Making the Carvin MTS3200 suck less

A few months ago, I purchased a used Carvin MTS3200 from a friend who had it sitting in his music store.  It had been there for a while and he didn't really understand quite why it wasn't selling.  I played it and thought it was a cool tube head.  After awhile, my Fender Twin was getting a little long in the tooth for what I was playing and the voice in my head told me I needed something with some high gain.  So, after some horse-trading, I called the MTS3200 mine.

I have owned a ton of different amps from different manufacturers and brands over the years, but I didn't have a whole lot of experience with Carvin.  I had friends that owned some and my father had some of their equipment back in the day, but that was about it.  For me, this was really an unknown.

Superficially, the MTS3200 Master Tube Series (also known as the Metal Tone Series by some) has some pretty decent specifications.  It's a two channel, foot-swtichable head with master reverb and presence controls.  Each channel has its own independent tone stack consisting of Treble, Mid, and Bass.  Channel one, the lead channel, is the high gain side of the amp with a volume and drive control.  Channel two is clean-ish with a single volume.  There is no master volume on the amp.  It's a pretty standard design on the front panel.

The specs on the amp list it as a 100 watt model, which is switchable to 50 watts.  It has two speaker cabinet outputs with a 4, 8, or 16 ohm impedance selector.  It has output jacks for a FS22 footswitch, cabinet voiced line out, and an effects loop.  It has a bias adjustment switch to allow the use of either EL34 power tubes or 6L6GC.

Brad, you're selling this pretty good but the title of the post implies a problem …

For being a tube amp, the entire thing is pretty sterile.  It doesn't have a lot of life out of the box.  I know this isn't just mine because after some research online, it's a pretty common complaint and really explains why this amp didn't really fly off the shelf.

It has a pretty respectable clean channel that takes effects well when they are thrown right into the input.  It is not a Fender clean by any stretch of the imagination, but it's competent for most things.  The big problem however is that lead channel.  Although it is a tube head, Carvin made a decision in the mid-1990s to use clipping diodes to drive the high-gain aspects of the amp.  This is an 80's trick that was employed in some pretty big name amps at the time, but it's a technique that is kind of like spandex, leg warmers, and big hair - it's time went and passed and is now better left in the decade of Gordon Gekko.

What this does is takes our nice tube gain and as our volume climbs, which means more voltage passes through the circuit (in a high-level sense), we begin to feed the second gain stage and drive it higher and higher.  This is good and gives us that ooey gooey tube goodness we love so much.  But then the current gets to a fever pitch when our drive is fully saturating the circuit and BOOM - clipping diodes work their solid state magic to block our distorted climax.  What should be an awesome sample of audio orgasm really turns into a cheap dirt pedal in the chain.  The sound becomes brittle and thin and that's just no good.  So what can be done about this?

Mod Part One

Disclaimer:  I am a crazy fool that has no problem poking around inside tube amplifiers.  However, the voltage inside these bad boys can kill.  There are big capacitors that are designed to do nothing more than get, hold, and release filtered power to the rest of the circuit.  Because of this, they hold onto power for a long time.  If you don't know what you're doing - don't do it.  Do not say, "But Brad told me to do it!" or have your next of kin try that either.  I'm crazy, but I will only play around with cold amps that have been properly discharged of voltage - I recommend you do the same.

I did some research on other mods people have done and came to a few conclusions.  We can see, in the image above, the path from V1.A to V2.A is pretty straight forwards.  We have a 47k resistor at r11, a .0022 400v capacitor at C14, four 1n4745 diodes at D1 - D4 that are in the path, and then pin 7.  Some people have approached this by simply removing D1 - D4, which creates a direct path from C14 to pin 7.  The problem is that within this design, the diodes act to regulate voltage hitting pin 7.  Simply removing them isn't the best idea.  The goal is to maintain the design integrity of the amp design while getting rid of these diodes.

In the picture above, this is the section of the board being worked on.  In the center of the picture is socket V2, which houses both V2.A and V2.B.  I've circled and marked pin 7, which is easy enough to find simply by looking at where the "1" is on the board and counting around, clockwise, until you reach pin 7.  Directly below that is C14 and directly below that are the four diodes (D1 - D4).

A voltage divider to the rescue.  For this, I removed D1 - D4 and installed a 150k resistor from the pin 7 side of C14 to pin 7.  This wasn't too bad, however be warned that the board will have to be removed from the chassis to do this neatly - especially C14.  Doing this gave me one half of the divider, which created a resistance between the two points.  To complete the divider, I added a 150k resister from pin 7 to ground.  For simplicity, after desoldering the diodes, an astute reader may notice D4 and D2 connect to ground, which can be reused in this modification.

This design works pretty well and is clean.  It doesn't eliminate the diodes and the attendant design implications they have, but it replaces them with pure resistance.  Part one of the mod is done, but there's a second part.

Mod Part Two

Many people who have modded this amp have reported delay and various oscillations in the gain stage when trying to eliminate the diodes.  Honestly, there are a thousand and one things that can be affected in the process and it's not my place or inclination to track them down.  While I could do the math, quite frankly, I don't really want to.  That being said, the most reasonable fix is to stabilize this adjusted current in the next stage, which is really the other half of the V2 12AX7.

Here, we can see the next stage, which essentially runs from after our first voltage divider, through the tone stack of the lead channel and then into the next tube.  Don't get too confused, we are still working with the same physical tube, but it's the other half of it.  Our first mod removed the diodes and replaced them with a voltage divider.  This means that while we got rid of the solid state nasty, we did modify the amount of voltage winding its way through the tone stack and into the next tube.  Obviously, this will change the tone profile of this channel, so we need to compensate.

Another voltage divider to the rescue.  For this, I simply lifted the ends of C22 and R17 and put in a 220K resistor between them and pin 2.  Note that pin 2 and C22 are across the board from each other and will require some tracing of the PC board leads to really visually see the path.  There are several jumpers spanning the distance and it may require some creativity to get it taken of.  My specific PC board was Rev C while the schematic posted is Rev E.  My portions were mostly the same, but to the reader, be wary.

In line with this first resistor, I put in a 680pf Cap in parallel.  This serves to attenuate some of the higher frequency response and bleeds off some of the brittle-ness of the overall post-tone stack signal.  This is really a preference thing for me and can be changed up to dial in the response by simply changing out values on capacitors.

Much like the previous voltage divider, I then inserted another 220k resistor between pin 2 and ground. Again, while the schematic shows some distance between the stages, this is the same tube socket, so those convenient diode ground connections are available for use.  It makes things much cleaner.

And that was that.

Once I was finished with the soldering iron, I metered all of my connections to ensure the signal path was showing the expected resistance and then I reassembled.  Happily, Carvin used a whole series of quick connectors to connect the board to ground, other boards, transformers, and switches.  This prevented a lot of unsoldering or fighting with wires, however, I HIGHLY recommend labeling each wire prior to disconnection.  It makes reassembly that much easier.

After reconnecting everything and giving the pots a good cleaning, I reinstalled the tubes and gave her a test run.  The distortion was chunky with a very smooth sustain.  The pots were very responsive and went all the way from a quiet grumble on low settings to over the top screaming at 10.  The bass was a bit more boomy after the mod, so I rolled that back slightly, but all in all it was a success.  My only gripe is that the amp is very sensitive to the guitar running through it.  My PRS sounds like a dream but my Strat, while still rocking, creates a lot of noise on the lead channel.  But after 15 years, maybe it's time to revisit my strat.

I haven't really had time to get to know this amp yet now that it's been retrained.  I'll see how it goes and update as necessary.