biasing preamp tubes?

[cbarrow7625]

Ahhh....now we're getting somewhere!

So, cathode biased, no feedback, verly large PI tail resistor, 56k cathode follwer resistor, 22uF cap on the second stage, no HF "emphasis" cap/resistor combos in the preamp and all ,022uF coupling caps.

With no negative feedback, EL84's and cathode biasing I would totally expect this thing to sound very midrangey & "honky". Going for a hi-gain amp, feedback can be your friend. An arrangement like this would typically be seen where you are trying to overdrive the power tubes into saturation (which I don't think you are trying to do - correct me if I'm wrong). Feedback around the power amp is going to go a long way to "flattening" out your sound and getting rid of some of the ugliness you are hearing out of your tone stack. If you add a "presence" knob (ala Marshall) you may find that you now have a very tight, balanced low end & an additional Upper Mid/HF tone control for your sound. I have a setting on my amp that allows me to tweak the feedback resistor and remove it altogether and I find that I really, really, prefer it in (at even lower than normal stock values, usually). Everything is just much smoother & more balanced.

I'm not really sure why you have a 56k tail resistor of the PI. Did you copy that from somewhere or was it the only value you had? You might want to try just copying a standard Marshall-type PI + feedback setup in this thing. I think you might like it.

For a cathode follower, you typically want to make the cathode resistor in the follower the same value as the preceding plate stage. This helps with symmetry of the voltage swings. A 56k here can affect the tone of the amp. Give that a try.

I find that having a cathode cap on the stage directly preceding the tone stack usually gives the tone a much-needed "boost". It will also lower the output impedance of the cathode follower by lowering the output impedance ofthe preceding stage. Put a 2-10uF cap (I usually find 5uF works well here) in parallel with that 1.5k Cap.

In doing that, you may find that you need to ditch the cap around the 3.3k resistor in the previous cathode.....speaking of which....for a hi-gain amp, 22uF in this spot in the circuit is almost always going to be way too large. No wonder you are getting blocking distortion going on. This very large cap (in combination with the large coupling caps & no HF emphasis) can explain more of the "midrangey" tone stack problems. You are emphasizing way too much lower-mids for a typical hi-gain circuit. If you want to keep that there I would recommend taming it down to 5-10uF at the most.

If you are going to use 0.022uF copupling caps in a hi-gain amp, you really need to emphasize the HF somewhere in the circuit or you will end up with the dull highs that you have been complaining about. Try just putting a 500pF cap around the 470k series resistor. How about 470k/470pf caps before one of the volume controls? That would help a lot as well. The fact that you are running each stage into a 500K load instead of a 1M-1.5M load (like is seen in many, many high gain amps) will tend to contribute to the lack of harmonic sizzle in the sound. To make nice harmonic distortion, tube stages like to "run free", that is they like to have a very high load to run into as to not load down the output of the tube (and create a voltage divider with the internal impedance ofthe tube itself).

If this was my amp (and these are totally just oersonal preferences based on trying out a lot of stuff over the years), I would have the 1.5K cathode resistor on the first stage & the 820 ohm resistor on the 3rd stage as well.

I feel like I'm telling you that everything is wrong & to start over from scratch....sorry. Without actually hearing it it is hard to make the right recommendations. I have tried a bunch of things over the years to see what I like & what I don't. I have tried preamp cirucits very similar to this & I have always had the same complaints about them that you are talking about here. Not enough HF presence, too much lower midrange, blocking distortion...etc., etc.

Take these thoughts for what they are worth. I hope they help send you down a different path where you end upwith exactly what it is you are looking for in the tone. Let me know what you think.

[daz]

Thats a lot of info ! I have a lot to try now so i'll be busy for a while. As to NFB, i tried adding a presence before but i really had no one telling me what would be best and the values i used or way i did it wasn't productive. as to the 56k tail, it was in the schematic i used. I'll look at a jcm 800 and swap to whatever value that is. thanks sooo much. I'll let you know how it goes.

[daz]

I've tried several things so far. Things that seemed to help were replacing the tail with a 15k, changing the 56k cathode follower resistor to 100k, and removing 2nd stage cathode cap. (which by the way was a 1 uf...my mistake on the schematic) replacing the 1st stage pot with a 1 meg seemed to be worse, but i will have to swap it out again and do more A/B testing because it seemed to give less gain which is opposite of what it should have done. Didn't have two 1 meg pots so i couldn't do the second pot. Also, putting a 500 pf across the 470k didn't seem good, and neither did a 470k/470pf before a pot. I assumed you meant to the input of one of the first 2 pots after the coupling cap. thats where i tried it anyways.

Of course there is always ear burnout and placebo effect. But at this point i thing some NFB would maybe be the thing to try. But i'm not sure what values and how to best implement it. So i may try a marshall copy. By the way, i used 220k instead of 470k on the PI for more headroom, which is something i was told to do when i asked elsewhere how to get a cleaner power section.

EDIT: tried nfb and didn't like it. I however then tried the hi pass filter you suggested (470k/500pf) coming out of the second pot and i liked it. Now the amp is sounding pretty good, but i think i may have to try nfb again

[Bruce / Mission Amps]

Wow Daz... are you aware that a cathode biased pair of EL84s is only going to need about 12-15v of drive signal to turn them on full blast and beyond?
You have quite a bit of gain structure there and I suspect the probability of preamp gain clipping has more of an effect on the actual tone then the tone controls. I mean there will be so much square wave energy that the whole thing will sound crunched and compressed with much less definition, or tonal spread, between the bass and treble.
In other words, I think you can smash the cathode follower driven tone stack with what you have now... and all that with no negative feedback.
I think the first thing I would do is limit the gain of this thing by about 50%.
And not just be reducing the gain of the phase inverter as you are doing with the 15K cathode and 56K tail resistors.
Reduce the gain of that first stage by a bunch to start.
Dump that 820 cathode resistor in favor of a 1k5 to 2k2 bypasses by a 1uF cap.
Try a little smaller value coupling cap too.
4n7 to 10nF comes to mind here and then use a 220K to 470K resistor in front of your gain pot... but smaller, maybe a 100Ka to 250Ka pot would be better.
On that triode in front of the CF triode... try a 4k7 cathode resistor instead of the 1k5.

etc...

[cbarrow7625]

Try a bunch of differnet values of NFB resistors. Remeber, the NFB resistor and the long tail resistor (the one that goes on the bottom of the PI tail that is not currently in your schematic) form a voltage divider for the NFB.

Just because you don't like one voltage divider setting doesn't mean there is not a combination that will work. We can only give you some basic guidelines here. It is up to you to take that info & run with it. Try a bunch of variations of things before saying you don't like it. Try a 100K series NFB resistor going to 5k to ground (at the bottom of your current PI tail). Remember, the .1uF cap & 15k tail resistor both go to the top of the 5k resistor (the .1uF doesn't keep going to ground, it gets "lifted" so that some NFB goes into the 2nd half of the PI).

Try different cap/resistor combinations in series with your volume controls. I can't list them all out here but suffice it to say that adding any combination of these won't hurt your amp so don't be afraid to just stick stuff in on a whim & try it.

[cbarrow7625]

Good advice Bruce. I assume that he is using it with the MV turned way down, which will work OK. If he opens up the MV it would probably not only sound bad but squeal like a pig.

Since it is cathode biased,it would be easy enough to put a voltage divider between the output of the PI & the EL84's. I'm thinking start with 150k series, 47k to ground to start (or just replace the 220K resistors with some 250K audio taper pots to find the right level, then substitute fixed resistors later).

[daz]

Well, it's obvious i'm no amp designer. But you guys are giving me a good education here ! Lemmie try some of this and get back.

[daz]

Heres the schematic as the amp is now after the last few things i tried. These are the ones that seemed to help in some way or another, tho i think lowering the 1st gain pot wasn't good and i may go back to 500k. The amp was better a couple tweaks ago i think. losing the gain seems to have hardened the mids even more, and i'm not sure what to do about that aside from going back a few tweaks. But there is also some good stuff happening that i can't quite put my finger on. So maybe i'm on the right track but in getting there it is causing other circuit values to need changing.

http://img142.imageshack.us/my.php?image=18bre8.jpg

[daz]

This is driving me nuts ! you can pretty much ignore that new schematic because i've changed a lot of things since then. I think i need to set everything back to where it was to begin with and start over again by trying things and slowly listening and digesting it all before i move on. this was all too quick and i think i did too much and sorta lost the magic i was hearing before even tho i was making progress in other ways.

One thing tho....does it matter which tap i use for feedback?

[Bruce / Mission Amps]

More techno-drivel

Go ahead and put it back where you last liked it and then just do one mod at a time...
My hat is off to you home builders that can and do this stuff without a good signal generator and O'Scope!! With one or two small mods at a time, just looking each stage, with a scope and clean signal injection can be incredibly revealing!!

If you have a feeling the basic tone was there but it was just to hard to control, you can also try a split load resistor on the first or second gain stage. That would be your 100K plate load resistors.
Try something like two 47K resistors with the coupling cap between the two or for less output to the next stage a 27K sitting on top of a 56K resistor. etc.
Or even better, stick a small 100K trim pot in place of the 100K plate load and connect the coupling cap to the wiper. Turn the trim pot to taste.
(I've done this many times only to find 82K to 100K was still the really good ones...)

With respect to your phase inverter, your 56K tail resistor is OK and it will help create better balance of the PI output.
But I'm not sure why you picked the 15K cathode biasing resistor in the phase inverter.
I think it might have been instinctive thinking on your part that it would help reduce the gain of the PI a little.
This really is easy to do but not the perfect way to achieve gain reduction.
The negative aspect to using both of these large value resistors is that the further you take that phase inverter colder in bias, by using your 15K cathode biasing resistor and holding the whole thing up further from ground by the large value tail resistor, the harsher and harsher that stage can sound and it does impose it's will on the output, plus it will also clip easier.
I think that stage really wants to be idling in fat solid class A for better tone with higher idle current... way better then higher bias voltage, cold idle current (nearing class B or C)... in such a way that it will run into cutoff.
Hint: The power tubes will follow what ever the phase inverter does.

Remember, with respect to where the triode's grids and grid load resistors, (see the bottom of the 15K cathode resistor), the common 15K resistor you have installed is sharing cathode current to both triodes.
Typically, when this is done you would see a lower value cathode biasing resistor.
You've seen it in many Fender and other amps who's dual triode's cathodes are sharing a common cathode resistor... 820 ohms vs 1500 ohms for each cathode.
So, try a 1k5 ohm to 2k2 resistor and work on gain reduction in front of the phase inverter.
The long tail phase inverter is a little tricky to really understand sometimes it's a good idea to work on that theory in order to see what you are doing with by changing the supporting components.

********
Adding to those thoughts, read up a little deeper on what the term "bias" really means with respect to the three basic elements of the tube and how there are different ways to change the bias!
It will reveal some interesting things to you about the circuit.
As an example, with that big 56K and 15K Rk resistor there (Rk means cathode resistor), you could be adding some other negativity to the total amp circuit, because the actual plate to cathode voltages are lower, creating less headroom of the PI and additional colder bias of those two triodes... which can result in more odd harmonic ugliness when drive so hard by your preamp stages.
With high relative bias voltage (with respect to the higher cathode to lower plate relationship)... both of the triodes can be driven right into cut off much easier.
In the wrong circuit I think this can sound very ugly and even worse if you then decide to use localized negative feedback in the phase inverter.

[daz]

Bruce, i wasn't sure at first, but now i am.......you think i know more than i do ! Pathetic, but true. I didn't understand much of what you said, so i don't know exactly what you were suggesting i do to the PI values. i've played with electronics in my amps and pedals and all for years, but i never built an amp till this one. I can easily read a schematic and know some basics. But theory is greek to me. I understand some basic things such as how NFB works, how filters work and such. But when you start talking about swing and cutoff and those kinds of things i'm lost.

I didn't do this for fun, and i never work on gear for fun. I do it because i am a slave to good tone ! I can't play without it because i cut my teeth on amps that were touch sensitive and without that i may as well be playing one fingered. So i never learned much, just what i needed to know. But building and tweaking an amp requires more knowledge and i guess i'll just have to take it a day at a time and hope i get lucky. The amp had so much magic at first, and i could have easily lived with the imperfections if i could have just gotten that smooth transparent midrange. But at this point i feel theres something inherent in it that won't allow that to happen. maybe the tranny or something, i dunno. I figured as good as it sounded at first it would only get better.

in any case, i will put it back to where it was assuming i can remember all the values, and i was thinking of maybe trying to reduce the level before the tone stack so i don't have to lose the amount of preamp distortion it had before. I can afford to lose a bit, but not 1/2 or more which is what happened. Then if i'm stuck with those hard mids i'll just have to deal with i suppose. Thanks much for your time.

[Bruce / Mission Amps]

More overly worded drivel

Bruce, i wasn't sure at first, but now i am.......you think i know more than i do ! Pathetic, but true. I didn't understand much of what you said, so i don't know exactly what you were suggesting i do to the PI values. i've played with electronics in my amps and pedals and all for years, but i never built an amp till this one. I can easily read a schematic and know some basics. But theory is greek to me. I understand some basic things such as how NFB works, how filters work and such. But when you start talking about swing and cutoff and those kinds of things i'm lost.

OK Sorry... this will be very general and so as not to upset the overly sensitive, hard core theorists here, I apologize now for the looseness of this.

All I was trying to point out was that as the input signal (a sine wave) is imposed on the grid, the plate output will be 180 degrees out of phase and when linear, should be a complete duplicate of the driving sine wave on the grid but at a much larger voltage swing and that pretty much is the definition of voltage amplification.
The swing is the total up and down of the signal from a zero point up positive to a peak, back to zero and then down to a negative peak and back to zero again.

In high gain amplifiers, with what is usually too many gain stages and little attenuation, ... it is super easy to get ambitious with "over the top" and unneeded gain. That builds in lots of problems with tone control, smearing and a buzzy sound that reminds me of hundreds of bumblebees in a paper bag.

At any given moment, a moment in time with respect to the signal being a sine wave riding on the actual DC bias voltage, the grid sees that as varying bias voltage.
The AC signal added to the DC bias is a momentary total bias voltage.
The difference between those, with respect to the actual bias voltage, is the audio component.
Positive going signal means less momentary bias voltage and negative means more momentary bias voltage.
But that change is what makes the cathode and plate conduct in variable DC, hence forth called AC or audio, and creates the AC voltage amplification on the plate which you want to pass through the low impedance path of the coupling capacitor... onto the the next stage's grid.
You've copied, or duplicated, the AC signal on the grid at a much higher output voltage on the plate, or if a cathode follower, in phase with the nearly the same AC voltage as on the grid, but
much much lower impedance and no gain.

If the input signal to your phase inverter is so large as to be higher then the tube's bias voltage, the output of the tube will be either cut off in the negative going part cycle or roundly flat topped and saturated on the positive going part of the cycle.
And that goes with the power tubes, especially!.... even though a little of that sounds really cool.

More "apparent" total negative voltage on the grid will drive that tube into cut off during the peaks... which means the tube is temporarily biased too far negatively and it stops amplifying.... to what is called cut off.
In other words, the tube gets driven negatively to a point where it stops conducting in that peak part of the negative cycle and there is no signal coming out anymore. Transition is UGLY.
The other part of the sine wave is more positive then zero level, and with heavy drive levels will push the output so far as to reach high current tube saturation. Not quite so ugly but not linear.
The negative going one will sound like a class B amplifier, which in most audio amps, is not really linear and can be very ugly sounding, while the positive one will still have output but be massively distorted with all kinds of even, odd and compliments of those frequencies... all mixed up which we call harmonics.
That can sound OK to some level but very dirty.
The power tubes will try and duplicate that exact same distorted waveform up until the driving level on it's grids does the same thing too and they get driven into cutoff or saturation. That's a complicated waveform now and the poor output transformer and speaker has to deal with it... nasty.

I guess what I'm really trying to say is, it is a matter of balance throughout the entire amplifier to get a great cleanish sound and still be able to kick it in the ass to get some of the lovely preamp and power tube distortion... especially where the tubes start screaming for mercy and want to take off on their own a little.
Other wise you just end up with an amp that does everything equally poor and has very limited use.

OK I just re-read all this and I'm rambling on and on.... sorry.

[daz]

No need to say you're sorry....you're spending your free time trying to help me and anyone looking at this whom it may help ! But still it's beyond my understanding to a large degree. i just need to know where my amp may be out of balance and what values to use where. at this point i went back and didn't like it. But i DID get that gain back. So what i did was to leave the cathodes close to how i had them, tho a bit less gain. then add many of the suggestions you both had for the rest like it was yesterday. at this point i like it best but it still needs tweaking and still has that hard mid sound. But it had that even after i did all of your suggestions, so it's something i guess i haven't altered yet.

One thing i wasn't sure about that you mentioned a couple times was that you questioned the use of a 15k cathode resistor in the PI, but i never had one noted there or used one that i recall. It was 56k, which was on the original schematic, then on cbarrow7625's advice i changed it to what i saw on a marshall schematic which was a 10k. Right now i have a 470ohm from the cathode to the same 10k to ground. that 470 WAS a 820, but i again changed it to reflect a marshall.

Anyways, it's not bad right now but i'm still desperately trying to find the source of that mid sound thats no where near as smooth and transparent as any of the marshalls i've had. NFB helped a tiny bit, but it also made the overall tone suffer. And i'm sure i wired it up exactly like the marshall i copied. (actually i tried 2 different NFB circuits i copied from 2 different marshalls)

[Bruce / Mission Amps]

See if you can spot the changes I made to your schem.
Try something more like this as a start and give me some plate voltages... it makes a difference.
I still think it will still have too much gain but from reading your posts, it sounds like that is what you want and the cleanish usable sound isn't that important.

[daz]

Geez man, you live and breathe this stuff dontcha?! i know you probably loved editing that, huh? Well, yeah....there are a ton of changes. what i wanna know tho is, what do you think that will sound like? i ask because i know you are into fenderish circuits, and i'm 110% marshall all the way. I lust after certain aspects of marshall tone and thats my #1 objective.

As to my gain fetish, you won't believe my answer. But the fact is that i like a lot of gain, but not so much for the high gain tones themselves. For whatever reason i have always loved the sound of a high gain channel turned only part way up. Say 1:00-2:00 on the pre, then turning the guitar down to get semi cleans with a single coil, and i use guitars with a hot hum in the bridge. So i can go to almost clean to cranking distortion without pedals. And the key for me is that for some reason i much prefer the sound of a high gain circuit turned down than one which i have to turn the pre to 10 to get the amount of gain i want. that may not make sense, but a high gain amp set for 1/2 the distortion it can generate sounds much different to me than an amp that with the pre gain on full generates the same amount. So while i don't need that much gain, amps that have that much just seem to have the type of tone i like even at lower gain levels. It works for me too, i assure you. And as far as turning it up, if i DO use it for gigs, which i rarely do anymore anyways, i would tuen the gain way down when the master goes up. so it should work even then. After all, the preamp only generates a ton of gain if you twist the knobs that way.

In any case, i may try some of those tweaks, but if i decide to do them all i would have to make another schematic of where it is now so mi don't lose that. It's got some tweaking to be done before i'd be happy, but it's close emough to be a good starting point. so i don't want to lose that. Took me all day to get it to that point from the way i left it yesterday. thanks Bruce.