Modifying the Fisher 400 with EFB

Fur -- I would use .2 mS on your sweep setting so we can see more detail of the wave form. I'm also beginning to wonder if the double trace image is related to the test conditions more than anything else. With literally EVERYTHING identical (scope channel, settings, cables, etc.), what happens if you test the other channel under the exact same test conditions? Does it show the same double trace wave forms and react to the volume control setting as these show?

Your transformer measurements are DC readings from your Ohmmeter. What is important to this discussion is the impedance of transformer, where impedance is a complex value expressed in Ohms, but includes resistance, capacitance, and inductance. It is often referred to as AC resistance as well. It is not hard or complicated to measure, and can be done with little more than your DVM -- but it is best done with the use of a variac. Do you have one?

Dave
 
Scope settings are set with the sweep setting at .2 mS. Trigger is in P.P auto mode and then using A&B INT to switch between channel 1 and channel 2. The division settings are adjusted in accordance to volume voltage reading taken at output transformers. Also switching between right and left probe channels.

Yes I have a variac.

I used 4 different voltage settings but when I got as high as 2.692 vrms it was getting harder to see the flaw in the signal.

While balancing volume voltages between right and left channels it just shows you how bad the pots can be.

Test pictures setup.

0167FFC3-4A92-4A64-8051-DB1F55CEC987.jpeg
0E594572-F971-4E47-8A04-60C4F8311086.jpeg 2977E2BE-6B58-47C3-9F07-C59DB7B09574.jpeg
164DB940-670A-49BC-A678-69F92C94DD8B.jpeg

Pictures below are 63. mv and division at 20 m. The first 3 picture down is in trigger ch 1 then trigger ch 2. When single trace is present the channel was switched to one or two. Top is right channel and left is bottom.

C248B6EA-915D-4706-B650-0078DF7C2A29.jpeg
1435B60A-CD03-4961-8F51-3432D11714BB.jpeg
EFFA09F9-E08C-4E7D-9A96-68B44DF14D1F.jpeg
BB6FD921-F167-44CD-AE5A-A43B1A0E1397.jpeg
To be continued.
 
150. mv division setting at .1. Same as last post last picture trigger ch 2.

9D725550-57E0-4EA1-9AB5-A2D1074A75FC.jpeg
8E7F0DCF-D337-47B2-8135-A52461A84D69.jpeg
D6A1D7E7-1AC7-43E4-8C8E-A93D17C297D7.jpeg
4AF3EBF6-EFE4-4660-B136-B17A94D279BC.jpeg

285. mv division .2 same layout with order of pictures.

10844414-C295-42AD-9336-E9B912BDD6F9.jpeg
52BD46E2-6EFE-49AC-AD61-CC4CE9404EC5.jpeg
2221F2AE-31B1-48DB-88FB-278EF945B6C6.jpeg
43440A0C-54AD-4424-AA0E-3CDEFD816C52.jpeg

Two more pictures at 1.930vrms and .2 mS and .5 mS sweep

4DB59C22-9C3A-4332-A019-69D1207B702C.jpeg
17ACEC34-A7BE-4635-8B15-16ACDF30B8C9.jpeg
 
Last edited:
Fur -- At that output level, you're down in the weeds of noise and any other junk that can occur with covers off, florescent lights on, switching supplies nearby, etc., etc. With vintage equipment, low level testing is most often standardized at an output level of 1 watt into 16Ω, for an output voltage of 4.0 vac RMS. At such a low output level as you were using, your scope is likely having trouble giving a clean sweep, particularly in dual trace mode. Until you can button things up, I'd surely used the standardized level going forward to rise above the noise floor.

The impedance of your OPTs is measured with no AC power applied to your Fisher. To measure the impedance of your output transformers, connect a 16Ω load connected across the 16Ω tap, and temporarily disconnect the red OPT primary CT lead from the B+ supply point it connects to. Tape it off. Now, use your variac and appropriate connections to apply exactly 100.0 vac (as determined by your DVM) across the two OPT plate leads. This level of voltage represents no danger to the OPT, as full power conditions has much more voltage than this applied to the primary winding. Now, with your DVM on the same scale as used to measure the voltage applied to the primary winding, measure the voltage produced across the full secondary at the 16Ω tap. Turn off variac and remove the connections applying power to the primary winding. Reconnect the red OPT lead. Determine impedance as follows:

1. 100.0 primary volts / 16Ω secondary volts produced = turns ratio.

2. Turns ratio squared X 16 = primary impedance.

Let us know!

Dave
 
Okay I keep forgetting about all that stuff in the weeds so to speak.

The right OTP came out with 26.5 so I got 11.236.

The left OPT came out with 3.789 so rounding it off I got 230.
 
Last edited:
It might be time to find a new hobby.

Had the right side hooked up wrong and had pin 5 attached instead of pin 9. So when I have both clips on the transformer leads I get close to the same reading as the left side. 3.768 squared equals 14.197824 x 16 = 227 rounded off.

88C196AB-CFC5-4B0F-BCA0-E75D29B6210A.jpeg

735A35D9-F788-4520-A4AA-26F121BE796B.jpeg

E8425177-5B27-4DBF-BD5A-18D05BB22413.jpeg
 
Don't quit. We've all BTDT at one time or another! It's part of the Learning curve, with lot's of Banana peels along the way. Just as long as you don't nail yourself or destroy the unit with a mistake, you're ahead of the game.
 
Larry, you’re right thanks and so far I haven’t nailed myself or destroyed anything yet well at least that’s my story.
 
Fur -- Good job on the measurements, but you missed a step. You first divide the primary voltage by the secondary voltage to get the turns ratio. Then, it is the turns ratio that is squared and taken times 16 to determine the reflected impedance. Using the voltages on your meters in the pic:

100.1 / 3.77 = 26.55 turns ratio.

26.55 X 26.55 X 16 = 11,278Ω Primary Impedance (when the 16Ω tap is loaded with a 16Ω load)

If the voltages shown on your meters are the correct ones with the correct connections, this represents the measurements of a later version OPT, so if both transformers measure this way, then you do in fact have a later version transformer set.

If you will, please generate some new 1 watt, 1 kHz scope shots using a 16Ω load.

But this still does not explain the low power output. Try this: Use a test frequency of 500 Hz, and measure power output again. But this time, use your DVM to register the voltage developed across the 16Ω load resistors at the onset of clipping. This will eliminate the possibility of scope calibration error.

Dave
 
it would also be worth making sure your meter is reasonably accurate at that frequency. I know not a thing about that Fluke, but some DVM's are really only good at mains frequencies, beyond 150 hz or so the accuracy starts to take a dive. The cheapie I've been using forever does that, its absolutely worthless for audio frequency measurements.
 
To Gadget's point, that is exactly why I asked for the test frequency to be 500 Hz when measuring the power output with a conventional DVM. My experience is that most DVMs are good to no more than 1 kHz, and then are toast in accuracy beyond that. Gadget's experience is that it's even worse than that. The point is to use a frequency where the meter in question is still accurate, but not so low as to be below what the OPTs can efficiently handle. Taking Gadget's comment into account, I would suggest that you could test at a frequency as low as 100 Hz.

Dave
 
mine is a really cheapo meter so I wouldn't neccesarily judge too much from it. I just know that even by 500 hz its not telling the truth and it gets worse beyond that point. To be fair, that one isn't quite dead-nuts at 60 hz either. I would expect any model of Fluke to be better than what was a cheap meter when I got it in 1999.
 
Thank you Mr. Gillespie for correcting my mistake on my calculation error on the transformers. So before I hooked the B+ back up to the transformers I took readings again and they are staying consistent as long as I hook it up the right way.

Looking at the Fluke 175 manual and if I’m interpreting it correctly I should be good with the 500 Hz and 100 Hz testing. Copied from the manual. (AC voltage from 30.0 mV to 1000 V. Frequency from 2 Hz to 99.99 kHz.) Thanks for the heads up gadget73.

After testing I took some voltage readings on power tubes. Plates were at 396v screens 290v and bias 24 mA with line voltage at 117 ac. I’m just apprehensive of running full power until I know she’s good to go thats all. Also my math used to convert rms voltage to watts is rmsv squared then divided by ohms and in this case would be 16.

Observations, the crossover distortion did not show up till after the top and bottom distortion appeared and it seemed the top distortion went away before the bottom did. The rms voltage difference between the 500 Hz and 100 Hz test was 0.12 for right channel and 0.15 on left channel. It was the 100 Hz with the higher number now if it was Operator error I don’t know. I do wear glasses so you know.

Pictures of 500 Hz to a 16 ohm load. Scope settings .5 ms on sweep and division is on 10. Outcome on right channel was 18.01 rmsv (20.27 watts) and left channel 17.98 rmsv (20.20 watts)

Right channel before onset of distortion.
4206E814-9440-4582-90BF-4E95A633EE09.jpeg

Left channel before onset of distortion.
52C7D167-AD94-4A9A-A83D-27B73DAE3BDA.jpeg
 
Whoa now! That there looks a whole lot better! As long as the amplifier is properly loaded, no need for any apprehension to run it right on up to clipping. As I recall, my 400 was producing 22.5 watts at the on set of clipping with both channels driven, but my B+ voltages were slightly higher than yours. In any event, it would seem that you're well on your way now!

Do you have any means to measure distortion? That would allow you to set the AC balance most accurately. In absence of that, setting the total cathode resistance in the phase inverter to a value of 45.2K will get you pretty close, assuming the output tubes are even reasonably well matched, and the plate resistor of the phase inverter is similarly close to 47K.

In the development unit, with both EFB regulators connected and operating correctly, 1 kHz THD at full power output (onset of clipping) routinely registered no higher than 0.15%, whether a single or both channels were driven.

Dave
 
Well it looks like my lack of knowledge of understanding testing procedures had me self doubt my parts swapping ability. Each channel was tested separately and when I bypass the variac I will pick up at least 20-25v more B+.

No I do not have any means to measure distortion. I had adjusted the phase inverter to the middle of the adjustment you designed into it and it just so happened to be very close to 45.2K Wink Wink.

Maybe it’s time then to let what is under this cranium of mine to be a distortion analyzer. And if all goes well it will be on to a buffer mod.

So thank you very much Mr. Gillespie, and yes it helped a lot.
 
Got to listen to Béla Fleck and the Flecktones and a Keb Mo CD today on the fisher 400 playing through Cornwall speakers and a Denon 2910 combo Blu-ray CD player. Have a little buzz coming out of right speaker but not going to be an issue for this listening session. I have the bass and treble at 12 o’clock position and the loudness switch is on the off position. These two CDs I feel are well engineered and always enjoy listening to them and today was no exception on the 400.

Starting the first song playing and by the first minute or so in I can relax and start doing some listening. The bass is not as strong as it usually is but without having done the tone bypass mod yet so not sure how much one way or the other that the tone controls are influencing the tone. I’m not feeling any urge to make any tone adjustments which tells me its sounding balanced from the top to bottom. After the first song on the Keb Mo CD I started to get a little crackling out of the right speaker and it ended up being the tape monitor switch.

I’m looking forward to doing the tone bypass mod not that I have any thing against them it’s just that when they’re available I spend a lot of time fiddling with them.

Thanks again.

DFF04F7D-1635-41B7-8A5C-866BC2F542DA.jpeg
 
Since this is a fairly recent thread, I hope it's ok to post my information and ask a few questions.

I have had a model 400 in storage for many years and recently decided to get it out and do a basic recap of the PS and get it running. I accomplished that using three dual section cans in the stock locations and moved the rest of the capacitance inside the bottom chassis. The values remain the same except the quad 40/40/40/20 is now 50/50 and 50/50. I also replaced the 2 diodes on the HV section to 8amp/1200piv HEXFREDS. I started searching for mods to make the bias supply adjustable and came across the recommendations for changing the 330k/.047uf combo to 200k/.1uf for use with the new EH 6878's. I changed the .047's for .1's and the new 200K resistors arrived today. While waiting on the resistors, I was doing some reading and found the full series on the EFB/Phase Splitter/ Tone/ etc....

After reading Dave's methodical and incredibly well explained development and testing process and the fantastic results he attained - there is no way I am not going to do these same modifications to mine! The last time I heard this unit run was in 1978 in my brothers college dorm room. I repaired it for him by replacing the full-wave bridge in the bias/heater supply and he also installed a new quad of Radio Shack Realistic Gold Series Lifetime 7868's. I still have that set and they were working when it was shelved. Most of the other tubes are original Fisher branded and the 12AX7's appear to be Telefunken smooth plates (diamond symbol in glass between pins). We purchased the receiver, a Garrard turntable, and KLH Model 17 speakers from a family friend that year after it quit working due to the a fore mentioned bridge failure. I'll always remember as a kid, making a beeline for the basement of the original owners home to "play" with that system whenever we visited. I'd never heard a home stereo that sounded so good (we had a small console at home that I now know was single-ended EL84 based and obviously, very low powered).

Here come the questions:
1) Dave left the stock .047uf/330K parts installed. I am halfway through that change and would prefer to keep the .1uf (AudioCap Theta) I installed, but not if it will compromise the design. Should I use .1uf/200K or .047uf/330K? I still have the original Ero Fol II's and they test good but I'll have to extend the leads which are now too short. I've read the change is ok, but which is BEST?
2) I have a late version 400 (s/n 56xxx) but I do not see any stamping on the OT bells. How can I verify I have the correct OT's for the mods? I do have an LCR meter if that can help.
3) I was also planning to install the 100ohm screen resistors, but verifying they are no longer needed?
4) I am planning to install the 10ohm cathode resistors for current measurement, but not planning on individually adjustable bias. If they measure too far out of match and can't be fairly closely paired, I will replace with a matched quad. Are the EH7868 OK with all these mods?
5) I see Russian 6P41S tubes on eBay selling as 7868 replacements. I know they are rated quite a bit below 7868 specs though. After EFB installation, might they work acceptably? Those things sell for under $3 each......
6) I've seen no mention of the "Center Channel" output. Does it harm channel separation remaining in the circuit? I'm guessing no based on the high resistance vs. the low impedance speakers. Could it be used to drive a sub amp assuming appropriate low pass filtering was used?


I'm sure I'll think of more to ask as this progresses. This is really exciting.......

Tom Kuhn
 
Last edited:
Hello Tom and welcome to the forum.

I left the .47 capacitors in my 400 so far and I like the the way it is preforming so maybe in yours you can try both. I did go with 220K resistors because one original was bad and I had the others.

The biggest sound change to my ears came from all the mods around the phase inverter and feedback loop from the output transformers.

The 100 ohm screen resister, I would ask myself if I’m going to pass down this amplifier. In my 400 I’m having trouble with one tube socket I believe and have replaced two 10 ohm resistors so amp is parked for time being on shelf.

Below is from Mr. Gillespie on post #64. I had around 11.2K primary Impedance on my transformers. Which made them late model 400 transformer.

The impedance of your OPTs is measured with no AC power applied to your Fisher. To measure the impedance of your output transformers, connect a 16Ω load connected across the 16Ω tap, and temporarily disconnect the red OPT primary CT lead from the B+ supply point it connects to. Tape it off. Now, use your variac and appropriate connections to apply exactly 100.0 vac (as determined by your DVM) across the two OPT plate leads. This level of voltage represents no danger to the OPT, as full power conditions has much more voltage than this applied to the primary winding. Now, with your DVM on the same scale as used to measure the voltage applied to the primary winding, measure the voltage produced across the full secondary at the 16Ω tap. Turn off variac and remove the connections applying power to the primary winding. Reconnect the red OPT lead. Determine impedance as follows:

1. 100.0 primary volts / 16Ω secondary volts produced = turns ratio.

2. Turns ratio squared X 16 = primary impedance.
 
Hi Fur Face - thanks for the reply.

I had read that method to calculate, but I don't have a variac. I wonder if the 100.0 input voltage is critical? I could use my 125v line voltage assuming the ratio would still calculate the same. Alternately, I wonder if I could use a simple non-digital analog light dimmer to build a simple tool for adjusting the input voltage?

As for your 400 using the original .047 paired with 220K resistors - my understanding is that the .047/330K pairing establishes a time constant that the .1/200K fairly closely replicates. If you are pairing .047/220K then you are altering that time constant. I honestly have no idea how the sound might be affected by that change though. Perhaps you are already aware of that, but wanted to mention it.

Tom
 
Dave did the calculations about 5 years ago when .068uf and .082uf were becoming a common value in caps and that a .071 or .072 was as close as you could get to the original time constant with the 330K/.047uf. Seeing as those values aren't being made his recommendation was/is 220K/.068uf or 220K/.082uf as these are more common now. The original 220K/.1uf was made when the only commonly available caps were .047uf and .1uf.

Screen Stability resistors are always a good idea on the 7591/7868/6gm5 family, no matter whether you are using EFB or NOT.
 
Back
Top Bottom