The End of an Era: The Fisher 55-A

That would actually be pins 4 and 9, which is the 6.3 volt circuit for the heaters that the octal plugs are tied into. Maybe some sort of preamp power plug......

Dave
 
I inherited two 55A amps and one 50A amp from my dad.
Been unused for probably 40 years.
Tested tubes - replaced 4 of the 5AW4 rectifiers (two missing and two tested bad).
All the other tubes tested good.
Brought up using variac.
The 50 A seems to work great.
One 55A has some distortion but both have high AC power draw.
The 50A draws about 1.5 amps at 120 VAC.
Both 55A draw at least 2.5 amps at 120 VAC.
Obviously high with a 3 amp slow blow fuse !!!
If I move the 1614 output tubes from the 50A to the 55A (replace the 6550 tubes) the current is the same as the 50A.
What is the nominal AC current supposed to be for both models - warmed up but no (or small) input signal?
Thanks !!!
 
I’m always in awe when I read a thread like this of the sheer technical knowledge and passion that exists within the confines of Audiokarma. I am a passionate collector and enjoyer of my hobby, but the engineering knowledge of those who laid the groundwork for originally designing, building, and now maintaining the marvels that so many of us enjoy on a day-to-day basis is humbling.
 
2020 UPDATE:

It's been 5+ years since I started this thread, but I now finally have two of the EL34 versions of this amplifier here for service and restoration, and have determined some new information that I feel important enough to pass along to the Fisher community for those owning these amplifiers.

These amplifiers represent the last version of the 55A amplifiers produced, in that they employ EL34s -- and -- they both have a 4Ω output tap. Earlier versions of the 55A -- even those employing EL34 output tubes -- did not have a 4Ω tap like these do, so they do indeed represent the final production form of these amplifiers.

One piece of basic new information concerns the output transformers. It was unknown if Fisher changed these from the standpoint of reflected primary impedance when they changed from the 6550 to the EL34 in these amplifiers. Now that I've measured the transformers on both the 6550 and EL34 versions, it is clear that there were no changes made to them when the change was made between these tubes. And of course, there would be no reason for this specification to change when the 4Ω output tap was added. By all performance indications then, these transformer are identical to the transformers on the original 55A amplifier with 6550 output tubes, except they include a 4Ω output tap.

METER: The meters did change appearance wise from the original to final version of the 55A, with most such meters being very inaccurate by now. The original meter had an expanded meter face section for setting the bias, and a more compressed section to indicate power output. In later versions, this was changed to allow for a more expanded power output section to aid in its usefulness, As long as the field winding is still good within the meter, these can often be restored to a more usable level of service than they typically represent by now.

These meters are quite unique, with the power output indicating section of the scale having non-linear graduations, and a dual sensitivity design built into the meter movement itself. The non-linear graduations acoss the power output scale are necessary because the relationship between output stage current draw and power output produced is non-linear itself. The dual sensitivity is necessary to allow the meter to properly respond to both the static current conditions (only) that exist when setting the bias, and the static plus dynamic current conditions that exist when producing power output. The split in the sensitivity of the movement occurs at the 0 indicator on the scale, with the movement being much more sensitive to current flow above the 0 point, than it is below it. That is, with a properly operating meter, it takes 125 mA of current flow through the movement's field coil to produce an indication with the pointer centered inside the black 0 section at the top of the red scale. This level of current flow represents the nominal quiescent bias current for the output tubes. Yet the meter will also produce a full scale indication with a total current of 300 mA flowing through it as well. In other words, with just 140% more current flowing through the meter than it takes to produce a 0 mark indication, the meter will produce a full scale indication, which represents a 700% increase in response above the 0 mark, versus the current flow required to move the needle from its resting position to the 0 mark. That is quite a difference in sensitivity! Today, most of these meters are typically non-functional, or obviously reading incorrectly. With a little luck however, you can usually restore the most important part of their functionality, and with a lot of luck, all of it.

A potentially significant issue are two enemies of any permanent magnet, which is a principle part of the movement within the meter -- time, and heat. Time always tends to diminish the strength of a permanent magnet, and heat does too. With enough heat, a permanent magnet can have it's magnetic properties permanently diminished. It's been 60+ years since these meters were manufactured which is the time element, and the combination of amplifier heat and heat from the illumination lamp that was installed right under the movement inside these meters is the heat factor, that just piles on to the problem. Heat from the bulb alone has discolored the meter bezel in some examples of these meters I've seen. Additionally, it's also possible that time and heat acts to harden the characteristics of the return spring tension as well. The result is that over time, the meters then begin to read low, or hardly at all. As long as the internal field winding is good however, then with a DVM, a single resistor, and a variable low voltage power supply, it's not too hard for those so inclined to re-calibrate these meters to compensate for the losses of time and heat.

First, a look at how many examples of these meters operate today:

BELOW: A 15 volt 1A variable DC power supply along with a 50Ω 10W resistor, a good DVM (set for DC mA operation), and the Fisher meter are all connected in series. With the power supply adjusted for 125 mA of current flowing through this circuit, it produced the indication as shown on the meter when the amplifier arrived here. The nominal bias setting is achieved when the pointer is centered on the 0 mark with no signal passing through the amplifier. If this meter were used to bias the output tubes in the amplifier as is, then the output tubes would surely be red plating with such an adjustment:
SAM_3161.JPG

BELOW: To correct this, carefully disassemble the meter, and you will find the adjustment that tensions the return pointer spring. It is the balance of return spring tension versus magnetic pull created by the field coil/magnet assembly that determines the calibration of the meter. With the circuit adjusted for 125 mA of current flow, careful movement of this adjustment will allow the pointer to again rest over the center of the 0 mark as it should. You will want to use a non-magnetic tool to make the adjustment, and preferably even a non-metallic tool if possible. Also, it will take some trial and error to achieve the best accuracy, since even the installation of the rear cover will change the calibration slightly. Patience then is the order of the day:
SAM_3162.JPG

BELOW: After calibration, the meter is reassembled now, with the same 125 mA of current flow causing the pointer to now rest in the center of the 0 mark. Camera parallax error shows the pointer as slightly off center within the 0 mark, but in reality, it is properly centered:
SAM_3166.JPG



Now, if you're really lucky, advance the current flow to 300 mA, which ideally produces a full scale indication, but this will not always be the case. Of the two meters from the amplifiers here, one will in fact produce a full scale indication with 300 mA total current flow through the meter, while also producing a centered 0 mark indication with 125 mA of current flow. When installed, this meter is also remarkably accurate in indicating the correct average power output of the amplifier as well.

BELOW: For the other meter pictured above however, I wasn't so lucky, with a 300 mA current flow indicating about 57 watts on the meter face as opposed to a full scale 65 watt indication. However, this is still way, way better than it was, and of the dual function features of the meter, the accuracy in setting and monitoring output tube current flow is by far much more important than the accuracy in monitoring power output, with the usefulness of that function being quite limited in practical use anyway. That's why of the two calibration points of the meter, adjusting it for an accurate 0 point indication at 125 mA current flow will produce the most useful results, and return the meter to the best overall operation for which it is capable today.
SAM_3164.JPG


Up next, the strange case of instability that both of these amplifiers arrived with which would not be unique to just these two particular amplifiers, and an age old problem of the original 50A series of amplifiers rises again: poor HF transient response. No wonder my client was very disappointed in the sound these amplifiers produced when he received them! Next time.

Dave
 
Last edited:
How was the dual sensitivity accomplished? Was there a second spring that became active above zero? Was there a change in the magnetic gap at that point?
 
Hi Fred -- These meters do not use conventional D'arsonval movements but rather, employ a fixed field coil. A magnetic disk has a shaft through the center of it, which is then suspended over the field coil so that nearly one-half of the disk resides inside the field coil. Each end of the shaft terminates as a point, that rides in the well at the end locating adjustment screws, which position the shaft so that the disk is located in the center of the field coil. The pointer is attached to the shaft near the front end of it. When the coil is excited, it causes the magnet to rotate according to the level of current flowing through the field coil. A standard coil spring is used to provide for the pointer to return to its resting point when the coil is de-energized, with its tension adjusted by the adjustment lever shown in the pic. There are no others springs in this movement.

In studying the meters, I can think of only two ways that the sensitivity of a meter of this type could be altered mid scale: Either the magnetic disk is manufactured so that its magnetism is stronger in one section of the disk than another (which seems unlikely), or the disk has uniform magnetic strength, but is shaved at the circumference in a certain section, so that there is less surface area of the disk residing within the field coil, which is more likely.

In one meter, adjusting the return spring tension caused both sections of the scale to then read with good accuracy. But in the other, the power output section was still somewhat weak when calibrated for accuracy at the 0 indicator mark. I can only assume that if the magnetic disk has lost some of its magnetism over the last 60+ years, that it was not lost uniformly over the surface of the disk, resulting is the disparity that now exists between calibration points. Either that, or it is a manufacturing tolerance issue, that had this meter performing this way right from the start.

I am certainly open to any other ideas as to how this was accomplished, but two things seem pretty apparent about these meters:

1. When working correctly, they actually work pretty well for their intended functions, but

2. When they fail, they always read weak in both sections.

A partially shorted field coil seems highly unlikely, since there is just a small fraction of a volt across the meter even at full scale, and that type of damage would likely cause a uniform loss of sensitivity anyway. A return spring with two different tensioning levels depending on how tightly it is wound would seem even more improbable, particularly since it would need to have more tension at first, and then less later. That leaves the magnetic disk as the only real way to provide for two different sensitivity levels within the same scale. And, its natural tendency to become weaker with age would fit with the symptoms these meters often display over time as well.

Thoughts?

Dave
 
only other way I can imagine it working is a dual spring setup, with the spring for the low sensitivity position hitting a stop that effectively disengages it from the movement below a certain point. I don't see how that would effect both high and low range sensitivity though.
 
The bias would only have to be calibrated to a single corresponding point on the scale, with the scale itself calibrated for power according to the design of the meter, if I'm understanding the quandry?
Yes, the reference magnetic strength in most iron or AlNiCo based magnets falls in time, afaik.
 
Pio -- No, it doesn't work that way. The rather small part of the scale that serves as the bias indicating portion has greatly reduced sensitivity compared to the much larger portion that indicates power output. If you tried to do this by merely calibrating the scale for a for a meter movement that had a uniform sensitivity, then you would end up with a scale that had 50% represent bias, and 50% represent power. As a result, the power output portion of the actual meter would be shrunk considerably, and the bias section considerably increased -- reducing the usefulness of the meter. By using a meter with dual sensitivities, the bias section can be made much smaller, so that the power output section can be made much larger, and therefore, much more useful. It is interesting to note however that it was only the 55A amplifiers that used these meters; the 50A before them, and the 200 after them did not use any meters at all. In fact, the 200 went back to the phone jack approach used by the 50A for use with an external mA meter. Fisher did play with these meters however, as those used on the original (6550) version of the 55A appear to have a meter more in keeping with that I described above if the meter's sensitivity were a constant: The bias section takes up at least a third of the scale, which makes the power output section therefore much smaller. No doubt, Fisher was trying to improve on the usefulness of the meter with those used on the later EL34 version.

Dave
 
FINAL 2020 UPDATE:

Those who have read many of my posts know that I am not a fan of cramming in large oversized components into a space that was never meant to accommodate such a physical size increase -- all in an effort to chase improved sound. Besides the issue of whether this actually improves sound quality (a completely subjective analysis), I have often warned of the increased chances for shorts when oversized metalized components are used, and the not-so-remote possibility of unintended consequences now happening due to component interactions with the (now) larger replacement components, that never occurred before. Case in point occurred with both of the two 55As just finished up. One of these two "restored" amplifiers actually operated, but my client said the sound was quite poor, and boy oh boy, it was. The other amplifier was temporarily repaired (enough to get it running) -- only to find out that it's performance was just as poor, even though it too had been restored. Here's a couple of snapshots of their performance:

BELOW: Amplifier #1 at the onset of clipping at 1 kHz. Power output is less than 20 watts RMS, from amplifiers capable of 50 watts RMS. Ever seen anything like this?:
SAM_3126.JPG

BOTTOM: OK, that was a little strange. Let's try the second amplifier under the same conditions. Oops, at even less power output, this amplifier produced:
SAM_3131.JPG

OK, this is really strange. Output tube bias is correct, operating voltages are close enough (at least not to be causing this nonsense), and all the tubes were quite good. Here's a pic of one of the amps as received -- but both looked basically the same, producing the scope shots above:
SAM_3160.JPG

See anything wrong? For those familiar with the 55A, a few things stand out -- but nothing obvious that would cause the problems noted above. It appears that the previous tech was either trying to improve sound quality by bypassing the power supply and output tube coupling caps -- or, they may simply have been trying to chase out the gremlins noted above. Either way, it didn't work. What was the problem?

BELOW: Take a look at this. It fits, yes. But boy oh boy, you couldn't get a hair on my head between the cap and the transformer, or the cap and the inside front wall:
SAM_3191.JPG

BELOW: By simply extending the leads on this cap and removing it out of its close confines, you get a nice 52 watt waveform at the onset of clipping:
SAM_3127.JPG

Think there's no collateral damage that can happen from the use of oversized components? You might want to re-think that one again. But even with this resolved, there was one other thing about these amplifiers that put them well off the performance mark of which they are capable.

BELOW: A 1 watt 10 kHz square wave presentation gives jeeeeest a little bit of a clue:
SAM_3125.JPG

This sucker is getting seriously close to becoming a sine wave!

BELOW: An examination of the schematic brings into question this cap:
SAM_3190.JPG

The schematic shows no cap belongs there, but yet, the 500 pF tubular cap installed there has every appearance of being factory installed, and is very much the reason for a response that was down several db by just 10 kHz. Removing it produces a very acceptable 10 kHz square wave, and even decent -- but hardly absolute -- stability. No pic was taken of that because ultimately, new NFB and HF stability networks were devised for these amplifier since the Z-Matic circuits were being completely removed. As an intermediate measure however, replacing this cap with a 100 pF silver mica cap will produce greatly improved results.

BELOW: The finished amplifiers sound superb with all the magic these amplifiers are so capable of, and thanks to the modifications installed, now produce just over 65 absolutely stable low distortion watts. And the meters work, too!
SAM_3203.JPG

Dave
 
1) Why would the large bypass/transformer caps cause this problem?

2) Are you worried about the heat from the large resistors under the chassis? I have a pair of troublesome 200a's and I couldn't find a replacement for the 15W ceramic resistor that pokes out of the chassis so I mounted it under the chassis with some thermal grease. It gets hot...
 
The large coupling cap was a problem -- not because it was directly shorting to anything -- but because with the full length of the body of the cap so close to the front chassis wall and the transformer mounting bracket -- which both represent ground -- it allowed a new capacitance to be inserted into the circuit between the body of the capacitor (that is, its metal container can which represents its negative terminal and is just under the blue plastic coating) that's connected to the control grid of an output tube (pin 5), and ground. The new capacitance -- that was never there in the original build -- unbalanced the push-pull drive to the output stage at higher frequencies, such that with any nominal amount of drive, the output stage oscillated. The cap was ultimately replaced with a proper sized component for the space allowed. But by simply removing the oversized cap from its close confines, that also would allow the circuit to operate properly because once again it was then balanced on both sides of the push-pull connection.

The original 50A amplifiers had the driver stage dropping resistor located under the chassis as well, replacing it with the chalk resistor -- not only to remove the heat from under the chassis, but also for improved dependability of the resistor. The original resistors had quite a history of failing due to the amount of heat they dissipate (about 12 watts), which the chalk resistor was supposed to improve upon -- except that they had their own dependability problems as well. So while they did get the heat up topside, they didn't really improve the dependability much at all. In fact, the one amplifier was DOA precisely because that chalk resistor was in fact open on that amplifier.

The new resistors therefore return the heat to the underside of the amplifier, which was never a problem before in the 50A to any of the components located there -- just the dependability of the resistor itself was in question. However, in returning the heat to the underside of the chassis and looking to potentially improve upon that scenario, it gives an opportunity to also address another issue that can be quite annoying with these amplifiers -- hum -- as in of the mechanical type.

When you first turn these amplifiers on, they are mechanically as quiet as a church mouse. But once the output tubes warm and start to draw current, then the ruckus begins. It's not the power transformer, and of course not the output transformer. It's the choke -- well, sorta. But it's not because the choke is lose and needs tightening up. It has to do with the type of filter circuit these amplifiers use.

Because these are Class AB2 amplifiers, the change is output stage current draw from quiescent to full power output is quite significant. Back in the day of the original 50A which these amplifiers herald from, all that existed were rather high drop rectifier tubes. If a conventional cap input filter design were used, then the B+ drop from quiescent to full power output would be such that the elevation of distortion and the loss of power output would both be quite significant. Therefore, a choke input filter design was used, as was the rule of the day for any Class AB2 design, since the regulation of such a filter is much much better in the face of a widely varying load. But, with a choke input filter design, it means that the choke then must handle very large ripple currents, which sets up a rather large magnetic field around the choke coil. With the choke mounted where it is (close to the bottom plate), then the bottom plate -- when installed -- acts like a resonator to the choke, as surely as the cone of a speaker voice coil reacts against the speaker magnet from the current passing through it. The result is the bottom plate resonating at 120 Hz just like a speaker cone. In a quiet room under quiescent conditions then, a pair of these amps can carry on quite a conversation.

I mentioned to my client that the real answer to both of these issues (the heat and the hum), is to have a couple of new aluminum bottom plates madefor the amplifiers, with more holes drilled into the bottom -- particularly over near and around where the resistors are now located. That allows for improved cooling of the resistors, and eliminates the resonator effect, since aluminum is non-magnetic. A total win-win solution.

Dave
 
I was thinking a rather more cheater fix of a rubber thing stuck to the bottom of the transformer to put a bit of pressure on the bottom plate so it can't vibrate so easily. Non-magnetic plate is a better fix though, and probably more effective.

That flat-top sine wave I was thinking was gross bias problem in the driver stage someplace, interesting that it was parasitic capacitance causing extra load. It makes sense though. The oscillation at the peak of the sine is not something I've ever seen before either. I imagine that had to sound terrible.
 
This is some great reading, thanks Dave. The physical hum of my 55a (el34s, 4 ohm tap) once warmed up is pretty evident but not so much with the pair of 200a (also el34). Does this match your experience? Do you think your client will have new bottom plates fabricated?
 
I've helped a number of folks with their 200As over the years, but never had any on my bench to be able to help with your question regarding the hum.

I think my client likely will have new bottom plates made, since these are to be his daily drivers.

Dave
 
For whatever reason, the 200As are quiet. In keeping with Gadget's idea, I wedged a bit of rubber between the bottom plate and the shelf below. Certainly not eliminated the hum, but improved. I am curious how much of this is physical and how much magnetic...it would be interesting if you could relate your customer's experience with the aluminum replacement.
 
Thank-you for supplying that link! It would be nice if one like that could be found for the later version with EL34s as well!

Dave
 
Back
Top Bottom