A Vintage Restoration of a Vintage Classic: Scott's Original LK-72

[dcgillespie]

A huge element of this story involves when it took place. As a result, those of all ages who are new to this hobby may find it enlightening. For those of us who were really part of vacuum tube audio at the time will remember all too well........

1993: I was given a beautiful brown face LK-72 by a friend. In addition to the typical problems that nearly 20 years of constant use followed by a decade of no-use storage creates, it had an elephant sized problem. But I'm getting ahead of myself.

1983: It was summertime, out in the GA country. A house was being built, and the roof decking had just been finished. The contractor wanted some tunes. Some good tunes. He brings his prized LK-72 that his dad gave him, a tape deck, and some speakers to the job site. The closest neighbor was a couple miles away. The perfect setting to let 'er rip! He asks his handyman buddy to run a makeshift power line from the temporary power pole over to where the Scott is set up and ready to go. He plugs the Scott in first and turns it on, which immediately pops the Scott's fuse rather violently, and produces a dull thud from under the unit. Replacement fuses gave no satisfaction. That's what happens when an LK-72 is plugged into 220 vac........ After the anger and emotions faded, the unit was stored hoping it could be fixed someday, but was ultimately forgotten about.

1993: I had come to know the contractor through work, and through that relationship, he learned of my love for vacuum tube audio gear. The story of the Scott's demise 10 years earlier came out. By then it was a muffled laugh and shake your head story, but he was moving and asked me if I wanted it........

At the time, vacuum tube power transformers were hardly available from Scott (who by then was gone anyway), or the local electronics part house, and the idea of finding a suitable junker/donor unit was truly pie in the sky stuff. What remained of this gear was still stored up in the attics or out in the barns and garages of The Greatest Generation, because they threw nothing away, and had long ago forgotten what all they had to boot.

Even though time would show that there was quite a stockpile of this gear waiting to be dug out of these hiding places, at the time, there was simply no way to find it: In 1993, there was no Internet, no eBay, no Craig's List, and no main line audio magazines anymore, either. At best, all we could do was check with the few remaining (but quickly disappearing) TV repair shops in existence, Good Will (who usually deemed whatever they did get worthless, and so discarded it), attend whatever hamfests we could, or place want ads in the back of Audio Amateur magazine that we sent in and received by snail mail. On top of that, GA was hardly a mecca of high quality audio gear either, so any real chance of finding any of this kind of equipment down here was exceedingly small, and that's being optimistic.

To make matters worse, when I got the Scott, one of the output tubes had gone to air, two had no getter left, and the one remaining one tested weak. So did the two 7199s. In 1993, I would have had better luck finding a suitable kidney replacement than either of these tube types. And a matched quad which the Scott required? Surely you're joking! Remaining NOS tubes were scattered hither-tither in hidden away tube caddies, just like the equipment itself was. Russia had no inkling of what a 7591 or 7199 was (when they did, they never did get the last one right), and there were virtually no tube vendors in business at the time, either. In 1993, this thing was a darn good looking door stop, with almost no hope for repair. Except, that's not the end of the story..........

 

[dcgillespie]

During the previous 25 years, I had built up a decent personal supply of various tubes from the end of the good 'ol days when they were still plentiful, from my continuous searching, and from servicing guitar amps. But not having any integrated amps or receivers at the time, that stash didn't include 7591s nor 7199s, which didn't see a lot of use in guitar amps, either. As a result, even at that early date, I had been experimenting with using the 6GH8As in my ST-70 with excellent results -- from a stash of which came in a box of tubes I bought from an old TV repair shop. But I still had no output tubes or power transformer.

In the course of determining the electrical specs for what a suitable power transformer might be (remember, schematics and service manuals were also the stuff of dreams), I also measured the turns ratio and impedance of the output transformers as well. The first one calculated to 7800Ω, which seemed high for the application. I tested the other transformer. 7800Ω. I used another meter just to make sure. 7800Ω on both. Hummmmm. Even with 440 volts of B+ from today's higher line voltages, there is no way this unit would put out much more than about 27-28 watts in each channel when driven individually -- let alone the 35 or 36 watts its rated for. But of course, things were rated with goofy watts back in the days when the LK-72 was produced. With both channels driven, it would likely manage something over 20 watts per channel, but that's about it. I had no power tubes or power transformer to prove it, so it was all academic anyway, but the math doesn't lie.

Learning that the TRA-11-1 OPTs were 7800Ω units however triggered two other thoughts: (1) That I had acquired the parts and pieces (literally) of an old 222C (including its power transformer) at a hamfest some years earlier, and (2) Leo Fender.

Fender had well established in one of his most popular units (the Deluxe) that 6V6GTA tubes can operate very dependably in a push-pull fixed bias circuit, with about 425 volts on the plates and screens grids. They are still made that way to this day. Into a 7800Ω load with a typical power supply, it can deliver about 22 watts RMS of clean power output. Since the 6V6GTA bottle is virtually the same size as that of a 7591, and requires an octal socket like the 7591, all of a sudden, some thoughts were starting to gel regarding the Scott:

1. Reconfigure the driver stage for increased output and maximum gain, using a 6GH8A tube. The original design calculated to about a 1 volt sensitivity at the input grid of the 7199, with about 20 db of NFB. With 6V6 tubes, a 6GH8A driver, and the same 20 db of NFB, the best I could do was a sensitivity of 1.3 volts at the input grid of the 6GH8A -- not bad since the 6V6s require double the drive voltage versus that of the 7591. Years later, I found that this sensitivity level was still good enough to keep the Aux input sensitivity within the Scott specification for the unit: At .45 vac, it just squeaks by the .50 vac published standard.

2. Rewire the output tube sockets to use 6V6GTA tubes operated with fixed bias as Leo did. I had plenty of those! With the available B+, bias, and load all virtually identical to that of the Deluxe, the R&D had already been done.

3. Use the 222C's TR-12-3-2 power transformer I got from the hamfest. It was a perfect fit physically, and actually offered a number of advantages to the new design:

A. The old design has a 30 uF cap connected right to the cathode of the rectifier tube. With the new design, I could use an 82Ω resistor between the rectifier and first cap, and still have 435 vdc available for the output tubes. The resistor helps to protect the rectifier tube in the event of any unfortunate hot switching events.

B. The new configuration allowed me to dispense with the underside 1.2K 10W screen dropping resistor and the high heat dissipation 8K 20W top side chalk resistor that supplied B+ to the driver and small signal stages. In their places, I could use a small choke I had to provide a much lower impedance source to power the screen grids (located where the old chalk resistor was), and a 7.5K 10 watt underside resistor to now electrically replace the chalk resistor. Since the new driver stages only draw about a 1/3ed of the current drawn by the original driver design, the dissipation from this resistor is now < 1.5 watts.

C. The quiescent cathode current draw of the output stage is 26 ma per tube -- within spitting distance of that drawn by the tubes of the 222C.

D. The TR-12-3-2's DC heater/bias supply winding works perfectly with the DC heater voltage needs in the LK-72, and the bias needs of the new 6V6 output tubes.

E. The heaters of the new output tubes draw < 60% of the current drawn by the output tube heaters of the 222C, so the overall current draw in this application is actually less than in the 222C.

As the original LK-72 offering, this unit has NO bias controls of any kind. Such controls would certainly be desirable for a 6V6 design, but their omission is potentially disasterous for a 7591 design unless a matched quad of factory chosen tubes are used. With everything else being revised then, a full Bias/DC Balance system was also installed, along with appropriate test points.

It was all surely a lot of work -- but when it comes to repair and/or restoration, those days were vastly different than they are today. The Internet, and the availability of quality tube oriented vendors and manufacturers has changed everything. However, I think the resulting product would have done Hermon proud given the available options at the time. It came off so well, that in detailed lab and listening comparisons to other brown face LK-72s of this version that I've restored for clients in recent years, it gives up nothing but about 100 mV of Aux Input sensitivity, and < 1 mV of Phono Input sensitivity, and that's it. Power, distortion, frequency response, and stability are all at least as good as the original. As a result, today, I've had no motivation to restore it to its original design, and continue to use it regularly as is.

Over the years, I've worked on it from time to time as I came up with better ideas to execute the build of the modification, and newer pieces now replace some of the older ones I installed to make for a neater installation both above and below the deck -- but the basic modification as described still remains in place, always done while trying to retain the integrity of the look, and the essence of what the original product represented in the process. Along the way, I also found a rather neglected brown face LT-110 to complete the set. Thankfully, it only needed an intermittent cap replaced, and a thorough alignment.

I'm hardly recommending this as a viable modification today, or even as an improvement modification at all. Rather, out of the love and respect I have for this old gear, I simply offer it to document the extreme measures that were sometimes required at the low point in vacuum tube history to keep the pieces we love alive and well. This LK-72 is still very much alive, and still very well! A few pics will follow.

Happy listening!

Dave

BELOW: The top side, now dressed up with some knobs for the bias controls and color coded test points. The red knobs control the bias, the black knobs the DC Balance. Note the little choke in the rear where the chalk resistor once resided. Note too the 222C power transformer that now motivates everything. 6GH8A tubes now reside in the 7199 labeled sockets. While I'm used to it now, it was a little strange initially getting used to seeing 6V6s in an LK-72.......


BELOW: The ticket that gets you into this particular ball game: The TRA-11-1 output transformer that has lop-sided end bells. Note that the one on the left is notably larger than the one on the right, which is unique to this particular transformer, and the earliest LK-72 offering. Later LK-72s and related units all used the bigger TRA-11-2 transformer, which also had a lower primary impedance, enabling it to actually develop the unit's full rated power output.


BELOW: The new look of the business end of the restored LK-72. A common complaint of Scott equipment is the lack of available ground terminals in the output tube area when bias and balance controls are being implemented. I resolved that by snaking a continuous piece of black 14 ga solid wire through the output tube area, with its insulation removed at appropriate locations for the various ground connections to be made to. Not only does this provide for plenty of ground connections, but it also maintains the integrity of the isolated ground circuit that Scott used in its amplifiers.


BELOW: One benefit that the 7591 simply cannot match: Other than the 6L6 family of tubes, there are few tubes that can match the beautiful blue glow that the 6V6 family of tubes displays under a full head of steam. The camera has accentuated it here, but the glow is still very obvious to the naked eye.


BELOW: In operation, if you didn't know the story, you'd be hard pressed to know that anything was different.


BELOW: Something to keep it company.


BELOW: A match truly made in audio heaven, still faithfully representing all that the Scott brand is known for.

 

[ab2forever]

Fascinating read.

I didn't know there was this difference between the TRA-11-1 and the TRA-11-2 because transcendar lists the 11-2 for the LK72 as does the scott info pages...

I'm lucky to have a good original pair to construct into an amp, so I'm all ears on this one, as that's an outstanding transformer spec.
Frequency Response: 10 Hz to 100 kHz +/- 1 dB @ 1 Watt
Primary Inductance: 90 H @ 120 Hz

The difference in sound between the 7591 and the 6V6**?

Are you able to try out a pair of Reflector/Foton 6P3S-E, I'm very keen to know the differences in THD/IMD between these variations?

I was hoping to use PPP GEC pass/stabilisor valve A21xx/N7x/CV40xx instead, as it's incredibly powerful/compact with gm of 9.5-10.5,- and PPP pairs will deliver close to 30W, - very close to 7591 in fact, & rated peak Ik of 140m/a for a pair.

(In PPP the gm effectively doubles which gives full output from only 20V pk-pk. Just a thought cos they are cheap, ultra compact and easily available for now.)

 

[Dandy]

Thanks for another interesting write up Dave. I would also be interested in the sound difference between 7591 and 6V6. I've never owned a 6V6 amp, but I prefer the EL84 sound to that of the 7591, all other things being equal.

 

[ab2forever]

I prefer the EL84 sound

The question is WHY?

I detest this valve, and they melt, always have, always will... typical Philips/Mullard.

 

[gadget73]

On the bottom side shot, am I seeing a relay added in there? Whats that doing?

Nice work, especially given the lack of "proper" tubes. Also interesting they'd run a 7.8K trafo on a 7591. I guess Sherwood with their 5K 7868 and Fisher with their 10K transformers weren't alone in the unconventional load territory.

 

[Dave451]

Enjoyed the story, Dave--thanks for posting. It provides fascinating context from the "dark days" and how much things have improved lately for someone who is late to tube audio like me. Coincidentally, I have picked up a champagne-face LK-72 and it's next for rehab, now that I've got a Sherwood S8000 IV off the bench that required a lot of attention. It has the can caps marked C1 through C4 on top with big sharpie letters, like a diligent kit builder would do and is in very good shape, right down to those beautiful TRA-11-2 output transformers. Can't wait to get it going and give a listen alongside my Fisher X-200!
Dave

 

[trainbuftony]

I thoroughly enjoyed reading this thread. I have found the same limitations on my scott gear, which all has the same power and output transformers. 299D 208, and LK72, they all make a solid clean 20 watts, but after that things fall apart. I have always wondered what could be done by replacing the power transformer with a larger one, adding a second 5AR4, and attempting to pull 36 watts out of both channels at the same time... Not knocking your work at all I love creativity and I love seeing what people of various skill levels have done to gear to keep the music playing... I'll probably never actually do this because all of my scott get sounds fantastic and rarely approachespecially 10WPC of actual power.

 

[dcgillespie]

Thanks Dave. Contrast was really the reason I put the post up. The actual piece of equipment could have been any one of a number of pieces that measures like that were needed on at the time -- not always to that extreme, and the 220 V accident was surely unique, but the basic story is the same. I felt that the Scott story just made for a more interesting read.
There are sooo many newcomers of all ages getting into (or even back into) tube audio today, that they have no idea just how good they have it -- through no fault of their own; they just weren't involved with vacuum tube audio at the time to know how we quite literally had to scrape something together to make it work. I have no doubt that the Internet has been a huge factor in reviving vacuum tube audio........

Gadget -- Good observation. I was going to comment on it, but decided not to, as I didn't want to stray too far from the gist of the story. The relay is a resolve for the power switch -- but not in the manner that you might think.

Over the years, I have seen all of the barrage of criticism for Fisher and their volume controls, and how Scott "got it right". Fisher volume controls require the user to turn the volume control fully down to turn off the AC power to the unit, while Scott had the push-pull power switch that required no rotation of the control to switch the AC power. Scott proponents suggest that the push-pull switch is the best way, as it dramatically reduces wear on the carbon tracks of the control. True enough. But what they fail to consider, is that by leaving the volume control in position when turning the unit off, it can allow some designs to produce a pulse in the output tubes as they try to amplify the collapsing voltages in the preamp section at shut down. When that happens, that is very hard on the output tubes as they are cooling down, yet trying to pass a strong pluse, which is very hard on the cathode.

In this unit, once it was up and running again, I noted that the cathode driven output tube would display a clear pulse in both channels at shut down with the volume control elevated. It wasn't a particularly notable event in the speakers (nothing that drew your attention), but the blue glow of the 6V6 tubes made it quite easy to see. The relay then is an instant off device that breaks the power to the output tube screen grids. It is a 120 volt AC coil relay that was originally connected directly across the primary of the power transformer. But the darn thing would produce a mechanical buzz that resonated throughout the chassis when it got warm. I nipped that real quick by still powering it from the power transformer primary connections, but now through a dropping resistor and bridge rectifier that supplies it with about 50 vdc. It's quiet as a church mouse now, and when you shut it off, instead of the blue glow fading away in an unequal pulse display (if the volume control is elevated), the blue glow simply vanishes instantly in all of the tubes. It's just one little detail to help make the tubes last. As far as the Scott push-pull volume control goes then, it allows me to have my cake, and eat it too.

We all have our opinions on sound, and any one opinion is just as valid as another. Thank heavens everyone is not like me! As far as the 6V6 goes, over the years, I think I've noted it to be a warmer sound, a little more tuby if you will, but I'm talking very small incremental differences that took years to notice. Again, opinions vary, but I believe that the sound of a given design is largely a part of design itself, rather than specific attributes of the tubes. As has been stated elsewhere, I've seen the same tube type sound great in one design, blah in another, and horrible in still others. The variable is the design, not the tube. That's not to say that there's not an absolute difference in sound between tube types. Let's face it, if my Cornwalls remain a constant, then 60 watts of large tube UL power is going to have a more defined bass presentation at a given power level than small tube pentodes will have. So are the large tubes better? You can't tell from this scenario because the power capability is so different between the two designs. Again, I suggest it's not the tubes. Retest with a UL quad-push-pull (8 tubes per channel) array of 6V6s against a 60 watt UL 6550 configuration and then see. For me, the difference displayed between the 6V6 LK-72, and other LK-72s that use the same OPT is pretty close to splitting hairs. Maybe if my speakers demanded more power there would be a greater difference notable. But even when I crank it -- and I do crank it -- the modified LK hangs right in there. The difference in power output produced at 20 Hz with these transformers between the 6V6GTA and the 7591A is less than a watt, with the .5 watt advantage going to the 7591. No real surprise considering how much bigger the cathode of the 7591 is. The real surprise is how well the 6V6 does by comparison. The modified amplifier produces just over 20 watt RMS at 20 Hz at the onset of clipping, with both channels driven, producing 1.8% (avg) THD. With the 7591, under the exact same conditions, it is just shy of 21 watts RMS. In both cases, power is limited by the transformer.

All tubes will melt down. Believe me, with my band back ground, I've about seen just about everything. EL84s go pop and die. Look at any Vox. EL34s literally melt. Marshall knows this. 8417s go out in spectacular fashion. Any bogen MX-60 or MO-200A proves that. In my experience, the 6L6GC (top honors) and the 6V6GTA are the toughest to kill, but the 6550 is right in that hunt, too. Abuse even these three enough however and they're gone. So much depends (once again) on the design. Conversely, all of these tubes can be made to last seemingly forever. As some may know, I've made a multiple decade study of power tube life, and what affects it. We all have our favorites. I'm old enough that it's changed 2 or 3 times in my life, not only for tube types, but designs, manufacturers, and even styles of music -- but have always stayed with tubes throughout, and likely always will. Get the design right for the tube and the application, and they win in my book every time!

Dave

 

[dcgillespie]

Train -- Other than Mac, virtually none of the vintage units will output rated power with both channels driven across the full audio bandwidth. The only way to understand it is to remember that there was a heck of a power war going on at the time, and no body wanted to deviate from the rating system that everybody else was using (i.e., to reality), and look inferior as a result. It was also likely a hold over from the mono days which weren't entirely gone at that point. To Heath's credit, they were one of the first "common man" manufacturers to provide really honest power ratings.

Dave

 

[ab2forever]

All tubes will melt down. Believe me, with my band back ground, I've about seen just about everything.

EL84s go pop and die.

Look at any Vox. EL34s literally melt. Marshall knows this.

8417s go out in spectacular fashion. Any bogen MX-60 or MO-200A proves that. In my experience, the 6L6GC (top honors) and the 6V6GTA are the toughest to kill, but the 6550 is right in that hunt, too.

My dad worked at Mullard and he held that the Mullard valves were rubbish, worst of all being the EL34 because of the dire G2 emissions.
They had them in the Mitcham works PA system.
Not a day went past without them reaching in and removing another dead EL34, then some clever bloke came in one day a bit fed up and swopped them all for KT88.
Not another second of downtime followed until the works manager found out and had all that GEC STUFF removed forthwith, whereupon the thing came back to its normal behaviour.

IDEM the EL84 which is a horrible pansy little valve designed for cheap SE radio sets to economise on driver valves.

As for the 8417, I don't believe the vast majority of stuff which passes for rumour.
With the wrong driver impedances it's well known large op valves will self bias and run away. ALL OF THEM.
I haven't melted an 8417 yet, and that's funny, cos it's approaching the absolute maximum ratings every time it's run.

I had enough of the 6L6/807 family to know how non linear they sound and the melting grids, glowing anodes and sparks of the 807s will remain a fun memory of teenage years.

Then I bought a couple of old submarine PA systems with PP KT88 in.
The output transformer were designed for a load of 0.5ohm as it fed a copper bus the length of the sub and speakers were all in parallel.
The old GEC KT88 are unkillable!
I tried hard enough!

Funnily enough I remember the 6V6 with fond memories of childhood cos my Dad's DIY record player system c/w nice microphonic 6BR7 & partridge/parmeko everything sounded so good.

It was him incited me to use up his remaining partridge inventory on first an ultralinear 6L6 amp which we tested at school down to about a strong enough 2-10hz as to make all the doors in the entire school buildings rattle.

We then make a whopping distributed port bass reflex thing with a 12" and 15" driver for bass guitar to use with the 807 AB2 amp.That thing was so darn efficient at 75W it shook the whole house.

'nuff said.
I don't believe half the stuff that's written these days, and neither the prices which have gone beyond scary!

 

[ab2forever]

Train -- Other than Mac, virtually none of the vintage units will output rated power with both channels driven across the full audio bandwidth.

Cmon Dave you know perfectly well that the good old MO200 did exactly that, but it was deliberately doctored to make the power tail off at LF. It's normal it's an industrial unit!

The Bogen claimed full power of 100W from 10hz, and if you change one or values it does exactly what it says on the tin... as does mine.

 

[Rob Thomas]

Really neat write-up Dave!

I remember seeing this little unit and commenting on it when I dropped off the TA. Your response was that there was a neat story behind it but never elaborated and I forgot to ask.

The longer I am into the vintage audio hobby, the more I realize that the personal connection and history of a piece is as or more important than the sonic signature. A restored Fisher or Scott sounds wonderful and receives a great deal of praise from vintage audio enthusiasts, and yet the back story and journey of each piece is always unique and adds so much more to its worth. If these old pieces could talk!

 

[gadget73]

Again, I suggest it's not the tubes.

And in a properly designed circuit, it should not be.

 

[ab2forever]

And in a properly designed circuit, it should not be.

Sorry, if you have read the Norman Crowhurst transformer design papers, he states very clearly that distortion caused by the valves is many orders of magnitude greater than that produced by well designed transformers.

That means the curves, current, non linearity and distortion characteristics of valves are all different.
There is a noteable sound difference between triodes, B-tetrodes and pentodes, and it's well recognised pure pentodes have lower objectional IMD that Beam tetrodes.

Non or well aligned grids and spacing are also absolutely critical, causing or limiting screen grid emission, which is why there is no doubt at all a KT66 sounds noticeably different from a 6L6GC, which sounds different from a modern late dev frame grid valve.
The latest TV valves and series stabiliser valves have huge cathodes which in some cases are so low impedance they can exceed amps.
That strongly impacts the winding ratio, matching and damping factors.
No way can it be otherwise.

 

[Patrice B]

Train -- Other than Mac, virtually none of the vintage units will output rated power with both channels driven across the full audio bandwidth. The only way to understand it is to remember that there was a heck of a power war going on at the time, and no body wanted to deviate from the rating system that everybody else was using (i.e., to reality), and look inferior as a result. It was also likely a hold over from the mono days which weren't entirely gone at that point. To Heath's credit, they were one of the first "common man" manufacturers to provide really honest power ratings.

Dave

Ok, it's why Heathkit states 25w/ch for the AA-100 I would guess...

 

[dcgillespie]

Patrice -- That's another amp that used a high primary impedance OPT as well (~ 7600Ω), which reduces the power output. But at least Heath got the power rating correct. I used a pair of the AA-100 OPTs to build a very successful clone of the Fisher SA-100. The transformers are excellent in that application.

Dave

 

[ab2forever]

Do you have a frequency response curve of that?

 

[dcgillespie]

Search for the thread on the build and I think the performance numbers are in there.

 

[ab2forever]

there was a heck of a power war going on at the time, and no body wanted to deviate from the rating system that everybody else was using (i.e., to reality), and look inferior as a result.

Ever looked at the lunatic chinese power amplifier figures lately?

What changed?