Fisher 400 Rebuild Project

Is it safe to assume that if I set the variac to 11.7vac, that all the internal voltages should read 1/10th of expected? The high voltages look reasonable, but possibly somewhat low, but the negative bias supply looks very low. I haven't done any troubleshooting yet, just curious if I'm starting out with a reasonable plan.
 
Tubes won't start conducting until about 60-70V.

Hi-Voltages on late model 400's (Serial # 48001 and up)being low is a common occurance. Usually about 20v-30V < than schematic nominal. Bias supply should be about -22v to -24v with an IBAM or IBBA board so you have enough adjustment for Old stock 7868's and EH7868's which need more negative voltage to make the same cathode draw. Did you put a resistor in btwn the rectifier and the bias caps? You can drop the value some to get in the range noted.
 
tek -- The correlation is there, but under the startup process, it's only there if you remove the phono preamp and line/tone stage tubes.

When applying about 12 vac to the unit, the B+ voltage will ramp up accordingly and unimpeded, since the heaters of the tubes are not warmed sufficiently to allow any of them to draw any current.

However, the DC Heater/Bias supply has an instant and significant load across it due to the non-linear temp coefficient of the phono and line stage tube heaters. Cold heaters show a very low resistance versus their warm (operating temp) resistance, which places a heavy load on the output of this supply at turn on. The actual surge itself in normal operation (relative to the supply's capability) is minimized by the 15Ω dropping resistor, and so is not a problem. But with a cold turn on at a low voltage, the heaters of these tubes do not warm up enough to any proportional resistance value, and so the voltage drop across the dropping resistor is very high, making the output of the supply look very low. Removing the tubes prevents this condition from happening, allowing a more normal voltage correlation to occur during startup testing.

I hope this helps!

Dave
 
Thank you both. I had not considered the heaters drawing more current when cold - that is undoubtedly what I was seeing because I did plug in the 2 line-stage tubes. The voltages did indeed look about as expected with no tubes installed. I'll keep this in mind as I ramp things up.
 
Was out of town last week, but anxious to get back at it this week.

I powered the receiver up again tonight and worked my supply voltage all the way up to 123.8vac. At that voltage, with no tubes installed, the main B+ was 452.8vdc and the EFB(tm) screen out was 328.3vdc and the bias out was -14.84vdc. That calculates to a ratio of 72.5%. Dave stated it should be about 73%, so I'm happy to report it seems to be working perfectly. I will adjust the bias up to about -16vdc to make sure I'm plenty negative for the first startup with the tubes installed until I can adjust for the actual 21ma cathode current.

I have examined the four 7868's that I removed from this unit when it was shelved. I can't bring myself to power the receiver on with them installed. There is virtually no getter material left in any of them, just an opaque brown stain with no silvery shininess at all. I have no way to test any of my tubes. I am shopping now for replacements, but unfortunately there will be a small delay. I did order and receive 10 of the Russian 6P41S which some sell as 7868 substitutes. They are rated for a lower output wattage, but I have read where people have successfully run them well beyond their ratings. Since this EFB(tm) circuit is designed specifically for the 7868 in this receiver, I have no idea if it would set the operating parameters correctly for lowest distortion on a 6P41S or not, but I suspect not. I'm reluctant to plug them in. They also appear to have the larger pins like the EH7868, and I don't want to stretch my socket contacts.

I hope my next report will be with tubes installed.....
 
I plugged in the four 12AX7's that run on DC for the heaters, and hit the switch. I let them run about 5 minutes and checked the voltages:

AC Line Voltage: 125.3
DC Volts feeding heaters: -25.07 - that's 6.27 volts per heater - perfect!
DC Volts prior to 10Ω dropping resistor: -28.03 - that's 7.01 volts per heater. Dave's mod is spot on.
DC Volts at pin 4 of op-amp: -15.88

That was using full line voltage with no variac. Here were the EFB(tm) voltages:
B+ in: 452.1vdc
Screen out: 327.7vdc (ratio still 72.5% of B+)
Bias out: -16.03vdc

I was shopping eBay for 7868 last night and saw a pair that looked promising and then noticed they are located in my small town! I messaged the seller and he responded immediately. He's just a few miles away. I hope to go make the purchase today and possibly strike up an acquaintance who might be willing to test my tubes for me.

One question. All 9 of my 12AX7 are the original Fisher branded tubes. 4 are made in Holland (assuming Amperex) and 5 are West Germany Telefunken smooth plates. I did not make notes when removing them. Do you think that the 4 Amperex were in the DC heater positions originally? I know it really doesn't matter, but sure makes me wonder if Fisher chose 2 different brands because they performed best for the job.
 
Last edited:
Your voltages look spot on!

As for the different tubes, please understand that it is really only a modern audiophile notion that different components sound differently. Back in the day, Fisher engineers didn't sit around and try out different components to chase a particular sound they were after. Oh, they may in fact have had a particular sound they were after -- but it was produced by engineering it scientifically -- not by subjective analysis of perceived sound qualities that different components are supposed to have. Back then, components were chosen based on the quality they represented versus the caliber of unit being produced. Price point was always a consideration -- but on a higher quality level for Fisher. The physical size of components were considerations as well, but sonic qualities were not. In fact, the very design of the units actively works to minimize any different sound qualities created due to variations in tolerance and construction. They couldn't have one Fisher sounding different from an other!

Sound quality can be shown to be almost entirely a product of circuit design. If a different component changes the sound of a given design, it is because the component is changing the way the circuit operates, rather than adding a unique flavor with the circuit itself unaffected. There are those who will disagree with this, and that's fine, as there's room for everybody. But scientific testing invariably proves that once the bias of anticipation and all knowledge of any subjective based changes made are removed from the equation, then the chances of accurately and consistently identifying the changes made by sonic signature identification under blind testing conditions is extremely small, if impossible.

The idea that components have a distinctive sound comes from guitar techs who have entered the high fidelity world. In the guitar amp world, no correction circuits are used, so that component and tube tolerances can and do in fact have an impact on the sound. But in the high fidelity world, by its very definition, every effort is made to remove any such colorations, so that the music is heard as it was recorded, regardless of component changes or tolerances. If Fisher installed both Amperex and Telefunken tubes, you can be rather certain that it was simply because Fisher was using products from both of these vendors to fill their production demands, and/or for keeping multiple sources on line in case one should cease operations or run short of inventory.

Dave
 
That sounds reasonable to me, so I won't worry too much about placement.

It occurred to me that all I need are V10 and V11 installed and the pre-amp should be functional, so I paired it with a Pioneer amp and gave it a try. There is a low frequency pulsing sound coming through both channels. It is controlled by the volume control (all the way to zero) and I can move it around with the balance control. It also occurs regardless of which position the source selector switch is in, although when I select the Tape Aux input where my CD player was connected, the sound changes somewhat, so I believe the source material was added to the mix. I tried a different pair of 12AX7 in V10 and V11 with the same results. Unfortunately, since those 2 tubes have their heaters in series, I can't troubleshoot by unplugging one since it will kill the other too. More troubleshooting tonight...
 
I still wonder about the polarity of the cathode caps on V10. They show negative to ground, but schematic shows cathode has -1 volts on it.
 
The negative sign apparently goes back to the original Fisher schematic, and is clearly an error. The tube, operating from a positive source of voltage relative to ground, cannot possibly produce a voltage less than ground at the cathode, as that would imply negative current flow relative to the positive source to which it is connected, which is simply not possible.

If pulling V11 eliminates the pulsing, then that clears the buffer circuit. One thing you can do is that with the unit operating, pull V10. The cathode of V11 will remain warm enough long enough to let you know if the problem still exists, or goes away with the removal of V10. That will help you to identify where the problem is originating at.

Dave
 
Thanks, Dave. I already eliminated the buffer circuit by connecting the pwr amp input directly to the center wiper pin of the volume pot and the sound is still there. It seems the buffer is doing a good job of duplicating the sound being sent to it :thumbsup:. I'll get back to testing the pre-amp side of things in a bit. Good suggestion on pulling V10 while operating.

In the meantime, I installed all the tubes on the output side and was rewarded with some nice clean sound when supplied by an external pre-amp. I have an issue of not being able to adjust the idle current low enough. With the bias voltage at -15.77 (the max I could get), B+ @ 423.3, and Screen @ 307.3, my cathodes range from .22 to .30 volts. All voltages were still slowly rising. I will rearrange the tubes and make sure the various idle currents follow the tube and not the socket. I haven't left the power applied more than about 3 minutes at a time because I didn't want the cathode current to get much above 30 ma. I still have all the FM-related tube sockets empty, so perhaps I need to install those tubes as well the get the PT load where it will normally be.

More to come.
 
Last edited:
OK - Here's where I am on troubleshooting the pre-amp pulsating sound:

- I pulled V10 while powered-on and the sound continued and then slowly faded as V11 cooled.
- I traced all the connections around V11 and haven't (yet?) spotted a wiring issue.
- I disconnected V11 from the center of the High Filter switch (both sides) to eliminate any possible issues prior to that point - problem still there.
- I measured voltages around V11 and found a problem. At point B just prior to the two 120K plate resistors, I measure about 350vdc (steady), but on the other side of the resistors at either plate pin 1 or 6, the voltage is swinging up and down. It's hard to tell the measurement extremes with a digital meter, but I'm seeing something like 195vdc up to 260vdc or so. When I measure V10 at the same points, all is steady with the 350vdc at point B and then about 192vdc at each plate.

I'm pretty sure now there must be a wiring issue or bad component around V11. I tried checking the two .047 coupling caps in circuit and measured about .051 for each. Also I measured the two 22uf cathode caps and they measure 25uf in circuit. Both 1K cathode resistors measure good, as do both 2.7M resistors, and both 120K plate resistors.

Next step is to disconnect anything forward of V11 that might be feeding back to it somehow.

BTW, on the output tube biasing issue, I rearranged the tubes and the various cathode currents did follow the tubes, so at this point all I can do is conclude that I have some very mismatched tubes, even though they were sold as matched.

Getting frustrated and need to step back and try again tomorrow.
 
With the oscillation appearing in both channels, that suggests that the problem could be in the power supply. Check that the bypass cap serving the tone control stages is good, and properly connected.

As for the output tube bias, when I designed the EFB control grid regulator circuits, it was based on the use of good, vintage US manufactured tubes. Foreign tubes notoriously require more bias voltage for a given current draw, so it may be that if you're using EH tubes, that is the issue. Did you include the bias control on the EFB board? And if so, did you adjust it for more negative voltage? If all is good and the voltage still cannot be brought into range for your tubes, then a slight component adjustment on the EFB board can address that.

Dave
 
Thanks, Dave. I will go back and check power supply connections, but V10 is powered from the same lead (daisy-chained from V11 actually) and it's plate supply is a nice steady 192vdc. Also, with V11 power lead showing a steady voltage prior to the dropping resistors, but an oscillating voltage after them, it looks like something is causing V11 to conduct, then not, then conduct, then not.

My 7868's are all vintage US-made. I have a pair of Zenith labelled and a pair of Realistic Lifetime labelled. All have identical construction and 7868 USA markings. Pretty sure they are Sylvania manufactured. Bought the Zenith pair yesterday from a local dealer - sold as NIB "matched". Do tubes require more bias as they wear out? Grid deterioration?

I built the EFB board as drawn - I adjusted the onboard bias pot for maximum negative voltage (about -15.75 is what I get).
 
Last edited:
Just found the problem - I have the grid connections swapped - embarrassing :(. I'll fix shortly and move on to the biasing issue.
 
Tek; I've found as tubes wear out you need less negative voltage (towards Zero Volts) to maintain some semblance of distortion free operation. I've got a quad of worn out 7591 (3 of 4 are EICO branded the 4th is a Westinghouse) that require between -9 and -10 volts to get them up to 30ma, which IMO is too low for the 7591 class tubes which includes the 7868. -15v or so should get you close to 40ma on a decent tube. If you're using the cathode draw to calculate dissipation don't worry so much about the voltage.
 
Thanks, Larry, that's good to know. Since I'm using Dave's EFB(tm) circuit with this, I'm trying to achieve the 21ma cathode current he determined to be the lowest distortion point for the 7868's (operating under control of EFB). I have a couple more spares that I will install to see where they idle at. If I can find four that all draw similar idle current, then I'll use those and work on the bias voltage to get the current where it needs to be.

My knowledge of vacuum tube circuit design is sparce, but I believe the reason we run the quiescent current so much lower than traditionally done for this class of output tube (and certainly for the Fisher 400) is due to the significantly lower screen voltages operating under EFB. Hopefully Dave will correct me if I'm wrong here.
 
In conventional amplifier design, the voltage provided to the output tube elements is only moderately regulated. Of these, the most critical is the screen and control grid voltages. Fixed bias operation helps to stabilize control grid voltage (over that of cathode bias), which the Fisher design uses, so that is a plus. But the screen grids can often draw 5 or 6 times their quiescent current level at full power output. When supplied by a typical dropping resistor then, the screen grid voltage can drop significantly in practice. When it does, the quiescent current setting point drops as well, since the screen grid voltage affects not only instantaneous current flow, but also sets the maximum plate current that can flow through the tube. When the quiescent current is incorrectly set, the "hand-off" of current flow from one tube to the other is not properly aligned, causing distortion to increase significantly. To combat this, manufacturers often (but not always) had to set the quiescent current at a much higher than optimum setting, to counteract the loss in screen grid voltage that typical power supply accommodations produced with increasing power output levels. The thought was that as power output increased, the screen grid voltage would fall, causing the quiescent current level to fall down to an optimum level, all in an effort to minimize distortion at full power output.

The use of EFB™ eliminates all of these problems. In the application of the Fisher 400, the EFB regulators act to hold the relative voltage levels of the plate, screen, and control grid elements tightly together. The main B+ supply will sag somewhat with increasing power output in both channels. But with EFB, the control and screen grids will sag in the exact same relationship with the main B+ sag, that is established under quiescent conditions. Because those relationships hold then under all power levels, the quiescent current can now actually be set at the true low distortion point, rather than one that is much higher that then "falls into place". As a result, the quiescent current level can be set much lower (at the true low distortion setting), allowing for cooler tube operation, and cooler operation of all the power supply components as well. Also, because the optimum hand off now occurs at all power levels, distortion is lowered significantly at all power levels, both low and high.

Finally, when Fisher inexplicably raised the reflected load to the output tubes from 6500Ω to over 10KΩ in later versions of the receiver -- without any corresponding change in operating conditions for the tubes to accommodate the new load -- it placed the output tubes in harms way with potentially serious screen grid overload. So in addition to adding EFB action to the output stage, by setting the screen grids at a more appropriate voltage level for the new loading conditions (creating a new voltage "relationship" for that element), it also resolved the potential screen grid overload condition while gaining all the benefits of EFB action all at the same time.

I hope this helps!

Dave
 
Excellent, thank you!

I had an interesting day. Yesterday I was thrilled to discover a tube dealer right here in my small town. I bought a pair of “matched” 7868 from him. When I installed them last night, I found that when I adjusted the bias to idle at 22ma on one, the other was close to 28ma. I contacted him today to ask about his matching process. He got pretty defensive and ultimately decided that he couldn’t support my “custom one of a kind amp”. Here is his last text to me after suggesting I should return them.....

“Yes, i think you should return them. I did not know you had a custom amplifier. I am use to selling to audiophiles with tube testers, who buy tubes to put in factory made, original amplifiers. They want good GM matched tubes and Care most about sound quality.”

He asked how I knew the tubes weren’t well matched. When I told him how I was measuring cathode current, he basically told me I didn’t know what I was doing and should be using a “real tube tester”.

So, back to looking for tubes.

I corrected my wiring issue on V11 and the pulsating sound and voltage oscillation is now gone - and I have very nice sounding audio ..........out of one channel. Pulled the oscilloscope out of storage and traced the audio until the signal stopped at one of the .02uf coupling caps on V10. I cut both caps out and they both measure the same, but quite low at .014uf. Resistance and capacitance both measured the same at the same point on both halves of V10 so I'm not sure yet what's wrong. I'm ordering replacement caps and tubes, so there will be a delay while I wait on parts.

My switches are quite finicky and require a lot of wiggling to stabilize the signal. I will probably look for replacements for them as well even though they won't have the correct burgundy color.
 
Last edited:
Deoxit products are about the best you can get to clean the existing switches, and do a credible job of doing so. You might try that before going to the work of replacing them.

Your experience with the recent tube matching fiasco illuminates the general lack of understanding by the seller. Tubes that are truly perfectly matched will be so at all points along the plate/grid curves produced by the tubes. More realistically today however, this rarely occurs so vendors endeavor to find tubes that at least draw the same cathode current at typical operating conditions (that is, one point on the curves) for the tubes. They support the static match with dynamic Gm values to seal the deal. But operate the tubes at alternate operating points and the static match can go right out the window as you found out. Further, Gm testing provides little useful information for power tubes, as the test just tickles the tube under test, and does not represent real world operation. In short, it is a test for basic characteristic, not condition or match over the full dynamic swing of the plate/grid curves.

You didn't say, but I'm betting that the caps that measured low were of the ceramic type. Low measurement of these type caps is common in vintage equipment. When they get replaced with an accurate value mylar cap, everybody says how much better it sounds, and therefore how crappy the ceramic caps sound, which isn't true. It's the installation of an accurate value component that made the change in sound, not the change from ceramic to mylar. I'm inclined to believe that most ceramic caps that Fisher used started life reading low and stayed low, rather than losing capacitance over time. As a result, I'm not even so uncertain that in some audio related locations, they were actually specifying a ceramic component based on its typical value, rather than the ideal circuit value required, due to the use of a ceramic specified component -- which was almost always driven by space requirements. In any event, it is very doubtful that the caps you removed -- if ceramic -- were actually bad, and therefore, were likely not the problem. The darn things live on forever.

The range of the EFB bias control is about 3 volts, which allows for a wide range and still relatively easy adjustment. However, if the range of the control needs to be re-centered to a value more appropriate for your tubes (needing more bias voltage), increase the value of the 39K resistor associated with the bias control to 47K. This will recenter the 3 volt spread of the control to a value that is about 2.5 volts more negative, which should easily then account for your tubes. I would encourage you to not change the value of the bias pot itself to a larger value, as then its adjustment will become very touchy.

Dave
 
Back
Top Bottom