Help with Fisher 660A voltages

derekva

This ain't no picnic...
Subscriber
I've been working on rebuilding / repairing a Fisher 660A console amp and have finally gotten to the point of firing it up to check voltages. While the B+ to the OPTs and the B+ to the 7591 plates appear to be close enough, the HV to the 12AX7 tubes (both known good) is significantly higher than it should be (381V instead of 320V). My thought is that I need to increase the value of the resistor just upstream (currently 2.2k) in the power supply, but I'm not sure as to how much to increase it. There is also a resistor just downstream, but it is effectively not in the circuit as there isn't a preamp connected to it. All components of the original power supply are in place (including those bits that aren't doing anything because of the lack of preamp).

Here's the overall circuit (somewhat hard to see):

picture.php


Here's the close-up of the power supply:

picture.php


Thanks in advance,

-D
 
How close is close enough on the the B+ on the OPT's and the 7591's. You may need to also lower the voltages proportionally before R123 and R126 to get them to the voltage shown which may bring the 320 line close to spec. (Within 15% is ok per FISHER notes on their schematics due to variances in parts). Math isn't my thing, so I personally would get some 5watt resistors in varying sizes between 1.5K and 4.0K oHms and try out combo's. It's probably what you don't want to hear, but it's an option if nothing else turns up.

Larry
 
Thanks, Larry.

B+ at OPTs = 415V (about 5% low)
B+ at plates = 389V (right on the money)
B+ at 12AX7s = 370V (about 15% high)
Junction of C98-C and R127 = 205V (about 5% high)

I'm not worried about anything downstream of R127. Bias is set at 44V on the cathodes (although I have wiggle room to go lower if needed). No evidence of red-plating, although R126 gets (IMHO) unhealthily hot - I'll have to replace it with a 10W or 15W component. Everything else stays at reasonable temperatures.

-D
 
Try 2.7 or a 3.0 @ R123 try to bring down the 12ax7 B+ to spec. If you are running this @ wall voltage of >117VAC, put it on a variac and set for 117VAC and recheck. That's a lot of voltage on the 12AX7's. About 14% over design max.

Design Max values for the 12AX7 and ax7A from the n7JP site are ;
Maximum Ratings (Design Center Values)Each TriodeSource ........................................ GE - 1973
Plate Voltage ................................. 330 V
Plate Dissipation ............................. 1.2 W

I'd want to get them down to NMT 320V @ wall voltage of 122. It'll be lower @ 117 too. Definately want to either variac it or use a bucking transformer to power it, unless you can get it down enough.

Have you figured out your dissipation on the OPT's? With it being cathode biased, can 10oHm resistors be put in on pin 5 to measure ma per tube?????

Larry
 
Try 2.7 or a 3.0 @ R123 try to bring down the 12ax7 B+ to spec. If you are running this @ wall voltage of >117VAC, put it on a variac and set for 117VAC and recheck. That's a lot of voltage on the 12AX7's. About 14% over design max.

Design Max values for the 12AX7 and ax7A from the n7JP site are ;
Maximum Ratings (Design Center Values)Each TriodeSource ........................................ GE - 1973
Plate Voltage ................................. 330 V
Plate Dissipation ............................. 1.2 W

I'd want to get them down to NMT 320V @ wall voltage of 122. It'll be lower @ 117 too. Definately want to either variac it or use a bucking transformer to power it, unless you can get it down enough.

Have you figured out your dissipation on the OPT's? With it being cathode biased, can 10oHm resistors be put in on pin 5 to measure ma per tube?????

Larry

Thanks, Larry!

I'll try a 3.0k resistor and see if that does the trick - is it safe to assume that going higher (e.g. 3.3k) would lower the 12AX7 voltage further? If so, I'll pick up a range of values and see what gets me to optimal voltages. Unfortunately, there's no room for a bucking transformer inside the amp and I don't own a variac at present.

Unlike a 500c / 800c, Pin 5 can't be wired to ground (which AFAIK means the 10oHm resistor trick won't work), so I'm simply going by the service manual's recommended bias voltage of 44V (measured at the positive end of C93-A).

-D
 
Unlike a 500c / 800c, Pin 5 can't be wired to ground (which AFAIK means the 10oHm resistor trick won't work), so I'm simply going by the service manual's recommended bias voltage of 44V (measured at the positive end of C93-A).

-D

Looks to be cathode biased, not fixed as the 500c. 10R resistors can be used but you'll need to run them to a common buss.
 
A 3.3K would bring it down further. You can series a couple to get the value you need (if necessary) for preliminaries, and once you've got it right, get the correct value.


I'd check the BIAS of the tubes @ PIN 6 and adjust to 33V-34V via the bias control in addition to the 44V @ C93A. If they aren't matched, bias the hottest one for 33V-34V as per pin 6. You've got 11K in resistors and a 5K pot directly between C93A and each tube, and with unmatched tubes each one is going to be different. If matched, verify 33-34V @ pin6. then set for the majority @ the set point. My thinkng on this might be off also as it's cathode biased, and looks to balance the whole shebang. In any event I'd still check at the tubes to see where they are.

Larry
 
Looks like I need right around 9.2koHms to get my voltages correct (the daisy chain of 5 power resistors was rather humorous). I'll check my local parts / surplus hut tomorrow morning, although I may have to go for a 10k resistor (which will put the 12AX7 voltage at 315V...a bit low, but still reasonably close). Thanks for the help, Larry!

-D
 
Last edited:
Derek -- For the voltages shown on the schematic, my rough calculations show the 2.2K resistor in the power supply should be replaced with a 15K resistor if you are not powering a preamp, but as I said, I am just working with very rough estimates, since no voltages are shown at the 12AX7 terminals. If you've come in at around 10K, that may be close enough. However.....

The bigger concern I have, is that this amplifier, as shown on the schematic, will not function without the preamp connected. That's because they are using the heaters of four small signal tubes on the preamp chassis (presumably two phono preamps and two line/tone stage amps) to effectively be the cathode bias resistor for the output tubes.

If you are trying to operate this amp in a stand alone fashion, you will need to install a resistor to take the place of those heaters. Otherwise, the output stage will never work properly.

Assuming R129 (I think that is correct) is a 1.3K resistor (hard to see), then a value of about 425 ohms should be pretty close -- close enough to get you in the ball park anyway. However, this resistor will be dissipating about 4.5 watts of heat, which means at least a 10 watt resistor is really the order of the day. Of course, it can be made up of lower value, lower wattage resistors connected in series as well, but either way, it needs to be installed.

On your schematic, it would be connected between pin #2 of the tuner connector, and ground -- which is also to say, it is in parallel with R129 (? -- again, can't see).

Once this is effectively dealt with, then you can readjust your power supply resistors as necessary to obtain the correct supply point voltages. Doing that before the output stage is drawing its proper quiescent current is really a wasted exercise, as the power supply will not be fully loaded in that scenario.

I hope this helps!

Dave
 
Derek -- For the voltages shown on the schematic, my rough calculations show the 2.2K resistor in the power supply should be replaced with a 15K resistor if you are not powering a preamp, but as I said, I am just working with very rough estimates, since no voltages are shown at the 12AX7 terminals. If you've come in at around 10K, that may be close enough. However.....

The bigger concern I have, is that this amplifier, as shown on the schematic, will not function without the preamp connected. That's because they are using the heaters of four small signal tubes on the preamp chassis (presumably two phono preamps and two line/tone stage amps) to effectively be the cathode bias resistor for the output tubes.

If you are trying to operate this amp in a stand alone fashion, you will need to install a resistor to take the place of those heaters. Otherwise, the output stage will never work properly.

Assuming R129 (I think that is correct) is a 1.3K resistor (hard to see), then a value of about 425 ohms should be pretty close -- close enough to get you in the ball park anyway. However, this resistor will be dissipating about 4.5 watts of heat, which means at least a 10 watt resistor is really the order of the day. Of course, it can be made up of lower value, lower wattage resistors connected in series as well, but either way, it needs to be installed.

On your schematic, it would be connected between pin #2 of the tuner connector, and ground -- which is also to say, it is in parallel with R129 (? -- again, can't see).

Once this is effectively dealt with, then you can readjust your power supply resistors as necessary to obtain the correct supply point voltages. Doing that before the output stage is drawing its proper quiescent current is really a wasted exercise, as the power supply will not be fully loaded in that scenario.

I hope this helps!

Dave

Hey, Dave - the more the merrier! :D

I added a 250 ohm / 20W resistor in parallel with C93-A (e.g. pin 2 to ground) right off the bat, based upon the assumption that it was using the heater string of 3 small signal tubes in the tuner for the bias (after reading up on this thread). However, I can increase this if it doesn't seem large enough to you, although I've not had any problem getting the voltage at the cathodes to measure the recommended 44V (with provisions for going lower if needed) using the bias adjustment pot. Am I safe in assuming that given that the other voltages are OK with addition of a 10kOhm resistor in the place of R123, that this should be sufficient?

Thanks!

-D
 
Tom's comments in that thread are dead on with respect to circuit operation, and you were smart to already realize that the heater issue had to be dealt with in this case as well. I would differ only in the assumptions made.

Tom's comments assume that Fisher designed the bias string to in fact provide 150 ma through the heaters of the preamp tubes.

I would be the first to say I hardly know everything. But in my experience, this was never true. Historically, Fisher only ran about 11 vdc across the preamp tube heaters when they used such schemes -- and they used it a lot.

From that thread, you have then taken the assumption of 150 ma for the heater string, and therefore assumed that three tubes were used to make up the heater string in the original tuner designed for this amplifier. Again, in my experience, Fisher always used four tubes to handle the phono preamp, and line/tone stage functions in their stereo designs that included on-board phono preamps -- unless you know differently for the tuner/preamp unit this amplifier worked with. I know of no designs where they only used three tubes for these functions when phono preamps were included. With the basic quality that this amplifier represents, I think it is a very safe -- and more accurate -- assumption then, that in fact four tubes were used in the heater string for this amplifier.

From my own assumptions then, I would therefore suggest that you are running the output stages rather hard using a 250 ohm resistor, as further evidenced by the fact that your main B+ voltages are rather normal in spite of today's higher line voltages. The increased quiescent current you are causing the tubes to operate at with a 250 ohm resistor would account for that fact.

Consider also that your 250 ohm resistor is effectively in parallel with R129. If it is a 1.3K resistor, that means that your 250 ohm resistor is now about 210 ohms. Adjusting for 44 volts across 210 ohms means that each tube is theoretically drawing just over 52 ma in this scenario. If we assume:

1. The output tube plates are currently operating with about 425 vdc on them, and

2. There is in fact 44 vdc on the cathodes, and

3. The plates are pulling about 50 ma of the total current through each tube (to account for screen current), then:

That equals about 381 vdc dropped across the plate and cathode elements of the output tubes. At 50 ma plate current, that means that each tube is dissipating just over 19 watts per tube, which is the maximum plate dissipation the tube can handle.

Fisher classically ran 7591 tubes to idle at about 35-35 ma per tube. This would in fact equate to a 150 ma string -- except remember than 1.3K resistor that is effectively in parallel with the heater string which is also passing current. With that resistor in place across a four heater string circuit, the output stages would effectively draw150 ma, yet only drop about 11 volts across each heater of a four tube string -- all very much in keeping with classic Fisher design standards.

I would therefore suggest, that your bias resistor should very likely be at least 375 ohms, rated ideally for 20 watts. Such a resistor -- after R129 is considered -- means that the bias resistor then is actually on the order of about 291 ohms (assuming R129 is on value). 44 volts across a resistance of this value then produces a total output stage current draw of just over 150 ma, which again is very typical of Fisher design.

Such a change will then cause your overall B+ voltages to rise, as is so common place with this equipment and today's line voltages -- but the operating conditions will be much closer to normal this way than with the output stages artificially loading down the power supply as they likely are now. You will also find that the 2.2K dropping resistor -- the one which started this whole discussion -- may in fact then need to be increased to around 15K to achieve the 320 supply voltage you are trying to achieve.

Best of luck with it!

Dave
 
Thanks Dave, that makes sense (I think...I'm very much a tube novice compared to you). I'll swing by Vetco surplus and see if I can dig up a 375-390 ohm / 20W resistor today as well as a 15k / 10W resistor and see where that leaves me.

-D
 
D - The 15K resistor only needs to be rated for 1 watt -- not 10 watts. That would be gross overkill!

Dave
 
Well that's 2 of us that learned some more today. Understood about 80%. I still have probelms with the math, but I worked on it bit by bit (pencil and paper) and it makes more sense now.

BTW Dave. The only units that used the 660A are the '62 FUTURA III , '62 Chatham, and the '62 Ambassador IV. The Futura and Chatham used the 480-T and the Ambassador the 470-T. Both had 4 tubes on the pin 2 to pin 10 string.
http://www.fisherconsoles.com/ambassador IV.html
http://www.fisherconsoles.com/futura III.html
http://www.fisherconsoles.com/chatham.html


Larry
 
Larry -- thanks for the confirmation on the 4 tube string. Given the design history of Fisher and how consistent they were with a given "approach" I was rather certain that would be the case, but without knowing for sure, I like to make sure my own assumptions are clearly put out there as just that -- an assumption!

The knowledge you have of Fisher gear, when it was produced, model differences, what consoles the different units were in, etc., constantly amazes me. Fisher affectionados are some of the most dedicated folks to their equipment of all the equipment from the golden age that was ever produced. That is has its own forum on AK only amplifies that point!

Derek -- The old 2.2K resistor was also supplying the current needs of the tuner/preamp chassis as well. Because that unit is no longer drawing current through this resistor, its value needs to be raised to achieve a similar voltage drop, and its rating can be reduced because it is dissipating less energy now.

As an estimate, the two 12AX7 driver stages will be drawing about 4.5 ma total, based on an estimated input stage plate voltage of 85 vdc. For the purposes of the B+ dropping resistor then, it needs to drop 389 vdc (on the screen side) to 320 vdc for the driver stages to operate from. This is a drop of 69 volts, but now, based on a current flow of only .0045 amp. Ohms law tells us that a resistance of 15,333 ohms will produce this voltage drop at this current flow, and will dissipate .31 watt in the process of creating that drop. While a 1/2 watt resistor would work, a 1 watt resistor provides ample rating reserve to account for higher under hood temps. Therefore, a 15K 1 watt resistor should be well in the ballpark. You can tweak from there if you like.

Of course, all voltages will normally be proportionately higher with today's higher line voltages. With the exception of making sure the output stage is still operating within proper dissipation reserves, the slightly higher voltages this produces on the small signal tube stages is not a bad thing, as their cathode bias resistors will automatically compensate for any changes in B+ presented to them. The higher B+ also allows them to operate with higher head room, and lower distortion as well. Therefore, trying to chase down a very specific operating voltage is hardly necessary with these stages, so that making sure that all the voltages basically remain proportionate to the new AC line voltage environment is the main point to observe.

This should be a very nice amp when completed!

Dave
 
Thanks, Dave. I was able to find a 350 ohm and 400 ohm resistor in 20W. Which would be better to use?

-D
 
I tried the 400 ohm and the net result was the plate and OPT voltages were higher (405V and 431V - Fisher states 389V and 434V respectively in the service manual) and the 12AX7 voltages were lower at 300V. Not sure if (1) it's safe to leave as-is, (2) I should swap out for the 350 ohm resistor or (3) order a 375 ohm resistor from Mouser (and wait 3-5 days).

Thanks!

-D
 
Well, I can tell the amp is happier just by the fact that the transformer buzz has gone way down (and the amp sounds fantastic when hooked up to my Heresy IIs).

-D
 
Derek. You can have an allowable voltage variance of 15%+/- according to FISHER. There should be a note concerning that in the schematic. This is due to variances in resistors, caps, input voltage, etc.

So the 389 can be as high as 447 and as low as 331. Same with the 434 which could be 499.1 and as low as 369. Of course you want ALL of them to be proportionately higher or lower if they are high or low. The voltages they show are under ideal conditions, (ie: laboratory power supplies running at 117V, with meters that had impedance loads that were correct for tube gear, etc.) Considering you are running without a variac and on straight wall voltage, and your meter has a different input impedance, Figure your voltages to be 5-10% higher than the "spec" in the manual, or at least proportionate to the difference in the "spec" wall voltage of 117VAC and your wall voltage. 117VAC + 5% = approximately 123VAC. As long as the B+ voltages are all approximately the same %age "off" spec, you should be good.

When I'm checking a new piece of gear, I'll check the voltages on direct wall voltage, and figure the %age difference between it and "spec 117VAC". All of the supply voltages should be off by the same %age +/- a volt or two. Anything glaringly wrong then gets attention. Small differences, I don't worry about.

Larry
 
Back
Top Bottom