So this has been a project I've been working on for a couple of months now, with the excellent help of some nice folks over at a different audio forum. I'm in possession of two of these amps and I've been working my way through the first one and finalizing my schematic before working on the 2nd amplifier. I'm pretty close to wrapping it all up, but there are some finer points that I really need some help with.
Attached is a before/after schematic. One completely stock, and the modified schematic which shows all my changes marked by a red dot.
The infinite possibilities in preamp design are a little overwhelming, and I find myself constantly questioning certain decisions regarding cathode, plate resistors, and bypass caps etc.
Here is a current list of planned changes: 1. Reevaluate bypass cap values. V1A & V2Ato 1.5uF(?). How to bypass caps interact later down the chain? Does is make sense to have larger bypass caps earlier in the chain and smaller ones later, or vice versa?
2. Similarly, reevaluate plate/cathode resistor values. V1A & V2A cathode to 2.7k? I had previously switched V2A's plate to 220k. How is 1.5k/100k different to 3.3k/220k from a practical, audible perspective?
3. I already increased R35 from 2.7k to 5.4k. What are typical negative feedback values? I have also read at Rob Robinette that running a 100-500pf capacitor around R35 can help me reduce ice pick highs, which are currently a bit of an issue. How does this work? What kind of roll-off should I expect from say, a 500pf cap there? Should I be looking at increasing R35 more dramatically?
4. I feel like I really need to do something with the bass and treble controls... they are not exactly practical for a guitar in their current form. The pots (2M) need to stay the same, or else I'll be drilling new holes for a power switch, and additional pots elsewhere. They are a stacked 2M/2M/switch pot. Any good practical mods to the capacitor values that will make these more useful?
5. I was going to put a bright cap on V1B a la Marshall Plexi. But I still need to tame my ice pick treble first. .005 is the standard value for a cap. I'm, assuming that should be just fine?
6&7. Is there any harm in increasing the value of C11 to say 500pf? Could that help tame treble? I have a .022 cap on my master volume in order to retain bias, what is the effect of changing this value between .01 and .05? I've had two different recommendations on that number.
Kind regards, Nat.
Attached is a before/after schematic. One completely stock, and the modified schematic which shows all my changes marked by a red dot.
The infinite possibilities in preamp design are a little overwhelming, and I find myself constantly questioning certain decisions regarding cathode, plate resistors, and bypass caps etc.
Here is a current list of planned changes: 1. Reevaluate bypass cap values. V1A & V2Ato 1.5uF(?). How to bypass caps interact later down the chain? Does is make sense to have larger bypass caps earlier in the chain and smaller ones later, or vice versa?
2. Similarly, reevaluate plate/cathode resistor values. V1A & V2A cathode to 2.7k? I had previously switched V2A's plate to 220k. How is 1.5k/100k different to 3.3k/220k from a practical, audible perspective?
3. I already increased R35 from 2.7k to 5.4k. What are typical negative feedback values? I have also read at Rob Robinette that running a 100-500pf capacitor around R35 can help me reduce ice pick highs, which are currently a bit of an issue. How does this work? What kind of roll-off should I expect from say, a 500pf cap there? Should I be looking at increasing R35 more dramatically?
4. I feel like I really need to do something with the bass and treble controls... they are not exactly practical for a guitar in their current form. The pots (2M) need to stay the same, or else I'll be drilling new holes for a power switch, and additional pots elsewhere. They are a stacked 2M/2M/switch pot. Any good practical mods to the capacitor values that will make these more useful?
5. I was going to put a bright cap on V1B a la Marshall Plexi. But I still need to tame my ice pick treble first. .005 is the standard value for a cap. I'm, assuming that should be just fine?
6&7. Is there any harm in increasing the value of C11 to say 500pf? Could that help tame treble? I have a .022 cap on my master volume in order to retain bias, what is the effect of changing this value between .01 and .05? I've had two different recommendations on that number.
Kind regards, Nat.
Attachments
68k is used with Fender's nifty voltage divider. 33k gives a bit more highs. If you use cathode bypass with extended range up front it can reduce hum. But if you plan on overdriving the next stage a reduced capacitor value makes sense. Fended values are 1.5k/100k, Ampeg 2.2k/220k (I think, cleaner I have been told), if you want earlier clipping try Marshall's 820R/100k 2.7k/100k. Depends on what you are trying to achieve.
C11, use ears after all mods done. Could just leave the 0.047uF's in for now and evaluate when completed. Tonestack calculator, use James tab for an idea on cap values. TSC C20, C6, 50uF, just leave it in, the added low bypass won't hurt as compared to 25uF. 0.022uF on cathodyne input negates the 1.0M when you have the master turned uo, just when you may need the grid stopper. I think this is dumb.
R22,23 looks like you wanted to change from 470k, possibly wanted 220k? That is what I would do. Grid stoppers of 220k on the 6L6's looks mighty high in value. You may loose some highs.
R24, dropping from 250R to 200R, leave it where it is, why do you want to run class A? Put stoppers on each output screen, 470R or 1k. Even Fender went to them eventually. R12/C8, might want to try jumpering across these two, don't see this being of value. Nothing set in stone, I would evaluate as a complete amp.
C11, use ears after all mods done. Could just leave the 0.047uF's in for now and evaluate when completed. Tonestack calculator, use James tab for an idea on cap values. TSC C20, C6, 50uF, just leave it in, the added low bypass won't hurt as compared to 25uF. 0.022uF on cathodyne input negates the 1.0M when you have the master turned uo, just when you may need the grid stopper. I think this is dumb.
R22,23 looks like you wanted to change from 470k, possibly wanted 220k? That is what I would do. Grid stoppers of 220k on the 6L6's looks mighty high in value. You may loose some highs.
R24, dropping from 250R to 200R, leave it where it is, why do you want to run class A? Put stoppers on each output screen, 470R or 1k. Even Fender went to them eventually. R12/C8, might want to try jumpering across these two, don't see this being of value. Nothing set in stone, I would evaluate as a complete amp.
Thanks for the reply, I think you are misunderstanding though... the schematic marked has all the changes I have currently made! Here's a video, it's sounding pretty good. I'm just trying to iron it out a bit at this point: YouTube
What are you referring to here? The only 33k I can see is R14.
I have been informed C11 is to prevent oscillation. My question is more if it has the potential to cut audible high frequencies also, and whether or not that's a good place to try and tame the ice-picks?
Thanks.
What are you referring to here? The only 33k I can see is R14.
I'm not sure if you noticed, but I took some inspiration from a Marshall Plexi with jumpered channels and similar preamp values. I did have V2A running 820/100K at one point. I may go back. Regardless, better to have smaller values later in the chain? V2A and possibly V2B were definitely getting changed to smaller values in my next parts order.
I originally went 820R on V1A and 2.7k on V1B but then I realized I have 220K plates from the factory, so jumped V1A back up a notch. Same goes for V2A. Any reason NOT to have 820R/220K? I assume it'll just clip to all hell and not sound very good.
I have been informed C11 is to prevent oscillation. My question is more if it has the potential to cut audible high frequencies also, and whether or not that's a good place to try and tame the ice-picks?
All coupling caps on the exception of C3 have been shotgunned to .022's. The bass was insane at first. It took a lot to tame it down.
I have it installed already, what values do I need to set for Zsrc, R4, and R5 in TSC for it to bear similarity to my circuit? I can't figure out how to get that program to help me.
They are both already 25uF. The old paper caps had drifted. Input 1 is still muddy and woolly sounding. I was going to at least swap C20 to a (probably) 1.5uF in order to clear up the rest of the signal path.
Now... I have been informed by multiple people that I need a small cap isolating the master from pin 9 of the PI in order to maintain bias. One person said .01, another, .05. I have also seen schematics saying it is unnecessary. Additionally, I need my grid stopper there. Nobody could give me a clear answer on exactly how to implement BOTH... so I ran them in parallel and it seemed to function nicely. If I don't need the cap I will remove it... but if it doesn't work properly... how do I wire in this bias maintaining cap?
R22, and R23 were actually 470k from the factory. See the stock schematic. I changed them to 220k, was advised against it, and edited it back. The amp is sounding pretty good with 220K grid stoppers and I'm still getting some ice-pick highs playing certain notes and bends on the guitar. That's part of what I'm trying to resolve. I have a feeling the Bass and Treble controls really need some work but don't know where to start.
I wasn't trying to run class A so much as run the tubes hotter, and get some additional headroom/volume. 95% of the amps stock circuitry struck me as being WELL on the safe side, and designed for clear flat frequency response with no distortion. It's a PA amp. I'm trying to make it sound good for guitar.
This seems to be a safety/protection feature... any effect on tone? I'll look into this!
I've been informed elsewhere that removing C8 might help with some of my treble woes... What is the function of R12?
Thanks.
The 68k resistor value was selected by Fender for their two hole input. You don't have a 33k but rather than a 68k on your inputs. When the 68k's are paralleled up in use you get a 33k, just saying 68k is not a magic value.
Again, it depends on what kind of amp you want to build, clean or dirty. As long as you have enough unclipped gain to make it through the amp to drive the output tubes fully. C11 is not normal in amps, might be in for stability reasons due to the layout.
38k will be fine for the treble. I would simulate the bass with it also and short out the resistor-capacitor in front of it.
With the pot in the position it is in you do not need isolation from the splitter circuit. If you had the wiper going to the grid and the bottom leg of the pot going to ground then you would have had to have the cap on the wiper.
The 470k value does not drain the electrons well enough if you plan on clipping the output to a great extent. It will work but the actual bias voltage on the grids can drift. The value of the grid resistor and the bias current through the tubes are related. 70% is not a magic number, as far as I can figure they used that as a target and the datasheet recommends a resistor value to use. People have bent the recommended values over the years. If you find the amp is eating output tubes in practice this would be a place to look at, if not don't worry about it. Oh wait, just realized it is a cathode bias, they spec a higher value, the 470k might be ok. On the grid stoppers, if it sounds good it is good.
The lower the resistor value the higher the bias current and the lower your headroom.
The amp will not be as stiff, it will sag a little on full output. It is a trade off to protect the screens from melting when the output section is overdriven.
The resistor reduces the signal level, the capacitor bleeds some treble around it. Without the need to have the increased treble that the cap provides the resistor is just a waste of signal level. Since you have icepick I would think scrapping the two would be a good idea.
Again, it depends on what kind of amp you want to build, clean or dirty. As long as you have enough unclipped gain to make it through the amp to drive the output tubes fully. C11 is not normal in amps, might be in for stability reasons due to the layout.
38k will be fine for the treble. I would simulate the bass with it also and short out the resistor-capacitor in front of it.
With the pot in the position it is in you do not need isolation from the splitter circuit. If you had the wiper going to the grid and the bottom leg of the pot going to ground then you would have had to have the cap on the wiper.
The 470k value does not drain the electrons well enough if you plan on clipping the output to a great extent. It will work but the actual bias voltage on the grids can drift. The value of the grid resistor and the bias current through the tubes are related. 70% is not a magic number, as far as I can figure they used that as a target and the datasheet recommends a resistor value to use. People have bent the recommended values over the years. If you find the amp is eating output tubes in practice this would be a place to look at, if not don't worry about it. Oh wait, just realized it is a cathode bias, they spec a higher value, the 470k might be ok. On the grid stoppers, if it sounds good it is good.
The lower the resistor value the higher the bias current and the lower your headroom.
The amp will not be as stiff, it will sag a little on full output. It is a trade off to protect the screens from melting when the output section is overdriven.
The resistor reduces the signal level, the capacitor bleeds some treble around it. Without the need to have the increased treble that the cap provides the resistor is just a waste of signal level. Since you have icepick I would think scrapping the two would be a good idea.
Replace R1 & R25 with 1M0.
Fit a 2k2 with a 2u2 16v cap in parallel between pin 3 and ground and again for pin 8 and ground both on V1.
Remove R29 and discard R30 etc.
Place a 220R pot in place of R15. Place a100pF cap in parallel with R35.
That will then become a presence control, giving you that Marshall sound.
Parallel R36 with 47R 5W W/W. To give more headroom. Leave the output valve bias as it is.
If you intend to drive the ouput stage to clipping, fit a reverse bias 1N4007 diode between the final anode and ground on both 6L6s. That will save the valves from arcing internally.
Fit a 2k2 with a 2u2 16v cap in parallel between pin 3 and ground and again for pin 8 and ground both on V1.
Remove R29 and discard R30 etc.
Place a 220R pot in place of R15. Place a100pF cap in parallel with R35.
That will then become a presence control, giving you that Marshall sound.
Parallel R36 with 47R 5W W/W. To give more headroom. Leave the output valve bias as it is.
If you intend to drive the ouput stage to clipping, fit a reverse bias 1N4007 diode between the final anode and ground on both 6L6s. That will save the valves from arcing internally.
Last edited:
That's easy then, retrofit the 220R pot.
Potentiometers | RS Components
Cheaper non commercial makes are available from most suppliers including Fleabay.
Potentiometers | RS Components
Cheaper non commercial makes are available from most suppliers including Fleabay.
Outside North America, pots tend to have values taken from the same E6 series as standard resistors. So you'll typically find pots with values like 100 ohms, 220 ohms, 470 ohms, 1k, and then ten times those values (2.2k, 4.7k, 10k), and so on.
In North America, for some reason that has been lost to history, pots have different standard values than resistors. So here we are more likely to find pot values of 100 ohm, 200 ohm, 500 ohm, 1k, 2k, 5k, 10k, and so on.
-Gnobuddy
Resistors, yes. Pots, no! (At least in my 30-odd years in North America.)
Elsewhere in the world pot values are also taken from the E3 range of resistance values (1, 2.2, 4.7). But in North America we're more likely to find 25k pots than 22k pots.
Considering their purpose, I doubt there is any functional difference. I also wouldn't be surprised if the exact same pot is labelled "25k" for the US market, and "22k" for sales to the rest of the world.
You just reminded me that in North America, power resistors also do not always follow the "E" series standard values.
In BC, my car speedometer reads in kilometres per hour and I buy gasoline in litres. But the air pressure in my car's tyres is specified in pounds per square inch, and the power output of the engine is specified in pre-industrial units coined in the eighteenth century, and based on the ability of a once-popular beast of burden.
Our units are a horrible mess, and it's unlikely that will ever get better.
-Gnobuddy
There is actually logic to it; the idea was to have successive resistance values roughly 20% apart, i.e. they are a geometric series with a common ratio of 1.2. In the era of +/- 10% resistors, this kept successive values from overlapping, even if both were at the extremes of their tolerance.
If you divide each successive value in the E12 series by the value just below it, the ratio between them, rounded to one decimal place, is 1.2 in every case except one: the value "1.5" should really have been 1.4. (See attached screenshot of a Libre Office Calc spreadsheet I threw together.)
It's not perfect (the ratio between many adjacent values is 1.22, others as low as 1.19), but it's not totally ridiculous, either. There is a reason for the imperfections: if you used an exact common ratio of 1.2, you don't get values that fit neatly into a decade - there is no power of 1.2 that gives you 10! So whoever created the original series tweaked some of the values just enough so you could repeat them as decade multiples.
There's nothing wrong with the 1 - 2 - 5 sequence at all (it works very well for money). The trouble is that it's hard to expand to include more values. Once you get to better than +/- 50% tolerance, 1 - 2 - 5 becomes inadequate. Then where do you go?
-Gnobuddy
Attachments
If I'm not imagining things, I used to see pots in all the E6 values when I was a kid - I seem to remember I could buy a 15k or 33k pot then.
In this era of the ubiquitous microcontroller, pots are going the way of the dodo. We seem to be down to just three values per decade now. Anti-log pots are virtually extinct, stereo pots are hanging on by the skin of their teeth, and even plain old 10% taper audio (log) pots are getting scarcer.
-Gnobuddy
In this era of the ubiquitous microcontroller, pots are going the way of the dodo. We seem to be down to just three values per decade now. Anti-log pots are virtually extinct, stereo pots are hanging on by the skin of their teeth, and even plain old 10% taper audio (log) pots are getting scarcer.
-Gnobuddy
It's a pity that you need to stick with these odd potentiometer values and the baxandallish tone control circuitry. No chance to get the power switch to an unconspicuous location, maybe at the backplate? This would allow for other pot values and for a Fender-like tone stack (with fixed mids, of course) and hence achieve 'the' sound.
Best regards!
Best regards!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.