The Nitro™ - a Homebrew High-Gain, Two-Channel Tube Amplifier

Updated 3 March, 2007

This amplifier is the culmination of a bit over two years of planning, designing, simulating (SPICE), dreaming, ordering parts, and, finally, busting knuckles and sniffing solder fumes.  The project started as a quest to build a low-powered amp (15 watts or less) with all the features of a big amp, and then some.  It's intended to be run guts-out into an attenuator and as such is heavily over-engineered in critical areas. The output transformer, for example, is a 40-watt output transformer which runs cool even when the amp is dimed.

I'm building three of these, two for a friend in Missouri and one for myself.  I powered up the first one last night (17 February, 2007).  To my complete astonishment it worked almost flawlessly right off the bat.  This is an extremely complex and high-gain amp and I expected to spend a week or so troubleshooting "issues."  The only issue was a problem between the post FX master volume and, I suspect, the large coupling capacitor on the output of the FX send buffer.  I added a resistor to the top of the volume pot which cleared that up right away.

I'd like to thank all the guys hanging out on the forum over at ax84.com.  Some of the success of this amp has to be credited to a couple of guys over there who I am smarter for just having "rubbed elbows" with - they know who they are.  I'd also like to credit Joe V. at amplates.com for the face plate, Paul Ruby for the cab, and Mike, the Kombo King for the chassis.  Also thanks to Paul Ruby for the idea to use zener diodes to prevent the bias on the power tubes from being driven too negative when heavily overdriving the power section (search the AX84.com forum for more info).

The rhythm channel is gorgeous and will do everything from a clean Fender bassman to a cranked plexi with the inputs strapped.  The lead channel is just sick with gain - if you can manage the fret noise it should do just about anything imaginable.  One thing that surprised me though is how well the lead channel does even with overdriven open chords.  Many amps turn to mush but on the Nitro each string is still pretty clearly defined even at crushing gain.  I'd like to snap my suspenders and claim some specialized knowledge and esoteric build techniques that caused this ("Yeah, we balance the capacitance of the shielded wire against the inductance of the wirewound resistors in the power filter to achieve that patented sound...").  Unfortunately, I haven't the slightest idea why this amp doesn't mush out chords the way some do - I guess I just got lucky.  I'll post clips when I have a chance to record some (or better, when I get someone else to record some - the world should not be subjected to my playing unless absolutely necessary).  In the mean time, here are a few pictures of the first build.

Here's a tip for tagging wires that originate from someplace inaccessible after the turret board is installed.  Use your computer to print some self-adhesive labels in a very small font.  Cut them out and wrap them tightly around the wires.  Cut them long enough to wrap completely around the wire at least twice, apply them starting at the blank end and the roll them around so the print is on top and the label is stuck to itself instead of just the wire.  Finally, shrink some clear heat-shrink tubing over them to protect them and hold them in place.  This works well even with wires that are quite thin - most of the unshielded wires in this amp are AWG-20 with thin teflon insulation.  Of course, you can also buy printable heat shrink but it is expensive and the last I was aware (1994) required a special printer.

After several days with the amp I decided that the lead channel was just too ridiculously high in gain.  I found that I rarely turned any of the gain controls above "3" and, while sustain was amazing, I just wasn't that happy with the performance overall of this channel.  Even with the gain controls fairly low there was enough compression from the extra gain stages that fret noise was a problem.  The clincher was when I discovered that while the amp alone did not oscillate it could be made to break into oscillation with certain effects inserted into the FX loop.

When I originally designed the amp I found myself with a left-over triode section and I added that section to the lead channel.  That necessitated adding a FET inverter before the FX loop to keep the two channels in phase.  So, I decided to try bypassing that extra triode stage and the FET inverter and FET FX loop buffer.  The result was much more pleasing, at least to my ears and playing style.  After removing the extra triode stage I wanted to add a bit of gain and "punch" back into the lead channel, so I replaced the 12AX7 for the NORM and BRITE input sections of that channel with a 12BZ7.  I also rebiased the input and mixer stages of this channel for a bit earlier/warmer overdrive.  This sacrifices some of the headroom but I feel that headroom is not a priority in the lead channel of a two-channel amp.

Everyone who has heard this amp in the new configuration has really liked it.  Unfortunately, nobody but me ever heard the original configuration so I can offer only my own opinion on the relative quality of the two versions.  I should have some good clips of the amp alone and in a jam session when I return from Missouri later this month.  The remaining two amps will be built in this new configuration.  In addition to sounding better this saves about twenty components since the entire FET buffer/inverter section and its +16VDC power supply go away.