Octave down

[Scruffie]

Eventually Stephen Giles will pop in and mention the EHX Deluxe Octave Multiplexer which I'm currently working on a PCB project for and replacing them pesky CA3094 with LM13700, as far as tracking goes it's about the best there ever was for analogue but it's certainly not a small build.

Having breadboarded it before, it can just about handle some simple 2 note chords and works pretty well all over the neck but even for its complexity it's not perfect.

The tracking circuitry does have both a compressor and gate function which helps a fair amount with keeping things clean and for lack of a better word, focused.

[Eb7+9]

for new ideas on down-octaving might want to consider a mixed-signal approach ... check out the SWR "mo-bass" schematics and guess at what the micro-controller does following the analogue rectifier stages ... might be a good discussion (new thread)

[ElectricDruid]

for new ideas on down-octaving might want to consider a mixed-signal approach ... check out the SWR "mo-bass" schematics and guess at what the micro-controller does following the analogue rectifier stages ... might be a good discussion (new thread)

Sound right up my street. Got a link for those schematics? I had a quick google and didn't find anything...

Thanks,
Tom

[Steve Newton]

Looks like the Mo_Bass schem is here on Fender's site - https://support.fender.com/hc/en-us/articles/115002392666-SWR-Amp-Schematics

And ... it looks like an analog signal squaring stage feeding a PIC(?) to process the square downward?

[PRR]

> feeding a PIC(?)

Just by the looks of those drawings, I'd guess it is MSI CMOS logic, many-stage counters and similar. I'm not recognizing the functions. If there is a part-list, we'd know what's happening.

[amptramp]

When you finally get a design, you could call it the

dsPICable downcount

[ElectricDruid]

> feeding a PIC(?)

Just by the looks of those drawings, I'd guess it is MSI CMOS logic, many-stage counters and similar. I'm not recognizing the functions. If there is a part-list, we'd know what's happening.

Yeah, I'm with PRR on this. I don't think it's a PIC either, despite what it says on the drawing. Couple of reasons why:

1) Power supply. +/-5V isn't a good supply for a PIC.
2) Wiring. Why take a load of 'unused' IO pins high and another load low? Doesn't make sense.

Comparing with the OC-2 schematic makes it look a lot more like some 4000-series CMOS, although it doesn't seem to be the 4013 that's common in these things. Like Paul said, some many-stage counter perhaps.

Still, it made me wonder. If you ran the signal into a PIC's ADC input, you could do a simple rectification and comparison digitally, and then output the control signal to the "signal flipper" (what *do* you call that op-amp stage with the CMOS switch on at the end?). It might save a bit of circuitry.

[anotherjim]

1) Power supply. +/-5V isn't a good supply for a PIC.

ECL logic?

[teemuk]

SWR "550" features the same circuit and there's a factory schematic for it.

Nope. The IC is not a "PIC", it's CD4027, so nothing but an ordinary dual flip flop. And the CMOS switches are doing the same thing the octave down circuits generally use a discrete FET for, that is synchronized phase inversion.

All in all, SWR's fundamental detector is pretty much generic "OC-2" circuit but built a little bit differently.

[ElectricDruid]

All in all, SWR's fundamental detector is pretty much generic "OC-2" circuit but built a little bit differently.

+1 agree. That's exactly what it looked like to me too. I hadn't managed to identify the device though, but 4027 sounds like exactly what I'd been expecting.

Tom

[PRR]

> ECL logic?

Chip ECL can run 300MHz. By running high current (among other tricks).

Sounds a poor choice for 5KHz work. (Agree that CMOS dual-flop seem likely.)

[Scruffie]

As digital and hybrids designs are being discussed, can I bring up the Deluxe Octave Multiplexer again.

Seems to me that starting with the building blocks of the fundamental extractor could possibly be a good route to go, it makes a pretty good sine wave from a guitar signal and the issue with it is that it's a complex bit of analogue circuitry, if it could be shrunk down to a one chip digital solution it would be a very useful chip for any circuitry that requires tracking.

All it is at its heart is a compressor running in to a state variable filter that tracks the incoming signal by comparing the response from the high pass and low pass through rectifiers and an integrator and then the low pass filter is fed in to a comparator to get the square that gets a gate fed from the compressors rectifier to stop it running wild at the tail end of the note.

Still wouldn't get to POG levels but hey, not bad for DIY, if such a thing is possible any way.

[Eb7+9]

you guys need to take a better look ...

https://support.fender.com/hc/en-us/articles/115002392666-SWR-Amp-Schematics

check out the 350X schematic for example ... it's got the added "Mo-Bass" PIC based circuit init

following two opposite polarity rectifier stages with LP filtered comparator groupings ...
that, in itself, is an interesting start to figuring out this puzzle

as for the PIC, it's obviously doing some form of threshold detection, threshold-pass counting divided by four, etc ...

https://www.manualslib.com/products/Swr-Mo-Bass-3338687.html

guessing at what the PIC does exactly or not is an interesting challenge in itself ...
couple of years ago I drummed up the inventor who's name I can't recall now (unf. since deceased)
but couldn't find a patent ...


yO!

[ElectricDruid]

you guys need to take a better look ...

No, sorry, but I don't agree. I know PICs well, and I can spot one at twenty paces on a dark and foggy night. Aside from the fact the schematic says "PIC" on it, what makes you think it *is* a PIC? (Although I agree that ordinarily that would be a pretty good guide!)

Compare with this schematic for a single flip-flop, for example:

https://www.electronicshub.org/jk-flip-flop-using-cd4027/

Notice that pins 5 and 6 are tied high. The equivalent pins on the other flip-flop would be 10 and 11. If you look at the 350X schematic you mentioned Pins 5 and 6 and 10 and 11 are all tied high. The two inputs are pins 4 and 7, the "set" and "reset" pins. Pins 1 and 13 are tied together, feeding the Q output from one flip-flop into the Clock input of the other. If it's a 4027 it all makes sense. If it's a PIC, it makes no sense at all. Why tie half the inputs high and the other half low? Why couple together two apparently random pins? In software, there *is* no point to that kind of thing - you'd link the things in software, not with a wire. It only makes sense if it's a hardware device, and that's been pretty clearly identified - the mystery chip is a 4027 dual flip-flop.

Ok, they tried to disguise it a bit, but that was just misdirection. They're being sneaky

[teemuk]

you guys need to take a better look ...

https://support.fender.com/hc/en-us/articles/115002392666-SWR-Amp-Schematics

check out the 350X schematic for example ... it's got the added "Mo-Bass" PIC based circuit init

Last page of that file, different revision of the schematic... not PIC .. CD4027. Why not just believe what I wrote earlier.

[Ben N]

All it is at its heart is a compressor running in to a state variable filter that tracks the incoming signal by comparing the response from the high pass and low pass through rectifiers and an integrator and then the low pass filter is fed in to a comparator to get the square that gets a gate fed from the compressors rectifier to stop it running wild at the tail end of the note.

Is that all? 

[anotherjim]

Also the rectifier "op-amp" feeding it says 339. It is not an LM339, pin outs are wrong  and the 339 is a quad comparator which would require pull-up resistors (and I'm not sure if comparators would function well as rectifiers). CMOS logic inputs wouldn't provide pull-up on its own, which was why I initially doubted the CMOS theory for the "PIC".
That said, if the part descriptions are bogus, then so could the pin numbers?

[Scruffie]

All it is at its heart is a compressor running in to a state variable filter that tracks the incoming signal by comparing the response from the high pass and low pass through rectifiers and an integrator and then the low pass filter is fed in to a comparator to get the square that gets a gate fed from the compressors rectifier to stop it running wild at the tail end of the note.

Is that all? 

You're right... it would definitely be better if the signal was split in to 6 copies that were band passed filtered

It's complicated and high parts count in the analogue realm but that's why I thought it would make a good stand alone digital chip because it is a great fundamental extractor, not my scope picture (credit goes to processaurus) but here it is turning a square wave input in to a sine:

[ElectricDruid]

All it is at its heart is a compressor running in to a state variable filter that tracks the incoming signal by comparing the response from the high pass and low pass through rectifiers and an integrator and then the low pass filter is fed in to a comparator to get the square that gets a gate fed from the compressors rectifier to stop it running wild at the tail end of the note.

It's complicated and high parts count in the analogue realm but that's why I thought it would make a good stand alone digital chip because it is a great fundamental extractor
<snip snip>

The state variable filter has been done digitally from a long while ago (it's in Hal Chamberlin's famous/fabulous book "Musical Applications of Microprocessors"), and a digital integrator is just a running sum, and a comparator is an "if" statement of some type. So it doesn't seem impossible, frankly, even on a low-end chip. I'll give it a whirl one day when I've got some free time (hahaha chance would be a fine thing ).

Tom

[Eb7+9]

you guys need to take a better look ...

No, sorry, but I don't agree. I know PICs well, and I can spot one at twenty paces on a dark and foggy night. Aside from the fact the schematic says "PIC" on it, what makes you think it *is* a PIC? (Although I agree that ordinarily that would be a pretty good guide!)

Compare with this schematic for a single flip-flop, for example:

https://www.electronicshub.org/jk-flip-flop-using-cd4027/

Notice that pins 5 and 6 are tied high. The equivalent pins on the other flip-flop would be 10 and 11. If you look at the 350X schematic you mentioned Pins 5 and 6 and 10 and 11 are all tied high. The two inputs are pins 4 and 7, the "set" and "reset" pins. Pins 1 and 13 are tied together, feeding the Q output from one flip-flop into the Clock input of the other. If it's a 4027 it all makes sense. If it's a PIC, it makes no sense at all. Why tie half the inputs high and the other half low? Why couple together two apparently random pins? In software, there *is* no point to that kind of thing - you'd link the things in software, not with a wire. It only makes sense if it's a hardware device, and that's been pretty clearly identified - the mystery chip is a 4027 dual flip-flop.

Ok, they tried to disguise it a bit, but that was just misdirection. They're being sneaky

alrighty, then ... YOU would know ED
thanks for taking some of the guess-work ut of it

so, even simpler then