Overdriven Fizz Sound

[tubeswell]

Saturation in the OT? (How big is the iron?)

[JMFahey]

when I hit a big chord, there's a transient fizziness on the top end of the overdriven sound that increases and decreases in volume, comes and goes, and finally fades away, leaving the overdriven undertone.

What you are describing is normal when overdriving multiple frequencies together (big chord) instead od single frequencies (solo picking) and is called "intermodulation" and yes ... it's normal in that case.

The comes and goes bit comes from "beating" , a new low frequency appears, which you did not play, it's the difference in frequency between 2 notes being played simultaneously when overdriven, and that low frequency will modulate sound in a way similar to tremolo.

It does not mean the amp is bad, it simply shows a physical fact.

That you like it or not is something else, I won't talk about that.

FWIW I also listened to your other soundcloud clip, and it shows about the same:
https://soundcloud.com/steeledriver/tube-amp-demo
it starts with very clean and sweet sound, which means your amp works perfectly; then you switch to the overdriven channel, proceed to overdrive and logically get lots of intermodulation.
Suggest others listen to both tracks a couple times for clarification.

here's the original one you posted, just for reference:
https://soundcloud.com/steeledriver/amp-fizz

[Theashe]

Tubeswell - the OT is a Hammond 1750JA (4250 Ohm, 40W) and the PT is a Hammond 272HX (186VA, 600vct @ 230mA).

JMFahey - when cranked up, the amp hums a little louder than I'd like it to. It's getting in at the reverb recovery, master volume, V1, and V2. I'm still working on sorting that out, but maybe IM between the hum and the notes I'm playing is causing some of the unpleasantness I'm hearing. When I pick individual notes hard, I can hear that fizz on top just like with chords.

[Littlewyan]

Listening to the second demo I see what you mean now with the fizz. It doesn't sound nice. Could just be an EQ thing though. Let me put your amp into a simulator and see what I get.

Edit: How have you done the power string? As you don't show that in your schematic.

Edit2: Can you measure any voltage on any of the grids of the Pre Amp stages?

Edit3: What position do you have the tonestack switches in?

Edit4: Is that tonestack supposed to be the same as a Dumble tonestack? If so I think its wired incorrectly.

http://ampgarage.com/forum/files/ods50w ... ic_170.pdf

Look at the bass control.

[Theashe]

Thanks Littlewyan. What simulator do you use? I test things out in LTSpice, but if I have more than a couple tubes in a model it can't do transient simulations very quickly. I've attached my LTSpice schematic and a model of the EQ in a zip to help you out, if you're using LTSpice. Might be easier to copy/paste.

The power string is at the bottom of the schematic. Unless you mean how is it physically laid out in the amplifier?

All switches are in the off position. The problem persists independent of the EQ switch positions and the EQ pot rotations.

The EQ is supposed to be very similar to the Dumble tone-stack. Even though it looks wrong, the bass pot is still wired the same if you trace it carefully. Sorry about the confusing differences - I made my schematic without referencing how the EQ is drawn in Dumble schematics. Though I did notice a difference in the Deep switch between my schematic and the one you've posted. I've simulated the wiring of my EQ in LTSpice and all of the controls and switches work as expected, and I can confirm that from playing the amp.

I'll check the grids when I get home in a couple of hours. I haven't tested them in a while, but I'm fairly certain I had no voltage on the grids the last time I checked. Then again, that was a million modifications ago.

[Littlewyan]

C34 is supposed to be like a bright cap on a volume pot. With how you've got it in the schematic its shunting a lot of high end to ground. If thats how you've got it wired in your amp I would change that to begin with.

Also do you have an oscilloscope?

And what is in that attachment? My Firefox won't let me download it. Try attaching the LTSpice sim you've got as I use LTSpice as well.

[Theashe]

Yes, I have an oscilloscope.

The forum won't let me upload LTSpice file types, but I can change the file extension to .txt and upload them. You'll just have to change the extensions back to .asc and they'll open up again in LTSpice.

Yes, you're right about the cap across the bass pot. I'll try swapping that out when I get home, although in simulations it makes the bass control interact with the mid and treble controls pretty badly.

Try dropping in this spice directive:
.step param B list 0.01 0.1 0.3 0.5 0.9

And decade plot it with the 0.01uF I left in there vs a 0.0047uF or a 0.001uF. I'm finding it shifts the curves into the 1-2kHz region.

Edit: I noticed the pot values themselves are different than what I used. I'll check that out in Spice.

[Littlewyan]

Cool, the scope will come in handy if we need to start looking at what stages are distorting but I think really your issue is just EQ. At the moment you have a real bass heavy amp so its bound to sound like that when distorting.

Look at the EQ in the images I've attached. You're dumping so much high end and having so much bottom end boosted. If you place that cap in the correct place it will help but I don't think it will completely solve the issue.

Compare it to my JTM50. When I have my tone controls set for the tone I like this is what the AC Analysis looks like. See the big difference between the lowest frequency and the highest and all the mid range it has?

Which Dumble tonestack did you copy? I'm not very knowledge when it comes to Dumbles but I've looked at a few schematics and can see the tonestacks vary a bit.

[Theashe]

It all started here, but it's changed over time. http://ampgarage.com/forum/download/file.php?id=20293

There are differences between cap values and potentiometer values which followed some conversations with people on DIYAudio. That's likely where I went wrong.

The amp doesn't sound bass heavy, especially since I lowered the C1 coupling cap to 4.7nF, put the 1.5k/1uF partial bypass on V1a, and replaced the 0.22uF coupling caps to the power tubes with 0.022uF ones. I could reduce bass even further by going to 1nF on C1.

I'm modeling the 3rd Gen ODS tone stack in LTSpice right now to compare its performance to mine. I should get pot models so I don't have to use two resistors and spice directives to simulate pot rotation lol.

[Theashe]

One major difference I've just noticed between the ODS 3rd gen EQ and mine is that when the Rock/Jazz switch is in the Rock position, the bass capacitor is shorted across.

I threw together a quick simulation with one gain stage, sweeping the bass control, and the cross-talk between the bass control and other frequencies is enormous. That can't be right. Simulation and simulation file attached.

[Littlewyan]

Looks pretty normal to me. In some tone stacks the controls can interact with each other. Common with the Marshall/Vox/Fender tonestack.

Did you try moving that capacitor?

[Theashe]

I'm afraid not - I'm in the middle of moving houses and I had to pack up my tube amp last night. I won't be able to do any physical changes and testing until Thursday. In the mean time, I can still postulate and simulate!

If I crank the first volume to overdrive V1b and cut the second volume so that just the V1b distortion gets passed cleanly to the rest of the amp, it still has that fizzy distortion. Do you still think that's an EQ problem/bass problem if that fizz appears to be happening before the EQ? I can try going even lower than 1uF on the first stage bypass cap, or reducing the first coupling cap to 1nF to reduce the pre-EQ bass.

[Littlewyan]

Doesn't matter which stage is distorting, you're still dumping a lot of high mids to ground which you need. Leave the cathode bypass cap as I don't think that will help. Change the tonestack

[Theashe]

I found more problems with my EQ, particularly where the Rock/Jazz switch is concerned. A couple wrong connections, I believe. Interestingly, it still works well as an EQ the way that it is. Thursday afternoon I'm going to tear apart the EQ and put it back together CORRECTLY. Thank you Littlewyan for noticing the problem when it escaped me for several weeks. I'll report my findings, though I'm still having a hard time wrapping my head around the idea that the fizz is coming from dumping high frequencies. Can you explain why you think that's the problem?

If the problem is still there after the EQ rebuild, I'll start trying my best to trace the problem with my signal generator, oscilloscope, multimeter, and signal tracer. I'll check the grids again for voltage, and come back with a table of voltages throughout the amp.

I mentioned before that I have a star ground scheme that I'm going to remove in favor of a bus-bar. Do you think the problem might be caused by interactions between the ground wires running to the star? I think it's the source of the background hum in my amp, which might be making the IM distortion problem worse at high gain.

[Theashe]

Here's a thread I found on MEF that's been diagnosing a very similar sounding problem on Vox Night Train 15 amps. Sound clips included.

http://music-electronics-forum.com/t27536-3/

More things on my list to try, as per this post:
1. Try putting a 500pF bleeder cap from grid wires to ground.
2. More interstage attenuation so that the overdrive is more distributed. There are four gain stages in this amp before the PI.
3. The FX Send tube is acting similar to a cathodyne, so I'll try boosting the grid stopper from 220k to 1Meg, or simply disconnecting the FX Send so it's a regular common cathode gain stage.
4. Put small poly caps across ALL of the power supply resistors to reduce ESR at high frequency.

(EDIT: for #4, I meant to say power supply capacitors, as pointed out by Littlewyan).