I'm working on installing an active (solid-state) effects loop in my 1969 Bassman. It was originally AC568, but over the last couple decades it's had some mods (AA864 phase inverter and a 2204-esque preamp on the Bass channel).
I gathered all the information I could get my hands on, and cribbed bits and pieces from different places to come up with the simplest circuit I could manage. It uses an IRF840 for the send buffer, and an LND150 for recovery.
I'm having a devil of a time with oscillation (90-120kHz) and motorboating (~4Hz). The buffer seems to be fine, so the problem appears to be on the recovery side.
I've tried multiple different tweaks along the way, including fixed bias for the LND150, lots of different values for Rs and Rd, lower Rg (down to 47k), high-value gate stoppers (up to 68k), etc. The closest I got to clean was using fixed bias at ~15V, but it still motorboated when the master got above about 7.
It oscillates even with volume, send, and return all turned down. Motorboating starts as soon as the return or send is turned up a little. There are some combinations that don't oscillate; in particular, turning the return all the way up makes it more stable (although there are still plenty of send pot positions that make it freak out again).
Notes:
I swear, everything single project I touch oscillates like mad. It's crazy-making. I spent eight hours today trying to resolve this. That's eight hours I would much rather have spent playing. I've easily spent three times as long chasing oscillations than I have doing everything else combined (designing, building, etc.)
I feel like I'm missing something dumb. What is it?
I gathered all the information I could get my hands on, and cribbed bits and pieces from different places to come up with the simplest circuit I could manage. It uses an IRF840 for the send buffer, and an LND150 for recovery.
I'm having a devil of a time with oscillation (90-120kHz) and motorboating (~4Hz). The buffer seems to be fine, so the problem appears to be on the recovery side.
I've tried multiple different tweaks along the way, including fixed bias for the LND150, lots of different values for Rs and Rd, lower Rg (down to 47k), high-value gate stoppers (up to 68k), etc. The closest I got to clean was using fixed bias at ~15V, but it still motorboated when the master got above about 7.
It oscillates even with volume, send, and return all turned down. Motorboating starts as soon as the return or send is turned up a little. There are some combinations that don't oscillate; in particular, turning the return all the way up makes it more stable (although there are still plenty of send pot positions that make it freak out again).
Notes:
- Some items have been omitted from the schem for clarity - 12V protective zeners on both MOSFETs, bypass switching
- It's currently built on miniature breadboards. I've left at least one empty row between all connections. They are sitting between the power transformer and the tagboard. Lead dress isn't great, but there aren't big gobs of wire floating around.
- B+ is sourced from the last preamp node. Yes, 220V is super low; the Bass-side preamp flows a lot of current, including the previously-unused triode.
- Voltages are at idle - had to disconnect the return to get there.
- I ran out of 100nF caps, that's why there are 150s in there.
I swear, everything single project I touch oscillates like mad. It's crazy-making. I spent eight hours today trying to resolve this. That's eight hours I would much rather have spent playing. I've easily spent three times as long chasing oscillations than I have doing everything else combined (designing, building, etc.)
I feel like I'm missing something dumb. What is it?
Last edited:
Good catch. Drafting error. It's in the correct place in the actual circuit. I fixed the schematic and for once I was able to edit the post and fix it there!
Last edited:
80KHz is probably layout. Pictures.
4Hz is surely power supply bypassing. Details?
Sharing two high-gain notes with a preamp B+ node seems a poor idea. Bring the power amp driver B+ out through 47k and 100uFd.
4Hz is surely power supply bypassing. Details?
Sharing two high-gain notes with a preamp B+ node seems a poor idea. Bring the power amp driver B+ out through 47k and 100uFd.
The bias divider on the first FET needs bypass cap - perhaps split the top resistor in two?
The symbols are the same yet the LND150 is a depletion mode device.
The high impedances mean unwanted capative feedback may happen if the layout is poor, some shielding may be needed between input and output. The supply rail must be decoupled too of course.
The symbols are the same yet the LND150 is a depletion mode device.
The high impedances mean unwanted capative feedback may happen if the layout is poor, some shielding may be needed between input and output. The supply rail must be decoupled too of course.
Are you sure that the problem is on the recovery side?
The IRF840 like ANY mosfet needs a gate stopper.
220R would probably do.
Cheers
Ian
The IRF840 like ANY mosfet needs a gate stopper.
220R would probably do.
Cheers
Ian
Okay. I'm a total idiot.
After fighting it for another couple hours (including splitting off a power supply node), I finally had to call it quits because I've got a ton of other things demanding my attention. So I backed off to a design that provides a buffered preamp out, and a passive power amp input, with a master volume just before the PI. It worked pretty well, and sounded surprisingly good - I'm not usually a big fan of pre-PI master volumes, but this preamp is so raw that it sounds great.
The preamp out was acting weird - turning down the pot actually raised the output - but since I'm not using it right now anyway, I left troubleshooting that for another day. After all, everything was working alright, and it sounded good.
And then, when I went to button it up, something caught my eye. I had star-grounded everything (including the preamp out pot) and run a ground wire back to the tagboard. Except I was off by one tag ... and I connected my circuit ground to the NFB input to the PI. So every ground in my circuit was actually connected to the secondary of the OT. Of course, I suspected grounding problems early on, so I checked for resistance to chassis ground ... which, via the OT secondary, was less than 1 ohm.
That explains the weirdness on the send, as well. To avoid any ground loops, I ran the pot's circuit ground to the star instead of to the chassis. Since the star ground was carrying the NFB signal, the pot was just balancing between my desired signal and the NFB signal.
So there you have it. I fixed the ground connection, and got the send working properly. I didn't have time to put the recovery stage back in, so I buttoned it up and left that for another day.
So there you have it. I managed to hook my effects loop ground to the NFB loop. Yay me. Thanks to everyone for their help!
After fighting it for another couple hours (including splitting off a power supply node), I finally had to call it quits because I've got a ton of other things demanding my attention. So I backed off to a design that provides a buffered preamp out, and a passive power amp input, with a master volume just before the PI. It worked pretty well, and sounded surprisingly good - I'm not usually a big fan of pre-PI master volumes, but this preamp is so raw that it sounds great.
The preamp out was acting weird - turning down the pot actually raised the output - but since I'm not using it right now anyway, I left troubleshooting that for another day. After all, everything was working alright, and it sounded good.
And then, when I went to button it up, something caught my eye. I had star-grounded everything (including the preamp out pot) and run a ground wire back to the tagboard. Except I was off by one tag ... and I connected my circuit ground to the NFB input to the PI. So every ground in my circuit was actually connected to the secondary of the OT. Of course, I suspected grounding problems early on, so I checked for resistance to chassis ground ... which, via the OT secondary, was less than 1 ohm.
That explains the weirdness on the send, as well. To avoid any ground loops, I ran the pot's circuit ground to the star instead of to the chassis. Since the star ground was carrying the NFB signal, the pot was just balancing between my desired signal and the NFB signal.
So there you have it. I fixed the ground connection, and got the send working properly. I didn't have time to put the recovery stage back in, so I buttoned it up and left that for another day.
So there you have it. I managed to hook my effects loop ground to the NFB loop. Yay me. Thanks to everyone for their help!
Yeah, DIYLC only has one MOSFET symbol. I keep meaning to get around to making an N-channel depletion symbol, but as you can see I'm too busy screwing up other things! 😀