Designing tone stacks.

[tubeamptech]

I've been working on a 2 stage Marshall design, with cathode follower, then tone stack. I have an issue where I am getting this annoying mid range tone coming through. I've tried typical Marshall values, then Fender values with treble = 250K, Bass = 250K, Mid = 10K. Caps are 270p, .1uF, 047uF respectively. I've tried 1 Meg for the bass pot, still can't get rid of the annoying frequency. In this config I can isolate the annoying frequency the most. If I turn the treble up, yes I get more treble, but you can hear this mid frequency coming into the mix. Treble at about 2 has the most attentuation of this frequency, but then I don't have enough treble. Changing the treble cap to 470p or 500p doesn't help. When wired up with typical marshall values, turning any pot doesn't get rid of it. I've tripled checked my wiring. Pre-amp cathodes have roll-offs around 88 Hz. The plate outputs have the typical 470K + 470p in parallel. Power tranny is a Hammond 270HX, and a 50 watt Marshall OT. I'm using EL34. Does anyone have any ideas?

[J M Fahey]

Speaker peak.

Read this typical guitar speaker frequency curve (Eminence Legend 1258):

[Chuck H]

The trouble with your scenario is that you actually need that upper mid from the treble cap value in order to achieve good overdrive modulation with this sort of amp. If you reduce the treble cap value in an attempt to rid the upper mids then there won't be as much of that frequency clipping the PI and power tubes. The result is a tone with less definition on the bass notes and shrill highs. Likewise with using Fender values in the tone stack. The mid scoop will be at too low a frequency and deeper as well as an abundance of lower frequencies that don't overdrive cooperatively for this type of amp.

IIRC the typical Marshall design only uses the 470k with a 470p jumper on one of the channels. The "normal" channel is generally a straight 470k resistor. This is usually not modified when a cascade mod is done to the design (if that's how you have it arranged). The "bright" channel (with a 2200p coupling cap) is usually stacked in fron of the "normal" channel.

To keep the amp jiving the way this sort of circuit is intended to I think you should have only the first stage with a 470p jumper and use the stock Marshall tone stack. With the possible exception that you might prefer a 250p treble cap. Move to the stock Marshall values for the cathodes as well. Their interactions have a lot to do with classic intermodulations that give THAT tone. Once you have the amp jiving as it should, with all the rich and tasty harmonics it should have, the best way to control the EQ is as late in the chain as possible. That would be the speakers. I don't know what speaker/s you have now but different speakers and cabinets have markedly different final EQ.

JM2C on this sort of design.

[pdf64]

Tone stacks aren't the 'be all and end all' of tone shaping in an amp; bright caps and presence controls may be seen to be as important, along with significant contributions from various high pass and shelving filters, speaker impedance-power amp interaction, the occasional low pass filter and, perhaps most important of all, the speaker's frequency response.

[überfuzz]

I have an issue where I am getting this annoying mid range tone coming through.

Have you scoped and plotted frequency responses prior to this (the question is not meant as an insult, just to rule out a thing or two)? I've done some Bodé plots on guitar amps and I'd tell anyone that hasn't, you're favourite amp has a wonky frequency response. A HiFi styled flat response will 'kill' the tone completely.

[tubeamptech]

The only thing I haven't tried, is another speaker cab. Just have the one 2X12 closed back cab. I've suspected it cuz I have a Celestion V30 with I think a Warehouse guitar speaker, that's voiced with more bottom end to compensate Could be the V30.
I'm going to go back to the stock Marshall tone stack, but I found if I had the bleeder resistor at 33K, or 47K, I had WAY too much bass. Couldn't go past 2 or 3. I also seem to have an abundance of treble, mixed with that annoying upper mid frequency. I'm close, but not quite there. A 100K bleeder seemed to work the best. The first stage rolloff is the typical Marshall value. I've also changed to voltage dividers between stages to get more into overdrive/crunch territory (ie less gain than an 800), so the second stage has cathode bypass, where the JCM800 has the 10K. Everything is manageable driving the second stage, no out of control grid clipping.

[Chuck H]

Ok... Now we're getting somewhere! So this is only intended to be Marshall-esque in design topology and not even necessarily a Marshall sound then.?.

Depending on coupling caps, dividers, how impedances affect knee frequencies and a whole host of other stuff it's very easy to get lost on what's actually going through the amp if you don't plot it for each stage, start to finish. I have one design that uses a 150k slope resistor and a 250k bass pot in what is otherwise a basic vintage style TMB stack. That just gave me what I wanted for the bass response. It's good to experiment with the stack, just don't rely on it to change the overall implied frequency character of the final tone too much if an amp is going to be clipping anything after the stack. A good way to get aggressive bass in check is with smaller cathode bypass caps. You won't hear much difference above 4.7uf (with a typical 1.5k resistor) and there's not much useful tone smaller than .1uf. Anything in between that gives you the response you want is fair game. If you're already close you might try a lower value presence capacitor. There's a ton of gritty upper mids in the stock .1uf value (with the Marshall type NFB circuit employed). This is a great sound if you're Judas Priest or Def Leopard but maybe not for you. Maybe try values between .033uf and .068uf. The Vintage 30 is a speaker I like BECAUSE it's got an upper mid kick. And if your 2x12 cab is shallow or of light construction that doesn't help. This is just my experience. Lose the V30. Speaker recommendations are always iffy. I only like a couple of Eminence models but they do have a neat listen before you buy feature at their site that allows some comparison for scale.

I think a couple of small tweaks and new speakers you like better with that amp could make all the difference. I'd start with speakers because they will change everything else. No sense making adjustments yet.

[tubeamptech]

I'm trying to get that great Marshall crunch sound, think Rush's first album, mixed with a bit of AC DC, with a bit of hair on top, and not too bright. Not to over the top, where you can roll back guitar volume, and clean up somewhat.

I find that I also get the best sound with a 150K slope resistor. My first stage utilizes a 1K5 + 1uF cap, giving you a roll off of 106 Hz. The second stage cathode uses a 1K8 + 1uF, with roll off at 88 Hz. I find that if it rolls off at anything below this, there's too much bass. I haven't played around with the presence cap. I have it hard wired with a resistance, rather than pot, I'm not a fan off presence, but it's there. I'll try some different values, and maybe hard wire so it's partially "on". I have an Eminence supplier, I'll check and see what they have. Which models do you prefer?

[Chuck H]

I mentioned that I actually like the upper mids. Maybe you do to but what you're hearing just doesn't sound "right".

1uf is only technically a roll off at those frequencies as a bypass cap. The thing to remember is that the cap only provides a 5 to 7dB bump in that circuit so the -3db point is a little misleading as to how the different values actually affect the sound. Just swap a few values at the input triode and you'll see. IME a 1uf value is exactly what you want to accentuate upper mids! Maybe try the classic Marshall .68uf for the first stage and a 10uf for the next. I know 10uf seems redundantly large but it's a damn site smaller than the stock Marshall circuit. The point is full bypass for that stage so you don't get a low treble/upper mid bump there. Make sure you have a 10% audio taper pot in the bass circuit. The wrong pot taper here can make you feel weird about only having the bass on two Maybe even switch to a 500k pot instead of 1M.

I still don't know if you're cascading stages like a channel stack mod. If you are then you actually have three gain stages before the cathode follower. I have a lot of experience with that circuit topography. I cut my teeth on a Marshall plexi by starting with a channel stack mod and then proceeding to change every preamp component and circuit many, many times over the next couple of years. So if this is what you're up to, well, I might be able to give you an idea of what some changes sound like before you do them.

Can you post a schem? Hand drawn is fine. It would give us a point of reference.

[J M Fahey]

I think you need greenbacks .

No Eminence has that sound, and speakers are the last filter affecting frequency response.

No filter in the audio chain has as much influence as them.

I cracked my head against the wall until I learnt that, the hard way.

Combined with the poor or nil variety of speakers available in Argentina I was forced to make my own speakers if I *really* wanted to get the sound I had in my head, go figure.

You have the advantage that you can order anything you want for a reasonable price, use it

To see how much speakers change sound, listen to this YT comparison.

Not as a definitive reference, here they (poorly) used a Rivera amp , clearly at low volume, pure preamp distortion very harsh and buzzy, but I suggest this test over others because the guitar plays a short phrase exactly the same over and over; the BIG defect of a lot (most?) YT demos is that musicians play a lot of different things on one speaker, generally play and talk too much, when they switch speakers and repeat you fully lost the sonic reference.
To boot some even play different stuff , switch guitars, add or pull pedals ... crazy.

This one is perfect in that respect, although as I said earlier, speakers which I know sound very good, here are almost unbearable.

But the point is the huge variation, keeping everything the same ... except speakers,of course.

Remember that "crazy equalizer" response all speakers have (unless they are Hi Fi, of course) , full of peaks and dips.
No tone stack can create or compensate that, except, maybe, a 31 band graphic equalizer.

[tubeamptech]

I just found out I had the wrong treble cap value. I pulled it out of the right bin, but it was the wrong value. That's what happens when you tweak, tweak tweak, you get dizzy after a while, and put parts in the wrong bins, and don't check. Anyway, I'm back to stock marshall values, with a 150K slope resistor, and she sounding pretty damn fine. Just going to tweak a coupling cap to the next smaller value, and I think I'll be just about right.

Thanks for that comparison video, I did like the greenback on the lead tones.

So yes, it is a cascaded design. What I've done is made the shunt resistor in each of the voltage dividers much smaller than typical. I'm trying to keep the signal smaller going to each grid, letting each gain stage add it's harmonics, etc. I've gone up and down on cathode and coupling cap values, many, many times over, and I think I am there now. I hear what your saying Chuck H about getting a flatter response in the second stage. I've tried the jcm800 approach of using just a 10K there, but I don't like the sound of it. I've kept the signal going into the grids much smaller, therefore I can shelf the frequency response in the second stage. As mentioned, I'm going for a good crunch sound, with getting into outright distortion territory. Sorry I don't have a scanner to get my schematic uploaded, but I think you get the idea.

[tubeamptech]

After absorbing all the fine suggestions, I went back, and re-tweaked. Basically I have cathode bypass and mid emphasis on the 1st stage. I removed mid emphasis, and the cathode cap on the second stage, and I am there! It's just a matter of fine tweaking the gain on the first stage, and maybe re-adjusting my voltage dividers. I am 95% there. Play a Cadd9 chord in 3rd position, and it all rings out beautifully, with nice articulation. I want to thank you all for putting the bug in my ear. A big thanks to you all!

My next build will incorporate a 3rd stage before the cathode follower, for a distortion based amp. My first thought is, I leave these 2 stages as is, add the next stage between these 2 stages and the 2 triodes in the cathode follower, and adjust. Any thoughts? Much appreciated.

[überfuzz]

... Play a Cadd9 chord in 3rd position, and it all rings out beautifully, with nice articulation. I want to thank you all for putting the bug in my ear. A big thanks to you all!

My next build will incorporate a 3rd stage before the cathode follower, for a distortion based amp. My first thought is, I leave these 2 stages as is, add the next stage between these 2 stages and the 2 triodes in the cathode follower, and adjust. Any thoughts? Much appreciated.

First things first buster, Id say you owe us a Cadd9 chords. So cough up a mp3 of it.

How much distorsion are we tankning about? You could start by buffing the resistance of the anode resistors in these two stages.

[Enzo]

Juan has mentioned it twice, and you even spoke to it. The speaker is the single most effective way to change the tone of an amp. No tweaking in the circuit will have near as much effect. You can take a speaker you don't like and tweak components to make it sound closer to something you do like, but really, borrow or even rent as many different speakers as you can to find one your really like, THEN tweak your amp to complement it.

[Tom Phillips]

Can you post schematics to go with the descriptive narrative? That would make the discussion more complete and easier to understand what you did.