buffered pedal cures oscillations in Jfet designed distortion pedal

[TheThingThatShouldNotBe]

Ive built a couple of "Dr boogies" with different layouts & screening methods , biasing etc but both still oscillated with passive pickup equipped guitars so thought Id try chucking one of my diy loop blender pedals in front ..oscillation disappears . Oscillation also disappears when I stick a digitech pedal infront and I expect any buffered pedal would be the same .
drboo-schem[1].jpg
Click on to enlarge .

Now I could just incorporate a buffer circuit on the input but Im wondering What is fundamentally wrong with the schem and so keep parts to a minimum .
Any suggestions ?
Its basically a scaled version of the mesa boogie preamp in a solo head I believe .

[TheThingThatShouldNotBe]

Sorry that image is pretty low rez try this link .
http://gaussmarkov.net/layouts/drboo/drboo-schem.png

[ICBM]

Yes, it does appear to be a close copy of a Mesa Dual Rectifier preamp, with many of the resistor values 'scaled' by a factor of 10, presumably to compensate for the different characteristics of JFETs relative to valves.

I actually doubt this is the right approach to designing a solid-state preamp, although I don't know enough about it to be sure - it just doesn't strike me that forcing a circuit designed for one technology to operate with another in that way is likely to be successful. Which might explain why it's unstable. At a guess the input network resistors - R18 and R19, possibly R1 - are too high value... that might be why driving it from a low-impedance source cures it.

It's also wrongly named - the Mesa Rectifier is a *Mesa*, not a Boogie. Completely different. [/pedant]

[marauder]

The schematic is fine, it will be the circuit layout causing the oscillation. Every diyer has had issues with the boogie, its a right of passage pedal, and will take a few attempts to work satisfactorily.

The amplified signal is being coupled back into the input wire, add this jfet buffer, as close to the stomp switch wire that goes to the circuit as possible, its output to the boogey circuit input. It changes the impedance of input making it less susceptable to picking up stray signals. Its the same as a buffered pedal in front, but still true bypass.

http://juansolo.demon.co.uk/stompage...Stripboard.gif

Try biasing 5.5-6.0v, sounds better with j201s, 6.5v with 2n5457s. Changing c16 to 10nf can also tame the fizzy highs.

My latest one has 2 j201s in first stages, and 2n5457s later, 10nf c16, its got less gain than an all j201 build, but sounds great, and the gain control is more useful above 50%. Built on a madbean chunk chunk pcb, the latest version of which has board mounted pots, which will help with any noise problems.

[TheThingThatShouldNotBe]

The schematic is fine, it will be the circuit layout causing the oscillation. Every diyer has had issues with the boogie, its a right of passage pedal, and will take a few attempts to work satisfactorily.

The amplified signal is being coupled back into the input wire, add this jfet buffer, as close to the stomp switch wire that goes to the circuit as possible, its output to the boogey circuit input. It changes the impedance of input making it less susceptable to picking up stray signals. Its the same as a buffered pedal in front, but still true bypass.

http://juansolo.demon.co.uk/stompage...Stripboard.gif

Try biasing 5.5-6.0v, sounds better with j201s, 6.5v with 2n5457s. Changing c16 to 10nf can also tame the fizzy highs.

My latest one has 2 j201s in first stages, and 2n5457s later, 10nf c16, its got less gain than an all j201 build, but sounds great, and the gain control is more useful above 50%. Built on a madbean chunk chunk pcb, the latest version of which has board mounted pots, which will help with any noise problems.

Thanks !
I quite like the insain gain + highs
I would like to separate the treble & presence a bit more but its a bit tricky identifying which caps govern which frequency from the schem until Ive mocked up a board fitted out with sil sockets for all the caps .I'd like to try different brands values etc. You couldnt shed any light on the EQ could you ?
Probably will end up buffering between the switch and board input as it obviously cures 90% of problems .
I noticed that the J201 trim pots are 100k ?! How low can they get away with ?

I actually doubt this is the right approach to designing a solid-state preamp, although I don't know enough about it to be sure - it just doesn't strike me that forcing a circuit designed for one technology to operate with another in that way is likely to be successful. Which might explain why it's unstable. At a guess the input network resistors - R18 and R19, possibly R1 - are too high value... that might be why driving it from a low-impedance source cures it.

Well Im glad you pointed that out . I'll be sure to give those a try when ive constructed an easier to live with sil board mockup of the real thing . TBH I dont think any one really expects this kind of project to sound as good as mesa engineering , I personally suspect that a few components were used and scaled down or even some that arnt really compatible with Jfets that shouldnt actually be there , who knows after ive stripped it back it may actually start to make more sense to me and why this schem is the way it is today . As I understand it ,it was pretty much put together by "gauss markov's" vision + a whole load of contributers .

[ICBM]

Don't take that the wrong way, I actually think that one of the reasons solid-state circuits don't sound like valve is because they aren't designed *enough* the same way - but I was very surprised to see how close a copy of the Rectifier preamp schematic this is, just with JFETS substituted for the valves and a lot (but not all) of the values scaled by a factor of ten. It strikes me as a "by luck and guesswork" approach rather than actually designing from the ground up for transistors, still using "valve-like" circuit elements.

A similar example in some ways is how Marshall copied the Fender Bassman, but made one simple and quite large mistake, I think by accident due to not understanding how to design a valve amp circuit fully but just copying the Fender schematic value for value. I could be wrong on that but it would be surprising otherwise - an engineer who really knew what they were doing would have automatically known to compensate for the different output impedance and altered the NFB loop to compensate, but Marshall didn't... and created a different (but rather good!) variation on the sound.

With this pedal, I would expect to see a lot more values tweaked by a lot more different factors than either ten, or not at all, for it to truly replicate the gain structure and tone of the Rectifier. I suspect this is why it needs so many trimmers, because it's teetering on the edge of instability anyway.

This is also not meant as a put-down, but I do think the circuit is very much of an accident that it works at all, and if instability issues are common, it shouldn't be treated as a finalised project circuit which just needs a few layout tweaks to make good... it shouldn't be as sensitive as that really. I suspect a lot more values need changing than just a single cap, in other words.

If I have the time next year I'll see if I can build one and whether I can find anything more definite than just this theoretical opinion .

[marauder]

The reason for the trimmers, is the massive variability of jfets. They are worse than tubes. No wonder they are now being phased out.

When i build a pedal with trimmed jfets, i put them in a little circuit which measures the cutoff voltages. I can then calculate the trimmed voltage on the drain, and not rely on my ears. It works well, pedals sound great, i dont even listen before trimming.

I know wampler pedals use similarly measured jfets, they dont trim, but batch the jfets, for use with particular resistors on drains.

Heres an article of jfet measuring. RG is now the designer at visual sound, but still contributes to the diy scene.
http://www.geofex.com/article_folder...h/fetmatch.htm

[jpfamps]

The oscillation is caused either by the inductance of your pickups interacting with some capacitance, stray or otherwise, forming a resonant circuit, or more likely by signal from later stages in the circuit coupling back to the input.

A buffer will either isolate the pick up's inductance from the input, or present a low impedance to shunt stray signal to ground.

You can usually tell which are the offending stages by getting the pedal oscillating and then touching different components. If the frequency of oscillation changes, then you are getting warm.

High impedance circuits are much more susceptible to stray signal than low impedance. If you do add a buffer before the first stage (as suggested) then you don't need the high impedance input to the first gain stage, so I would reduce the input impedance of this stage; 10k would seem a suitable value.

[TheThingThatShouldNotBe]

Its been a couple of months now since I last opened up my Dr Boogey . After installing a buffer circuit I decided that there is something else at play here because it didnt work as well as having a buffer pedal infront in a seperate box . So I stripped it back and just rebiased the gain stages by ear and multi meter until the oscillation dissapeared . One thing I did notice is that Jfet 4 (the last one) controls the overall noise floor which is good because the first 3 seem to cascade the oscillation more . It doesnt need a buffer and its much quieter . Its actually a balancing act more than anything , you can get it to sound very quiet indeed if you sacrifice some of its insanity .
Mines full fat !

Im looking into making a direct to record speaker emulation box to go with it now. Already tried one schem but the dual op amps were clipping quite badly and I couldnt really hear much emulation going on underneath .

[marauder]

Glad you got it working better, its a great circuit when it works right, and I dont think anyone builds one which sounds quite like another one. JFETs just vary too much unless tested before hand.

Here's a demo of one using our mods and biasing, into a clean Princeton Reverb RI.

http://www.youtube.com/watch?v=e2tDoqKvH4E