Unlike silicon chips, which go out of production soon as they're no longer fashionable, valves never go out of date. Almost all popular audio valves are still in full production and readily available. Good news for lovers of valve amps!

But what about the exceptions -- the ones with rare and extinct valves? Are they destined for the scrapheap?

Classics like this gorgeous Luxman 38FD are in serious trouble... their Compactron output valves have been out of production for decades. The precious few that are left are rapidly running out and are hideously expensive.

By the 60's, valve technology's dominance in comsumer products such as black-and-white TV sets was slipping away fast. The introduction of the Compactron series of valves was pretty much an attempt to claw back some of this lost ground by offering manufacturers smaller, more cost-effective valves for specific applications.

If anything, Compactrons just made valves less attractive by muddying the water... production didn't last as long as the cheap TV sets they wound up in and the species went off to join its friends, the Wooly Mammoth and the Yeti.

Few Hi-Fi manufacturers were seduced by the ill-fated Compactron - unfortunately, Luxman was one of them.

We've seen forum threads about modifying 38FDs to use available valves but they balk at the prospect of changing transformers and the limited choices in 12-pin valves.

But if the voltages and the sockets were taken out of the equation, the 50CA10's electrical characteristics would not be unlike those of a triode-connected 6L6 - which is one of the commonest valves in the world... But then you'd have to find some octal sockets that fit the chassis and then re-wire the output stage and then somehow shoe-horn in an extra transformer for the heaters...

Oh look... here's one we prepared earlier!

So, now we're wired for mainstream valves, let's look at the selection criteria... we need valves with dissipation similar to the original Compactrons, smaller envelopes to fit in the tight space, readily available, reasonably priced, reliable and of course, premium audio valves.

These superb TungSol 5881s easily tick all those boxes and are available in matched sets. They're a robust little performer that just oozes quality and sounds lovely. The other ancient Japanese valves get swapped with some nice new electro-harmonix EF86's and 6AQ8s.

After some testing, we replace a handful of components that have gone off spec, re-fit the timber sleeve and apply a little furniture polish.

The 38FD comes to life and it is immediately apparent that it has regained all of its former glory and then some! After years of gradual deterioration, it's easy to forget just how good these old Luxmans used to sound... but this unit is now wide awake and ready for work, surprisingly fresh and dynamic but still with that legendary Luxman musicality.

Not all endangered valve amps can be modded for available valves in this way, but this one's done and ready to make beautiful music for the next 40 years!

The Jadis Defy 7 -- Impressive -- Heavy -- Lots of metal.

There are two ways to make good valve amps; you can either draw upon the legacy of classic designs that have worked superbly in countless products for many decades, perhaps even improving on them (e.g. Conrad Johnson)...

- or a talented engineer could develop completely new and patentable topologies that spawn a whole generation of new audiophile products (e.g. Audio Research).

- or you could defy logic and make an amp like this one.

There's no nice way to say it... This thing has issues.

The owner was a bit disturbed by the unit's failure. It made an 'expensive smell' and then smoke came out.

Thirteen components showed either complete or incipient catastrophic failure with several deep burns in the board. Further, it was obvious that this was not the first time this unit had failed in this manner.

Repair (i.e. just fixing what was broken) was therefore not an option as there was obviously an underlying instability which at first impression, seemed to be inherent in the design. Further research would confirm this impression.

A little Googling turned up several accounts of Defy 7s blowing valves, fuses and smoke. After the schematic was downloaded, a comedy of errors unfolded. The driver stage was overrated and overloaded, the output valves incorrectly biased. Below every pentode, an incorrectly rated fuse failed to adequately protect a badly designed cathode circuit. Almost every other valve amp in the world uses a sacrificial resistor and/or a B+ fuse. The fact that this amp has a dozen replacable fuses in holders indicates awareness of a problem but not of its cause.

Every part that's blown, fried or just wrong gets removed. Carbon deposits are removed and the board gets patched.

The old driver relied heavily on global feedback for linearity and could scarcely control the Miller Effect of the pentodes resulting in grid-positive runaways and red plating.

A new front-end and driver topology is selected and then optimised by SPICE modeling software. The new topology (loosely based on a Hafler/Keroes style phase splitter but employing dual 12AT7 SRPP gain/driver stages) is laid in point-to-point with some teflon/silver links.

This provides plenty of clean, linear drive with buckets of headroom and contol making it possible to correctly bias the output stage for long-term stability and reliability.

Twelve sets of upgraded, corrected support components are installed around the pentodes. Cleaned up, patched, re-loaded and re-configured, the board is re-installed. A new heater elevation circuit is added. Initial tests yield results consistent with SPICE software modeling... a little tube rolling plus a tweak or two and the unit comes up dead-centre, rock-solid.

Listen Test: Powerful and competant. Very much in control of the load without losing valve amp detail and sweetness - if anything, it has gained in this respect... Imaging, depth, transparency - all in line with high-end expectations.

Conclusion: Above and beyond the scope of repair or upgrade, this job amounts to a comprehensive design review. The question of cost/benefit ratio arises. In this case, the unit was purchased for a reasonable price and the work equated to approximately 30% of the purchase price. IMHO, the total outlay vs. the end result represents a cost-effective solution. After six months of flawless operation, the owner reports that it still sounds stunning and very significantly better than it ever did before the modifications.

This amp quickly found a spot in my personal 'Top Ten'... after it was fixed of course! An oldie, but a goodie, it's still one of the largest, heaviest monsters around.

The DM-100 is rated at 100 Watts which is conservative to the point of silliness, really. Despite the fact that it's 100% class A, this amp has seemingly bottomless reserves of power and never seems to be working hard no matter what you throw at it.

This one had been serviced overseas and operated for a few months and then mysteriously failed.

There's a triple decker sandwich of PCBs mounted to each of the large heatsinks. You can't see the lower one (the current amp) in this shot of the right rear corner, but you can see the voltage amplifier PCB which has its own small heatsink and the voltage amp's power supply which sits alongside the large white capacitors on long brass legs.

This PCB had obviously been damaged due to improper re-assembly when the amp was serviced overseas.

The lower part of the the PCB had been badly deformed to the extent that the components on it had been crushed up against the small heatsink on the voltage board. The fact that it operated for several months in this condition is a testimony to the quality of this product.

It's very unusual to see a fibreglass PCB with such a huge warp in it. The reason for the deformity does not become obvious until the unit has been disassembled.

After lowering the big heatsink on the right side of the amp, the reason for the deformity becomes clear... the corner of the board has gotten snagged on a clinch-nut securing one of the feet to the bottom panel (see magnified insert).

How on Earth did this amp go out in this condition? It's just improper re-assembly, pure and simple - amazing it even worked, let alone pass inspection!

After the PCB is repaired and straightened, each of the blocks (main power supply, current amplifiers, voltage amplifiers and voltage supplies) are inspected, serviced, cleaned and individually tested prior to re-assembly.

The amp sections are then carefully lifted (as opposed to shoved!) into position, observing mechanical clearances.

As a final touch, the bussbars are removed and polished prior to testing. After this amp has been on for an hour or two, it's like a barby full of hot rocks but it sounds every bit as good as you'd expect a big Gryphon to sound.

Question: How would you connect a state of the art active or semi-active speaker or sub to an all valve system?
What if it was 10m. from the preamp?

Here's a no-compromise, custom solution which converts a valve amp/semi-active combo into a true bi-amped system.

A VTL preamp and monoblocks drive a pair of Avantgarde Duos which have mid and high horns with an active bass box. The valve/horn combo may be ideal for the mid/tops but the problem is that the bass gets amplified twice; once by the valve amp and then again by the speaker's internal amp. The Avantgarde's bass box is naturally tight and fast but a valve power amp's bass is typically slow and sloppy by comparison. The Active Bass Interface brings the bass back into sync with the horns. In this case, the Interface is styled to compliment this TL-2 preamp.

The rear panel contains two high impedance RCA inputs, two low impedance XLR outputs and a pair of stepped gain controls designed to trim out any net imbalance in the sensitivity of the the two circuits. In this way, we can run the four power amps in parallel instead of series.

As this individual unit is custom built to team up with the TL-2, the layout rationalises the signal path. Unused tape record outputs are commandeered and a short RCA cable links the boxes. The ABI's bridging impedance presents a negligible load to the TL-2 and the balanced outputs will drive any length of cable with ease.

Avantgarde Duos have speaker terminals on the rear plate of the bass box as well as auxiliary balanced line inputs.

Well, the current production units do anyway...

These Duos were an early version which didn't have XLR inputs fitted so the backplates had to come in for mods as well. Balanced line inputs and input select switches were fitted and the panels tested with the ABI and some custom made 10 metre balanced interconnects.

XLR inputs are also fairly commonplace in high-end subs.

A good valve preamp is a thing of beauty but their higher impedances and limited connectivity doesn't have to limit your choices - a simple bi-amp arrangement, or even a good subwoofer can give a valve system a huge upgrade if you can resolve the impedance and connection issues.

Another 'Top Ten' candidate, Burmester 911. This one had been the centerpiece of a reference system for some time and had convincingly wiped the floor with several high-end contenders in A-B shoot-outs.

After many intensive listening sessions, a day came when this 911 lost its edge... it still worked - even sounded good but that 'Je ne sais qua' was missing.

Difficult to discern what exactly was going on here - so let's start with a general service and 'shampoo'.

There was no denying that the performance had dropped off - it wasn't a massive change, but as it was enough to warrant investigation. As this phenomenon was common to both channels, the logical place to start was the power supply.

Close inspection of the underside of the PCB revealed signs of mechanical stress in the soldered connections. After re-soldering, the power supply is remounted and the rest of the amplifier undergoes similar assessment.

One thing that was very obvious about the Burmester, was its ability to reveal subtle differences in interconnects and power cables in any system it powered, so it's reasonable to expect that this would extend to internal connections as well. On inspection, some oxidation of internal connections is noted, along with some loosening of crimp terminals.

Once all the connections had been cleaned and coated, re-tightened and/or re-soldered, the amp is tested - and yes, it is once again spine-tingling, as it should be.

Audio Research D-79 - I don't think there would be too many audiophiles of my generation who haven't dreamt of owning one of these at some point in time. Apart from being one of the nicest amps to listen to, it looks like a very important piece of equipment with those big groovy meters, knobs and fuses.

The thing is though, there's just not that many of them in the world - and if you've ever pulled one apart, you'll know why - it has to be the most complicated valve amp ever made and must have cost a fortune to manufacture.

This one was blowing fuses - lots of 'em and not just any old fuses, the big ones that cost $15 ea!. The owner had only had it a short time but it had clearly had a checkered history. The first issue was that someone had tried to use the wrong fuses and damaged the fuseholders so the front panel had to come off to replace them. In the process, it became obvious that this amp has been repeatedly fiddled with - repairs, 'upgrades' and component substitutions were evident and really only making things worse.

The major issue however, came down to leaky capacitors... the original capacitor bank consists of large can electrolytics mounted on a PCB in a central compartment formed by two metal partitions. There's just no point in trying to source replacements for these... they would have been out of stock for maybe 30 years and the whole array only adds up to about 3,000 uF anyway.

Translating the capacitance and voltage requirements of this circuit into current-production component values, we find that 28 modern capacitors will do the job nicely. Of course, they're a different size and shape so a new mounting arrangement has to be devised.

A special mounting bracket is fabricated to fix the capacitors to the partition plate and the assembly is completed with some old-school point-to-point wiring.

The new capacitor bank is fitted into the chassis. When finally assembled, the other partition plate will help to keep fingers away from the capacitors which will hold a lethal charge when the unit is energised.

The new caps have slightly higher capacitance than the old array and are superior, simply because of improvements in the manufacturing process. While the amp is in this state, several previous dodgy repairs are reworked in line with the original standards of workmanship (looks like this amp's made more comebacks than Rocky!).

Underneath the chassis, all the wires that attached the old capacitor bank were carefully labeled so the final phase of re-assembly is mostly about redressing the wiring looms to suit the revised layout. When the wiring is finished, the unit is powered up without valves to check the unloaded socket voltages. Valves are then fitted and the unit is given a preliminary test, alignment, burn-in period, realignment and finally a listen test. Worth the effort to do it properly, this old classic should sound lovely for years to come.

After a session of tube-rolling, the D-79 (above) began to intermittently blow plate fuses. The problem was due to seriously worn out, gnarly valve sockets.

Branded "American Phenolic" (Amphenol?) these sockets look older than the amp and have seen untold insertions and withdrawls.

The output valve sockets are part of a subassembly that sits like a bridge between the sides of the chassis. The sockets were soldered to a PCB (above) that contained some support components. The chances of finding an exact fit for both the bridge and the PCB were not good -- but as the PCB wasn't in great shape, some tagboards and ceramic sockets get wired in point-to-point.

The rebuilt valve bridge installed back in the chassis. The ceramic sockets make excellent contact with all valve pins and are much stronger than the originals.

The original PCB had survived several repairs with burns and patches all over. It really wasn't worth saving.

Converting this stage to point-to-point pays off in terms of sound, reliability and also provides a cleaner air-flow.

This is a nice little amp... it's a well balanced, competent performer which drives most loads without a fuss and looks the part as well. The Mark Levinson 331 has a comparatively low resale value and can often be found online at a bargain price. Team it with the right preamp and it will do the job nicely in most systems.

So, what's the bad news?

There's a common issue with these, which appears to relate to a defective batch of capacitors... which is possibly why they tend to a bit 'bargain priced'.

Flip a 331 over (watch it, they're heavy and pointy!) and you'll likely spot some dried-up, yellowish goop near one or more of the round feet on the bottom panel.

The culprit... one of four large blue capacitors in the power supply. These components contain a chemical, composed mostly of water, sealed inside the can. If, for any reason, they overheat, pressure builds up inside the can until the relief valve (the little round bit at about the 3 o'clock position in the photo) pops and the cap disgraces itself all over the joint. All four caps had dropped their bundle and it looked like The Three Stooges had a pie fight in there!

After removing the big heatsink on the left side of the amp, we can get access to the power supply for left channel. The two smaller blue capacitors usually don't fail - the big ones are the chunder-caps so the next task is to remove the clamps and bussbars holding them in place.

Then comes the fun part - cleaning out the dried-up goop... and then you get to do it all over again on the other side!

After cleaning, we can replace the caps but the big blue ones are very expensive and difficult to get... but why would you want to put those back in there anyway?

This array of smaller caps provides similar capacity, has a higher thermal and voltage rating and their response time is faster so they sound better. Oh, you want more? Sixteen of these caps cost less than ONE of the big blue guys!

The finishing touch is to remove any traces of electrolyte spillage from the baseplate. No rocket science here, just lots of hot, soapy water and elbow-grease required.

This is a labor-intensive repair but the good news is that what we save on parts nicely offsets the labor and you wind up getting a sonic upgrade at no extra cost.

Overall, a worthwhile effort and a satisfactory end result.

"Hi Dallas,

Thank you for your great work on my Levinson. It sounds very good, even better than I expected. You are right, I probably had not enjoyed its full potential for a long time, if ever, since I bought it second hand."

Richard

The legendary Phase Linear 700 Series II. Its performance isn't exactly what you'd call 'State-of-the-art' but its bold, gold panel and those oh-so-sexy Nightrider power meters have undeniable appeal - and at 700 Watts, this baby still demands attention. This unit was in fine condition but had blown its outputs once and then a second time after an inexpert repair attempt.

The choice: Should we jump through hoops to source the obsolete parts required to restore it? - or - should we give it a total makeover with some 21st. Century technology?

Just kidding... we already knew the answer to that one! The 700 wasn't the only super amp of its day but it was probably the one that finally crushed the notion that valves would always dominate in high wattage applications. This large array of transistors is what delivers the grunt from that massive power transformer. This unit blew a second time because the array had not been properly repaired... it's not enough to replace just 1 or 2 shorted transistors, the entire Right channel cluster should've been replaced.

The old output transistors were originally intended for use in electric golf-buggies and were only used because they were the biggest thing they could get their hands on back in the '70s. Quasi-Complimentary topology was dictated by the fact that the transistors only came in one 'sex' but that's not an issue these days, so the chassis gets rewired to accommodate a Complimentary-Symmetry array of super-matched audio transistors which are heaps faster, tougher and more linear. The old gal now has the output of a modern high-end amp!.

After modernising the output array, the drivers get an upgrade which involves fitting some small flag heatsinks because the new devices are a different physical shape to the originals. While we're at it, the gain stage may as well get in on the makeover so the old bootstrapped quasi-current mirror gets replaced with a constant-current loaded Wilson configuration (seen mounted on the custom heatsink across the rear) and the bootstrap caps can now go straight into the trashcan. This faster, more linear setup is more in keeping with the upgraded outputs and drivers.

Shakedown time. Prior to refitting the panels, the 700 gets a major test and tweak session followed by a 48 hour burn-in... and then it's time to rock the house!

These amps always reminded me of the muscle-cars of the same era - heaps of straight-line grunt but no finesse but this one has lost that 'industrial grunge' and deftly presents a deep, wide soundstage with good spatial separation and dark silences. Its resolution of fine detail is surprisingly eloquent... but it still kicks like a crazy horse!

"The amp powered-up and in my system sounds stunning. Very quiet, very detailed. Many thanks"

Bruce

OK - I'll admit it - I looked at this and said: "No way!"

This ugly little pile of scrap was supposed to be a Dynaco Stereo 120. Now, these were a popular kit and so there are a lot of them around and sometimes the construction is a bit amateurish... but this was not a pretty sight.

Plus, this one had been visited by the naughty technician who had perpetrated some pretty dodgy modifications.

There wasn't any part of this amp that was unmolested - but the owner was genuinely fond of it and was serious about bringing it back to life so the job was on.

It was a bloody miracle this thing ever worked at all with broken tracks, fried components and globs of solder everywhere - eeeew!

The only thing for it was to strip it down to components, assess what was reusable, replace what was not and then to build a new Dynakit 120 from scratch.

The PCB was littered with cuts and breaks and links and blobs which were obviously someone's idea of upgrades but I couldn't see any evidence that these changes were in any way improving things.

After studying the original Dynaco manual, I decided that the best course of action would be to restore the amp to original spec and to do it right.

I mean... who on Earth has designed more amplifiers than David Hafler? OK - this is some of his early solid-state work but... who on Earth has designed more amplifiers than David Hafler? I can't think of too many!

Rebuilt in accordance with the original design but using some 21st. Century semiconductors (plus a few tasteful tweaks and some nice gold connectors), the Stereo 120 comes to life.

After a 48 hour burn-in period, this amp sounds fine - it's a classic 60's sound that is very likable and about as good as a result as could be expected from this chassis.

"Dallas
Just thought I'd write to you to tell you how happy I am with the Dyna re-build, you've done a great job! The amp was never going to be a world beater but it sounds musical, smooth and listenable. It sins are of omission, it doesn't add any harshness etc. Anyway thanks again and next time I need any of my equipment modded or repaired I'll call you."

Jon