From: Kiwi [flying.kiwi@paradise.net.nz] Sent: Tuesday, September 19, 2000 6:11 To: chrish@ax84.com Subject: RE: Moonlight philosophy May I post this discussion on the Moonlight Page? >> Certainly. As this is the closest I'm likely to get in the next few months to writing a tech page for it, I've rephrased a few bits so ( I hope) they say more closely what I intended. Also added a bit about one possible effect of cath followers between PI and O/P grids ( which I would have put in the original note to you - if I'd thought of it! ) If you want, you *can* put up the original discussion verbatim - but I think this revised one is more useful for people reading it. It reads alright to me, but if you think something is a typo, give me a shout. > -----Original Message----- > From: Kiwi [mailto:flying.kiwi@paradise.net.nz] > Sent: Monday, June 19, 2000 3:30 AM > To: Chris Hurley > Subject: RE: Moonlight philosophy > > > >.....[or if I might] go with a different preamp anyway... I'd be > >interested to hear your opinion on the subject. Some of the > answer >might > be contained in why you chose the preamp that you did... Ok.... I wanted to be able to overdrive the output stage in a fairly small and simple amp, and for this to be the primary source of compression and/or distortion. It's been done before with small s/ended beam tubes or pentodes, but I was curious to see how it would go with a small *push-pull*triode* output stage. In the real amp, some distortion is contributed by the PI, and probably some more from the o/p grids loading down the PI when they are pushed hard enough to start drawing grid current, and their input impedance drops sharply. (BTW, inserting cathode followers between PI and output grids would stop overdriven output grids loading down the PI. With this factor removed, distortion form the PI itself might appear more significant) I hoped the overdriven sound would be reasonably like an old Marshall 50. I'm not sure how close I came to the last bit; but I got it to the stage where it sounded nice to me, then ran out of project time - and put up the schem in public domain for other people to play with. About a year later, the idea took off...... Don't assume this amp was designed from theory. It was a thrown-together idea which sounded promising, and then had a lot of time spent tweaking components, playing guitar, listening, and watching a scope to get an idea of what was going on. I did check that the 6/12SL7 was not running significantly over it's data-book plate dissipation, but just about everything else was fine-tuned by ear. *Your* ears may prefer this amp tweaked differently - or even extensively modified: that's fine with me. Originally, I had one triode preamp stage and an even simpler treble tone control - which didn't have quite enough gain to drive the later stages from some guitar pickups. As the intention was a tiny tube amp which *could* be pushed to the wall, I considered other options for preamp gain. Two cascaded triode stages had more gain than I needed - so much that in my rats' nest layout it would go unstable at higher gain settings. (I don't know why I never tried a 12AU7 here - but anyway, I didn't.) My absolute favourite low-power amp uses an (rf) pentode as the first stage. It's very difficult to find samples of 6SH7GT which (a) sound right, and (b) are not too microphonic for a guitar amp. However, this probably gave me the idea to try a pentode in the Moonlight. It turned out that every old EF86 in my box rang like a bell ( they used to be renowned for it.) I happened to have a few EF184's. They were available quite cheaply; and none of my samples were microphonic....which looked promising. ...Looked at the curves, decided on 1 to 2mA and about 100k plate load, and then tweaked screen resistor against cathode resistor until it had something of the sound quality I liked in my other pentode input amp. (Which turned out to be a low screen voltage and a low value cathode resistor - something I had found before, though I did this one quite by ear.) You are dealing with three or more variables here, so it's a 'fuzzy logic' problem. I knew the sound in my head. At moderate guitar input levels, it's a kind of clarity, without being significantly more toppy - There's a point in adjusting tube operating conditions when the mud clears - but you have to listen really closely to nail it. I've found the same effect with other pentodes; haven't tried it with EF86 - It may occur, but probably at some different combination of plate load and screen and cathode voltages. [Even for the same plate or cathode current, different combinations of screen voltage and cathode resistor sound subtly different, because you are running on different curves.] This is not the highest stage gain you can get, but I hit a combination which sounded good and had as much gain as I was looking for, and it sounded really nice with no tone control in circuit at all ! However guitar pickups and speakers and room acoustics do vary, so I wanted at least *some* treble adjustment up and down a bit. I couldn't find a tone stack design which could get back to that *no-EQ* sound at some point within its range; and I didn't need a *lot* of adjustment. So from straight theory, I worked out the EQ shown in the schem. In its electrically 'central' position it is essentially a balanced bridge, with no voltage across the cap.... It is very nearly phase-flat as well as amplitude flat ( and might be of passing interest to some HiFi tube amp folks) This EQ *did* retain the 'no-eq' tone in its central position, and I simply tweaked the cap value until the EQ sounded good when boosting or cutting treble. As I already had the amp sounding about right without EQ, I chose the 'bridge-network' component values for a relatively small range of tone variation: in this way, it also has relatively small through-losses. And happily, it turned out that after including the (15dB ?) mid-position loss of the tone control, I still had enough gain in the chain..... Nu......that had the tone balance that pleased me. And I always wanted a small 'back-porch' amp which would just sound right, with as few technical distractions (a.k.a. "knobs"), as possible. So that's where I stopped. Someone else may want more tone range. Or the characteristic Fender mid-notch. That's fine....but you'll likely have more through-losses in a fender stack, and you'll likely need more gain somewhere in the preamp to make this up. ( If you also want to drive a self-PI 6SN7 and dump the PI, then you'll *very* likely need more gain) With a 3-knob tone stack, two triode stages may be just about right for gain. But (though I'd enjoy it) I don't have time to try this for you. You'll have to wing it yourself. I think you could probably lift one or other of the PI preamps entire, and paste it in. ....Which is why I have also released a longer tagboard design in Mayura format (MoonMODboard) with extra tags at the preamp end. Should be enough tags there for a two-triode preamp. ** The intent of the Moonlight amp was nice *Output*-stage distortion and compression, at domestic sound levels....The combination of PI and output stage, also the low plate load on the 6SN7, plus the frequency response and magnetic curve of that particular OT, and possibly also the sag time-constants built into the PS. ** .....IMHO that's what makes it a Moonlight. If you want to make one with a fender stack and more preamp gain, then go for it! I can't predict how the OT will respond to a signal with the bass EQ turned up. I don't think you'll break anything, but I just don't know how it'll sound - so you'll have to try it. At worst, you might have to use some smaller coupling caps somewhere along the line. That's about all the ....er....philosophy....behind the design (that I can remember) It *was* two years ago So go and have some fun with it. Cheers, Stephen. (update from Stephen: Sept. 2k) When I can find some spare workshop time (Ha!) the eventual plan is to rebuild it on a proper chassis, in a nice box, possibly with uprated transformers and a 6BL7 output tube. It'll just be a little back-porch amp for my pleasure and I'm not likely to push it into serious grid blocking. So even with all the good discussion on this subject, I don't know if I'll go as far as inserting cathode followers. But I may try a 12AU7 PI with higher idle current and lower output impedance. - SD.