Minimizing intermodulation distortion In an amp built for overdrive tones.

[tony321owen]

I have built a few amps that suffer in the hi-mid to hi frequency range from loud and unpleasant intermodulation distortion. This is the kind of dissonant distortion that occurs when you play more than one note at the same time and you hear low out-of-tune notes that are the sum of the two notes played and the clashing of their harmonics. As far as I understand this kind of distortion is a normal part of guitar amp function and overdriven tone but I am curious if anyone has tips about how to mellow the effect. Maybe I should aim to increase headroom in the phase inverter/ preamp stages and bias to minimize distortion?
Im curious of any tricks people use when designing amps to create sweet overdriven tones that don't sound out of tune while playing guitar lines with harmonized intervals.
maybe I just have to walk the tight rope between dirty and clean?

My Amps that suffer from this in varying capacities are AC-15 top-boost type build, AC-15 top-boost style with 6v6's, and a little 5f2A style build though a lot of the issues cleared up when I put it in a little 8" champ cab instead of pairing with my 12 bench test speaker.


 

[tubenit]

Things that have helped me lessen unwanted distortion and add smoothness to overdrive:

A good layout and a good job in wiring and soldering things up help.

Not ever using EL84's help. (my least favorite tube & I've had zero success getting a smooth tone with EL84's)

Using the "enhance cap" across the "entrance" LTPI plate resistor.

Using certain resistors and caps if the amp is geared towards an overdriven tone. (Not as critical in a clean tone amp).

Using smoothing caps.   (plate to cathode)

Using a 5751 or 12AY7 in V1

Using shielded wiring where needed.

Correct speaker match for the tone I want.  Some speakers just sound awful with some amp designs. 

Using an EQ pedal to determine what frequencies help or hurt the tone I want and then adjusting tone caps and coupling caps accordingly.  (An example would be in my 3 amps, I won't use a post LTPI coupling cap great then .047. And in my Carolina Overdrive Special, I found that I needed to have some mids scooped a little more then I originally had designed to get the smoothness I wanted.)

Using lower value cathode caps such as 5uf or 2.2uf instead of 22uf or higher. Highest I use is a 10uf on V1a on a clean amp tone.

Having a mindset on anything that's like a volume pot (such as an active FX level pot) that it doesn't have to be dialed up above 5 IF it sounds better lower then 5 & I still have all the volume I need.  This is a mental thing and matter of discipline for me.  For example with my Carolina Overdrive Special, I get the "best" smooth overdrive tone with my volume pot around 3 or 4. The COS essentially has 5 gain stages & if I crank up the volume pot to something like "7", it simply adds grittiness that I don't care for.  I don't need to "fix" that as though it's a problem.  I just need to accept and enjoy setting the volume pot at a lower setting.

In some situations on overdrive type amps not using silver mica caps but using ceramic caps instead for those small values like 500p or less.

Accepting that sometimes an amp is great with any guitar I have ………….. and sometimes an amp prefers a certain guitar I have to sound it's best.  That may not be directly related to the question of unwanted distortion but it is a factor for me in getting the smooth "tone" I want.

There are probably other things that help lessen unwanted distortion but these are factors in my thinking specific for the tone I want.

With respect, Tubenit

[pdf64]

Generally, to avoid nasty distortion artifacts, there’s some degree of low frequency roll off prior to the clipping stage (which, with non master volume amps such as yours, the most significant will be the power tube g1-k diodes).
However, I wonder if the extreme low frequency roll off of the typical Vox brilliant channel (which is the used with the TB stage) may be contributing / exasperating your issue? ie the bright tone and absence of low freq content may make the IM products generated by the clipping stage more apparent.
I suggest to replace the 470pF coupling cap to the vol control with one of much higher value, 2n2F - 10nF.
Adjust controls to taste and reassess.

[kilowhat]

Dont know if this also help intermod dist between notes in signal (as you describe), or just between signal and supply ripple. Either way:

"I add a common unbypassed resistor in the output tube cathodes of approx 25% of the value you would use for cathode biasing. This generates some common mode feedback in the output tube pair which helps supress some 3rd and other odd harmonic distortion but is particularly effective at supressing intermodulation between the signal and the power supply ripple. I learned this from HiFi Amp design but it works well in Git. Amps too."
LINK: https://www.diyaudio.com/forums/instruments-and-amps/266220-connect-plate-choke.html#post4154882

/kw

[jjasilli]

Intermodulation can happen at the speaker cone. To test that in an amp is above my pay grade.  I think this needs spectrum analysis which can now be done with software on a PC.


Can we sure what you're hearing is actually intermodulation distortion?

[shooter]

is above my pay grade

 
My scope will do frequency based, a single note (frequency) I can get, add a chord and all bets are off 

[pdf64]

Strictly speaking, IM distortion is a different thing to harmonic distortion; IM generates sum and difference frequencies, which probably won’t be harmonically related to the input signal frequencies.