re-checked sound. Hard to say with bench speakers, but sounded more vigorous at a given volume level.

I re-set the phase inverters with the above method, re-checked the cathode current for 21 mA quiescent and re-measured power. It appears I lost a couple of watts in the transition?:
-Left and Right Channel both--13.0 VRMS for 21.1 WRMS.

I also did some measurement with the scope and confirmed that the left channel has more THD than the right (next post).

That said, I put the bottom plate back on and took it upstairs for a listen through the Klipsch KG4's. It sounds marvelous, numbers notwithstanding!

More on the measurements in the next post.
Dave
 

Attachments

  • IMG_2699.JPG
    IMG_2699.JPG
    93 KB · Views: 42
  • IMG_2700.JPG
    IMG_2700.JPG
    96.9 KB · Views: 38
Here are scope shots before 'cutting the noose' on the PI. The first picture is LEFT channel. My calculated number notwithstanding, you can see that there are a number of harmonic peaks, with the biggest being 3rd, 5th and 8th. Note that 2nd order is gone because of the way I adjusted the phase inverter.

The second is the RIGHT channel. Getting a little less power before clipping, similar pattern of harmonics. The differences in amplitude of the harmonics come out better with 50 or 100 mV per division scale on the blue FFT channel, where I did the quantifying.

The next two are AFTER removing the PI resistors. I didn't re-calculate the THD numbers. Note that (unfortunately) the x and y axis scales are different that the pre- pictures, but you get the idea.

The third picture shows the LEFT channel after removing the PI resistor at the onset of clipping at 13 VRMS. The PI was re-set using the same method as the first set of shots. You can see the left channel has very significant 2nd, 3rd, and 4th order peaks and a significant 5th.

The RIGHT channel (fourth) shot shows the same 13 VRMS at incipient clipping, but only a 'shoulder' 2nd order peak, and significantly reduced 3rd and 4th order peak and few others.

Thoughts appreciated on why this would be?

NEXT, I took some square wave shots in the following post.
 

Attachments

  • IMG_2701.JPG
    IMG_2701.JPG
    109.9 KB · Views: 33
  • IMG_2703.JPG
    IMG_2703.JPG
    106.9 KB · Views: 28
  • IMG_2708.JPG
    IMG_2708.JPG
    100.8 KB · Views: 31
  • IMG_2706.JPG
    IMG_2706.JPG
    115.6 KB · Views: 29
And, finally, a couple of shots of square wave input at full output power. The measurements are made with the tone controls visually centered, at 1 kHz one channel at a time with the balance control full over to the channel being measured.

The first two are the LEFT, then RIGHT channels at full output BEFORE removing the PI resistor "noose." I'm not an expert by a long shot in reading these (or the prior ones, either), but I think you can see a fair amount of HF roll off. You can make the wave almost perfectly square by rotating the treble control toward full.

The second two are the LEFT, then RIGHT channels at full output AFTER removing the PI resistors.

To my rookie eye, it looks like there is significantly less roll off in the AFTER shots, compared to the before. Recall that both these are with EFB installed at 21 mA quiescent current on each tube. The difference is the removal of the PI 'noose' resistors. I HAVE NOT made any further mods to the power amp yet, beyond EFB/IBAM and 'removing the noose.' I also realize that I have not brought down the plate voltage of the audio amp stage, as per Dave G's recommendations yet, so I'll need to do that as well.

Well, I may have this all totally screwed up but, as I said before, the thing still SOUNDS really good right now and this is how we learn, I guess.

I WOULD APPRECIATE ANY AND ALL COMMENTS ON THESE OBSERVATIONS, PARTICULARLY 1) WHY I'M SEEING LESS POWER AT CLIPPING AFTER REMOVAL OF THE P.I. RESISTORS AND 2) WHAT COULD ACCOUNT FOR THE DIFFERENCE IN THD BETWEEN THE CHANNELS and 3) IS THERE REALLY AS MUCH DISTORTION AS I THINK I'M SEEING, OR AM I JUST READING THINGS WRONG?

Thanks much!
Dave
 

Attachments

  • IMG_2705.JPG
    IMG_2705.JPG
    100 KB · Views: 18
  • IMG_2704.JPG
    IMG_2704.JPG
    107.4 KB · Views: 17
  • IMG_2709.JPG
    IMG_2709.JPG
    107.1 KB · Views: 18
  • IMG_2707.JPG
    IMG_2707.JPG
    107.1 KB · Views: 17
Dave -- Since I published my original comments about the Fisher phase inverter noose, I have come to work with it more over the past year or so, and can report these additional findings:

1. The stock circuit with the noose in place often does not work well with modern (at least) Russian tubes. The noose causes the tube's operating point to be located right on the edge of producing any reasonable linearity based on the characteristics of the original American piece. While the Russian tubes are very good tubes when operated at any reasonable operating point, their performance at (at least) the lower end of the grid curve seems to move towards cut-off faster than it does in the American pieces, making the performance of the Russian tubes in the Fisher setting quite iffy.

2. I had originally surmised that the noose was an output tube protection measure, but I have now come to believe that that product of the noose is just a desirable by-product. The real issue is that Fisher was trying to maximize the gain of the AF amplifier stage in the power amplifier section, which would automatically increase the NFB and therefore improve performance, and all without increasing parts count. This required using an elevated plate voltage for the AF Amplifier stage -- which -- because it's direct coupled to the inverter -- caused bias problems for the inverter stage. In short, if the plate and cathode resistors of the inverter stage are made small enough to help provide a low output impedance from the stage, then with the high plate voltage direct coupled into the inverter's grid, it cause the inverter grid to actually become positive over the potential of the cathode, producing grossly non-linear operation and distortion.

Enter the noose. It works to carry some of the current load passing through the inverter's cathode resistor, that otherwise the tube would have to pass in full -- but if it did, that wouldn't leave any current flow capability left within the tube to develop an audio output when a signal is applied. By the noose carrying some of the current load then, it allows the inverter section to bias itself normally. But of course, the noose being what it is, in the process of it allowing the inverter stage to bias itself properly, it also acts to very much reduce available output from the inverter as well. So with the noose, it throws the operating point all the way to one end of the scale and the problems that produces. But if you remove the noose, then it throws the operating point all the way to the other end, with the problems that produces as well.

3. The real answer is to remove the noose (restoring adequate drive), partially bypass the cathode resistor of the AF Amplifier stage (to achieve the increased gain level), and power the AF Amplifier stage from a lower B+ voltage source. That will then lower the plate voltage of the stage. This causes the inverter section to reduce its current draw and bias itself normally -- all as was done in my original modification for the 400 receiver. In the 400's bigger brothers however, the overall B+ is slightly higher, meaning that if the noose is removed but the AF Amplifier stage B+ supply is not reduced, then the inverter simply runs out of gas well before full power is reached, causing significantly higher distortion -- and reduced power output.

All of this is the long way of saying that by reducing the B+ to the AF Amplifier stage (by moving it to the next lower voltage tap), it should resolve both problems at once in your 500C: Distortion will definitely go down, and power output will return.

I hope this helps!

Dave
 
Dave,
Thanks very much for taking the time to provide a rich answer to my questions and observations. Your thoughtful feedback is really helping me understand the theory behind these improvements and tube operation in general, and is much appreciated.

So, I'll go back in and re-read and then complete the power amplifier/driver mods for my 400 described in your 2013 post called "Improving the Fisher 400." I'll take a new set of measurements and post here. Please let me know if there's anything else I need to consider as I make this next set of mods, in addition to the 2013 post and the above.

Again, thanks Dave!
Dave
 
Finished the DCG-recommended PI/Driver and Power Amplifier mods. Pictures below are 1) PI/Driver/Output area with the EFB/IBAM board rolled and 2) with the board in place. If you look closely, you'll see that I rotated the C52 replacement cap one tab counterclockwise. I did this inadvertently and had soldered a tab down before I realized it, so everything is adjusted accordingly on the hookups. Mods completed are:

1. Changed B+ source for 12AX7 Audio driver plates (V12/V13 pin 6) to the next lowest voltage stage in the string; i.e. from 320V on the original schematic (355V on Dave's schematic) to 310V source (mis-labeled "390V" on original schematic). NOTE: as Dave notes, you DON'T want to do this for the phase inverter plates at pin 1 on V12 and V13. You want to leave these on the "355/320 source". Since these were tied through their respect 47K resistors to the same tie strip point as the audio stage plates, I had to clip the PS end of the 47K plate resistors and wire them directly to their original 320/355 supply point with a new wire for one and directly for the other.
2. Removed 500K AC balance pots (R107 and R108) and associated network 120K resistors (R111 and R112). Added 5K trim pots and new paralleled 56K/180K resistors (1/2 W) between PI cathode pin 3 and chassis ground on V12 and V13.
3. Added 22 uF 16V electrolytic caps across audio driver stage 1.2K cathode resistors.
4. Removed 2 pF disc caps (C89/90) across Audio/PI grids (between pins 7 and 2)
5. Removed 18 pF stability disc caps C97/98
6. Replaced 47K audio stage input resistors with 10K precision metal film units (1/2 watt). (NOTE: I have not yet rewired the 'space expander' jacks in the unit, so no 1 meg resistor across the audio input).
7. NOTE: I had already removed the 150K "noose" resistors R115 and R116
8. Removed NFB network components, including 2.7K R113/R114 and 560 pF disc caps C91 and C92, as well as 330 pF RFI disc caps C99 and C100. Replaced with 1.2K 1/2W and 390 pF 1KV mica caps for each side connected to 4 ohm output tap using new wire connected to the tie strip. Re-wired the 16 ohm OPT output directly through the original wires to the output terminal.
9. Added new step network (27K 1/4W + 100 pF mica caps in series) across the 390K audio stage plate resistors (pin 6, V12 and V13).
10. Replaced with EFB/IBAM board and triple-checked the connections. A lot of changes in this section of the 400!

Sound checked the output stages into the bench speakers. Sounded fine. Watched the bias levels adjust with big swings in output. No distress on the plates and everything played well for 30 min. Going to do some measurements (power, distortion, square waves) before moving to the next stage.
Dave
 

Attachments

  • IMG_2719.JPG
    IMG_2719.JPG
    161.1 KB · Views: 49
  • IMG_2722.JPG
    IMG_2722.JPG
    180.9 KB · Views: 45
OK, I put the amplifier on the scope after completing the power amp/PI/Driver mods as above. I set the phase inverters using the "resistance" method Dave G outlined for 400 receivers with his mods (set V12 & 13 PI pin 3 cathode to ground resistance at 96.2% of the 47K measured PI pin 1 plate resistors). During measurements, I found that the left channel still had noticeably higher THD than the right channel, but considerably lower than the pre-mod levels. I switched the V12 and V13 12AX7's (Fisher label Telefunken) and the left channel distortion went down, so I decided to try a new set of near-matching Sovtek 12AX7LPS tubes. These gave notably further reduction in THD in both channels, so I left them in.

The first two shots are the LEFT channel before and after the mods/Sovteks and the next two are before/after of the RIGHT channel. All measurements are 1 kHz, single channel driven to clipping. Completing the mods did restore the power levels back to a hair under 23 WPC; 13.5 VRMS for each channel at clipping. The scales for the distortion FFT display (blue) are the same in these shots. Considerably less distortion seen for both channels.

The next two shots are LEFT CHANNEL 10 kHz full-output square waves, followed by a 5.0 VRMS 10 kHz square wave. The next two are the same for the RIGHT CHANNEL. The signals are entering through the AUX/TAPE inputs with the tone controls centered.

DAVE GILLESPIE: Do these seem OK for the mods we've done so far?

I'm going to take another listen before proceeding with the buffer stage mod.
Thanks!
Dave
 

Attachments

  • IMG_2701.JPG
    IMG_2701.JPG
    109.9 KB · Views: 13
  • IMG_2749.JPG
    IMG_2749.JPG
    117.5 KB · Views: 12
  • IMG_2703.JPG
    IMG_2703.JPG
    106.9 KB · Views: 10
  • IMG_2747.JPG
    IMG_2747.JPG
    117.7 KB · Views: 10
  • IMG_2736.JPG
    IMG_2736.JPG
    90.5 KB · Views: 12
  • IMG_2737.JPG
    IMG_2737.JPG
    107.2 KB · Views: 10
  • IMG_2728.JPG
    IMG_2728.JPG
    100.7 KB · Views: 11
  • IMG_2729.JPG
    IMG_2729.JPG
    93.5 KB · Views: 10
Dave -- When I was developing the new power amplifier section, I did that as a stand alone piece, applying the generator to the new "Power Amp IN" jacks. This is the customary way of testing/measuring power amplifier performance, as inserting the test signals at an earlier point in the signal chain will not allow you to pin-point where different performance criteria is being generated within the circuit.

For example, while using a 10 kHz square wave -- at low levels (typically a 1 watt output) -- can tell you many things about the power amplifier section when directly injected into that section, it will tell you little about the tone control section as few if any tone control designs will pass a 10 kHz square wave fundamental with any kind of faithfulness at all. I have often written that it takes a circuit response that is typically from 1/10th to 10X the fundamental frequency of a square wave to pass it faithfully. That means that with a 10 kHz fundamental square wave frequency, you are asking the tone control section of your 400 receiver to display a flat, phase neutral response to 100 kHz -- which it will never do. For testing the HF response of tone control circuits, a fundamental of 2 kHz is much more useful, and represents a fundamental square wave frequency that the circuit should be reasonably expected to pass with a good degree of faithfulness.

Also, high power square wave testing is not a particularly useful tool in testing power amplifier performance. In fact, it can make an otherwise poor looking square wave (one with excessive overshoot) look much better than it really is, by effectively clipping off the overshoot, and therefore giving an improved appearance. Square wave testing is most useful in determining HF response and HF transient response -- where here they are extremely useful. For power and distortion, sine waves are still the best conventional way to measure distortion -- either as single tones, or complex multi-tone (IM) test signals.

Finally, if the buffer stage has not been installed yet, then when testing control section response/performance (either as a stand alone section, or overall performance), overall circuit response will be notably reduced at settings below a full volume setting due to the effects of Miller. Therefore -- and it is always good practice -- that unless you are specifically testing for the effects of different volume control settings on performance, you make control section tests with the volume control set full up. You will find that once the buffer circuit -- and the resulting Preamp Out/Power Amp In jacks -- are installed, testing of the individual sections will be a breeze!

I hope this helps!

Dave
 
Excellent stuff,Dave! Thanks for the insight on testing and using the scope to judge what a circuit can do. I'll study this and take another crack at it in a couple days. Traveling for a bit now.

Dave, this work is about learning for me, as well as having great gear to use listening to music so your insights really help.

Any references you care to recommend on testing and evaluating audio and/or receiver circuits using generator/scope?

Thanks.

Dave
 
Continuing listening tests for the modified 400 while working on some other projects. Just restored a pair of AR-4x speakers and giving them a listen on the 400 and the X-200. Really nice sound for their size; not up to the KG-4's, but quite nice. Had to de-construct (getting the grills off without destroying them took a lot of patience and research) and dis-assemble and clean the treble pots (badly corroded wire-wound pots), re-cap and re-assemble with velcro for the grills. The AR-4x's are considerably less efficient that the Klipsch (about 85 dB vs. 92 or so for the Klipsch). Have to drive them harder with the amps. The corroded pots are a known problem that disables the tweeters and, of course, they didn't sound very good with only the woofers going.

The 400 continues to perform well and sounds very good, indeed. Couple of observations, just confirming what Dave G has reported with the 400 EFB and power amp mods in place: 1) the amp runs cool--PT warm, but you can easily leave your hand on it after a couple hours operation and OPTs barely warm and 2) The HF roll off is definitely there compared to the X-200 and my Scott 299B. The sound of the 400 is warmer and the responds to the treble control, but does not have the level of detail of the other amps. So, that's what the buffer amp mod is for! Going to get this moving next day or so.

Also read a very helpful thread from 2014 by OP njcanuck called "Improving a Fisher 500B Power Amp Section--Advice?..," since a 500B restore is on the near horizon for me as well. Re: above posts on scope testing, the 500B thread contains a lot of good dialogue between NJ and Dave G on using a scope to performance test and I'm going to read it carefully.

Dave
 
The 400 is back on the bench. Prep work underway to construct and install the buffer amplifier mod. Pictures show moving the 'speaker off' load resistors and arrangement of components for the perf board. Hope to begin soldering soon, but the light over my work bench gave out, so had to correct that before proceeding!
Dave
 

Attachments

  • IMG_2792.JPG
    IMG_2792.JPG
    115.8 KB · Views: 49
  • IMG_2793.JPG
    IMG_2793.JPG
    58.3 KB · Views: 48
Success! Just finished installing the buffer amp unit and initial sound checks were good to go with clean audio out of both channels and no apparent ground loop hum, etc. Going to do some measuring using the scope and audio generator and finish everything up before moving up for sound checks using the good speakers. More to come (including pictures)....
Dave
 
Here's the Buffer Amp board mod installed in my 400. Two views of the installed unit, then a 'step back' of all the restore and DCG mods (I elected to keep everything under the chassis), including:
1. All new cans and bias/heater electrolytic caps
2. New coupling caps in the tone control amps and power amp; various resistor upgrades
3. Power amp mods for late model 400, including new PI adjust, NFB and stability mods
3. Combined EFB/IBAM board installed under chassis with cathode current test points on the sockets
4. Buffer amplifier mod installed near level control with re-purposed -15V/12AX7 heater power supply.

Everything seems to be working well. Next post has scope shots. Going to do front end/IF alignment (maybe MPX) before taking it back to the listening room, but amp is sounding good into my (indifferent) bench speakers. Output power holding at 22.8 RMS WPC at clipping. EFB/IBAM is keeping the tubes each at 21 mA quiescent cathode current. Distortion looks quite low, as best I can tell on my FFT display. Buffer went in like a dream, Dave!
Dave
 

Attachments

  • IMG_2805.JPG
    IMG_2805.JPG
    108.4 KB · Views: 63
  • IMG_2807.JPG
    IMG_2807.JPG
    121.1 KB · Views: 61
  • IMG_2812.JPG
    IMG_2812.JPG
    105.8 KB · Views: 62
Back
Top Bottom