Taming the HH Scott 208

dcgillespie

Fisher SA-100 Clone
Subscriber
I spend much of my time hanging around Fisher gear, but that doesn't mean I don't appreciate quality pieces from other manufacturers as well. Besides Fisher -- Eico, Heath, Mac, Dynaco and Scott pieces all populate my listening room. Within my Scott stable, I have a very nice LC-21, LT-110 (brown face), LK-48B, LK-72B, original LK-72 brown face (some may remember my efforts with that one here):

http://audiokarma.org/forums/index....vintage-classic-scotts-original-lk-72.783989/

and I even have a very nice 335 Multiplex Adapter I can press into service when needed. But one piece of Scott equipment I didn't have and always wanted was a 208.

To my mind, for the type of service the 208 was primarily intended for (console power amplifier duty), it stands without equal where 7591 power packs are concerned. Yes, Mac made the well revered MC225, but it was hardly a console piece, meaning that for me, and other than the Mac, all other 7591 basic power amplifiers bow to the 208.

If memory serves, neither Heath nor Eico produced a 7591 based basic power amplifier model, and Dynaco never even broached the 7591. Pilot produced one, and while I have nothing against Pilot, their output transformers often lack the performance from similar pieces by other manufacturers. Fisher produced a couple of 7591 based console destined basic power amplifiers. But they require modification (more than just bridging the power switch terminals) to operate in stand alone mode, and the output transformers used are a clear step down from the superb transformers used in their 7591 integrated and receiver offerings. One version is even an 8Ω output only.

Against this field of contenders, Hermon produced the 208. A ready made fully stand alone basic stereo power amplifier, the 208 sports his very best 7591 output transformers (TRA-11-2), as used in the best versions of his 299 and LK-72 integrated models. It was designed to be hidden away, but is plenty good looking just as it is, and always draws an eye. By the end of the vacuum tube era, I know of no product in its service class that rose to its equal.

With that backdrop then, I recently acquired a very nice looking example of the 208, of the older(?) three can variety without center chassis support bracket underneath. It had new JJ power tubes, and a previous owner had painted the transformers. Underneath, one of the coupling caps had been replaced, as had the power resistor coming off the rectifier tube, and two of the 400K output tube grid resistors as well. Oh yeah. The leads to the power transformer were rather long, not at all in keeping with how the factory would have installed it, so the power transformer that was installed had clearly been done so after the unit left the factory. Hummmmm. But the transformer was of the TR-13 variety (-2-1) which was the transformer class used in Scott's 7591 offerings, so with it being a bonafide Scott transformer, I moved on to trying this puppy out. It was offered for sale as good to go........

A precheck of the (original) power supply caps showed that in fact they were well formed and drew no significant leakage current. When I plugged this baby in then everything acted normally -- until the rectifier tube warmed up. No sparks or fireworks -- just a power transformer growling in protest. The B+ to the output tubes was pretty much dead on 470 vdc as called for, so the growling was investigated. The rectifier tube tested very good so no issues there, so the rectifier tube wiring was investigated. That revealed what surely was a "fix" installed somewhere along the line with this unit for a hardly unheard of Scott complaint in some of their models -- excessive B+ voltage: The pin 6 HV lead at the rectifier tube socket had been attached to pin 5 (effectively connecting it to nothing), and pin 6 cleaned up so that everything looked quite normal to the casual glance (other than the long transformer leads). Connecting this lead to its proper pin 6 connection point eliminated the transformer protest (the whole amplifier had effectively been running off of 1/2 of the power transformer and 1/2 of the rectifier tube), but produced a new problem: B+ at the rectifier tube cathode was now nearly 520 vdc, and the output tubes were clearly showing protest with visible plate color in a well lit room. The chalk screen grid bleeder resistor was none too happy about the matter either, reaching roasting temperature in short order. There is little doubt then that the miss-connected transformer lead had probably been miss-connected on purpose somewhere along the line as a remedy to the excessively high B+ voltage produced when connected properly.

So what to do? Had the transformer that's now installed been so because the previous transformer also produced excessive voltage and this was an attempt to fix it? Had the previous transformer simply gone bad or was it used to fix another unit (maybe with high voltage!) and this was the result of the swap or the only available Scott replacement? The unit had multiple previous owners, so its doubtful that the answer to that question will ever be known -- although it's always interesting to know how a unit ends up in a given condition. In any event, what was known was that it was apparently the correct transformer for the unit (by part number anyway), but simply produced too much juice. But there was another factor.

About 10 volts of the 520 volt B+ was due in part to operating the unit from a 121 vac line -- except that at this line voltage, the heater voltage to the output tubes was 6.40 vac -- meaning that if the line voltage was reduced by a variac or multiple current limiters to tame the B+, then the output tube heaters get short changed. So while there was room for a little AC line voltage adjustment, there was nowhere near enough to bring the B+ in line.

So next, makeshift output tube cathode resistors were installed to get a better picture of what was going on. The tubes were well matched, and drawing just over 40 mA on average, so with about 38 mA of plate current and 495 volts of plate voltage, the color in the plate was certainly understandable -- the tubes were dissipating right at their rated plate dissipation level of 19 watts -- and these were JJ tubes! The bottom line then was that if the quiescent current was backed down to save the tubes, then the B+ would go even higher, and clearly overshoot the rated voltage of the power supply electrolytic caps.

With little downward movement in AC line voltage possible, and reducing the output tube quiescent current to sane levels guaranteed to make things worse, something had to be done if any kind of long term dependability was to be expected without this sucker reaching critical mass in my listening room. And yet, there was still one other factor that would add further insult to injury.

Next time. For now, a couple quick pics of the patient.

Dave

Below: One of the better looking 208s I've seen. They aren't all that common, and some look pretty pathetic. This one took a smack on the power transformer, indenting both end bells, and the chassis under it as well. Some creative coaxing undid most of the chassis indentation, while the end bells were deemed acceptable enough.
SAM_2232.JPG
Below: Other than my temporary correction of the HV lead to its proper connecting point at pin 6, this is the unit as received.
SAM_2231.JPG
 
Last edited:
I'm officially jealous! I've always lusted after one of those but have been unwilling to pay eBay prices. I'll be following.

Btw, the Fisher 500B power amp that you helped me build is still singing beautifully. :thumbsup:
 
I'd love to have a 208 to mate with my 130. I love the Scott sound. I'm currently listening to my 340B. That's a very roomy chassis.
 
Dave, nice write up. I don't have a 208 though I have a lot of the Scott "stuff". One thing that I appreciate is the use of the term that I remember and that is calling it a basic amp. I don't know who coined monoblock but I hate it.

BTW, do you think that possibly the power transformer is off of a different amplifier and the bells from your amp were placed on it? If so, that may explain why it appears to be the correct transformer but produces too much voltage. Who knows...
 
Other Issues

One of the things that Scott did with most of their designs, was to use an output tube screen grid bleeder resistor. This does two things:

1. It improves performance by helping to regulate the screen grid voltage when the screen grids start pulling more current with increasing power output. Fisher also did this with their SA-100 and X-202 models, which are highly regarded pieces in part because of this design feature.

2. It cooks everything around the resistor within a 3 block radius.

To help minimize the issues associated #2, Fisher broke the resistor up into a number of smaller resistors to spread the heat around, while Scott ultimately located the resistor above the chassis. Either way however, the screen grid bleeder circuit in the 208 is dissipating nearly as much as a 5th output tube would. The concept does improve performance. But it is also a fact that these types of designs run extra HOT as a result. The SA-100 and X-202 do, and it is quite common to see Scott amplifiers with charred wiring, broken power resistors, and carbonized/cracked bakelite terminal strips -- all from the effects of excessive heat, and time.

Aware of these concerns about Scott amplifiers, I had already planned to install an EFB(tm) Screen Grid Regulator into the 208, which is much more effective than the bleeder regulator system is, and by comparison, dissipates hardly any heat. In the 208, the screen grid series dropping and bleeder resistors collectively dissipate 14.65 watts. Against that, the total heat dissipation for the EFB Screen Grid Regulator is just 0.66 watt under quiescent conditions, and a maximum of 2.5 watts under full power conditions in both channels. With nearly 14 watts reduction in heat under normal use conditions, and far superior regulation, it's a win-win modification to install. Other benefits include increased power output and lower distortion as well.

Due to the improved efficiency that the new regulator brings however, it means that if implemented, there is now less current draw from the power supply (allowing it to run cooler as a result), meaning that executing the EFB modification will elevate the B+ as well. With a makeshift bias control temporarily added to bring output tube dissipation down to more sane levels (the 208 only includes DC Balance controls), it was found that this adjustment, along with the implementation of an EFB Screen Grid Regulator, allowed the voltage at the rectifier tube cathode to rise to and settle in at a stable 540vdc, with the output tubes set to dissipate 15 watts from the plate of each tube, while operating from an AC line voltage of 121 vac. This equates to a cathode current of 32 mA per tube, for 128 mA drawn by the output stage. Add to this another 16 mA drawn by the two driver stages and less than a couple of mA for the EFB Screen Grid Regulator, and you have a B+ power supply that produces 540 vdc when 145 mA is drawn from it. This is simply nuts, with the supply basically producing about 100 volts more than needed or as shown to be required by Scott's other 7591 designs.

Scott was clearly counting on the extra draw from the screen grid bleeder circuit to help pull things down, plus no more than 117 vac applied to the unit. Add to that having the output tubes draw max safe current draw levels and the drop across the 80Ω resistor off the rectifier, and the B+ to the output transformer will still be 20 volts higher than those shown on the schematic -- which are already notably higher than Scott shows their integrated 7591 designs to operate at. Talk about a quagmire. Against all of this is the fact that the amplifiers operate in Class AB1, meaning the current draw from the power supply increases dramatically as power output increases. Therefore, while increasing the value of the rectifier dropping resistor, using a higher drop rectifier tube, or installing multiple AC line current limiters will all in fact lower the quiescent B+ voltage, it will also cause the B+ voltage to droop much more than it normally would as power output increases, which both reduces power output and increases distortion significantly. But still, if the 208 is to operate dependably and properly, some sort of stable reduction in B+ is needed if the amplifier is ever to be enjoyed for the performance it is otherwise capable of. What is needed then is something to provide a fixed voltage drop regardless of current flow, that will then bring the quiescent B+ voltage levels down, but not let it drop any more than what the B+ supply itself drops as power output is increased. What this describes, is a BIG Zener diode. But even that presents significant problems.

The amplifier draws a steady 145 mA under quiescent conditions. But at full power output, the amplifiers are attempting to draw a combined ~ 320 mA, assuming the screen grid bleeder resistor has been ditched and the EFB regulator installed. Under quiescent conditions, a Zener dropping the B+ by 50 volts would be dissipating just over 7 watts, while under full power conditions, it would be dissipating as much as 16 watts -- and this assumes the Zener is placed after the first filter cap. For safe operation of the Zener, a 30 watt device would be the minimum acceptable when de-rating is considered to account for underhood temps. If the Zener is placed before the first filter cap, a 50 watt Zener would be more appropriate to account for the ripple current it is passing. Such a Zener is not readily available, meaning that multiple lower wattage units (likely 5 watt rated) would need to be strung together to get the job done. While this would work, it none the less adds significant heat back into the amplifier, which if possible should be avoided.

Adding regulators on an external sub-chassis with an umbilical cord to the 208 starts to turn it into something it never was, so that was ruled out from the get go. Redefining what we need then means what we really need is something that will handle a high continuous current flow, provide a fixed voltage drop regardless of current flow, and do so while performing a useful function for the 208s basic operation, without adding any unnecessary heat back into the amplifier. That's a tall order. But in fact there is a way to accomplish all of these things in the 208, that produces not only the stated goals, but provides a few extra benefits as well, with no significant down side. Next time. Although I'll give you a hint.........

It glows.

Dave
 
Last edited:
My 208 came from the dumpster and it had had the power transformer swapped before I got hold of it. The last 1/2" of each lead was still attached to where it went, indicating the tech simply clipped out the old and swapped in the new. When I got the amp there was a service tag from a shop in the bay area stuck to the top. I didn't notice any excessively high voltage I installed an inrush current limiter for good measure nevertheless. My power transformer is a TR13-3. 352717 the serial number on my Amp is 127966. Was there maybe a Technical service bulletin on this topic, which has been lost to the sands of time, or has yet to surface?
 
0A3 ? Those won't pass enough current though. You'd need what, 300ma or so at full scream? Gas regs are good for something like 50ma. Something like a 6AS7 as a pass tube might do it with both sections in parallel but thats a bit of a stretch on the supply I'd think.
 
Sounds like a beefy pass tube setup with a Zener dropper on its grid, so that the Zener provides the constant voltage drop with respect to the cathode at a few milliamps, but communicates that voltage drop to the grid which the tube then mirrors that voltage drop at the high current needed...?

That's the only thing I can really think of anyway that would do it at the current needed...

But where to install it? In place of one of the can caps maybe?
 
Following this thread as I have a 299C with what looks like a 208 w/preamp and have high B+ issues.
PT is a TR13-2-1.
 
At this point this is starting to sound like quite a lot of effort/brainpower to preserve a PT model number. Wouldn't the elegant solution be to replace the PT with something more appropriately spec'd?

I do appreciate the exercise in circuit wizardry...
 
I have a 299C and a 340B receiver, I have always been curious as to why the tower resistors on my 299C get so hot but on my 340B they only get warm. For a receiver my 340B runs relatively cool.
 
Train -- There have always been lingering reports of folks with Scott equipment battling notably high B+ levels. It's hardly every piece -- but there are enough cases that it's out there that the issue is real. Within my Scott stable, the B+ is perfectly normal on my LK-72B -- but not so on this 208. You raise a good point in wondering if there was ever a Technical Bulletin addressing the issue. I have postulated that either a transformer vendor just blew it and Scott was not aware or that fact until the units were already produced and out in the field, or Scott just blew the specifications -- but either way, any issues issues from these transformers seems to have been handled on a case by case basis. One way or the other however, the problem is with the build of the transformer, as the other three windings all produce quite normal voltages -- even at today's higher line voltages, but the HV winding does not. I also note that the issue seems to only be centered on transformers designated as TR13-2-1, with a manufacturing code of 138xxxx.

Selmer -- Now there ya go taking all the fun out it! Such common sense should not be allowed!! :)

I appreciate all the responses. Let me get started on the last post........

Dave
 
Resolve

One of the things that always puzzled me about the 208 was the bias supply. It was built like a DC Heater supply rather than a bias supply. I mean really: A bridge rectifier, 18Ω 1 watt dropping resistors, and a four section can filter cap to supply what, maybe 1 mA of current to the bias voltage divider? Really? A single diode half-wave rectifier is perfectly appropriate along with just a couple of filter caps and 3 half watt resistors is all that is needed, and would surely be less expensive to produce. I always knew that if I ever got my own 208, and given the great potential for hum problems that all the small signal tri-pent tubes have, that I would use that supply to also provide DC power to the filaments of the 6U8/6GH8 driver tubes, in addition to the output tube bias it supplies -- that is, until a cooler head prevailed, and excessive high voltage issues arose.

That supply can't power the driver tube heaters because while it produces plenty of voltage to do so, the heaters of those tubes draw .45A each -- three times that of the .15A 12AX7 heater string it would normally power, which would overtax the supply. But going down that road a bit further, I then searched for a similar driver tube that would only draw .15A. Investigation showed that a 19KG8 had exactly the same characteristics as the specified 6GH8, with the only difference that its heater requires 18.9 volts at .15A, rather than the 6.3 volt .45A of the specified tubes.......except........it has a slightly different pin out. Further investigation however showed that within the usual players of the tri-pent tubes, there in fact exists an 18.9 volt version that does have the same pin out, and is a very close match to the 6GH8 -- closer than the 6U8 tube is in fact. Enter the 19EA8 -- 18.9 volt version of the 6EA8.

Wiring the heater of two such tubes in series would then require (basically) 38 volts at .15A -- perfect for DC operation from the strangely designed bias only supply. But then I thought well heck, instead of letting the bias supply power the heaters of these tube, use the 38 volts they require to chew up some of the extra B+, which became a very neat solution -- but it did require some thought to implement properly.

Such an approach has multiple benefits: Besides eliminating any hum problems from the use of DC to power the heaters, it also neatly takes 5.7 watts of what would otherwise be wasted heat being pumped into the underside of the chassis, and converts it into above chassis heat used to heat the cathodes of the driver tubes. Under quiescent conditions, the amplifier circuit draws about 148 mA -- or basically exactly what the 19 volt driver tubes require to fully heat their heaters. With the heaters then absorbing a large portion of the needed voltage drop (~50V), and conveniently expelling that heat from the top of the chassis (rather than under it), the only issue left then is to determine how to maintain the proper voltage across the heaters as power output increases.

Using the voltage drop across tube heaters to provide a useful benefit is hardly a new idea. Fisher and others (Scott even) used this approach to cathode bias the output stage of many a design to not only provide the bias, but also power the small signal tube heaters with DC voltage to minimize the generation of any hum. But those efforts were always dancing on the edge of output tube destruction, because besides the current drawn through the heaters to light them, there was also a large power resistor placed in parallel with the heaters and drawing current as well, to minimize voltage elevation through the heaters when the Class AB1 output stages draw more current with increasing power output. In this instance, the voltage drop across the driver tube heaters would not be providing output bias. But as the output stages of the 208 are also Class AB1 designs, the current through the driver tube heaters will still increase with with elevated power output. A parallel power resistor could be used to minimize this voltage elevation as was done by Fisher and others. But a much more attractive approach is to not use a resistor, but rather, a regulator.

Such a regulator could take the form of a 38 volt Zener diode placed across the heater string. Under quiescent conditions, the Zener would not conduct any current, and therefore dissipate nothing. Under full power conditions however, it would dissipate nearly 8 watts of power (this in addition to heat dissipated by the heaters), and add this under the chassis. While this is a simple approach that would certainly work, Zener diodes are often imprecise, and don't regulate perfectly, with their Zener voltage dependent on the current flowing through them. A better approach uses the Zeners to chew up some of the current, but includes an active element to make the regulating action more precise. The easiest way to add an active element is to call on an old friend, the 3 terminal regulator.

By using a negative 3 terminal regulator in a configuration very similar to that when used to provide EFB operation of an output stage via the cathode connection, a very effective active shunt regulator can be devised that will hold the voltage drop across the driver tube heaters at precisely 38 volts at any amplifier current draw from quiescent to full power conditions in either one or both channels. Of course, to achieve the 38 volt drop to begin with, the output tubes need to draw a minimum of 35 mA each, but with the reduced B+ now achieved, a 35 mA current level now causes the output tubes to each operate at exactly 80% of rated plate dissipation, which with proper ventilation, posses no problem for the tubes. With a new B+ level to the OPT of 475 vdc, this is undoubtedly the original operating point intended by Scott. And, because the regulator causes no significant dynamic resistance to be introduced into the B+ supply, the full power output of the amplifier as originally intended by Scott is maintained -- unlike the reduced performance that would result if larger dropping resistances or higher voltage drop rectifier tubes were used to achieve the reduced B+ voltage. The final 475 volt main B+ level is also the product of placing one CL-80 current limiter in the primary winding lead of the power transformer, providing about 10-12 volts of drop when fully warmed. The important thing is that its inclusion still allows about 6.32 vac to be applied to the output tube heaters, and 5.02 vac to power the 5AR4 rectifier.

Since the current draw of the output stage is important in achieving proper performance, maintaining the health of the output tubes, and now, providing proper heating of the driver tubes, it was decided to go ahead and install a full set of test points and bias controls for both channels while I was at it. Unlike back in the day when a matched quad of tubes of the proper characteristics were available from Scott, Fisher, and others who (as a result) provided no such controls, today's vacuum tube audio environment is very different, making such controls mandatory. In the process of doing this, the output stage was upgraded to the DC balance circuit including control grid stopper resistors as shown in the 299D-C1 schematic dated 8-64. Screen Stability resistors were also installed, as were 10Ω 1% .5W MF resistors in the cathode circuit of each output tube. A EFB Screen Grid Regulator powers these grids, eliminating the high heat dissipation and much poorer performance of the 1.2K 10W and 15K 20W screen grid dropping resistors. Finally, the 208 has an output stage configuration worthy of the performance that Hermon intended for it. Other than the use of EA8 type driver tubes, the input and driver circuit maintains the original Scott design for this version of the 208.

The other half of the EFB equation (the Control Grid Regulator) has not been added, as at this point, I was just trying to get the unit safely and properly operational with the least amount of heat generation possible. The elimination of the screen grid dropping resistors and reduced current draw from the power supply all paid off in spades too as now, after 6 hours of operation in an open air 76F environment, the power transformer finally reaches 140F -- right next to the rectifier tube. On the outside side of the power transformer, the laminations struggle to reach 130F, which is a very safe operation level in deed.

At this point, the 208 makes a wonderful addition, and for me, completes my Scott stable. Very solid, enjoyable, and powerful sound that Scott is known for, and for this unit, dependable performance now, too.

Pics show the changes to the unit as the project advanced.

Dave

Below: Showing the original rectifier tube HV connections as received.
SAM_2213.JPG
Below: From the inside, test points were added to check cathode current, connected in the same order as the output tubes appear on the chassis.
SAM_2234.JPG
Below: New Bias Controls were added for each channel, and all the old cathode, control grid, and screen grid wiring was removed as well.
SAM_2235.JPG
Below: The space between the two T-Strips supporting the OPT secondary wiring was cleared out to make room for the new 38 volt active shunt regulator.
SAM_2239.JPG
Below: And the new shunt regulator installed. The small on-board 9 volt transformer (actually supplying 14 vac) powers a voltage doubler that biases the 3T regulator at 39 vdc, allowing it to regulate at 37.8 vdc below ground level. Behind the transformer, you can see the 3T regulator mounted to the chassis to provide an effective heat sink. Worst case dissipation from the device is 2.28 watts when it is passing 150 mA above quiescent current flow.
SAM_2253.JPG
Below: On the side wall, the new silicon bridge rectifier -- still serving only to provide output tube bias -- and the EFB Screen Grid Regulator mosfet are mounted in the old selenium rectifier holes. Worst case dissipation from the mosfet is just 1.4 watts.
SAM_2254.JPG
Below: Hefty resistors now make up the original 80Ω 10W rectifier dropping resistor. This prevents the original super hot single 10 watt resistor from carbonizing the T-Strip it connects to. In the foreground is the EFB Screen Grid Regulator circuit, and in the right corner, the bias supply, and elements of the heater regulator.
SAM_2255.JPG
Below: The finished 208. As I've done before, I use a 14 ga buss wire to provide a ground rail for all components needing a ground connection in the output stage area.
SAM_2259.JPG
Before: Operating in the listening room, with the new 19EA8 driver tubes now powered by the B+ supply, chewing up most of the originally ridiculously high B+ voltage in the unit.
SAM_2260.JPG

This is simply a wonder piece from Scott -- but at least my unit required significant work to bring out its best. But it is work that is well worth it!

I should be able to post a concept schematic later this evening.

Dave
 
Last edited:
How does the heater-cathode voltage work out now that the tube heaters are run off the B+ line, or have I missed a detail in there somewhere that covers that?

Side thought, could such a regulator arrangement be used with setups that use tube heaters as a cathode resistor in order to stabilize the bias? Thinking this might be a bit of an end-run around the shifting bias problems my Fisher 600 has that causes the two channel performance to suffer considerably compared to the single channel performance.
 
Back
Top Bottom