Moonbeam: Therapeutic vacuum tube reconstruction for a guitar amplifier and a man

I constructed this tube guitar amplifier using the chassis and speaker cabinet of a solid-state combo amp with several objectives in mind:

Resume building tube gear after a five year interruption due a serious outbreak of real life
Thin out the stock of tube amp parts I have accumulated in the past 7 years
Recycle the chassis and cabinet of a solid-state guitar amp that sat unused in someone's closet for the past 10 years
Do something useful with the prototype turret board I built for my Moonlight amp projects
Build my younger daughter a guitar amplifier to take to college

It's been almost five years since I completed my last tube construction project, a dual channel clone of the Pultec MB-1 microphone preamplifier. I didn't intend to stop building, but real life found ways to keep my attention directed elsewhere. I finally have some time to build again, and decided to start out with a relatively low-intensity project to reacquaint myself with the rhythm of the DIY process.

I enjoy all aspects of tube DIY design and construction: the circuit, the layout, the cabinet, etc. However, I don't consider myself a natural craftsman, so it takes me a lot of time and effort to perform the electrical and mechanical work to the standard I set for myself. I didn't feel up to taking on the whole load for my first project back, so I decided to improvise a useful amp out of stuff I just had laying around.

The main ingredients for this project were:

A Peavey Envoy 110 solid-state combo amp that sat unused in my friend Gary's closet for 10 years, that wished to become reincarnated as a real tube amp.
A "Deluxe"-sized power transformer, choke, and output transformer from Weber that I ordered about 5 years ago
The populated turret board I built to prototype the AX84 Moonlight amp back in 2003.

At the start, I was thinking I'd try a Moonlight with a push-pull 6K6 octal pentode output stage in place of the 6SN7 triodes of the original to get that classic pentode sound with about 9W power. But I expected that the Deluxe-style power transformer secondary voltage would be too high and the output transformer impedance ratio would be too low for 6K6's, so I decided to go with a pair of cathode-biased 6V6's (as in a tweed 5E3 Fender Deluxe) for the output section. Hence the name Moonbeam - a Moonlight preamp and phase splitter driving a pair of beam power tubes.

Construction details

The Peavey provided a chassis suitable for tube-amp reincarnation - a simple 16 gauge steel U-channel with a minimal number of holes in the bottom. I made 4 holes on 2" centers for the octal sockets with a Greenlee 1" conduit punch, and a 3/4" hole for the miniature 9-pin socket with a Unibit. Once again, my friend and co-worker Arpad made quick work of the power transformer cutout with the Bridgeport mill. I could have made the cutout with a drill, hacksaw, and file, but it's nice to have a friend with the tools, skills, and willingness to help. One note about drilling/cutting steel: it's a lot harder than aluminum. I use the minimum speed on my drill press, lubricate the drills/Unibits with oil, minimize the pressure, and pause between step-drill steps. Using minimum pressure also reduces the size of the burr that has to be cleaned up on the exit-side of the drill holes.
Moonlight layout and construction

The front panel of the Peavey had a few hundred holes for all the pots and switches typical of the knob-happy solid-state amp genre. I used the Unibit to enlarge the ones I needed for the full-sized volume and tone pots, toggle switch, and pilot lamp I would use.
empty chassis front

The back panel wasn't quite as perforated, but I did need to make a hole for the external fuse holder. The fuse for the original amp was inside on the PCB.
empty chassis rear

The populated turret board and the tube sockets were lifted straight from the prototyping chassis (shown below), where they had lain dormant most of the past 5 years. When I originally made this turret board, I accidentally flipped the pattern left for right from what I designed for my Velleman chassis layout. It turns out to be the correct orientation for this chassis.
moonlight prototyping rig, circa 2004

The circuit is essentially the same as my first Moonlight, with a power and output transformer appropriate for a push-pull 6V6 output stage. I hope you will forgive the lack of an as-built schematic. Here is a text description of the power supply and other important circuit elements:

Connect 540 VAC (red-white) pair from PT to 5Y3 recto plates. The 680 VAC (red) pair and the bias (blue-red) lead from the PT are unused, insulated, and tucked away. First stage filter cap is 22uF/450V. Note: 5Y3 can't look into too much capacitance, the current inrush stresses it. The tube spec says 8 uF, but modern designs seem to go up to 32uF.
Connect first stage filter cap to standby switch. Other side of standby switch connects to output transformer center tap (red) and input to choke (black). B+ to the center tap is 314 VDC.
Output side of choke connects to second stage filter cap (47uF/450V). The screen supply for the 6V6's connects here. Each screen is fed through a 1500 ohm 2W current limiting resistor. B+ to the screens is 302 VDC.
Third stage filter cap (33uF/450V) connects to the second stage by a 6800 ohm 2W resistor. The phase splitter plates are fed here. B+ to the phase splitter is 273 VDC.
Fourth stage filter cap (10uF/450V) connects to the third stage by a 10K ohm 1/2W resistor. The preamp plate is fed here. B+ to the preamp is 252 VDC.
The 6V6's use a shared 330 ohm 5W cathode bias resistor bypassed by a 22uF/25V cap. Cathode voltage is 19.7 VDC. Idle plate current to each tube, measured using the transformer shunt method is 32 mA. The static plate dissipation is (plate voltage - cathode voltage) x plate current = (302V - 20V) * 32 mA = 9W.
This power transformer has plenty of filament current and the 680 volt HV winding, so it would be no problem at all to run 6L6 outputs instead of 6V6's, or to run the 6V6's at higher voltages. To run 6L6's, the cathode resistor rating should be increased to 10W. A 6U4GB rectifier should replace the 5Y3GT (pin compatible, and the 5V filament winding has the 3A the 5U4 needs). Of course, the higher voltage transformer secondary will probably exceed the 450 VDC rating on the first stage filter cap, so we'd want a 500VDC rated cap there.
Substituting a 5U4GB rectifier for the 5Y3GT increases all the B+ by 20 volts.

The Peavey power transformer, power switch, and mains cord attached to the original PCB with Molex connectors. I decided why the heck don't I do the same with the new PT? So I Molexed the input side of the the PT and 6V filament outputs so the transformer can be removed without desoldering anything. The down side is now there are more connections to go bad.

Here are the guts all wired up with transformers and switches
chassis wired inside

Debugging issues

Because I already knew that the turret board and sockets were wired correctly, I didn't spend as much time as I usually do checking my wiring before applying power. I checked the power supply up to the standby switch, ramping up the AC gradually on a variac and watching the voltmeter - no problems. The whole amp worked on the first try, but with fizzy distortion at low volume. I tracked it down to a loose resistor I hadn't soldered on the turret board. That solved, I cranked the volume and was quickly rewarded with the dreaded sound of high-pitched oscillation with the tone control anywhere near the treble side. I spent a while sifting for cold-soldered joints, bad grounds, and shorting grids and cathodes to ground with a cap, swapping tubes, and being perplexed. In desperation, I connected one side of the output transformer secondary to ground, and the oscillation disappeared. On the bench, I had connected the speaker to the OT's flying leads with test clips and no ground. I had planned to connect the speaker directly to the OT flying leads in the cabinet, but decided to obtain the necessary ground connection by running the leads back inside the chassis to a 1/4" phone jack mounted on the back panel. That seemed slightly less of a hack than running an extra lead back from the speaker to a terminal screwed to the chassis. Note the speaker jack is right next to the preamp tube socket, with the OT secondary leads close by. These are the kinds of things I normally go to great pains to avoid in my "careful" layouts, and I am happy to get off scott-free this time.
chassis wired outside

Tubes, from left to right: 5Y3GT rectifier, 2 x 6V6-GT output, 6SL7GT phase splitter, and EF184 preamp. The power transformer is the previous edition of the Weber W025130, with 120V primary, 540/680 VAC@150mA HV windings, 45V bias, and 5VAC@3A and 6.3VAC@5A filament windings. The output transformer is the Weber W041318 6600 ohm primary / 8 ohm secondary. The choke is the Weber W022699 9H@120mA.

Here is the rear view of the complete package. The cabinet is nothing special, cheesy MDF covered with a decent tolex job. It arrived for re-animation in pretty grubby condition, with rusty hardware and covered with a quarter inch of dust and grime. The shop-vac and Formula 409 took care of the grime, but it shall carry its rusty screws onward into its new life. It's only rock and roll. The speaker is a Weber Signature 10f that I bought for just such an occasion all those years ago.
amp assembled rear view

What does it sound like?

I'm relatively pleased with how it sounds. I like pentode preamps because they respond well to changes in the guitar volume control, and are pedal-friendly. This amp has a good amount of clean headroom, which the original Moonlight sorely lacked. I thought the problem was caused by the output tubes being overdriven almost all the time, and that the cure to was to use a lower gain, higher headroom output stage. It turned out that it's much easier to get clean headroom by using a lower gain phase splitter tube, e.g. a 6SN7 instead of a 6SL7.

In this amp, both strategies are effective. The 6V6's have plenty of headroom even with the high gain phase splitter, and can really accept a lot more useful (non-flabby) bass than a pair of 3.5W triodes. With the low-gain phase splitter, there's a couple more numbers of clean on the volume control for jazz-style playing.

It's not as bright as a blackface Princeton or Deluxe, which I wouldn't expect with cathode bias and the 320V B+ I'm running. It might sound 5E3-ish. When cranked, the distorted tone has the bluesy growl and singing sustain that makes it easy to get well-controlled fundamental and one octave higher feedback all over the guitar, something that I found it much harder to do with blackface-style amps.

It is conventional wisdom that pentode preamp tubes, e.g. EF86, EF184, can have microphonic issues in combo amps, because of the vibrations in the cabinet. It doesn't seem to be a problem for this circuit, though, probably because the pentode gain is set for the lower end of the possible range.

What next?

I'm almost afraid to get caught up tweaking with this thing, because I may end up wanting to play it instead of giving it to my daughter. But she needs an amp to take to college, and it sure as heck is not going to be the '67 Princeton Reverb I lent her to play at home. I'll give myself a month to play with the higher voltage 6V6 (and maybe 6L6) options. The present B+ is probably low enough I could get away trying 6K6's, though I might need to adjust the cathode resistor or screen supply voltage. I have a 16 ohm 10" alnico in my parts stash that would give the correct impedance ratio with the Weber output transformer.

Cosmetically, a new control faceplate, speaker baffle, and grillcloth would de-Peavey-fy it to the point that someone might even think about stealing it. I'm not sure I can call it complete without attending to the cosmetics, but it's far enough along that I can start working seriously on the next build. And that was the main objective.

While doing the research for this project, I found a USA-made multi-tap output transformer to run 6K6's at the proper impedance ratio at Triode Electronics. They also sell a complete line of Fender-style power and output transformers from the same company (Magnetic Components of Chicago) at really good prices. While I decided to stick with the Weber transformer set for this amp, I couldn't resist picking up the similarly-sized set from Triode, along the a pair of JJ 6V6-S's. The Magnetic Components transformers arrived and look really beefy. I think they will be mounted on my next prototyping rig in the near future. Of course, buying another transformer set nullified the objective of reducing my amp part inventory, but you can't have everything. Where would you put it?