Greg Robinson's mini tube amp

[Randy Roberts]

Greg,

I've not understood a single word of this entire thing, and yet it's incredibly captivating.

Having no electronics backround whatsoever, your whole thread makes me feel like I've discovered a time warp and I'm eavesdropping on the Manhattan project or something. What cool terminology, it's like you're writing the dialog for Spock for the next Star Trek movie.

[Greg Robinson]

Thanks Steve.
I think I should be able to stay under budget, I've got something more creative than expensive in mind for finishing.

And for you and Randy, both seeming a little overwhelmed, I'll take that for the compliment you both intended it as, but this is really quite a simple amp, and like any other instrument design/construction related skill, it just takes time and practice to learn, I'm sure either of you would be able to make sense of it if you had stared at as many schematics as I have.

If anyone want's me to explain anything in more detail, or to give a bit of theory of operation, please let me know! I'm happy to share.

I've been doing some testing, and found that the noise when the boost is engaged (which isn't too offensive really) is really all down to power supply ripple effecting the LND150, with it's ridiculous amounts of gain (400+ need to measure it properly). So I might regulate the entire high voltage supply. I don't have any high voltage transistors or darlington pairs on hand, so that might come later.

[Jason Rodgers]

If anyone want's me to explain anything in more detail, or to give a bit of theory of operation, please let me know! I'm happy to share.

I'm with Randy: there's not enough room here to explain what I don't know... "So... what's a capacitor do?" You just keep up the good work, we'll smile and nod.

[Bryan Bear]

Like others, I have no idea what you are talking about but I do know awesomeness when I see it!

[Samuel Hartpence]

For a finish Idea, my brother-in-law makes sculptures with all sorts of material including metals, wood, stone...anyways...he uses various acids to oxidize metals in different ways then just hits it with a clear-coat sealer once he gets the looks he was going for. I think for this type of build, some sort of distressed finish could be economical and fitting for a "reclaimed radio-shack sneeze" amp.

[Greg Robinson]

Samuel, you might just be a mind reader.

Thanks Bryan and Jason!

[Samuel Hartpence]

I apologize if you intended that to be a surprize. I'm really excited to hear this in action.

[Greg Robinson]

No worries Samuel, I'm actually having a little trouble with the idea. As the enclosure is an aluminium alloy, it's not easy to distress or artificially age. I'd prefer to get some sort of realistic patina rather than just paint something. I guess some sort of acid will work. Any suggestions?

[Samuel Hartpence]

Well... Corroded aluminum looks like....Aluminum

I think instead of attempting to get a corroded look, you'll have to just go for an etched look, or a distressed look. You can use a salt/vinegar slurry and you'll get a matt white residue. You could also use gun aluminum blue (I'm not sure how easy that stuff is to get down under, but it's cheap (<$15) here in Wyoming), or a different strengths of hydrochloric acid. All the while, you use varying roughness scratch pads to knock it back and do it again until you get the look you want, then you can leave it as is, or hit it with a coat of clear if you want to preserve the look.

My brother-in-law said that he's killed two birds with one stone by throwing aluminum parts in a concrete mixer, letting it tumble with the sand/concrete, then he pulls it out while pouring the patio, then throws is back in the tumbler with just sand and water, but this would require disassembly (but if you need some flatwork done it may be worth it)!

[Greg Robinson]

Thanks Samuel.
Yes, of course aluminium oxide (alumina) is the surface we see, this happens almost immediately on exposure to the atmosphere. I'm just hoping to get some sort of nice patina, like an old cast aluminium pot or something. Some sort of etching and discolouration or something is exactly what I'm after.
I'm not sure what alloy the enclosure is, I was hoping to get some interestingly coloured oxides.
I've tried some scuffing and treating with acetic acid (white vinegar), but nothing interesting has happened so far. Guess I'll have to try something stronger!
Thanks for the suggestions!

[Samuel Hartpence]

You could try some liquid drano or something like that as well. If you use the foaming stuff, that may create some depth to the finish.

[Beate Ritzert]

Greg, please let me add a few suggestions for the schematics:

- increase the anode resistor of the first valve stage close to its limit, use, say, 420 kOhms in order to maximize clean gain of that stage.
- decrease it a little bit into the range of 68-82 kOhms in the 2nd stage.
- add a switchable 2.2 kOhm resistor in parallell to the cathode resistor of valve stage 2, yielding a cathode resistor of 820 Ohms as found in some Marshall designs.

These mods will enable You to overdrive valve stage 2 to at least soft crunch.
I did all that in my Epiphone Valve Junior and can overdrive its 2nd stage evene despite the loss of 7 dB of the PI-tone filter after stage 1.

I would also suggest to reduce the values of the coupling capacitors to a lower corner frequency of 70 Hz. You do not need more, even if You play bass over such a small amp (and in the latter case can even run it with more power in the musically important frequency range)

[Greg Robinson]

Hi Beate,

Thanks for the suggestions, but I don't think I'll be making any further changes to the circuit other than maybe the power supply.

increase the anode resistor of the first valve stage close to its limit, use, say, 420 kOhms in order to maximize clean gain of that stage.

As the first stage is fully bypassed, centre biased, and there is no interstage attenuation between the first and second gain stages, it already provides plenty of gain to overdrive the second stage. Increasing the anode resistor value would only marginally increase the (voltage) gain from around 70 to about 80. Also, it would limit the current through the tube, and I find that this usually does not sound the best.

decrease it a little bit into the range of 68-82 kOhms in the 2nd stage.

I'm not sure what your aim is with this suggestion, it would not significantly alter the gain or other operation of this stage. 100k is just about a perfect anode load for a 12AX7/ECC83, there's rarely good reason to deviate from this value.

add a switchable 2.2 kOhm resistor in parallell to the cathode resistor of valve stage 2, yielding a cathode resistor of 820 Ohms as found in some Marshall designs.

This makes the previous suggestion make a little more sense, in that it would more closely centre-hot bias the tube again, but there wouldn't be much difference between the two switchable values - around 0.3V bias voltage difference. I'm not convinced that there would be a significant change in tone.

I did all that in my Epiphone Valve Junior and can overdrive its 2nd stage evene despite the loss of 7 dB of the PI-tone filter after stage 1.

Just by the way, the Roy Bean tone stack I've used here is very low-loss - around 3-4dB. And with it placed after the second stage anyway, trying to wring more out of the first two stages won't achieve much - plus, I've got the LND150 boost!

I would also suggest to reduce the values of the coupling capacitors to a lower corner frequency of 70 Hz. You do not need more, even if You play bass over such a small amp (and in the latter case can even run it with more power in the musically important frequency range)

Well, the -3dB point for the interstage coupling is around 150Hz, and the amp is already almost too bass heavy, in fact I was considering lowering the value of the coupling caps to tighten it up a bit (I assume that above you meant increase their value, not reduce, as that would raise the corner frequency). Keeping the value of coupling cap low is also important in combating blocking distortion, which is especially important when you're trying to squeeze this much gain out of only two stages. Bass frequencies don't need to be overdriven as much to sound pleasing anyway.

As it stands now, the amp has a great bluesy grind to it, think Clapton Beano tones - even when only pushed with Strat single coils. If I want to get my metal-on, the boost is more than enough even on about 2! The Roy Bean stack takes some getting used to, but is more flexible than the common FMV you see in nearly everything - I like it.

Anyway, nice to chat with someone else who knows what I'm talking about, thanks Beate!

[Andy Birko]

That thing is so super cool Greg! Although the piece of paper I got from the university says I'm an electrical engineer, I'm continually amazed at folks who have learned things like tube amps to the point where they can look at a schematic and truly understand what's going on there. There's no way I could sift through a junk box of tubes and transformers and come up with a working amp and I'm also certain that a large percentage of the EE's out there couldn't do that either.

Excellent work!

[Greg Robinson]

Thanks Andy!
I've never studied electronics formally (and tubes/valves aren't something you're likely to study in much detail these days unless you're into high powered radio transmission or a few other eclectic fields anyway), but I've worked around another person who's used them in high quality studio equipment, and from this I've taken an interest and persued my own path.
I have quite a few peers that have studied/are studying electronic engineering (I'm still quite young remember, only 26), and for the most part they can follow a datasheet for a switchmode controller chip and come out the other side with a power supply, but most can't make heads or tails of an amplifier or have any idea where to start with component level repair. They have very different skills, but I'm very happy with what I've achieved/am achieving. I have no particular interest in most of the applications of EE (seemingly all SMPS and software), and feel I've got more practical skills than most of those peers who have studied, although it would be nice to have a piece of paper!

I've been in the process of developing a few amp designs that I eventually hope to release commercially - they're quite a bit more complex and involved. It was a nice break to design something like this that was so much simpler!

[Beate Ritzert]

Hi Greg,

despite Your belief You will hear differences in voicing die to changes like those i suggested to try out.
Reason is the nonlinearity of valves. By reducing the kathode resistor, for example, the signal will be amplified slightly more asymmetrical and thus even order harmonics will be increased.

Analogously the value of the anode resistor has a really noticeable effect on nonlinear distortion - it increases if You lower the anode resistor and it decreases if You raise its value. Because the effective (dynamic) anode resistor is always lower than the static one the gain will also be lower - You might expect something between 55-60 für Ra=100k.

Reducing its value will lead to a somewhat more "compressed" tone which is, admittedly, more suitable for bass amps.

With the 1.5k / 100k default (actually a tradition; the optimum for linearity is at somewhat larger values; the german tradition has been 2.2k / 100k) You will always obtain a "fenderish" tone. This is, of course, not the worst to target, but there are other possibilities, and these are applied in actual practise to control the voicing.

best

Beate

[Greg Robinson]

Hi Beate,

Not saying there won't be any differences, I just don't think that they'll make a very significant change - or change in the direction that I'm aiming for.

By reducing the kathode resistor, for example, the signal will be amplified slightly more asymmetrical and thus even order harmonics will be increased.

Yes, but as I mentioned, with your recommended change in both cathode and anode resistors, the bias point barely shifts, about 0.25V hotter, so yes, slightly more assemetrical, but this stage is already heavily overdriven, so it will not effect the character of overdrive much except at the beginning of breakup, and is also at the expense of output swing to overdrive the power stage - which I think is more important to the character of this amp.

Analogously the value of the anode resistor has a really noticeable effect on nonlinear distortion - it increases if You lower the anode resistor and it decreases if You raise its value. Because the effective (dynamic) anode resistor is always lower than the static one the gain will also be lower - You might expect something between 55-60 für Ra=100k.

Reducing its value will lead to a somewhat more "compressed" tone which is, admittedly, more suitable for bass amps.

Looking at the loadlines, and having compared circuits with very similar values in both auditory and oscillioscope tests, I think you're perhaps ascribing more differences than are actually present. With multiple cascaded stages, these non-linearities accumulate, and more drastic changes to operating points are certainly more audible - amps like the Soldano SLO and all it's derivitives are prime examples.
Case in point - one of the projects that I'm working on is a "high gain" amp intended for modern metal sounds, but using 12AU7's/ECC82's instead of the more traditional 12AX7's/ECC83's seen in most amps of this type. The tube is lower gain mu=17 compared to 12AX7 mu=100, and slightly less linear. By exploiting these non-linearities and using non-linear circuits like badly designed (by textbook standards anyway) DC coupled cathode followers, I've been able to achieve tones that sound just as "gainy" and distorted, without using nearly so much brute-force gain to achieve cutoff/grid-current limiting (overdrive) and the inherent noise that comes with that technique.

With the 1.5k / 100k default (actually a tradition; the optimum for linearity is at somewhat larger values; the german tradition has been 2.2k / 100k) You will always obtain a "fenderish" tone. This is, of course, not the worst to target, but there are other possibilities, and these are applied in actual practise to control the voicing.

I'm aware that the values I have chosen for cathode and anode are "traditional", but I'm not trying to simply be derivative. I chose them specifically for a balance between assymettric clipping and output voltage swing. I have, in fact played with the voicing, and as it turns out, I'm happiest with the values I originally chose. About the last thing I would describe this amp as sounding like is "Fenderish".

I'm enjoying this discussion! Thanks for your input Beate!

[gene downs]

Hey Greg.

Been away from the mimf too long. I didn't even know this contest was going on!

Great project. I just discovered this thread and haven't read it yet, but I wonder why you went with such large cathode bypass caps. I haven't built an amp with a 12au7 as a power amp, but it seems that the usual rules would apply. I would think the amp would be extremely bassy. For my own personal taste, I never go larger than 5mf and usually go with a 1 or 2mf. I hate a muddy amp. You'll get a lot of gain with a 50mf, but the little 'uns sound so much sweeter. To each his own, in any case. Nice job!

[Greg Robinson]

Well, about time I checked in here!
I would call the project finished - actually, I had it finished before the deadline, but paying work has gotten in the way of posting final photo's and soundclips, I'll get on that ASAP.

Thanks Gene,
Remember, fully bypassed cathodes don't accentuate bass, they provide equal gain over the whole frequency range. Using partial bypassing accentuates higher frequencies or midrange, depending on the capacitor/resistor combination. The formula is:
f = 1 / ( 2*pi*Rk*Ck )
Where:
f is high pass half boost (-3dB) frequency
pi is 3.14...
Rk is the cathode resistor
Ck is the cathode capacitor

So with my values of 1k5 and 50uF, my half boost point is 2.12Hz, about three decades below the audible range, so fully decoupled for audio purposes. Remember I was just using parts I had lying around too, so I didn't select this value for anything much more than convenience.

However, the bass response can also be tamed with the interstage coupling capacitors, I used 22nF.
The math is a little more involved if you want to do it properly (you have to calculate the output impedance of the driving stage), but as I've used very standard values, you can make some easy assumptions. The formula is otherwise the same:
f = 1 / ( 2*pi*Co ( Zo+Rg ) )
Where
f is high pass half boost (-3dB) point
pi is 3.14...
Co is the output coupling capacitor
Zo is the output impedance of the driving stage (assumed to be typical value of 39k here)
Rg is the grid resistor (load resistance) of the following stage

With my values of 22n and 1meg, the high pass frequency is 6.96Hz, still well below the audible range.

In a higher gain amp, you would probably want to limit the bass response quite a lot more in order to prevent blocking distortion, but here the levels of gain are really only just enough to overdrive successive stages (discounting the LND150 boost stage), and the Roy Bean tone stack allows the bass to be rolled off significantly, so while you certainly can set the amp to be muddy if you want to, you can also make it almost overly bright by rolling the tilt control the opposite direction.


Ok, guess I'd best go get some photos and soundclips organised!

[Samuel Hartpence]

Ok, guess I'd best go get some photos and soundclips organised!

Yes Please! I'm excited to hear it and see what you did for the patina.