Harsh sounding high notes (first tube preamp)

[jaywalker512]

Hey guys, I posted here about my first tube preamp I was planning to build a while back, and that build is now underway after plenty of weeks of studying and experimenting. I'm in the process of designing/messing around to find a sound I like (going for an 80's metal type sound, think Paul Gilbert with Racer X and Mr. Big), and I'm having a bit of trouble. I've been able to get some decent sounds out of various designs, but all of them have one very annoying characteristic that I cant seem to get rid of. The high notes I play are very harsh sounding, almost like a crappy SS amp would sound. Chords and lower notes sound alright (mind you, not amazing because its still a work in progress), but high notes are very harsh, buzzy, or "gritty". It happens even with very little distortion. Here's a schematic of what I'm working with at the moment:



I drew it in LTspice since that's what I've been using to analyze various different ideas and whatnot. The tubes I'm using are about 50 year old NOS Raytheon 6247 Subminiature triodes I bought on ebay. They are similar to 12AT7's as far as their specifications go. The power supply is a 350vAC CT transformer using a 2 diode rectifier and a 100uf filter cap for 250vDC B+. Heaters are wired in parallel at 6.3vAC. Here's a recording of the preamp so you can get an idea of what I'm talking about. It's a little harder to notice in a recording, but It's very obvious in person, especially if I use a gainbooster from my guitar. I've tried it through different speakers/power amps and other amps into the speakers I've been using, so I'm pretty sure its not a problem with my speakers/power amp, because the sound remains. I've played around with 3 gainstages, 4, 5, now back to 4 and about a million different cathode resistor/bypass capacitor/plate resistor combinations, and I'm really hoping you guys can help me figure it out. I hope its not just ugly sounding tubes

http://68.83.42.160:8099/ampclipharsh.mp3

[pdf64]

I'd be looking a trying 5F6A component values in the tone stack to bring the low mids up a bit / high mids & treble down a bit. I not tried it without a cathode follower driving it, but if your tubes have a lower plate impedance than a 12AX7 it may not make much difference anyway.
If that doesn't help, try rolling some more treble off early on - increase that 1n across the 1 stage to 2n2, or try it direct to ground rather than the cathode, or try a small cap 100-470pF to ground at the 2nd stage grid, then as the gain is increased, more treble is rolled off. Peter.

[booj]

I'd try replacing the 2nd stage cathode cap with either no cap, or a larger value, like 4.7uf.

[Gtr_tech]

He's got the right idea by using the small value cathode bypass caps. Things start getting mushy and undefined when too much low end is pushed thru the gain stages. I don't like the first stage coupling cap.....I'd use a 10n there to get better mids. C12, now I'd shove that sucker straight to gnd to get rid of more fizz. Adding a 470p to gnd at the second stage grid is a good idea too for more fizz reduction.

Now,why the 1m series resistors after stage 2 and 3? Seems like overkill.....470k should be plenty.

I've not worked with those little acorn tubes before. Could be they just don't take well to being overdriven. A 3 stage design with 12ax types will get you to 80s hair metal land quite easily.

[black_labb]

i've read that 12ax7's have the softest clipping characteristics and 12at7's have the harshest out of the common tubes. have you tried other tubes in the amp or the tubes in another amp.

[Gtr_tech]

C12, now I'd shove that sucker straight to gnd to get rid of more fizz. Adding a 470p to gnd at the second stage grid is a good idea too for more fizz reduction.

Typo alert....the 470p should go on the *third stage* grid, or just after the coupling cap.....whatever is easier.

[guitician]

A good way to smooth the "edge" of the clipping is to run the tube at a lower voltage. That will put the tube in a more "Non-linear" region and give it a "rounder" or "brown" sound.

[Gtr_tech]

A good way to smooth the "edge" of the clipping is to run the tube at a lower voltage. That will put the tube in a more "Non-linear" region and give it a "rounder" or "brown" sound.

It would be a trade-off tho. Response and general tone will change, as well as clean headroom. Well...as clean as it gets with a 4 stage design.

Could be these little subminiature tubes (I called them "acorn" tubes before, they are even smaller than those....just slightly larger than a neon indicator lamp) just don't work well for audio. I think they were developed as utility tubes for the military.

[jaywalker512]

Sorry it took me a long time to reply guys. I've been working on various things but the good news is I did pretty much solve the problem (thanks to all your help and a combination of the suggestions ).

He's got the right idea by using the small value cathode bypass caps. Things start getting mushy and undefined when too much low end is pushed thru the gain stages. I don't like the first stage coupling cap.....I'd use a 10n there to get better mids. C12, now I'd shove that sucker straight to gnd to get rid of more fizz.

I used the 2.2n there because any higher and it started to get really muddy. Also I tried the 1n straight to ground and at the cathode and it hardly made a difference, so I just left it where it was because it was easier to connect.

Now,why the 1m series resistors after stage 2 and 3? Seems like overkill.....470k should be plenty.

Well, you were right, it was overkill. These tubes overdrive very easily so I needed massive voltage divider between stages to avoid too much distortion. Wasnt thinking at the time to just lower the bottom resistor That also helped in getting rid of some extra bass contributing to muddyness.

Typo alert....the 470p should go on the *third stage* grid, or just after the coupling cap.....whatever is easier.

Did help kill some treble, but didnt get rid of the fizzy nastyness.

I'd be looking a trying 5F6A component values in the tone stack to bring the low mids up a bit / high mids & treble down a bit. I not tried it without a cathode follower driving it, but if your tubes have a lower plate impedance than a 12AX7 it may not make much difference anyway.
If that doesn't help, try rolling some more treble off early on - increase that 1n across the 1 stage to 2n2, or try it direct to ground rather than the cathode, or try a small cap 100-470pF to ground at the 2nd stage grid, then as the gain is increased, more treble is rolled off. Peter.

Changed the tone stack, did improve the mids. I think I put a small cap across the 2nd stage grid at one point, I'll probably put it back though because it is still very trebly.

i've read that 12ax7's have the softest clipping characteristics and 12at7's have the harshest out of the common tubes. have you tried other tubes in the amp or the tubes in another amp.

The tubes I'm using are subminiatures, so even if I did own another tube amp I couldnt test them in it

A good way to smooth the "edge" of the clipping is to run the tube at a lower voltage. That will put the tube in a more "Non-linear" region and give it a "rounder" or "brown" sound.

This seemed to help the most. At first I tried putting a 3K series resistor before the filter cap to increase the power supply impedance and thus drop the voltage at the plates, and that didnt work (mainly because the voltage hardly dropped at all), and then I remembered that simply increasing the plate resistor dropped the voltage AT the plates AND put it into the non-linear region. I had tried this before on a few of the tubes but I ended up needing it on ALL of them to kill the nastyness. 220K resistors are on there now, I would bump it up to 330K if I had any around.

It would be a trade-off tho. Response and general tone will change, as well as clean headroom. Well...as clean as it gets with a 4 stage design.

Could be these little subminiature tubes (I called them "acorn" tubes before, they are even smaller than those....just slightly larger than a neon indicator lamp) just don't work well for audio. I think they were developed as utility tubes for the military.

Well the frequency response didnt change a whole lot as a result of the above, but the feel of the preamp certainly did (in a good way). Also the way I re-biased them allows me to get a pretty clean signal out by rolling down my guitar volume, before I couldnt get a clean sound out at all. Yep, they were designed for military use, says so on the datasheet and that they were made specifically for low microphonics.

Thanks a lot guys, all of your contributions helped solve the original problem and overall it sounds a lot better now. There's still a slight hint of the fizzyness on high notes and its still pretty trebly, but its not a horrible painful type of sound like it was before. I'll record a clip of it if anyone wants to hear it tomorrow, but it is a little late for me to record one tonight.

Here's the updated schematic with all the current changes:

[guitician]

I'm glad to be of help. Yeah 30k is way too small, a 470k would have worked the same as 4x220k. Those 6247 don't draw much current. Another thing you could try is to put resistors between r9, r14 and r17 and maybe one to ground on the end, like 470k or 680k to help drop voltage. Since your not driving a power section, keeping the voltage supplying the plate resistor low will give you a more manageable output voltage.