This page contains the day-by-day narrative of my first guitar amplifier build in over thirty years. The overall summary of the amp is documented separately.

Contents

  1. The Build
  2. Playing & Tweaking
  3. Schematics
  4. Measurements
  5. Revision History

The Build

With schematics and parts list in hand, I started by ordering my supplies. I bought virtually all of the components from Antique Electronic Supply and the tubes from The Tube Store.com. I used ground shipping: after a four-year gestation period, there really isn't any great urgency.

When I was building projects as a kid, my tools were limited to what I could find around the house or buy inexpensively. Cutting holes for tube sockets was one of my most arduous tasks. I'd scribe the radius, drill a series of small holes just inside the radius, and use a round file to shape the hole. This time around, I decided to splurge and buy some Greenlee chassis punches. That, along with other tools that I needed to have, was a significant portion of my initial outlay. I really haven't needed to do anything mechanical in ages that I can't do with a screwdriver.

Preparation

First things first: I unpacked the parts, compared them against the order, and sorted the small components into trays for easy access. One capacitor was wrong: I received a 50V part rather than a 450V part. An email to Antique Electronic Supply resolved that.

Before I started any work, I had to work up a chassis layout. Prior to placing my orders for supplies I did some initial sketches using a drawing program on the Mac. I used those sketches to choose the chassis size based upon the size of the transformers and other components. I knew roughly where all the components would sit on the chassis, and had a tentative plan for how I'd lay out the components on the terminal strips and tube socket lugs. But now I needed to come up with a dimensioned drawing showing the location of every hole in the chassis.

With the chassis and all the components in hand, I finally started working in three dimensions. I used the tube sockets, components and terminal strips to decide on the relative placement of the sockets and strips. Since the components are supported by the terminal lugs on the sockets and strips, the strips have to be close enough to the sockets so that the leads reach but not so close as to make assembly unneccesarily difficult. The input jack and volume and tone controls were placed next, along with a terminal strip for the tone-stack components. Once all the sockets and terminal strips were positioned, I finalized the transformer placement. In addition to proper orientation (power transformer end-bells facing front and back, subsequent transformers oriented so that their laminations are at right angles to each other to minimize flux coupling), it's wise to leave some space between the power and output transformers to minimize hum at the output. You also need to keep in mind that the presence of a transformer on the chassis surface precludes any other hardware; you can't have a terminal strip under a transformer. So I put some space between the PT and the OT and used that space inside the chassis to locate the power supply rectifier diodes, filter caps and dropping resistors. Finally, I placed the power switch, line cord, fuse holder, and output jack. I located a terminal strip on the chassis end wall for the safety ground and AC neutral wire (the hot wire goes directly the the fuse holder).

Day 1

With all the prep work done, Day 1 was a breeze. A trip to the hardware store for nuts, bolts, lockwashers and a few more tools was followed by some marking and drilling, then by another trip to buy a decent variable-speed drill. My little battery-operated drill did fine drilling small holes in the 20-guage steel chassis, but it was clear that it wouldn't be up to handling the number of holes that I had to make nor would it be able to power the bit for the larger holes.

The chassis punches, if you've never used one, are simply magical. You drill a starter hole, thread the punch, crank about four turns on a wrench, and pull out the slug leaving a perfect hole. Man, I wish I had had some of these as a kid...

I learned the value of keeping a clean workspace. The metal filings that accumulated on the bench left a couple of scratches in the paint on the face of the chassis and embedded themselves in the soles of my shoes.

I did have to file a few holes despite having the chassis punches. I use rubber grommets to pass the transformer leads through the chassis. Two of those grommet holes are too big for my drill and too small for my punches, so I had to file those out to the proper size. And the strain relief for the cord expects an odd "double-D" hole shape. Punches are available but are extremely expensive compared to the round hole punches. So I punched a round hole and then filed it out to the proper shape.

With all the holes drilled and punched, I cleaned up the edges of the drilled holes, washed the cutting oils from the chassis and proceeded to mount all of the chassis-mounted components. At this point, it's starting to look like an amplifier!

January 22, 2005

Day 2

The next morning, I realized that my spatial skills need a bit of honing. I had intended for this chassis to hang inverted from the top of the combo cabinet, with the input jack on the left. Well, I got it backward - the input jack will be on the right. No big deal...

After I recovered from my spatial anomaly, I temporarily placed a few discrete components to check critical dimensions. Then I started wiring the chassis. I cut and tinned transformer leads and attached them to terminal lugs. I also laid in the heater wiring, a few wires for ground, B+ and signal, the common cathode connection between the output tubes, a wire from the fuseholder to the power switch, and a short shielded run from the input jack to the preamp. There are far fewer wires in an amp built this way than in one built using tag boards or turret boards. I soldered all of the connections that are complete; the rest are awaiting the discrete components. The wiring is stable at this point because all connections are mechanically secure, with the wires wrapped around the terminal lugs.

Day 2 was a bit more work than Day 1. My supplies included some cloth-covered hookup wire that I bought because it was relatively inexpensive and because so many vintage-amp afficionados seem impressed by cloth-covered wire used in modern reproductions. Well, if I build another amp I'll probably just buy some color-coded wire with normal plastic insulation. That cloth-covered wire is really fussy to work with, and doesn't look as "neat" as plastic-covered wire.

There were a few moments when I wished that the chassis was a bit larger, or that my eyes were much better, or that I had some tools that were a better match to what I was doing. All things considered, I'm happy with the way this build is going so far.

Here are pictures taken at the end of Day 2:

Front viewday 2, front view
Underside, rear viewday 2, underside back view
Underside, front viewday 2, underside front view
Power wiringday 2, power wiring
Preamp & driver wiringday 2, preamp & driver wiring

January 23, 2005

 

Update: The tubes arrived the day after I took the pictures. I didn't realize when I placed the order that The Tube Store is a Canadian company. The package had to go through customs, which helps to explain why the UPS tracking page gave the appearance that the package was taking a tour of the USA.

Day 3

I'm still waiting for that last capacitor to arrive. It's the final power supply filter cap that will feed the EF86. On the other hand, I'm restless and I want to see whether I can make this amp make some noise. So I wired all the components that I do have. For now, I've bypassed the final filter section; the dropping resistor is shunted with a piece of wire and the cap is missing.

I'm reminded once again that this is a tight chassis as I connected the components between the tube socket lugs and the terminal strips. But the layout works as I envisioned: I'll be able to access all of the lugs for soldering and testing. After all the components are fastened, I worked from one end of the chassis to the other soldering every connection.

I put the fuse in the fuse holder, ran a speaker cable from the output jack into a Hot Plate set on its Load position, and double-checked everything. Are the caps and rectifiers in the right way? Does the wiring make sense? Are there any connections that I forgot to solder? Are there any lugs shorted on those 9-pin sockets where space is tight? It all looked good, so I put a pair of 6V6s in their sockets, made sure that the power switch was off, plugged in the line cord, held my breath and flipped the power switch. I was rewarded by the faint hum of the power transformer and the glow of the 6V6 filaments. More important were the things that were missing: there was no smoke, nothing smelled like it's frying and the 6V6s didn't redplate. So far, so good. I measured some voltages and calculated the plate dissipation: the idle power is about nine and half watts per tube. That's fine.

That was the easy part, but it was encouraging. I turned off the power, plugged a speaker into the Hot Plate and set the Hot Plate for zero attenuation. The next step is to bring up the driver stage. I plugged in a 6BM8 and double-checked all the connections in the driver section. I applied power once again, breathed another sigh of relief that there were no spectacular failures, and checked the critical voltages. Dissipation was well within spec. Tapping the input of the driver with the meter probe gave me sound from the speaker. This is good! But it seemed like it was inconsistent, so something was not quite right. I grabbed a portable CD player to use as an ad-hoc signal injector and applied a signal across the volume pot. Nothing... I brought out out the meter again: there was no voltage drop across the plate load resisor. I unplugged the power and found that the driver's cathode resistor hadn't actually been soldered to ground. It was the first component attached to a busy solder lug, and I missed it while soldering. Easily fixed. I reapplied power and was able to hear the CD player through the speaker. Way cool!

Last up was the EF86 preamp stage. I inserted the tube and checked voltages. The plate was near zero and with about two milliamps flowing in the cathode resistor; this tube was saturated. I wasn't surprised. I was unable to locate a set of characteristic curves when I did the design, so I guessed as to reasonable values of the plate and screen resistors. I guessed wrong. A very short session of trying different values for those two resistors got me the plate voltage I wanted to see. At that point I figured I should have a working amplifier, so I cracked the volume control and pressed on the input jack, breaking its ground connection. I was greeted with the expected buzz and squeal, so I plugged in a guitar. And it worked! Nice breakup, easy to back off into a cleaner sound... Heck, it even behaved like a guitar amp. The tone control did its thing with the midrange frequencies, just like the SPICE model predicted.

This amp was suprisingly easy to bring up, especially considering that I designed it from scratch rather than copying an existing design. I remember from my past experiences with building tube gear that it's pretty difficult to build something that won't pass a signal. The circuits are simple. So long as you do the math to make sure you're not exceeding dissipation limits, choose components to handle the worst-case voltages, plan your gain staging and don't make any mistakes in wiring, the worst that can happen is that you'll bias a tube into saturation (as I did with the EF86 on the first try) or cutoff. The rest is just a matter of tweaking...

I applied labels to the rear of the chassis and set it aside. Now I have to sketch a design for a cabinet to hold the amp and a Jensen P15N speaker and hand it over to Mary-Suzanne, who has much more experience with the table saw than I do.

 

Update: While I was writing this entry, the missing filter cap arrived. I installed it and tested the amp once more just before taking the pictures. I think I'm going to like this amp...

 

Here are some pictures taken at the end of Day 3:

Internals, power endday 3, internals, power end
Internals, preamp endday 3, internals, preamp end
Chassis rear viewday 3, chassis rear view

January 29, 2005

Day 4

A cabinet was the last hurdle to putting together a finished prototype. I wanted to do this myself despite not having any real experience in woodworking. I actually spent several days sketching cabinet designs before coming up with the final candidate. My objectives were:

  1. to create an open-back cabinet that
    1. provides optimal support for the chassis.
    2. orients the chassis for optimal free-air cooling.
    3. is large enough to offer some low-mid extension for the speaker.
  2. to have the total weight of the amp between 35 and 40 pounds (preferrably closer to 35).
  3. to be easy to construct.
  4. to use no more than a half sheet of plywood.
  5. to be sufficiently rugged for use as a gigging amp.
  6. to serve as a validation of the cabinet design.

Here's the end result:

Front panel, from top: input, volume, tone, power
Front viewday 4, front view
Speaker: Jensen P15N
Tubes, from top: EF86, 6BM8, 6V6, 6V6
Rear viewday 4, rear view

The overall size (excluding feet and handle) is 18.5" high by 26.25" wide by 10" deep.

Because this is a prototype cabinet, I tried to work with materials on hand or locally available. Most of the wood was cut from scraps left behind by the previous owner of our home. The cabinet sides are half-inch textured exterior siding, giving the amp kind of a "rural" look. The speaker baffle was another piece of half-inch plywood cut down from the remains of my pedal board. The rest of the materials were purchased from the local hardware store.

Tools were the biggest issue for this part of the project. We have a creaky old benchtop table saw that was just barely up to the task. The table was small enough that Mary-Suzanne had to make most cuts without a fence. My friend Stephen loaned me his router to cut the speaker opening. For the final prototype cabinet, I may farm out the cutting and do the assembly myself. Further down the road, who knows... ?

(Stephen got to play through the amp that first night using his Parker. We cut the baffle board, did the final assembly, and fired up the amp. The Parker sounded really nice, with plenty of top-end swirl. Even the piezo sounded good.)

The cabinet is assembled using butt joints and cleats, held together using fine-pitch drywall screws and carpenter's glue. Originally I had cleats only in the corners and for the baffle board, but I added one more along the top rear edge of the cabinet for strength. Now it's rigid enough that I can stand on the cabinet.

The handle is a heavy-duty door pull. It fits well with the look of the prototype, and it's comfortable.

I didn't want to spend a lot of money on grille cloth for what is essentially a disposable cabinet, so I used black fiberglass door screen at a fraction of the cost. I painted the baffle board flat black after installing the T-nuts, then stretched and stapled the screen just as if it was grille cloth.

I stapled a strip of double-sided Velcro to the bottom of the cabinet to hold the power cord during transportation. The final cabinet will have an L-shaped trim panel on the back to hide the tubes and to keep the unplugged power cord inside the cabinet.

The weight of the completed prototype is about thirty-six pounds. I haven't yet decided whether to try another cab made of poplar plywood; that should shave another four or five pounds from the total weight. As is, this amp is a joy to carry at half the weight of my Vibro-King.

February 10, 2005

Playing & Tweaking

Now the fun starts...

Day 5

Today I fired up my oscilloscope, a forty year old Tektronix generously donated by my boss. It's in usable condition, modulo some possible calibration issues and a lot of flaky controls that need to be revived with a shot of contact cleaner.

The first thing I wanted to do was to see how the output stage was behaving. I confirmed that the 6V6 grid voltage does go positive with respect to the cathode. That's good, since I had designed the driver to have enough power to push the output into class AB2 operation.

The second thing I did was to check the outputs of the driver and the preamp. The driver was OK, but the preamp had a really obvious asymmetry. The preamp's plate voltage was still a bit too low (see the description of EF86 trial-and-error adjustment in Day 3) and the stage was saturating on the negative-going peaks. I changed R3 from 220K to 100K and used the scope to confirm that the clipping was gone with the input pushed hard.

I spent the rest of the day playing through the amp and sketching out ideas for some tweaks to try:

  1. Boost bass toward the CCW position of the tone control using one of
    1. a 470pf cap across R9 (this slightly reduces high treble), or
    2. a series 47K resistor and 0.001uF cap across R9, or
    3. reduce R8 to 330K.
  2. Introduce slight preamp asymmetry by changing R3 to 150K.
  3. Reduce low-frequency roll-off by increasing C4 to 0.1uF and C9 to 4.7uF.
  4. Ground pin 1 of output tube sockets to give EL34s (yes, the power transformer will handle them) a suppressor grid connection.
  5. Experiment with the value of the output stage cathode bypass resistor.
  6. Raise preamp gain by increasing both R3 and R4.
  7. Float the heaters on the output stage cathodes.

February 12, 2005

 

Update: The parts arrived for the tweaks. I ordered some extra caps to have additional values on hand for tone stack changes. I replaced the 2A SB fuse that I'd been using with the as-designed 1A SB fuse. I also bought some grommets and washers to try shock-mounting the 9-pin sockets. All that's going to have to wait until I have some spare time. I'm starting to line up other guitarists to play and comment on the amp. That'll help guide which tweaks I actually try, if any.

Day 6

Mark T. and Chris S. came to try out the amp. Mark brought his Tele and Strat and Chris brought his super-Strat. The amp responded nicely to all three guitars; it brought out gorgeous ringing harmonics with no harshness. That was a pleasant surprise, since I hadn't heard any really bright guitars through the amp since Stephen played his Parker right after the amp was assembled for the first time. I had made a change in the preamp since that first night, and was concerned that it might have an adverse affect that wasn't apparent while I was playing my Koll and Gibsons.

One of my concerns has been that the amp could use a bit more bottom end. I'm sure it will fit well as-is in a band mix, where the low end is supplied by other instruments. It's even OK playing solo fingerstyle guitar, just not as full-sounding as other amps can be with the bass dialed up. Well, we plugged the amp into Mark's cab, a lightweight detuned 2x12 about the size of a 4x10. This cab, which Mark designed and built, is loaded with Celestion speakers: a G12H30 and a Vintage 30. I've heard this cab before, and it seems to make everything sound good. In this case, it filled out the low end without unduly compromising the highs.

We also tried Chris's cab, a Dumble-style open-back loaded with an EVM 12L. This cab, relative to the Jensen, brought out more of mids and eliminated the break-up tones. I was pleasantly surprised that the amp sounded as good as it did through the EV - my experience is that this speaker is very unforgiving of "idiosyncracies" in the amp. An EVM 12L paired with this amp would be great for lead guitar; too bad about the weight, though...

We experimented with using a four inch wide strip of plywood as a diffuser in front of the Jensen. This tamed the highs, but not enough to really help the low end by comparison. I think I liked it better without the diffuser, because I can imagine the extra high-end extension being useful on stage with drums or in a room full of bodies. The guitar's tone control can tame the amp's top end, but nothing will bring back top end that has been suppressed by the amp.

I need to get feedback from more players. For now, I'm thinking about possible cabinet mods to help the low-end response without altering the amp's voicing. I may also experiment with speaker changes if I can't get anywhere with a cab redesign.

One thing that was obvious to everyone is that the amp really needs a bit of a gain boost to help single-coil pickups get a satisfying lead sound. I personally prefer humbuckers and like the way the amp is voiced. I'm a bit worried that adding gain will sacrifice some of the definition that this amp lends to humbuckers even when played flat-out. I may experiment with converting the cathode-follower to a low-gain (3 to 6 dB) stage. I think I also have some room to raise the gain of the EF86. Or maybe I'll just leave it as-is and let those who want it get their additional gain from a boost pedal. As is, the amp really brings out the unique character of each guitar. I'd hate to kill that by tuning the amp toward a specific sound.

February 27, 2005

Day 7

What can I say? I'm impatient... I decided to try a few tweaks tonight.

I modified the tone control to slightly pull down the mids and highs as the tone control is turned CCW. The difference is only a few dB: not nearly enough to turn this into a traditional treble-cut tone control, but enough to take a bit of the edge off a bright guitar, if that's what you like. I added a series 47K resistor and 0.001uF cap across R9.

The tone control mod doesn't do anything to fill out low-end response. It's strictly a tone-shaping option. I still wanted to do something to extend the low end. I took a look at the corner frequencies of the coupling and bypass networks in the amp, and came up with this chart:

Rev 0.2 LF Voicing
ComponentsFunctionCorner Frequency
C2, R5preamp stage cathode bypass35 Hz
C3, R4preamp stage screen bypass0.75 Hz
C4preamp output coupling35 Hz
C8, R12driver stage cathode bypass3.5 Hz
C9, R11driver stage screen bypass50 Hz
C10driver output coupling10 Hz
C11, R15output stage cathode bypass1 Hz
 Note: Screen bypass corner frequencies do not account for screen resistance. Actual frequencies will be a bit higher.

Those two 35 Hz corners are just over an octave below the low E string's fundamental. And that 50 Hz rolloff is perhaps a bit too close. The three rolloffs compound for an 18 dB per octave rolloff below 35 Hz and there will be some loss at higher frequencies. I tacked a 4.7 uF cap across C9, which brings the driver's screen bypass corner frequency down to around 2 Hz (actually slightly higher due to the contribution of the screen resistance). This change made a noticeable difference in the low-end response, and also slightly increased the overall gain. Not bad for a one-component change...

With the change in C9, the two 35Hz corners dominate the low-frequency rolloff. They'll combine to be 3 dB down at about 56 Hz.

I'm going to live with these changes for a while to see how well they hold up in different situations. It takes time to get acclimated to changes: paying too much attention to short-term results can lead you to creating a hyped-sounding amp which overemphasizes something attention-grabbing rather than presenting a balanced sound. If these changes survive scrutiny, I'll do another revision of the schematic.

February 28, 2005

Day 8

I've had a number of players come in to audition the amp, and still have a few more either scheduled or on my to-invite list. One thing I've found interesting is that I'm getting some contradictory comments; there are no absolutes for guitar tone, and every player brings his own expectations.

It has become clear that this amp "prefers" humbucker guitars. Vintage-wound single-coil pickups don't have enough output to drive the amp hard. Heavily-overwound single-coil pickups have plenty of drive, but sound a bit honky when combined with the amp's shallow midrange scoop. By way of contrast, all of the humbucker pickups have sounded good so far, regardless of their output level or voicing.

After having lived with the modified tonestack and driver screen bypass for two weeks, I've decided to undo the former and retain the latter. I ended up not liking the high-end rolloff introduced by the tonestack mod. Its effect was not enough to tame the highs of a single-coil guitar. On the other hand, the loss of highs weakened the chime of the humbucker guitars and made the effect of the tone control less obvious (given the combined yet still-subtle mid-scoop and treble-cut adjustments).

While inside the amp today I made two more changes. I floated the center-tap of the heater winding on the cathodes of the output tubes. This did not have any noticeable effect on reducing hum (the amp is already very quiet), but I decided to leave it that way in order to have a bit more margin on the cathode follower's heater-to-cathode voltage. I also increased the preamp's plate resistor from 100K to 150K in order to drop the plate voltage slightly, increasing the potential for generation of second-harmonic distortion in this stage.

At this point, I'm happy with the design and will leave it unchanged for the forseeable future. I've updated the schematic to reflect this "final" design.

I had intended to try shock-mounting the 9-pin tubes by adding grommets to isolate the sockets from the chassis. I was hoping that I'd be able to unbolt the sockets, pull them slightly away from the chassis, and drill out the holes. I abandoned that plan when I realized that I'd have no room to drill the existing hounting holes out to the necessary quarter-inch diameter.

My one remaining issue is microphonics. The 6BM8, in particular, is bad and getting worse. I had similar problems with my Vibro-King: I went through a dozen 6V6 reverb driver tubes in order to find one that didn't develop a serious rattle within a week or two. I'll probably stock up on 6BM8s to see whether they're all this susceptible to rattling. Ironically, the Russian EF86 (which has a reputation for microphonics) has been fine in this amp. Perhaps I got lucky... I'll probably rotate several more tubes through that socket to find out whether they all hold up.

March 12, 2005

Day 9

I spent some time tonight with the scope and my "signal generator" - a CD player looping a test-tone track. I decided to make a couple more minor tweaks. I rewired the output transformer secondary tap for 16 ohms (still driving the 8-ohm speaker) to provide a lower plate-to-plate impedance for the output tubes. I also added a resistor between the bridged-T filter and the volume control to limit grid current to the 6BM8 at high drive levels.

March 14, 2005

Day 10

The rattling 6BM8 finally got to me. I have a spare, but I don't want to rattle that one to death. I decided to try shock-mounting the 6BM8 socket using the rubber grommets I had bought a while ago. Rather than tearing down the driver circuit, removing the socket, drilling out the mounting holes to accomodate the grommet and finally rebuilding the driver circuit - a process which would have taken far more hours than my evening (or patience) would allow, I decided to take a shortcut. I removed the existing hardware, wrangled the socket out of the way just enough to clear the drill bit, and drilled the holes with all the wiring intact. I carefully removed all the metal chips, pushed the grommets into place, and remounted the socket. While drilling, I did manage to nick the driver plate resistor - a "cement" wirewound unit - and cracked the resistor's casing. Fortunately, the element was undamaged and the amp still functions.

I also put the OT secondary wiring back to its original connection. I found that the bass was too loose with more of a load on the output tubes.

While I was checking to see how the cracked resistor had survived, I checked the plate voltage on the driver and found that it's pretty low. Where the original design called for a 7.5K resistor, I substituted a 10K because that's what was available when I bought the parts. I'm going to revisit that decision to try to give the driver a bit more symmetry.

How did the shock mount work? Great! I played the riff to John Abercrombie's "Timeless" - those low stacked fifths are great at exciting rattles - and the tube rattle was gone!

March 15, 2005

Day 11

I spent some time tonight with `scope, meter and a small stash of resistors checking out the behavior of the driver. It turns out that the screen voltage sags a lot when the driver is hit hard. That's good: I designed it to behave that way. But the amount of sag caused R11's dissipation to exceed its rating. (It was showing signs of overheating.) So I changed it to a two-watt resistor; I also increased its resistance for added margin.

While I was in there, I replaced the driver load resistor that I cracked on my last outing. It turns out that the manufacturer of the "cement" power resistors does not make a 7.5K ohm part. So I stuck with 10K, which is OK. With the screen sag, the driver doesn't actually saturate as I would have expected given the low plate voltage.

Finally (actually a couple nights later, but it was a quick fix) I added the grommets to isolate the EF86 socket from the chassis. I figure that even though I'm not having microphonic problems with that tube, I may as well do this to extend its life.

March 22, 2005

Day 12

I must be close to being finished with the tweaking. I'm not finding further improvements.

Today I tried adding a loudness compensation circuit. This would boost the low-end response at lower volumes to compensate for the ear's change in perceived loudness. I did this by using a negative feedback loop around the driver, injecting the feedback at the low end of the volume control. The lower the volume, the more feedback. The feedback network rolled off the lows; this has the effect of boosting the lows when used in the feedback path.

The effect seemed fairly natural in practice. There was a slight bass lift at the lowest volumes which decreased as the volume increased. However, I decided not to leave the loudness compensation in place. I preferred the uncompensated response at lower volumes. Also, the overall negative feedback slightly reduced the gain and made the distortion just slightly less smooth. It might be useful to add this compensation to a version of the amp specifically intended for clean playing, along with other changes to reduce distortion and compression.

Another change I tried today was to add a shunt capacitor across the driver transformer secondary to reduce high frequency response. I tried a 0.001uF cap for a 10KHz rolloff and a 0.0022uF cap for a 4.8KHz rolloff. Both caps added a harshess to the distortion sound. In retrospect, that's not terribly surprising: it was either slew-limiting the driver or otherwise interacting badly with the already discontinuous load presented when pushing current into the output tube grids. I still think that the best way to roll off the highs is to use the guitar's tone control.

Simpler really is better.

Finally, I grounded pin 1 of V3 and V4. This is a no-connection on 6V6s, but provides the suppressor grid connection on EL34s and the base shell connection on 6550s and KT88s. The power transformer (and now the socket wiring) will handle 5881, 6550, 6CA7, 6L6, 6V6, EL34, KT66, KT88 and KT90 tubes. I'll have to check the bias to be sure they're all OK for long-term use.

One thing I concluded while testing tonight with 6L6s and EL34s is that the output tubes do add their own signature to the amp's sound and behavior. In fact, the EL34s may be just the ticket for the players who want more gain.

At this point I'm ready to move on to the final stages of testing the amp with different brands of tubes and perhaps with one or two other speakers.

March 26, 2005

Day 13

I vetted the amp for 6L6s and EL34s. I made sure that the output stage cathode resistor (R15) and power transformer were operating within specs for all tube types. I also increased the idle dissipation of the output stage. I thought about adding a switch to reduce power output by increasing the value of R15, but the difference in loudness was negligible and the tone suffered with lower idle current.

While measuring voltages using the different tubes, I noticed that cathode current would increase up to about 9:00 on the volume control, then decrease slightly. It took me a moment to realize that this was due to the screen compression the driver. That's really cool: sag without ripple.

I experimented with changing the center frequency of the tone stack (by changing R5) and eventually decided to leave it along. Likewise, I returned the circuit to its original values after experiments which increased the sweep of the tone control by reducing R9.

I thought about adding a "cut" switch consisting of a switch and capacitor in series across the volume control (VR2). I tried caps ranging from 100pF (barely noticeable) to 500pF (very noticeable). I didn't like the effect of any of them enough to complicate the amp with another switch.

While using SPICE to preview tone stack changes I noticed that the source resistance had a noticeable effect (about 3dB) on the balance between bass and treble. A lower source resistance emphasized the treble. This is because a certain amount of the high frequencies are shunted to ground through C5/VR1/R9. I realized that I'm driving the stack with a much lower impedance than its design specifies. So I removed R16 from between the tone stack output and the volume control and moved to between the cathode follower and the tone stack input. This change does improve the bass to treble balance and gives the impression of less (relative to the bass) treble without killing off the upper harmonics.

(The higher source impedance driving the tone stack means that it'll be really easy to turn that triode stage into a gain stage if I ever decide to make a version of this amp having more gain.)

Finally, I added a 5KHz low-pass filter (R16/C13) just after the first gain stage. This is out of the range of all but the uppermost guitar harmonics, so it doesn't really have a noticeable effect on the tone. What it does do is to tame the highest harmonics of picking "click" and reduce their amplitude going into subsequent gain stages. That seems to give a slightly more pleasant sound when playing with a pick.

Here's an update on the low-frequency voicing of Rev 0.8:

Rev 0.8 LF Voicing
ComponentsFunctionCorner Frequency
C2, R5preamp stage cathode bypass35 Hz
C3, R4preamp stage screen bypass0.75 Hz
C4preamp output coupling70 Hz
C8, R12driver stage cathode bypass3.5 Hz
C9, R11driver stage screen bypass1 Hz
C10driver output coupling10 Hz
C11, R15output stage cathode bypass2 Hz
 Note: Screen bypass corner frequencies do not account for screen resistance. Actual frequencies will be a bit higher.

I think I've finally run out of things that I want to tweak in this circuit. It's amazing how close the final design is to the original on-paper design. Ignoring the components changed prior to the build because of availability, only three values changed. In later revisions, two components were added and one connection was moved.

So what's left? I'd like to make a better-looking cabinet. I may try out a Jensen Neodynium-magnet 15" speaker if I find that I need either less weight or more volume. I intend to try some, perhaps all, of the other compatible tubes: 5881, 6550, 6CA7, KT66, KT88 and KT90. (I've already tested 6L6, 6V6 and EL34.)

Of course, the next big "project" is for me to use the amp. My chops have gotten a bit rusty over the past two months. I'll get back to studying the jazz method books that I picked up in mid-January, record a few clips to add to this page, and perhaps find someone to jam with.

April 2, 2005

Day 14

I stopped at a local shop and picked up some more 6BM8s to try out: a couple of Ei (Yugoslavian) and an NOS of indeterminate brand. All of them have spacer micas that look like they have nice tight fit to the envelope at about a dozen points around the periphery, unlike the Sovteks which have only four points just barely touching the envelope. So far the new tubes (I tried one of the Ei and the NOS) have absolutely none of the rattle that plagues the Sovtek tubes. I have hopes that these tubes will hold up for the long run.

I ordered (from an online store) a couple more NOS 6BM8s to see how they work. I also ordered pairs of KT66 and KT88 power tubes to try.

April 17, 2005

Day 15

Today I was playing the amp along with my Leslie 147, using a pan pedal to go back and forth between the two. I decided that I like the tone of the amp as-is for rock, but for jazz I'd like the tonal balance to be more like the Leslie, having more low-end and less high-end. I added R17 and S2 as shown in the Rev 0.9 schematic. With S2 closed ("Rock"), the amp behaves as before. Opening S2 ("Jazz") adds a 100K resistor in series with the feed to the tone network. This does two things: it reduces the mids and highs by about 6 dB while preserving the bass response, and it moves the high-frequency corner due to C13 down to about 1.7KHz (the corner is at approximately 5KHz with S2 closed).

May 7, 2005

Day 16

The amp still doesn't sound "right" with some Strats. Mark Terrien noted that the mids were congested, and suggested voicing the amp with a deeper mid scoop. I increased R8's value to increase the depth of the midrange scoop (which also shifts the center frequency slightly lower) and it does sound more like a Fender amp. So I changed the Jazz/Rock switch to alter the value of the bridging resistor in the bridged-T network, and went back to a fixed source resistor. I also removed C13, since it does have an audible effect on the presence frequencies. The resulting schematic is now at Rev 1.0.

Rick Davis spent a couple of hours playing the amp tonight. He observed that the Jensen speaker has an early breakup, and that there's some pretty obvious ghosting (due to power supply ripple). His Strats and Tele both sounded pretty good, especially with the EL34 power tubes.

I'll need to solicit some additional player comments regarding the new voicing.

May 24, 2005

Day 17

In an attempt to further tweak the tone stack, I added C13 to give a slightly darker sound (revisiting one of the mods made on Day 7) and R18 to slightly increase the bass response. Both of these changes are most effective at the CCW end of the tone control.

I'm going to live with this tone stack for a while longer. I'm tempted, though, to simplify it to a Gibson/Fender tweed era tone control.

I now suspect that the interstage transformer is responsible for attenuating some of the low end. This is OK for humbucker guitars, where too much low simply adds mud to the overall tone. However, the lack of an extended low end makes single-coil guitars sound a bit thin.

June 14, 2005

Day 18

I finally gave up and threw out the bridged-T tone stack. It was too subtle, and didn't offer enough adjustment range to accomodate different guitars. I replaced it with a Big Muff tone stack, as described in the updated Circuit section.

I also repurposed the Rock/Jazz switch as a Slow/Fast switch. On the slow setting, this inserts a series resistance in the power supply to the output stage to simulate the sag of a tube rectifier.

June 15, 2005

Day 19

The last change introduced distortion caused by overloading the cathode follower. It's important to note that the low cathode follower output impedance only applies to small signals. As the peak magnitude of the output approaches the cathode follower's resting voltage, the load resistor becomes unable to drive a capacitive load. If you look at the output on a scope, you'll see that the negative going portion becomes slew-rate limited. This isn't exactly an unpleasant distortion, but it's something that I wanted to eliminate.

I lowered the cathode follower's load resistor (R6) to provide more drive capability. This required a change in the power supply dropping resistor (R7) to compensate for the increased current draw. The values chosen are a compromise to ensure that the cathode follower doesn't exceed its plate dissipation limit and also to ensure that R6 and R7 operate within their power dissipation range when V1 is pulled and the input of the cathode follower is pulled to the positive rail by R3.

I also raised the input impedance of the tone stack to reduce the load on the cathode follower. This required a change in C5's value to preserve a useful sweep of the tone control. The new stack has a deeper notch. This is an interesting sound, but I wanted to also have a midrange-rich sound available. I changed the function of S2 yet again, this time to a Mids switch. When closed, a portion of the tone stack's input is shunted to its output via R16 to reduce the depth of the notch. (The Fast/Slow function of S2 was too subtle to keep.)

Finally, I raised the value of the load resistor (R14) on the secondary of the driver transformer. This slightly extends the frequency response and adds a bit more headroom.

June 18, 2005

Day 20

I decided to run the tubes a bit cooler. I almost always run 6V6s in the amp and they were running near (perhaps even a bit beyond) their dissipation limit. This was making the sound a bit mushy and reducing headroom.

I also tweaked the tone stack in three places:

  1. I added R18 to make the bass end of the tone control's range more usable.
  2. I added C13 to make the treble end of the tone control's range more usable.
  3. I modified the midrange notch by reducing R16 and adding R17.

May 27, 2006

Day 21

Simpler is better. I completely removed the tone stack. There's a lot more gain on tap, and the tone is actually improved. Crunch and feedback starts at about half volume with my guitar. I'm thinking that there's probably enough gain now to satisfy Strat players.

June 3, 2006

Schematics

Here's a history of the amp changes in schematic form. Click the Rev x.x links to see the schematics:

Rev 0.1: 2004-01-11
Original Design
Rev 0.2: 2005-02-26
First build. Changes are:
R3 was 470K
R4 was 220K
R10 was 6K8 5W
R13 was 2K2 5W
C5 was 0.005 uF
C7 was 33 uF
C8 was 220 uF
C11 was 100 uF
Adjusted cap voltages
T1 was 270EX
Exchange R9 and VR1
Rev 0.2.5: 2005-02-28
C9 was 0.22 uF
Rev 0.3: 2005-03-12
R3 was 100K
Float filament winding CT
Rev 0.4: 2005-03-14
Add R16
Rewire OT secondary to 16-ohm
Rev 0.5: 2005-03-15
Return OT secondary to 8-ohm
Rev 0.6: 2005-03-22
R11 was 15K 1W
Rev 0.7: 2005-03-26
Ground pin 1 of V3 & V4
Rev 0.8: 2005-04-02
R15 was 680 5W
R16 was between R8/C6 & VR2
R16 was 100K
Add C13
Rev 0.9: 2005-05-07
Add R17 and S2
Rev 1.0: 2005-05-24
Remove C13
Move R17 (was 100K) and S2
Rev 1.1: 2005-06-14
Add C13 and R18
Rev 1.2: 2005-06-15
Major changes to tone stack affecting C4-C6, R8 and R9
S2 is now the Slow/Fast switch; affects R16
R17, R18 and C13 are no longer present
Rev 1.3: 2005-06-18
R6 was 100K
R7 was 22K
R8 & R9 were 22K
C5 was 0.0047 uF
R14 was 15K
S2 is now the Mids switch
R16 was 300 10W
Rev 1.4: 2006-05-27
R15 was 360 10W
R16 was 470K
Add C13, R17 and R18
Rev 1.5: 2006-06-03
Removed tone stack consisting of C5, C6, C13, R8, R9, R16, R17, R18, S2, and VR1
R8 is now 22K
Changed designation of VR2, now VR1
C5, C6 and R9 do not exist

Measurements

Voltages measured w/DMM (Rev 0.7 build; R15 = 470 ohms)
180 mV 1 KHz sine input, 120 V line, tone control full CW
8 ohm resistive load
ValueIdle (no input)Load (half-volume)Full-clip (full-volume)
output stage B+390 V380 V377 V
output stage cathode33 V52 V61 V
driver B+333 V325 V325 V
driver screen164 V147 V102 V
driver plate126 Vn/an/a
driver cathode12 V11 V11 V
preamp B+280 V273 V271 V
preamp screen71 V67 V67 V
preamp plate101 Vn/an/a
preamp cathode1.3 V1.3 V1.3 V

 

Voltages measured w/DMM (Rev 1.3 build)
400 mV p-p 1 KHz sine input, 122 V line, tone control centered
8 ohm Hot Plate load
ValueIdle (no input)Load (half-volume)Full-clip (full-volume)
output stage B+392 V387 V375 V
output stage cathode29 V30 V48 V
driver B+331 V327 V318 V
driver screen157 V154 V135 V
driver plate127 V127 V137 V
driver cathode12 V12 V11 V
preamp B+266 V263 V256 V
preamp screen66 V64 V63 V
preamp plate94 V95 V93 V
preamp cathode1.3 V1.3 V1.3 V

 

AC Measurements
Voltages measured with scope (Rev 1.3 build)
122 V line, 1 KHz sine, tone centered
8 ohm Hot Plate load
ValueHalf VolumeFull Volume
input400 mv p-p400 mv p-p
cathode follower70 V p-p68 V p-p
volume hot10 V p-p9.6 V p-p
volume wiper1.4 V p-p9.6 V p-p
driver plate40 V p-p150 V p-p
power tube grids28 V p-p100 V p-p
output12 V p-p30 V p-p

 

Output Tube and Cathode Resistor Combinations
Voltages measured w/DMM (Rev 0.7 build; R15 = various)
180 mV 1 KHz sine input, 120 V line, tone control full CW
Cathode Resistor
(R15)
Output TubesValueIdle
(no input)
Full-clip
(volume @ 3:00)
360 ohm6V6B+381 V371 V
cathode28 V46 V
6L6B+375 V364 V
cathode34 V50 V
EL34B+377 V360 V
cathode29 V53 V
400 ohm6V6B+382 V370 V
cathode29 V49 V
6L6B+378 V368 V
cathode35 V53 V
EL34B+379 V364 V
cathode30 V56 V
470 ohm6V6B+382 V372 V
cathode30 V55 V
6L6B+319 V371 V
cathode36 V58 V
EL34B+382 V366 V
cathode30 V61 V
680 ohm6V6B+388 V377 V
cathode33 V61 V
6L6B+383 V373 V
cathode40 V67 V
EL34B+388 V373 V
cathode33 V69 V
1K ohm6V6B+392 V382 V
cathode36 V70 V
6L6B+388 V379 V
cathode43 V76 V
EL34B+390 V379 V
cathode34 V79 V

The output power (Rev 0.8) is approximately 15 watts at onset of clipping and 21 watts at full clip, measured into an 8-ohm resistive load. The choice of output tubes does not significantly affect power output.

Revision History

Revision History
2005-06-20Split the original document into a "Reader's Digest" version, plus this account of the step-by-step construction and tweaking details.
2006-05-27Revision 1.4
2006-06-03Revision 1.5