Couple of modified BJFs
- Thread starter HamishR
- Start date
That construction looks very Menatone. Except neater!
I drew my own vero layout for the Baby Blue because I didn't like the ones I found online. I have developed my own anal peculiarities about layouts.
I quite like the Baby Blue but only with a Gibson. I think it's extremely ordinary with a Strat, for example. It compresses more than I like but at low gain with a Les Paul it has something I like.
I drew my own vero layout for the Baby Blue because I didn't like the ones I found online. I have developed my own anal peculiarities about layouts.
I quite like the Baby Blue but only with a Gibson. I think it's extremely ordinary with a Strat, for example. It compresses more than I like but at low gain with a Les Paul it has something I like.
Hey Music 6000 - how do you like the Baby Blue? Do you use it much?
I just altered the schematic for the Model G because I had a 4K7 on the output instead of 47K. It's correct now. And just for the record, here is the 1N4001 version. I'm finding this one even better now, and it saves the 2N5952s. This overdrive suits me very well - it's amazing.
I just altered the schematic for the Model G because I had a 4K7 on the output instead of 47K. It's correct now. And just for the record, here is the 1N4001 version. I'm finding this one even better now, and it saves the 2N5952s. This overdrive suits me very well - it's amazing.
Chuck D. Bones
Circuit Wizard
I ran some traces of three JFETs, a BJT & three Si diodes, then plotted them for comparison. The MPSA18 was tested two ways, just the emitter-base junction and with the base & collector tied together. Vf for the JFET gate junction was fairly high. Note that JFETs are not designed to work this way, but as long as the gate current is not too high, they'll survive. I have seen JFETs used as ultra-low leakage diodes in precision circuits. The JFETs, MPSA18 E-B junction & the BA283 PIN diode have the highest Vf. I think the high Vf is what drives some pedal designers to use them as diodes. The BA283 or MPSA18 would both make good substitutes. As always, let your ears make the final decision.
Chuck D. Bones
Circuit Wizard
I use this:
https://forum.pedalpcb.com/threads/my-transistor-diode-test-station.1609/
I export the I-V data as a CSV file, import it into EXCEL, do some filtering and offset correction, then plot it. The wiggle in the traces are quantization error. I need to mod the curve tracer or put a preamp between the curve tracer and scope to boost the signal and get better resolution.
https://forum.pedalpcb.com/threads/my-transistor-diode-test-station.1609/
I export the I-V data as a CSV file, import it into EXCEL, do some filtering and offset correction, then plot it. The wiggle in the traces are quantization error. I need to mod the curve tracer or put a preamp between the curve tracer and scope to boost the signal and get better resolution.
Wow... (head exploding quietly)... So it looks like in some ways a BA283#2 could be a good replacement. Except Mouser calls them obsolete! Oh well, I have plenty of MPSA18s.
Thanks Chuck - this is really interesting. Do you think the Vf is the main characteristic here? Could the lower Vf of the 1N400x be why it dirties up faster and has less volume? I'm also wondering if the lower Vf of the 1N400x - if responsible for more and earlier clipping - is why it sounds a bit brighter and raspier? And I don't mean raspy in a bad way. It's slightly less smooth and that can sound good too. I suspect the version with the 1N400x diodes would stand out in a mix slightly better. Mind you, we're talking quite subtle differences here. Audible, but subtle.
So when you show "MPSA18 B-E" you mean use just the base and emitter legs?
Thanks Chuck - this is really interesting. Do you think the Vf is the main characteristic here? Could the lower Vf of the 1N400x be why it dirties up faster and has less volume? I'm also wondering if the lower Vf of the 1N400x - if responsible for more and earlier clipping - is why it sounds a bit brighter and raspier? And I don't mean raspy in a bad way. It's slightly less smooth and that can sound good too. I suspect the version with the 1N400x diodes would stand out in a mix slightly better. Mind you, we're talking quite subtle differences here. Audible, but subtle.
So when you show "MPSA18 B-E" you mean use just the base and emitter legs?
Chuck D. Bones
Circuit Wizard
Vf influences into how much clean headroom is available before the dirt takes over and the max volume. The shape of the curve has some effect too, although they're all pretty similar in that regard. The CA3130 can swing the output right up to the rails, so the peak diode current in the hard clipper is roughly (Vcc-Vref-Vf)/1K = 3.8mA. I say roughly because the 1uF cap has some impedance and a little bit of the current sneaks around the diodes thru the 10K and 12K resistors. If Vf is larger, then the peak diode current is lower.
Yes, MPSA18 B-E means the curve tracer is measuring the current and voltage between base and emitter. The collector lead is not connected to anything.
The BA283's are hard to find, but they're out there. I have some BA482's, I should run curves on them too.
Yes, MPSA18 B-E means the curve tracer is measuring the current and voltage between base and emitter. The collector lead is not connected to anything.
The BA283's are hard to find, but they're out there. I have some BA482's, I should run curves on them too.
Well despite my best efforts I am actually learning something here. Honestly mate, I have learnt more from you in the last few months than I have ever learnt in my life about electronics. Chuck, you should work in electronics.
Another thing I learnt today is that the G works incredibly well with my silicon Red Rooster. Well most ODs do, but it seems to mesh particularly well with the RR.
As it happens I am liking the 1N4001 G more than ever, and there is one last thing I'd like to address... What I am calling the bass knob is not really a true bass knob. It seems to affect low mids and never gets rid of the low end completely. I'm thinking that if I reduce the lowest of the lows before clipping I can get a clearer sound on the bass strings. Right now I can clean up the bass strings pretty well by keeping "bass" below noon but there is still a little noise there. I tried making the 4u7 at Bass 3 smaller but it didn't have the effect I was chasing. What about the 220nF cap to Vref there? What would happen if I reduced it to 47 or 100nF? Maybe it's the 1µF at pin 6 I should be looking at? I love the full range sound of the pedal so I'm reluctant to mess with it much further!
Another thing I learnt today is that the G works incredibly well with my silicon Red Rooster. Well most ODs do, but it seems to mesh particularly well with the RR.
As it happens I am liking the 1N4001 G more than ever, and there is one last thing I'd like to address... What I am calling the bass knob is not really a true bass knob. It seems to affect low mids and never gets rid of the low end completely. I'm thinking that if I reduce the lowest of the lows before clipping I can get a clearer sound on the bass strings. Right now I can clean up the bass strings pretty well by keeping "bass" below noon but there is still a little noise there. I tried making the 4u7 at Bass 3 smaller but it didn't have the effect I was chasing. What about the 220nF cap to Vref there? What would happen if I reduced it to 47 or 100nF? Maybe it's the 1µF at pin 6 I should be looking at? I love the full range sound of the pedal so I'm reluctant to mess with it much further!
I’ve been enjoying following this thread and admiring your clean vero work.
For the bass, one thing you could try is increasing the value of the bass pot. 10k may not be enough to completely take it away with the knob at minimum, though I’m not sure if that’s your problem.
Yes reducing the 220n like you mentioned would reduce some low-mid content. The 1k and 220n determine the cutoff frequency of the high pass filter (F=1/(2piRC) = 723Hz). So reducing the cap would raise this frequency. Reducing the 1k also has the same impact on frequency, although it also adjusts the gain range. For the cap, I’d probably start with 150n (1000Hz) and Experiment from there. You could also try a trimmer where the 1k goes to help you dial in the cutoff frequency exactly (like how the voice knob works on a zendrive).
One more thought based on your description, maybe your happy with the cutoff at 723Hz, you just want a steeper roll off. (It’s currently a 1 pole high pass which can have a shallow effect). I think by reducing the 22n on the input, you could have 2 separate HPF around 723Hz. Using the 470k to Vr as the “R” for the calculation, I get that 470p would put that filter at 720Hz (Not sure if that’s close Enough, do I need to consider the impedance of the op amp for the calculation, chuck?). The problem with this method is it permanently removes some bass. So with the bass knob down you’d get 2 pole HPF, with the bass knob up you’d have 1 pole high pass. (Please correct me if I’m thinking of it wrong). Of course you could also put those caps on a switch to have all the options.
reduicing 1uF at pin 6 could also work. I just thought since there is some soft clipping in the op amp feedback loop, maybe you’d want to cut bass before that.
For the bass, one thing you could try is increasing the value of the bass pot. 10k may not be enough to completely take it away with the knob at minimum, though I’m not sure if that’s your problem.
Yes reducing the 220n like you mentioned would reduce some low-mid content. The 1k and 220n determine the cutoff frequency of the high pass filter (F=1/(2piRC) = 723Hz). So reducing the cap would raise this frequency. Reducing the 1k also has the same impact on frequency, although it also adjusts the gain range. For the cap, I’d probably start with 150n (1000Hz) and Experiment from there. You could also try a trimmer where the 1k goes to help you dial in the cutoff frequency exactly (like how the voice knob works on a zendrive).
One more thought based on your description, maybe your happy with the cutoff at 723Hz, you just want a steeper roll off. (It’s currently a 1 pole high pass which can have a shallow effect). I think by reducing the 22n on the input, you could have 2 separate HPF around 723Hz. Using the 470k to Vr as the “R” for the calculation, I get that 470p would put that filter at 720Hz (Not sure if that’s close Enough, do I need to consider the impedance of the op amp for the calculation, chuck?). The problem with this method is it permanently removes some bass. So with the bass knob down you’d get 2 pole HPF, with the bass knob up you’d have 1 pole high pass. (Please correct me if I’m thinking of it wrong). Of course you could also put those caps on a switch to have all the options.
reduicing 1uF at pin 6 could also work. I just thought since there is some soft clipping in the op amp feedback loop, maybe you’d want to cut bass before that.
Thanks for the suggestions. It helps because it makes me feel that finally I am getting some of this stuff! I'll try your ideas one at a time and see what happens.
Anyway, the main reason I am enjoying the G so much is that it is almost all there - it has a complex, natural and quite amp-like sound, it's full and responsive, sounds like a great old small-box plexi with chords, warts and all. It has some imperfection like any great old amp does. As much as I like a clear low-end I worry that if I achieve it I might lose what I like about the G. But I will have to try it and see.
Anyway, the main reason I am enjoying the G so much is that it is almost all there - it has a complex, natural and quite amp-like sound, it's full and responsive, sounds like a great old small-box plexi with chords, warts and all. It has some imperfection like any great old amp does. As much as I like a clear low-end I worry that if I achieve it I might lose what I like about the G. But I will have to try it and see.
Chuck D. Bones
Circuit Wizard
The opamp's input impedance is HUGE, so we don't need to consider it in the calculation. I agree with all of Phi1's ideas. I don't think I'd dial down the input cap that far, and I'd only do it as a last resort because as Phi1 says, it's non-adjustable. You should be able to get the BASS control to do what you want by reducing the 220nF. To find out if the BASS pot is big enough or not, disconnect one of the leads and see how much of the bass goes away. If it's still too bassy, then increasing the BASS pot won't get help. Where are you setting the GAIN pot?
How about adding reference designators to your schematic?
How about adding reference designators to your schematic?
Sorry about the lack of reference designators. Drawing schematics is still quite new to me. I will do so now! Actually I was worried in case I had made some rookie schematic mistake, having never been schooled in the art. It appears I have!
And yeah, I was reluctant to decrease the input cap value. I'll try a few things and report back.
And yeah, I was reluctant to decrease the input cap value. I'll try a few things and report back.
Chuck D. Bones
Circuit Wizard
Yeah, you're gettin' there. Don't forget the semiconductors. 
Cool! I have tried a few things as discussed - swapping C2 (as shown conveniently numbered above) to 100nF and it did clean things up but made it a little sterile and kinda scooped mids more than extreme lows. So that's back at 220n.
What seems to be working is using a 470nF for C5. Maybe it's because it's limiting the lows a little before the hard clippers? I dunno. But it sounds really good and I think is an improvement on the overall punch and clarity. I'm keeping the first G with the 2N5952 diodes-as-clippers as per the first schematic shown here and just modifying the 1N4001 version. And i think I will leave it with the 470nF at C5 for now. It seems to be doing everything I wanted. There's still plenty of thump if I crank the bass knob - too much really. So I keep it around the noon mark mostly.
Thanks for the help guys! This forum is amazing.
I have tried a few things as discussed - swapping C2 (as shown conveniently numbered above) to 100nF and it did clean things up but made it a little sterile and kinda scooped mids more than extreme lows. So that's back at 220n.What seems to be working is using a 470nF for C5. Maybe it's because it's limiting the lows a little before the hard clippers? I dunno. But it sounds really good and I think is an improvement on the overall punch and clarity. I'm keeping the first G with the 2N5952 diodes-as-clippers as per the first schematic shown here and just modifying the 1N4001 version. And i think I will leave it with the 470nF at C5 for now. It seems to be doing everything I wanted. There's still plenty of thump if I crank the bass knob - too much really. So I keep it around the noon mark mostly.
Thanks for the help guys! This forum is amazing.
Chuck D. Bones
Circuit Wizard
I was gonna suggest that next. That's exactly what it's doing. C5 doesn't have much effect on the soft clipping, but it keeps some of the lows out of the hard clipper.
