Project: Building a Pass Aleph J Amplifier

After the success of the ACA amplifiers, I decided to tackle something a little more difficult: the Aleph J, which is yet another Nelson Pass design that he released to the community at DIY Audio. Like the ACA, the Aleph J is single-ended design but instead uses two output devices in parallel, along with a current source. Power output is in the 25W range, which is good enough for my small listening area, even with 86dB efficient speakers. Except for the input capacitor on the negative leg of the input, the entire amplifier is also DC coupled which means the output offset must be nulled out.

Nelson Pass has a manual on the Aleph J here.

DIY Audio Build - with better pictures than mine - is here. Also includes a schematic.

I first stuffed the circuit boards, including the power supply PCB, with the passive parts:

And next was the installation of the Jfets and FETS, along with the placement of the circuit board on the heatsink. I ran into some problems that the DIY Audio community graciously helped me with. Turns out that I have too much of a soft touch with the soldering iron and had a few cold solder joints. It's that fear of burning up those rare 2SJ74 jfets that make me nervous. Also note the Keratherm output transistor insulators which are a lot easier to install than the old "white goop" thermal grease, and also make removing the board easier for troubleshooting. 

  

I'm obviously skipping a lot of the build steps - power transformer installation, power switch, and the wiring. So with the magic of my time machine, here is the end result of most of my labor: some magic glowing blue LEDs. I do have to say that the DIY Audio Store supplied chassis and circuit boards made this project a lot easier - everything, with the exception of the power transformer, just fits without having to drill holes or battle with placement.

Here is a picture of the completed amplifier :

Note on playing the Aleph for the first time in the main system: I'm quite impressed - lots of little detail on records that I hadn't heard before, super fast but delicate sounding - ie, not aggressive. Excellent bass control. Some actual depth and wide imaging. They would make excellent mastering/studio monitoring amps.

The sound is a tad "clean" - lacking some of that bloom that I hear with the best tube amps. So like digital television versus 70mm film. Perhaps not the best analogy but the only one I can think of. It does need some more break-in time so I'll be patient. But - at least in this system - it is the best solids state amplification I've heard. I think a really good SE tube amp would be preferable but at half the cost, the Aleph is certainly in the running.​

01/04/19 Update: The Aleph J is one of the more interesting amplifiers that I've ever experienced.  It has such a lack of apparent "character" that it just blends into the background.  I'm not trying to say it is boring, quite the contrary; you can hear the differences between recordings.  That is to say every record sounds different, as to be expected.  And, for a solid-state amplifier, it has some of the best body and definition I've heard.  I'm really beginning to hear the limitations of my front end now!  The bottom end is also tightly controlled too, giving quite a bit of boogie factor even with the small woofers of the Denton speakers.

Now the Aleph still doesn't sound like a tube amp - but it also doesn't have that upper-midrange glare or treble edginess I hear with lesser solid-state amps.  Highly recommended for its neutral character and dynamic ability.  This amp will be running in my system for a long time while I start to explore different turntable and speaker upgrades.  Stay tuned!

Project: Building the Pass ACA Amplifier

 

The ACA amplifier is a Nelson Pass design created to teach basic concepts about amplifier design. It is low-powered (6-8Ws per channel), run in pure Class A, capacitor coupled on the output, and is even single-ended.

There is a long, and very detailed article from Mr. Pass which, to put it bluntly, far surpasses my more vacuum-tube orientated knowledge. I see a single jfet driving a current sourced Mosfet, capacitor coupled on the input and the output, along with some modest feedback.  Anyone who wants to know the nitty-gritty of this amplifier should take a gander at the link for more information. For those who are a glutton for punishment here is the schematic:

 

My own requirements are modest wattage and an attempt to capture some of the magic that I hear with single-ended tube amps. The ACA amplifier can also be bridged, via an XLR input, to double the power, something I will need with my 86dB efficient Wharfedale speakers that are in a moderately sized listening room.  I won't need gobs of power for reasonable listening levels. Since my Classe Five preamp has XLR outputs, I will be going this route, building two stereo amps to use as bridged monoblocks.

Enough babbling - let's get building the first amplifier:

Here you can see the passive parts and the circuit boards. It's best to populate the small parts first - the resistors, followed by the caps and solid-state devices.

Once the circuit boards were stuffed it was time to start assembling the amplifier chassis for PCB fitment, along with RCAs, XLR, switches, and binding posts.

Here you can see the stuffed circuit boards attached to the heatsink. The Mosfet outputs are mounted too.

And I did some more work on the back panel, which has all of the connections.

Powering up the amp for the first time is always exciting. But I ran into a snafu - the power indicator LEDs would begin to flicker. Measuring the voltage from the power supply brick, I could see it was fluctuating from 24VDC (good) down to 5VDC (bad). A check on the DIY Audio forum indicated this is the lap top power supply going in protection mode if there is short.

To troubleshoot I removed a power supply lead from the boards, one at a time.  Once I found the board with the problem, I disconnected it from the heatsink. And then the problem went away. This was a clue that there was something wrong with this board.  I ended up having a cold solder joint that was easily fixed.  I guess those years of point-top-point wiring has come back to bite me.

Here I am biasing the output mosfets, each side set to 12V:

And then once that was done, I was able to - FINALLY! - give it a listen:

Initial Impressions: lower gain than the old workhorse Adcom 545. Instead of 9-10 o'clock on the dial, I have to turn up the volume to 1 o'clock to get the same sound levels. The ACA is very smooth, a little lacking in bottom end thump but still very articulated. Jazz - like Chet Baker's album Broken Wing - sounds fantastic, while harder rock albums seem to lack some of the macro dynamics. Of course we are only talking 6-8Ws here with 86dB Wharfedale speakers (6-ohm impedance) so not an optimal match. I wouldn't say the ACA sounds like a tube amp - some similarities: with a non-fatiguing top-end - but it also doesn't sound like your stereotypical SS amp either. Obviously some break-in may also be needed.

Luckily the second amp was a lot easier to build, taking almost half the time and with no problems to troubleshoot.

In this picture I still have to install the front power indicator LEDs and the top and bottom part of the chassis, but I did manage to listen to them bridged, using the XLR output from my Classe Five amplifier. Very impressive.  Neil Young - Live the Cellar Door sounds good on just about any stereo I've owned, but this may the best I've heard it, even compared to the days when I had an all tube chain and big UREI 813A speakers. Lots of micro dynamics, very realistic, and engaging. The amps sound way more powerful than the 12-16Ws it is making.

Some further listening - The Bee Gees - Trafalgar - revealed a slight metallic sheen that seemed to fade with further break in. Very clean but not lean. Detail, studio reverb, etc was all there. Frequency response was very even, no warts, and less "lossy" than some of the lesser tube amps I've owned. Single ended Mosfet is interesting, to say the least, not quite tube-y, nor is it "classic" solid-state (whatever that means). Just different, even more so than Class T (Tritpath) designs I've built. The old, and much need of a recap, Adcom 545, when put back in the system, sounded more indistinct, gray, and uninvolving, though it does have superior deep bass reproduction (and better damping and more power).

Gain is still a bit lower than I would like. Deep bass isn't as prominent as other amps, but again, 6-ohm and fairly inefficient Wharfedale speakers with 5" woofers here.


Listening Update: After a few weeks in the system, I'm quite impressed.  Very neutral with a touch of sweetness.  Some of the best depth I've heard out of a solid-state amplifier, along with very good detail retrieval.  Some of the better tube amps I've build or owned are slightly better at separating voices and instruments, but the ACA amps do a very passable job, just not quite as much "breath" to the performers.  Whether the tube amps are revealing realism or a coloration is an interesting question.  Considering the capacitor coupling, low output - even bridged, and the low damping factor the ACAs still managed to control the Wharfedale Denton speakers quite well, with just some minor loss of bass control.  These amplifiers just make me want to build an Aleph J amplifier that much more - stay tuned!

New Project: The Command 1625 Deuce Tube Amplifier

It's been a few years since I've built a DIY amplifier, so I approached this project with some trepidation. With my new house with no dedicated listening room, I went from a multi-kilobuck system to something considerably more down market. But as always, the itch to get back into tubes was too much for me.

So what to build? I'm an audio cheapskate at heart - not because I don't like spending money - but it just irks me to throw down too much money for output tubes. Even the price of an old EH 300B has risen a lot the past few years. So with that in mind I decided to use the 1625, which is a 12V filament version of the venerable 807. Most available 1625s were made during WW2 and used for fighter plane radio sets. There are still scads of them out there and the prices are cheap. I've also used this output tube with another amp and liked the results. So why not use it again?

And for those keeping track, this is the second time I've built an amplifier using the 1625, so I added Deuce to the Command 1625 name.

I know I wanted some power which gives me greater flexibility when it comes time to replace my B&W Matrix 805 speakers. So I decided to do something different and parallel the two output tubes. Guesstimated power output would be 16-20Ws if I run the 1625 in pentode (ok Tetrode!). So I scoured the web for ideas, checked out different schematics and asked a few questions. And I came up with the following design (no schematic yet):

12HG7 pentode driver, parallel 1625 output tubes into a 2500 ohm Hammond 1627SEA, a real beast of an output transformer. No global feedback around the transformer, instead I decided to experiment with plate-to-plate feedback.

Screens for the input and output tubes are regulated via gas tubes. Raw power supply is a simple CLC filter using a 5AR4, large polypropolene capacitors, and a 6H choke.

Off on switch, gain controls and banana jacks for 4-8-16 ohm speakers. Also current meters to see the health of the output tubes since in the past I've had gassy 1625s that went into runaway.

Once I was done with the design, it was time to start gathering parts and figure out top plate dimensions and where to drill the holes. For a clean, non-DIY look I went with Front Panel Express.

Soldering everything together was a matter of patience - and my usual sloppy spaghetti point-to-point wiring. I would only devote an hour or two a day to minimize mistakes. I also ran into some fitment issues that required some uh, judicious modifications.

The first time I turned the amplifier on, I was a nervous wreck. As I said it had been a few years since I built anything. Was I going to get a ton of hum? Distortion? Some problem I couldn't track down? Or worse, a whole bunch of smoke?

Well there was no smoke but I did get some hum and a lot of distortion. And so began some very frustrating troubleshooting. Long story short, plate cap tubes made for radio transmitting like to make a lot of RF. Two tubes in parallel only compounded the issue. I'm no expert HAM radio guy, but the differences between the two output tubes may have caused a tank circuit, or something weird like that. Luckily there was a solution - add carbon composition anode stoppers to the output tubes. Since these are plate cap tubes the resistor and small hand-wound coil will have to be added to the lead. Once I did this the distortion and hum dropped to the point where they were no longer audible.

Now I have to build another monoblock and make some cleaner plate cap leads since I'm no fan of exposed voltage.

Sound? Since with this amp I can only listen through one speaker I can't say much about soundstaging depth. But using an extra DVD player and a single B&W 805, I heard a very clean 'n' clear presentation with a good top-end and good bass control from the 4-ohm tap. Not very "tubey" warm sounding but more neutral than that

Stay tuned for Part 2. where I build the second monoblock, give a schematic, and do some listening tests.

Project: Mult-Valve: The Universal Pentode Amplifier

Introduction:
It seems like a long time ago, but I remember when the first Sound Practices magazine came out.  For many American audio nuts, the idea of a using an ancient triode like the 300B in single-ended mode was a new revelation.  Of course this was merely revisiting an old technology, spruced up with some modern twists and a hearty dollop of audio hipsterism.  I easily fell into this new movement, making my first ever piece of DIY gear, a 71A linestage using dual-mono mil-spec potted transformers, gas tube regulation, and some vintage oil caps on the output.  After that, it was a series of amplifiers, including PP 6B4Gs and single-ended 2A3s.  All this on a college student budget.  Once I had a real job, I built was my first 300B amplifier using Thorsten Loech's SV83 pentode driver.

But after a while, my interest in this triode movement began to wane.  It was a problem of speakers:  Low power amplifiers require efficient drivers and my experience with the lower priced offerings, including vintage coaxes and more modern designs, left me unfulfilled.  My first experience with audiophile gear included the wonderful Quad ESL-63 electrostat speakers, and nothing I bought seemed to meet that detail, speed, and snap that I remembered so fondly.  I did, however, found myself willing to compromise, thoroughly enjoying the UREI 813A loudspeaker, which ideally requires more power than I could get out of a common single-ended triode running at sane voltages.

With the UREIs, the best all-around amplifier I've heard to date are the Eico HF-60s.  It's a classic 1950s ultralinear design, but runs the EL34s in heavy Class A.  There is a dynamic ease and sense of unlimited power with these classic 60W monoblocks.  Of course some modern components, the addition of a choke in the power supply, and a "battery bias" mod on the EF86 really brought these ancient wonders up to a new level.  But a close runner up was the 1625 amplifier I built, another Class A wunderkind that sounds much better than one would expect considering the budget build cost.  I began to wonder if I could make an amplifier that would make enough power to drive the UREIs, provide class A output, and have the ability to roll in multiple types of output tubes.  The concept of Multi-Valve was born!

The Design:
Most single-ended amplifiers use Cathode or self-bias.  This is an easy way for the circuit determine the idling current of an output tube.  But since I wanted to roll in different sorts of valves, I decided to go with fixed bias.  This decision complicated the power supply since an adjustable negative voltage is need to put across the grid of the output tube.  This also changes the requirements of the driver tube since output tubes using fixed bias need a smaller grid leak resistor.  That meant my favorite pentode driver was out of the question.  Instead I went with the 5687 dual triode which is a low mu tube that would work well cascaded.



For the power supply, I was limited by the small size of the chassis I had in mind.  So I went with a a simple CLC filter using a 5AR4 rectifier, a Solen, a Triad Choke, and then a larger motor run polypropolene capacitor.  RC filtering with some nice Panasonic capacitors and wirewound resistors provide the lower voltage for the driver stage.  Filament is all AC, but I referenced it to ground using a pair of 100 ohm resistors.  The bias supply is a standard Dynaco-type half-wave solid-state rectifier and some CRC filtering.  A pair of linear potentiometers provided the necessary adjustment, while current meters let me dial in the current of the output tube.  With Pentode amplifiers, a reduction in distortion can be found by the use of a regulated screen power supply.  To meet this requirement, I went with a pair of VR tubes connected in series.

Parts:
Instead of Edcor output transformers with their single output tap, I went with potted James 6123HS, which have the outputs for 4, 8, and 16 ohm speakers.  They also look really nice.  However, for the power transformer, I went with Edcor, but bought plain steel bell covers to match the rest of the metallic look I was going for.  Except for the octal sockets, all the small tube sockets are NOS.  A motor run cap suitable for 500VDC was selected, along with some small current meters, miniature pots for the bias control, and some Tocos volume controls.

The most difficult part of building a tube amplifier is the metalwork and chassis.  Instead of the common wood sides, I wanted something different.  Searching through Ebay, I managed to find a nice Chinese unit that included a bottom and top plate, holes drilled in the back for speaker binding posts, an IEC connector, and a pair of RCA jacks.  It's about the size of a Dynaco 70, but with the size of the output and power transformers, I knew it was going to be a tight fit.  To make the top panel, I did a lot of measuring and made several different versions using the Front Panel Express software.


Testing:
Once I had everything wired up - and it was a tight fit getting everyting inside that small chassis - it was time for some testing.  I first tested the bias controls, making sure that the maximum voltage was set.  I then plugged in all the tubes - GE 5687s, Sovtek 5881WXTs, a GE OA2, and a GE OB2 - except for the Shuguang 5AR4.  Leaving the rectifier tube out allowed me to check that I hadn't messed up on the filament wiring.  Once that checked out okay, it was time for the smoke test.  After I plugged in a pair of cheap test speakers and a running CD player, I had the DMM ready to go with the ground lead already attached to a ground point.

Is there anything more exciting than turning on a brand new tube amplifier?  Okay, don't bother to answer that question - of course there is!  But I still get a little rush even though I've been at this still hobby for 24 years now.  To my surprise there was no smoke or flames shooting out the top.  Instead there was a nice purple glow of VR tubes and that warm orange filament color.  After some quick voltage measurements, I adjusted the bias upward and music started flowing out of the little 4" woofer of the Pioneer test speakers.  Ah, it sounds like magic.


Listening:
After some further measurements and some very minor fixes, it was time for some listening.  The Eico HF-60s were unplugged and disconnected from the system.  Up went the Multi-Valve amp and on went the Cardas wiring.  I replaced the Sovtek 5881s with a pair of vintage Tungsol 6550s which required a bias change.  I started with a Schiller CD on the Sony player, adjusting the gain knobs until I found a suitable match with the Quicksilver preamplifier.  After I was happy with that, it was time to spin some vinyl.  I started with some Tom Waits, checked out some Willie Nelson, and ended up with some Police.  Initial impressions was an amplifier that sounded a little lean, but had plenty of detail, very good bass control, and a suprising amount of dynamics for such a relatively low-powered amplifier.  There was a real 3-D sound with the vocals projected in front of the speakers.  A good beginning...

Of course this amplifier was still incredibly fresh.  Except for the tubes, all the parts were brand new.  So over the course of the next few nights, I used this amplifier to provide background music as I worked out, played video games, or started to brew another batch of beer.  As the hours went on, the lean sound began to recede and was instead replaced by something quite neutral.  I also tried different pair of output tubes that I had lying around - Mullard and Shuguang EL34s.  Both provided very pleasing results, but the vintage Tungsol 6550s seemed to sound the best.


No More Games!
Now it was time for some serious listening.  For this, I invited my fellow local audiophile friend over.  He has a great stash of quality records and also had some different vintage output tubes to try out.  I also replaced the rectifier with a 5AR4 made by Mullard.  Here are some notes from this session:

The first pair of output tubes was a pair of vintage Mullard XF2 EL34s.  Listening to Neil Young - Live at Massey Hall, the darkness and classic warmth of this venerable pentode was obvious.  There was a slight loss of detail, but the overall effect was quite musical.  Some of the macro dynamics seemed a touch restrained.

Next up was the Tungsol 6550.  This pair was from the early 1960s with the grey plates with three holes.  Dynamics really improved here with tight and well-defined bass.  There was also a sense of ease to the music, making the amplifier sound much more powerful than the estimated 15Ws.  Treble detail and extension wasn't the best I've ever heard, but it was still very good, but perhaps lacking that last bit of shimmer and liveliness.  Robert Ludwig's mastering of The Band - s/t album was the best I've ever heard it, managing to find clarity in this darkly recorded album.  On the other hand, Donald Fagan - The Nightfly still had speed and the latent fingerprint of a digital recording.

For the next record, a pair of 1950s Dutch Philips EL34s was selected.  These metal base tubes are quite expensive these days, but there was some real magic here.  Listening to Cat Power - Jukebox was an amazing experience, pulling me into the music in a way that made it seem real and alive; as if I were sitting thirty or forty feet away from the band in a small hall.  There was an evenness to the music from the top to the bottom without any  part of the frequency spectrum sticking out in an unnatural way.  Dimensionality and soundstaging was some of the best I've ever heard.  Even more surprising was the controlled but taut bass through the 15" Eminence woofer.  Very impressive.  These tubes appear to be made for this amplifier.

The following experiment was not much of a success.  A fresh pair of early 1950s Tungsol 6AR6s was selected, along with an octal socket adapter.  The sound on Steely Dan - Greatest Hits and The Who - Tommy, became very uninvolving and flat.  Perhaps some more break-in time was needed or a different bias point.  It was time to move on and try something different.

1970s Gold Lion KT77s (with black bases):  These sounded much like the Tungsol 6550s with excellent bass control and incredible dynamic range.  It was, however, a little richer, hewing somewhere between the dynamic Tungsol 6550 sound and the classic richness of the EL34.  Neil Young - Tonight's The Night delivered all the tortured darkness that I expected with a sense of realism that was inviting.

For a more budget choice, I also tried a pair of SED 6L6GCs, one of my favorite modern production tubes.  Though there was a slight loss of the last bit of detail, the overall sound was very pleasant with only some slight overlaid grain over the music.  The explosive dynamics on Blue Nile - A Walk Across the Rooftops were really a surprise, while the Police - Synchronicity kept the rhythm and pace of the music together in a bouncy and engaging way.


The Scorecard:
Bass: Normally a weakness of tube amps, and even more so for single-ended, the Multi-Valve was a real surprise here.  The lowest notes were deep, well-damped, and had excellent control without the slight tubbiness I normally associate with tube amplifiers.  This is a testament to the James output transformer and the feedback loop controlling the woofer.  Not only was there depth, but detail, whether it was the smack of a tom-tom, or the playfulness of a skilled bass player.

Midrange: Clear, defined, and transparent are the first words that come to mind.  The music is served in an even-handed fashion without tube bloat or an overly analytical presentation.  The quality here is quite dependent on the output tube used, ranging on the scale from amazing (Tungsol 6550s and Philips EL34s) to a more middling sound (Mullard EL34s or SED 6L6GCs).  But at no time, except for the 6AR6s, was I actively disappointed by the music coming out of the speakers.

Treble: This again was dependent on the output tube selected for use, with varying levels of quality.  I preferred the Philips EL34s and the Gold Lion KT77s here, while some listeners, depending on the speakers used, may prefer the warmth of the Mullard XF2 or the qualities of the Tungsol 6550s.  At its best, the Multi-Valve amplifier treble is perfectly integrated into the midrange, not calling attention to itself, but not lacking in detail or extension.

Soundstaging: Whoah!  Images are way beyond the edges of the speaker, making my listening space sound larger than it is.  There is also an incredible amount of depth.  Vocals are also projected forward in a nice fashion, keeping separate from the rest of the music.  This clarity comes from the lack of sonic mud, allowing the listener to "see" deep into the musical performance.

Intangibles: Considering the limited power that single-ended can deliver, one of the biggest revelations was the handling of macrodynamics.  Even at lively listening levels the amplifier never failed to deliver the wattage when needed.  This, along with the various microdetails and instrument shadings, gave a unique presentation that makes so many other tube and solid-state amplifiers sound cloudy or unfocused.  The Multi-Valve isn't stubbornly analytical either, bleaching out warmth or lacking humanity.  Instead the presentation is incredibly balanced.  An amplifier like this would be a great tool for mastering or component evaluation.


Conclusion:
There really is something special about single-ended designs, whether they are triode, ultralinear, pentode, or even solid-state.  It's the Class A power that really matters in the end, along with the uncomplicated driver circuits and simplified power-supply requirements.  A pentode, at least with regulated screens and negative feedback, gives a different presentation than, for example, the classic WE91 300B design.  Of course listening preferences, speaker selection, and source components will greatly effect the best amplifier to use in a given system, but with mine, the Multi-Valve is perhaps the best I've ever heard.  Really.  It's also great fun to sample different output tubes and the unexpected differences between them.  Highly recommended.

Parts list and Front Panel Express layout available upon request.


Review System:
VPI Aries with JMW 10.5i tonearm and SDS Power Supply
Dynavector 10X5
Cardas Cross 1M interconnects
Quicksilver preamplifier with Mullard short-plate 12AX7s, RCA 12FQ7s, Amperex 12AU7
Cardas Quadlink 5C 1M interconnects
EICO HF-60 monoblocks with Mullard XF2 EL34s & 5AR4s, GEC CV4085s, and GE 6SN7GTAs.
Cardas Hexlink speaker cable
UREI 813A monitor speakers
VTI BL503 equipment rack

New Project: Universal Single-Ended Pentode Amplifier

I've recently built two budget orientated single-ended amplifiers, one using the WWII-era 1625, and the other using the 6CB5A connected as triode.  Both were built to a price point - Edcor output transformers and parts sourced from Mouser.  The sonic results for these unusual designs were quite gratifying, considering the sub-$700 price range.

Sonically, I actually preferred the sound of the 1625 amplifier which was used as a pentode with screen regulation.  It had a pleasant forward sound compared to your average triode, and also seemed to develop much more power than expected, along with greater speaker control with the use of (yikes!) negative feedback.  This project made me rethink an older amplifier I built using the EL156 output tube connected in Ultralinear.  I began to wonder how a true pentode assault on the "state of the art" would sound.

With that in mind, I came up with the following concepts to guide me through the design process:

• Ability to use octal socket output tubes: 6550, EL34, 6L6GC, KT66, etc
• Fixed Bias - allowing the maximum range of bias 
• Plug-in Plate cap: for the weird tubes
• Pentode operation - possibly with a triode switch
• Tube rectification using damper diodes
• Screen regulation using gas tube shunt regulation

Perhaps "state of the art" was the wrong term to use since I won't be spending hundreds of dollars on Teflon coupling caps, but I will be using higher-level James output transformers, Rel-Cap polysterene couplers, and some really nice Nichicon electrolytics.  And I have a pair of vintage TungSol 6550s that are just crying out to be used!  Stay tuned for the schematic, build pictures, and listening results.

Building a single-ended 6CB5A amplifier

Introduction:
After the success and sale of my single-ended 1625 amplifier, for my next amplifier build I decided to return to the venerable triode.  However, I wanted a bit more power than your average single-ended 2A3 amplifier, but still didn't want to use the pricey 300B.  With that in mind, I looked at several push-pull designs using the 6B4G or 2A3.  Once again I ran into expensive tubes, even for Russian or Chinese versions, and pricey interstage iron.  However, after reading several threads on diyaudio.com, I came across Thomas Mayer's blog describing the use of the 6CB5A - a TV vertical deflector tube - in single-ended mode to make 7Ws.  He touted it as a budget alternative to the 300B, and at ~$5 to $6 a pop on Ebay, it certainly would cut down on the output tube cost.

Circuit:
Once I had the signal schematic in hand, I went and designed my own power supply circuit around a Edcor power transformer.  Note that the 6CB5A uses a mighty 2.5As of current at 6.3V.  This required an additional filament transformer for the 6N7 driver tubes, and, in order to reduce costs, I stuck with a 5V rectifier tube since the filament tap was available instead of the recommended TV damper diode which would require an additional transformer..

Parts:
Once again, this is a budget build, so Mouser and Allied Electronics were used heavily.  This means electrolytics used in series to get a high enough voltage rating, Cornell-Dublier metallized polypropolene coupling capacitors, plastic speaker binding posts, and plenty of KOA resistors.  Output transformers are 3.5K/6-ohm units from Edcor.  The power transformer was also sourced from them.  For the top plate I used Front Panel Express.  Wood chassis is from Valab, an Asian Ebay seller.

Construction:
This was an exercise in point-to-point wiring, which can be frustrating if one is not experienced.  A single star ground is used near the RCA input jacks.  Wiring is all plated solid-core.  Due to the simplified power supply and lack of regulation, this unit was much easier to finish than the 1625 amplifier, my last project.

Listening:
After a quick voltage and current check, I hooked this amplifier up to a pair of Pioneer BS-21 speaker that I use for test purposes.  At 84dB efficiency, these aren't exactly a great match for single-ended amplifiers, but at least I can tell if music is being made.  The sound, via my VPI table and Quicksilver preamplifier, was very smooth without any noticeable hum or other noise.  With my medium-output moving coil cartridge, I had to turn the volume control way up to get any decent playback level.  Of course such a low gain amplifier will prove to be useful for much more efficient speakers than the test Pioneers units.

After everything checked out, it was time to listen to the 6CB5A amplifier with the KEF iQ30s.  Once again my big UREI 813As, which would be a much better match, are in storage so I had to make due with what is on hand.

This amplifier is smooth, coherent and almost touches the better 300B amplifiers I have built.  The soundstaging in uncluttered with a deep and immersive soundfield - all traits of a triode amplifier.  The bass was also surprisingly well-controlled, lacking that fat underdamped sound that plagues some tube amplifiers.  Treble extension was smooth with a nice shimmer and swirl to high hats.  Detail was also very good, providing plenty of definition - eg, recordings sounded different and weren't congealed into a "it all sounds the same" blob.

Switching to the less efficient B&W 805s and I noticed the amplifier would run out of steam on big peaks.  Obviously not a good match, but at lower listening levels it was quite the pleasurable listening experience.

Conclusion:
Single-ended amplifiers - even using triodes - require careful speaker matching.  Many speakers are too inefficient and have wild impedance curves that require more power and low output impedances. However, the 6CB5A coupled to the Edcor iron did a decent job driving any of the speakers I have on hand.  Those into rock or heavy orchestra would benefit with more power or horn speakers, so keep that in mind if you're interested in building something like these amplifiers.