The Bluesman Dynaflow

18W Mono Guitar Amp Project

by Bob Richards 12 9 23


The goal with this amp was to make the physically smallest amp that should be able to keep up with a typical rock/blues band in a typical bar venue. Also to use only the most conventional tube types. I've noticed that most of the bands I see in the Portland area use the Fender Deluxe Reverb, or an approx. equivalent. Small enough to be practical to carry in and out of venues, but powerful enough to do the job well in most cases. The stages are usually cramped, and carrying a big heavy amp is less fun as you get older. So here's my version of a Fender Deluxe guitar amp.

Fender pushes their 6V6GT output tubes pretty far beyond the max B+ voltage ratings, according to the "official" tube manuals, in order to get the 20watt output power rating. I did that in one of my other amps, no problem (with JJ brand 6V6's). In this amp, I chose to go with 18 watts output power, because I needed a 10 watt power resistor for part of the RC power supply filtering, and that's the value I happened to have (180R). I exceed the max voltage rating for the 6V6's, but not by as much as the Fender Deluxe.

I've heard that due to the popularity of the Fender Deluxe amp, todays tube manufacturers are designing their 6V6GT tubes to handle the higher B+ voltage.

Here's the final circuit I came up with. The first gain stage uses part values similar to the Dumble amp circuits. That's the only thing experimental about this amp. I wanted to see if there would be any noticable difference. It has lower quiescent current and higher gain, relative to the more typical topology you find in most Fender amp circuits. More gain and lower current in the first stage suggests lower hiss noise. It's not really a problem in any of my amps either way, but it didn't hurt anything. It makes it easier to overdrive the 2nd stage if desired, while allowing the Master volume to keep the final output level lower.

I initially set up the second stage 12AX7 the same way, but the amp had too much gain. If you overdrive the output stage by way too much, it sounds shitty IMO. So I went in and reduced the gain of the 2nd stage, and even still had to feed the Master Volume pot with a 6dB loss 1M resistor (embarrassing, but OK). It still goes WAY the Fuck LOUD for being an 18 watt amp (although I haven't yet tested it in an actual jam session with a full band).

For a Tone Stack, I wanted to have Bass, Mid and Treble. The difference between the Fender and Marshall stacks appeared to be "academic". Slight topology difference, perhaps to avoid copyright issues, but no significant operational difference. This is a slight variation of a Marshal Tone Stack. If I had it to do over again, I'd go with a variation of the James topology, which is lossier (by about 10dB), but has more range in all directions. Plus then I wouldn't need that 1M resistor feeding the Master volume pot. But this is good enough.

The Phase Splitter (the 3rd 12AX7) is a "Cathodyne" topology, which could be argued to be less high end (compared to the differential ("longtail")or Paraphase splitters), but the asymetry at higher frequencies produced by the output Z differential, creates a slight distortion that is arguably desireable for a guitar amp. It's what is used in the 1957 Fender Twinamp that Eric Clapton has been using for the last at least 25 years. I've tried all 3 phase splitter circuits in my various amps, and they all work fine. This one MIGHT sound slightly "juicier", but there are too many other variables between the different amps for me to make any real claim here. My personal favorite might be the Paraphase, because it has a pot that adjusts the exact amount of asymetry between the two outputs, which shapes the distortion spectrum better (specifically, a slight asymetry brings the 2nd harmonic distortion product up in amplitude, which sounds more "organic" or "natural". Less metalic, less "fingernails on a blackboard".

The output stage is is a Push-Pull 6V6GT circuit, with a fixed bias arrangement. Cathode bias would have been fine too, but I already had the power tranny I would use (which costs over $100), which has less secondary voltage than I wanted, and I already had the bias supply built (pulled from another amp I built that didn't measure up), so I went with the fixed bias route to get the full 18 watts. If I'd gone with Cathode bias using this power tranny, I'd have lost 20-30 volts of B+ across the cathode bias resistors, and ended up with something like 10-12 watts final output. Voltage is a squared term in the power formula, so a small loss in B+ will give you a big loss in output watts.

One reliability issue in tube type output stages of all guitar amps is where the impedance of the speaker goes way up at frequencies above the audio frequency range (typically above about 12kHZ depending on the exact speaker driver). When you overdrive the output stage, the clipped waveform generates energy above the audio frequency range. The output transformer becomes barely loaded at these higher frequencies, and can turn into a spark coil. Many amp circuits put reversed bias diodes off the plates of the output tubes to combat this. When the output tranny is effectively unloaded, it can turn into a spark coil, and cause arcing in the output tubes and/or in the output tube sockets. Rather than using diodes off the 6V6 plates to limit this, I put a series R and C across the speaker driver, so the impedance as seen by the ouput tranny is always limited to a somewhat safe amount. Since this RC combo doesn't do anything below about 15kHZ, it does no harm and makes the amp significantly more reliable when pushed to it's limits. The Fender Deluxe ignores this issue all together, so maybe not necessary in a low power amp.

The power supply is all pretty straight forward. I don't like using chokes because the are physically large, they add significant weight, and they're really not needed now that higher value electrolytic capacitors are easily available (they weren't always available and/or cost effective back in the 1960's and 70's). But a two stage RC filter does get rid of a significant amount of potential B+, so you need to take that into acount when choosing the power tranny secondary voltages. This circuit loses about 15 volts there, while a choke might only lose a few volts. The Bleeder resitors (to help discharge the filter caps when the unit is turned off) also acts as a voltage divider that is used to "float" the filiment supply at 37 volts (in this case), to reduce the amount of potential issue with the cathode to heater max voltage spec. I exceed that spec to some degree, but most, if not all Fender, Marshall, and most other amps don't seem to care about exceeding that spec by quite a bit in their designs. It's especially important when using cathode follower topologies. I do this for improved reliability.

Rf (Radio Frequency) and Digital Noise... When supersonic energy is delivered to the input of the amp (with the guitar acting as an antenna, and/or a pedal board with digital processing and/or switchmode power supply), at some point in the circuit that energy will get "demodulated" and generate Intermodulation Distortion, which generates sum and difference frequency products, which are not musically related to any notes you're playing on the guitar. This will create noise and distortion in the audio frequency range. Hence, the RC low pass filter at the input jack where the guitar plugs in. This Rf energy can also come in on the AC power cord. Especially if you are in a venue that has light dimmers, which are usually switch mode circuits that generate lots of Rf endergy. Hence the 0.01uF 3kV ceramic cap across the AC line connector. Most amps I've seen don't bother with either of these Rf reducing methods.

One thing that is common to all the amps I've designed, is that they don't use any distortion reducing negative feedback (like the Vox ACXX amps), from the output to an earlier stage. This distortion reducing feedback causes the amp to "clip" during overdrive with sharper corners, which means the distortion products go way further out in frequency, and happen significantly more abruptly when the amp is over driven. I prefer an overdrive that happens gradually, for a better bluesy sound. Using cathode follower buffers in the 1st and 2nd stages helps keep the amp very clean (low distortion) unless you specifically adjust the two gain pots to overdrive the 2nd stage, or the final output stages after the Master volume pot.

Here are some pictures I took during the build process:

First I plan the physical arrangement of everything on graph paper (oldschool and proud of it). That becomes a template for fabrication of the aluminum chassis. I scotch tape it onto the chassis, center punch all the holes that I remembered to include on the template (I always forget a few till after the paint job). Then pull off the paper template, drill all the holes to proper sizes. wash the chassis with hot water and dish soap, then sand the outside with fine grit sandpaper (500ish). Then wash it again with dish soap real good, dry it thoroughly, then spray paint the outside with Aluminum Primer (most primers don't work worth a shit - the primer must specifically say "Aluminum Primer". Then color paint maybe twice, then do the dry transfer lettering (my preference on one-offs), then at least two coats of matt finish clear spray to protect the fragile dry transfer lettering. Then it's time to populate the chassis, with all the crap.

I always guess at what and how many lug strips to mount in the chassis, rather than having an intensive pre-plan showing where every single part goes (the professional way). I like the challenge of guessing (since I've got alot of experience with this), and I'm lazy. If I was seriously going to produce quantities of one of my amps, I would go the pro route for this.

This is when I'll have to drill more holes in the chassis after the paint job, which can cause the paint to chip if not careful.

I usually start the wiring process with the power supply. then the front end stages. Then the phase splitter, then the output stage.

Whatever devices (transformers) have the longest most annoying wires get wired early on.

The chassis is held into the cabinet by steel bolts at each end, and then there's also angle irons (in photo above) for further support.

The one mistake I made with this amp (very embarrasing) is that when I went to install the finished chassis into the cabinet, there was no way for me to get my hand in there to put a nut on the screw for one of the angle irons (the one by the power tranny). You try to predict every part of the build process, so you can optimize everything, but sometimes you overlook something, and then you get to learn... what to do better next time. I should have positioned the angle irons so I could get at them on the bottom side in the cabinet. Oh well.

I like to somewhat "calibrate" the heater voltage so it's very close to 6.3 VAC fully loaded.

Most amps seem to run that pretty high IMO, which ages the tubes faster. Hence the above power resistors.

I'm a huge fan of a simple control panel, in case you want to adjust something in the middle of a jam. It's much less distracting.

Since I'm a pedal board guy, I don't bother puting Tremolo and/or Reverb in any of my amps. All my guitar friends are also pedal board guys or gals.

Newer juicier pedals come out all the time. Built in effects can end up redundant, and clutter up the control panel.

Above are some waveforms: Just short of overdrive on the left (plenty clean). Somewhat over driven in the middle. WAY overdriven on the right.

Don't read too much into these photos, since this is just one frequency (around 1kHZ) and there are many other variables in the real world sitch.

These photos are with the amp driving into a 10 ohm 100 watt Load resistor, rather than an actual speaker driver. Another significant variable.

This is overdrive of the output stage. There's also the option of over driving only the 2nd 12AX7 stage by setting the Input Gain and Master volume controls to do that.

The tweed covering came loose at the edge due to shitty spray contact cement, so I had to re-glue it down. I clamp it with a scrap board and wax paper.

On the right is an approximation of what the tone controls do.

That's about it. Now I just need to see how it integrates into a full band situation. This is a very basic amp. Nothing particularly fancy.

The speaker driver is a Jensen P10R; the same driver that is used in the coveted 1959 Fender Bassman amp. A classic rock/blues sound.

The Alnico magnet helps keep the weight down, and allegedly has a softer over-drive charactoristic, compared to ceramic magnets.

It sounds real good in my living room at home. No significant hum or hiss-noise issue at all.