Gilding the X-202B

dcgillespie

Fisher SA-100 Clone
Subscriber
Following in the footsteps of previous efforts with my 400 receiver, an X-1000 and X-101C, this time it's an X-202B being presented for modernization. However, unlike those projects, this one also includes a full electrical refurbishment, with the thread being offered then simply as a record for referral for those with their own 202B projects.

First up is the laundry list of items addressed, to give some perspective to the work presented in the pics. Those items include:

1. A complete recap of all power supply and audio circuits.

2. Installation of an appropriate AC line current limiter.

3. Installation of cathode sampling and screen stability resistors in the output stage.

4. Installation of test points to allow adjustment of DC bias and balance controls without removal of the bottom cover.

5. Convert the output stages to operate with the speaker common leads referencing ground level.

6. Install EFB for improved output stage performance and extended tube life.

7. Improve phase inverter performance by removal of the inverter "noose".

8. Convert the old Reverb In/Out jacks to new Preamp Output/Power Amp Input jacks.

9. Remove the Aux 2/Tape Monitor Bleed circuits.

10. Remove the two permanent LF filters from the original design. The switchable Low and High Frequency filters remain.

11. Fix any other issues noted along the way.

These are all modifications made to the previous units mentioned, but installing them is never the same in any two models, so it's always a slow go on the first go round. Mix the holidays in there, and its been a solid two month project. Some of the modifications are rather easy, while others are quite involved.

The one section of this unit that was not being addressed was the phono preamps. They had previously been recapped, and were otherwise not the priority of this unit. Therefore, they were addressed only in terms of operational needs.

This unit was initially in rough shape. Missing its bottom cover, it did work after a fashion, but had very poor previous repair work done to it, and needed a lot of TLC: Most of the indicator lamps were bad, the Mode Switch was frozen, the chassis service controls all had a "grind" to them when rotated, the grommets for the phono preamp tube sockets had all but rotted away, and the preamps themselves didn't work. Plus, the thing was just filthy. Still, an initial evaluation indicated that the basic goods were in tact, which allowed the project to move forward. Tubes were performance checked, replacements obtained as needed, and work began.

At this point, the pics present the finished work, with a battery of listening and lab tests on tap later this week.

It will probably take two or three posts to present the pics, with a brief explanation of each provided.

Pics include:

1. The HV power supply area. The 202B has a busy underside to begin with, and with the modifications made, things got a whole lot busier. In this pic:

A. The hi side doubler can has been replace, and the old cardboard protector removed from the old cap, and installed on the new.

B. New Ultra Fast HV diodes now replace the old ones.

C. The low side doubler can has been replaced with discrete caps which have been mounted in place with silicone.

D. Since this unit had already had the Control Plug internally jumpered, that was just taken one step further, and all the wiring to the Control Plug removed. This freed up space, and allows the Control plug to now perform test point duty, making that feature completely non-evasive.

E. In the middle of the power supply area is the EFB Screen Grid Regulator circuits, and good eyes will spot the power Mosfet associated with this circuit near where the AC line enters the unit.

F. A CL-80 rounds out the power supply modifications. This device provides for safe turn on, while still allowing full power output to be developed.

2. The Low Voltage DC Heater/Bias supply area. Here the original can cap was retained, since there is no known replacement for it. Discrete caps now do the work, along with an upgraded dropping resistor as well (10 watts versus 7 watts). Also two small circuit boards can be seen which will be detailed in other pics. Note also the recapping of the tone control amplifier stage. I do not use the snip and tac method of component replacement, as part of what makes a Fisher a Fisher was the original build quality, which I make every effort to leave in tact.

3. The EFB Control Grid Regulator is built on a board that is mounted to the line stage shield box. It works in conjunction with the original adjustment controls, so that the process of setting the bias and balance remains unchanged, except for the quiescent current level target itself.

4. In converting the output connections so that the Common output terminal operates at ground level, it requires the removal of the wiring to the center speaker level control switch (earlier models had a line level volume control in this position), the speaker on/off switch, the off position load resistors, and the headphone jack wiring as well. Since I was in there, the channel identity of the headphone jack was also corrected. On the left, the caps on the loudness switch were replaced as well, and strangely, this unit seems to have been built without an AC switch cover. Since there is no longer a Center Channel Speaker output, the Center Channel indicator lamp now stays at full brightness all the time now.

5. The output stage has had individual cathode current sensing resistors and screen stability resistors installed, with the old cathode resistors removed. Additionally however, all of the OPT wiring was removed, leads extended as required, and reinstalled/reconfigured to finish the Common output terminal modification. This modification causes the unit to reverse absolute phase now, but is a small price to pay for having the inputs and outputs now all reference ground. This can easily be resolved by reversing the speaker leads if that is of concern. Also note on the back panel the absence of the components making up the bleed circuits between Aux2 and the Tape Monitor settings.

More in the next post.

Dave
 

Attachments

  • SAM_1127.jpg
    SAM_1127.jpg
    138.5 KB · Views: 414
  • SAM_1142.jpg
    SAM_1142.jpg
    127.9 KB · Views: 332
  • SAM_1129.jpg
    SAM_1129.jpg
    101.1 KB · Views: 310
  • SAM_1124.jpg
    SAM_1124.jpg
    78.2 KB · Views: 333
  • SAM_1126.jpg
    SAM_1126.jpg
    133.4 KB · Views: 342
Last edited:
You're not kidding that its busy under the hood. Yikes. And I thought my Sherwood was tight.

Did this originally have that weirdness where the 4 ohm tap was chassis ground and the "0" ohm taps became the center channel output or somesuch ?
 
Continuing

More Pics --

1. The phase inverter section. Here, one of the two permanent LF filters was removed, the inverter noose removed, and the HF tailoring circuits have been adjusted for optimum performance as stand alone amplifiers. So converted, this section now operates with 7247 tubes for maximum drive/minimum distortion capability. While the original power amp section displayed a response that was down 1 db at 20 kHz, the new configuration is down just .1 db at 20 kHz (virtually flat), and down .5 db at 60 kHz. The reduction in response is smooth and gradual after that point. Stability is absolute.

The power amplifier section of the X-202B was the best 7591 amplifier Fisher produced. It can be driven to full power with 1.8 vac rms, with the amplifier now in fact delivering a full 35 watts RMS per channel with both channels driven from 28 Hz to 20 kHz, with distortion not exceeding 1% throughout that range, thanks in large part to the addition of EFB. This is all achieved with the output tubes operating with a quiescent current of just 32 ma, producing a level of performance the original design could not achieve with the tubes operating at the factory specified 40 ma. Output tube plate dissipation is 13.5 watts, which represents 71% of the tube's rating. The input impedance of the new power amp section is 450K.

2. For the preamp section to connect with the outside world, it needs to provide a low output impedance, that the original design did not offer. The addition of a small non-inverting unity gain buffer amplifier fills the bill nicely, operating off of the DC heater power supply.

The basic guts of the X-202B's preamp/control section is nearly identical to that of the 400CX-2 preamp, so the ability to use it separately now gives added versatility to the unit. When the unit is operated at such, it is now possible to simply remove the four output tubes, and operate the unit long term as a very close cousin to it's preamp brethren. The preamp section is now capable of outputting 4.0 volts RMS at very low distortion, and is capable of driving amplifier loads as low as 100K ohm, making it suitable for use is SS amplifiers as well. The line stage has a gain of 17 db with the tone controls set at flat. Output impedance is 100 ohms.

3. The line amp section. Here, the other permanent LF filter has been removed, allowing flat response down to 20 Hz now. However, extending the response of a high gain amplifier built into such close quarters as the 202B can result in collateral damage -- and it did in this case as well. Under certain conditions, the amplifier could be made to oscillate, which is completely unacceptable of any high quality piece of equipment. The problem was traced back to coupling from the headphone jack wiring back to the right channel line amplifier. As a result, you can see the extended shielding piece I added to the original shield box, which completely eliminated the problem. If you ever wondered why those partitions were installed in the first place, that's why -- to prevent circuits from interacting that shouldn't.

4. Same song, second verse. With the speaker phase switch located closely next door to the Tape Monitor switch, unwanted interaction resulted. After all, the Tape Monitor switch operates nearly at line input levels, while the speaker phase switch is at speaker level. Therefore, a small shield was added between these two switches to resolve the issue here as well. The reason that Fisher got away without using a shield in either of these locations, is because the HF response was much more rolled off in the original designs, and, because the output wiring was operating 180 degrees out of phase from what it does now. In any event, the shields solved the problem.

5. Time exposure of output tube operation is a surefire way to pick up any tendency towards hot spots or abnormal operation that the naked eye would not normally pick up on. The tubes all hold a steady bias, operating comfortably under the control of EFB.

Next time, I'll wrap it up.

Dave
 

Attachments

  • SAM_1120.jpg
    SAM_1120.jpg
    50.5 KB · Views: 279
  • SAM_1138.jpg
    SAM_1138.jpg
    86.6 KB · Views: 228
  • SAM_1143.jpg
    SAM_1143.jpg
    122.4 KB · Views: 222
  • SAM_1141.jpg
    SAM_1141.jpg
    92.5 KB · Views: 218
  • SAM_1132.jpg
    SAM_1132.jpg
    109.8 KB · Views: 237
Final Pics

A final underside pic is provided showing the completed work. One last item was addressed just this morning which was to remove a low level (but still annoying to me) buzz that could only be heard with your ear up to my Cornwalls. It was a buzz that shouldn't be there, as was ultimately traced to the Fisher shielded cables running between new Buffer Amp board, and the rear panel. These cables are really great for most purposes, but work on a similar principle as the old spark plug wires did: The ground and signal conductor travel through carbon impregnated tubing. Since the ground conductor is bare, it contacts the tubing over the length of the cable, and provides an effective shield for most applications. The operative word is "most".

These particular cables ran close to some HV power supply components, for just enough length to pick up the noise. So, out they came, and in went a cable that contains two individually and tightly hard shielded conductors, that takes a slightly different route in the unit as well. No more noise.

Work to the phono section amounted to installing new socket grommets, and repairing the preamps to work properly from someone's previous effort to recap that section.

A shot of the unit in operation with its face plate mounted up is also provided as well. The jewels are in the wrong sequence, going left to right as green-red-amber-red-green. It's a strange one though -- the jewels look to be authentic Fisher items, unmolested from manufacture. Were the color and sequence ever changed for these jewels?

The entire 202 series of Fisher amplifiers represent some of the best of the best for their day. They are truly wonderful units, and continue to be so to this day!

Dave
 

Attachments

  • SAM_1146.jpg
    SAM_1146.jpg
    156.7 KB · Views: 271
  • SAM_1147.jpg
    SAM_1147.jpg
    55.6 KB · Views: 372
Another truly masterful piece of work Dave :thmbsp: I'm sure, as I pour over the pics and comments, I'll have a number of questions concerning transferable modifications. In the meantime, BRAVO!
 
Dave, I assume you used the (now?) standard EFB topology as described on your other Fisher rebuild threads (such as the X101-C, 400, and 500)? What were the final EFB parameters chosen (screen and bias voltages)? Does this Fisher model use a 6.6K-ish secondary transformer?

I am impressed with your layout/solder skills. It looks mighty tight under there.
 
Also, I noticed the mosfet you use in the EFB circuit is now scheduled for obsolescence and will be discontinued by the manufacturer, although Mouser still has 400+ of them in stock. Is there an alternate you would recommend? I'm not sure how fast these sell, but by the time I get to adding EFB to several of my amps, I wouldn't want there to be a shortage of one of the key ingredients in this circuit.
 
K -- With the 202B, the basic EFB parameters are already established right in the RCA tube manual: 450 V Plate, 400 V Screen, 38 ma quiescent current, and 6600 ohm load. The Fisher transformers are slightly lower in impedance (6400 ohms), which is of little significance. These are basically the operating parameters that Fisher was also trying to emulate, specifying 40 ma of quiescent current per tube. However, with a simple dropping resistor to supply the screen voltage, and a screen current that rises over 300% from quiescent conditions to full power output, those ideal operating conditions get destroyed real quickly -- on top of deterioration cause by droop in the plate supply, and the bias supply remains stable all the while. As a result, 1 kHz THD rises to over 1% at full power in the stock design with both channels driven. Fisher advertises distortion as .5%, but this is based on only one channel being driven.

I deviate slightly from these parameters, frankly because with EFB, you can, and still get very good performance -- much improved over that of the stock design -- while getting the benefits of increased tube life and cooler operation. At 32 ma quiescent current per tube using EFB, power output is increased nearly 25% over that of the stock design, but more importantly, at that increased power output level, 1 kHz THD with both channels driven is just .15%, which is nearly a 10 fold drop in distortion over that produce at a lower power with the stock design. When distortion is measured at the maximum power output level of the stock design, distortion is dropped over a 10 fold amount. Additionally, Fisher operated the tubes at 82% of rated plate dissipation, where as the performance improvements with EFB come with the tubes operating at just 66% of rated dissipation. Tube life is extended, and operating temps reduced. Had Fisher biased the output tubes that cool, then it never could have met its single channel distortion rating, let alone maintaining any dignity when both channels were driven.

So, once again, EFB does not provide ridged voltage regulation but rather, maintains the relative voltage relationships of all the elements established under quiescent conditions, and does so under all conditions of power output. In essence then, order is maintained at all times, versus the "every man for himself" results of the stock design. The only thing better would be to use tightly regulated supply voltages for all elements -- but the only improvement this would offer is a slight improvement in power output, since it is assumed that such an arrangement would have no droop in plate voltage.

As for the Power Mosfet -- I got notice as well that the specified part was scheduled for obsolescence. The specified piece is really a great device for the job, being rated for 500 volts, 30 watts of dissipation, 9A current rating, and having built in Zener and Back Diode protection -- and all in a plastic case for easy mounting: A dab of grease and go. There is little reason that the part should ever fail in this application.

I haven't done a search to see what else is out there, but it sure seems that everything is going SM these days, so you wonder how much longer leaded components will be around for any device. As you can see however, I like to specify any SS device with extreme overkill, which is why I've never had any failures of these devices. In reality, most any N-channel Power Mosfet that can handle the voltage, dissipation and current will work, although those without internal protection should have external components hung on it to do that job. And, it may be that something is available, but not in plastic. The specified part is not that expensive. You could buy a dozen or so and still get a burger for lunch.

The design of the EFB circuit for the 202B is very similar to that which I published for the X-101C, although because of the CT winding in the 202B's power transformer, a rather conventional doubler arrangement can be used, all of which is included on the small EFB Control Grid Regulator board shown in one of the pics.

Thanks for the interest!

Dave
 
Thanks Dave for the detailed answer.

The only thing better would be to use tightly regulated supply voltages for all elements -- but the only improvement this would offer is a slight improvement in power output, since it is assumed that such an arrangement would have no droop in plate voltage.

And presumably at more expense. I'm pretty sure it would cost significantly more to build fully regulated and bullet proof plate, screen, and bias supplies.
 
Absolutely correct, not to mention the whole host of other problems that approach would entail to implement:

1. To maintain the same level of power output, a higher voltage power transformer would be required to allow for the voltage drop that conventional regulators would require.

2. Within the confines of the existing physical structure, there simply is not enough room to allow for the space that conventional regulator circuits would require.

3. The added heat those circuits would generate would be a significant issue to deal with.

4. It is a virtually certainty that the look of the original Fisher product from such an installation would be forever changed -- reducing much of its elegance along with it.

Against that approach, the installation of EFB did not require a single physical modification to be performed to the unit, nor is there any outward appearance of its installation, save for a single small screw head appearing in one of the vent holes behind the power transformer (which is all but hidden by it) to mount the power MOSFET.

By ridgedly maintaining voltage RELATIONSHIPS rather than absolute voltages, all of the problems and issues resulting from the installation of traditional voltage regulators are avoided. The main B+ supply can do what it will as varying audio and AC line conditions impact it, but the operating voltages for all of the elements of the output tubes will remain in relative lockstep together regardless of those conditions. For the original power supply offered then, adding EFB maximizes power output, and reduces distortion to the same levels that fully regulated supplies would produce, without any of the problems associated with that design approach.

Dave
 
Finishing Touches

Over the past couple of days, this project moved into the documentation phase, where all the performance parameters are checked and measured in the lab. Being able to split the two sections apart makes this very easy now, and having the Common output terminals operate at ground level allows for both L and R presentations on the scope (at the same time) to check for performance variations between the channels.

POWER AMPLIFIER SECTION:

In this section, preliminary performance tests had already forecast expected results. But since the output stage had now been stuffed with a fresh set of NOS Westinghouse 7591s, a recheck was required. The results were obtained after a 15 minute warm up period, operating from a 121 vac line, and are based on BOTH channels being driven. Quiescent output tube current = 32 ma. AC and DC Balance controls adjusted for optimum performance:

At 1 kHz = 35 watts RMS (per channel at the onset of clipping).

Distortion @ 1 db down: Ch A: .08% THD Ch B: .10% THD.


At 20 kHz = 33.1 watts RMS

Distortion @ 1 db down: Ch A: .75% THD Ch B: .88% THD


At 28 Hz = 33.1 watts RMS

Distortion @ 1 db down: Ch A: 1.0% THD Ch B: 1.10% THD


At 1 kHz at 2 watts RMS = .005% THD both channels.


Frequency Response: +0/-.1 db @ 20 kHz, +0/-.5 db @ 60 kHz.

Stability: Absolute, with rapid settling under pulsed conditions.

Noise: -94 db below 35 watts

Sensitivity: 1.80 vac for 35 watts output.

Input Impedance: 450K

Output Impedance: 1.0 Ohm.

This is quite impressive performance from any power amplifier -- and particularly so from one in an integrated design. Midband distortion has been lowered on the order of 12X over the stock design, while at the frequency extremes, it has been cut to <1/3 that of the original figures, and all while producing greater power output, and a flatter power bandwidth. Distortion performance is unchanged from that produced when only one channel is driven. While all of the gains in power and distortion are due to the EFB circuits, the low distortion levels are also made possible by the use of a 7247 driver tube. The revised NFB and stability circuits have provided a 3 fold increase in HF response.

PREAMP/CONTROL SECTION:

Other than the addition of the Buffer Amplifier stage and removal of the permanent LF filter in this section, there were little circuit changes in this area, as opposed to those in the power amplifier section. However, removal of the original wiring for the Control and Reverb jacks, rerouting the Recording Output jack wiring, and removal of the filter LF components can always invite changes to the HF response of this section. When initially tested, the 1st line stage (feeding the Recording Output jacks) had a rising response reaching +2 db at 20 kHz, while the 2nd line stage (starting at the Tape Monitor inputs) were down about 1.5 db at 20 kHz. These response characteristics largely canceled each other out, but still presents response errors at the Recording Output jacks, and can lead to overload issues as well. Ultimately, it does no good to make the power amplifier section so wonderfully flat, but not have the preamp/control section complement it.

As a result, the HF compensating caps (C9, C10, C19, C20) were all adjusted for each stage to present a flat response within the audio bandpass. The result is that now the complete line amp/control section also produces a response that is flat within +0/-.1 db from 20 Hz to 20 kHz. It doesn't matter whether the signal is taken from the Recording Output jacks or the Preamp Output jacks, or whether the signal is applied to the selectable inputs, or the Tape Monitor inputs. The response is very flat now in all of these scenarios, and thanks to the buffer amplifier, unaffected by any setting of the volume control.

Other data includes:

Channel Balance: Within .15 db, 20 Hz to 20 kHz.

Gain: 230 mv at the line level inputs produces 1.80 vac at the Preamp Output jacks.

Maximum Preamp Output: 4.0 vac.

Preamp Output Impedance: 100 ohms. Capable of driving 100K ohms load.

S/N Ratio: 80 db below a 230 mV input.

RIAA Accuracy: Within .75 db 20 Hz to 20 kHz both channels, which is much better than the usual +/- 1 db allowance.

Phono gain: 2.7 mV produces 1.80 vac at the Preamp Output jacks.

Also noteworthy, the Volume, Bass and Treble controls all displayed very acceptable tracking between the channels (within 1 db) as well. All in all, this represents a very high level of performance, certainly as good as the 400 CX-2 that the preamp/control section of the X-202B shares its heritage with, if not better. By simply pulling the output tubes, this unit can now operate indefinitely as a very high quality stand alone preamp/control center.

Pics include:

1. Power amplifier only, 10 kHz square wave, left is the upper trace, right lower, normally loaded.

2. Power amplifier only, L channel normal load, R channel, .25 uF only. The right channel is utterly stable, with the well damped waveform showing how the amplifier is first dumping current into the cap to charge it, with the discharge cycle following, maintaining complete composure during the process.

3. Hard to catch a pic of. This is the amplifier delivering 33 watts RMS in each channel at 28 Hz. There is no tendency towards any parasitic oscillations, which again speaks to the high level of stability produced.

4. A 2 kHz square wave, as viewed at the Recording Output jack.......

5. Which is little different than the same waveform viewed at the Preamp Output jack. This is with the tone controls centered, showing the on-center accuracy of (primarily) the treble control.

6. A 200 Hz square wave as viewed at the Preamp Output jack. Tone controls centered, showing the on-center accuracy of (primarily) the bass control. There is some slanting due to phase shift, but the flatness of the wavetop speaks to the flat response down to 20 Hz now that the permanent LF filters are removed.

Operating together, both the preamp/control and power amp sections make for a formidable performance package when the jumpers are installed, or when removed, superb separates, worthy of the best external pieces possible.

And with that, this baby is done!

Dave
 

Attachments

  • SAM_1156.jpg
    SAM_1156.jpg
    39.9 KB · Views: 49
  • SAM_1157.jpg
    SAM_1157.jpg
    50.1 KB · Views: 50
  • SAM_1155.jpg
    SAM_1155.jpg
    47.9 KB · Views: 54
  • SAM_1159.jpg
    SAM_1159.jpg
    48.2 KB · Views: 49
  • SAM_1158.jpg
    SAM_1158.jpg
    49.9 KB · Views: 45
  • SAM_1160.jpg
    SAM_1160.jpg
    48.9 KB · Views: 44
Last edited:
Jeeez - I go away for a week and Dave's whipped through another amp and made it look like a cakewalk.

Beautiful work as always. :yes: I'm jealous as hell!! :tresbon:
 
Back
Top Bottom