Anyone have built succesfully (no bleedthrough signal) a CMOS Nyquist Aliaser?

[gigimarga]

Hello,

I reload on my table a CMOS Nyquist ALiaser made using this layout: http://www.mrdwab.com/john/AliaserCMOS.pdf
It sounds amazing, but it has a very strong bleedthrough signal.
I soldered the frequency pot on the PCB, I used shielded cable for the input and for the output, I tried 2 or 3 types of CD4016, but i had no luck!
Anyway, the sounclips that i found sounds OK, so has anyone built one which works OK?
If yes, which is the tip?

Thx a lot all!

[gigimarga]

 

[Taylor]

I want to know, too. I built one using the same layout and also have carrier bleedthrough. I built mine on perf, I wonder if there needs to be a ground fill on the board to avoid hearing the oscillator. Maybe this is a clever ploy by John Lyons to get me buy yet more of his (excellent) boards.

[gigimarga]

I've built tonight a Robovox (using a 4016, too) and it has a bleedthrough signal, too...

[frequencycentral]

Never built the Aliaser, so I can't comment on the difference between one and a ring mod, but they're doing a similar job right? I did once breadboard Colin's Analog Bit Crusher, which also suffers from bleed. In a true ring mod, no signal passes at all unless the carrier is present. I've build a couple of these ring mods which use a MC1496, they can be adjusted for zero modulator bleed. I know the values are hard to read, but it is possible, and it's a great ring mod, though you will need to add an oscillator of some description.

http://fa.utfs.org/diy/roland100m/schematicsshow18.html

[Taylor]

The aliaser sounds kind of like a ring mod with a pulse wave carrier. I've also built a Ring Stinger, but the aliaser does a sound that the RS can't do.

[nelson]

Yep, put a gate on the oscillator.

[gigimarga]

Thx a lot all!

The Maestro Ring Modulator and Red Stinger are too complicated for my available time...maybe in a holiday

Anyway, the sound of the aliasers is unique and i like it a lot at high frequencies.

Maybe someone will solve this problem...i'm still hoping...

[Taylor]

I think I'll try Nelson's Wolf Bagger. For some reason I thought it was using the envelope to modulate carrier frequency, which is a sound I find tiring. Now I realize it's got a gate, which is just what we need.

[gigimarga]

I think I'll try Nelson's Wolf Bagger. For some reason I thought it was using the envelope to modulate carrier frequency, which is a sound I find tiring. Now I realize it's got a gate, which is just what we need.

I tried it and it sounded very good (take care at parts list: R10 is listed as 47K, instead 1M as is right regarding the schematic...and i think that is one more little mistake in that list...), but after i modded it following a Nelson advice it stopped to work (a mistake of mine, i think...), so i abbandoned it...good luck!

[gigimarga]

 

[Taylor]

The only person I know of who said there wasn't noticeable bleed was John Lyons, hence my comment that maybe the ground fill on his PCBs has something to do with it. I've heard some mention that keeping analog and digital ground separate is necessary, but I don't really know how that's possible, since they have to tie together at some point, right?

http://www.diystompboxes.com/smfforum/index.php?topic=54135.0

[gigimarga]

The only person I know of who said there wasn't noticeable bleed was John Lyons, hence my comment that maybe the ground fill on his PCBs has something to do with it. I've heard some mention that keeping analog and digital ground separate is necessary, but I don't really know how that's possible, since they have to tie together at some point, right?

http://www.diystompboxes.com/smfforum/index.php?topic=54135.0

You're right...i hope that John Lyons will be so kindly to tell us how to separate the grounding (as i saw in the picture he used the same PCB).

[gigimarga]

 

[The Tone God]

Bleed through on these style of aliasers is like noise with boosters, it just a nature of the beast. What I did with my aliaser is put an envelope detector that would shutdown the clock when the input level dropped below a user set level. It ends up being a nice feature too as you can set a threshold where the aliasing takes over.

Andrew

[gigimarga]

Thx a  lot Tone God...i think i will retry your Crazy Larry and Nelson's Wolf Bagger again...succesfully this time, i hope

[brett]

Hi
I haven't built the NA, but I've built many different sample and hold devices using a CD4016 or CD4066 as the switch.
Minimize switching noise is about a range of factors.  One of which is making sure that switching charges are small (relative to the signal) and transients are fast 20kHz or more, so that they can be easily filtered out).  The second part is easy.  These chips can switch in about 1 microsecond if driven via a low output impedance device.  Therefore, filtering everything over 30kHz will smooth the transitions.  I *always* drive a substantial cap (>0.01uF) on the hold side.  This keeps the "held" charge large, even for small voltages.  IIRC the aliaser doesn't have such a cap, so the charges are tiny (the input capacitance on the hold side is used, and this is deliberately kept very small if it's an inverter).  That's a potential weakness.

My idea of an ideal circuit would be two JFET op-amps with a CD4066 switch and hold cap (or both sample and hold caps?) between them.
just my 2c

[gigimarga]

Hi
I haven't built the NA, but I've built many different sample and hold devices using a CD4016 or CD4066 as the switch.
Minimize switching noise is about a range of factors.  One of which is making sure that switching charges are small (relative to the signal) and transients are fast 20kHz or more, so that they can be easily filtered out).  The second part is easy.  These chips can switch in about 1 microsecond if driven via a low output impedance device.  Therefore, filtering everything over 30kHz will smooth the transitions.  I *always* drive a substantial cap (>0.01uF) on the hold side.  This keeps the "held" charge large, even for small voltages.  IIRC the aliaser doesn't have such a cap, so the charges are tiny (the input capacitance on the hold side is used, and this is deliberately kept very small if it's an inverter).  That's a potential weakness.

My idea of an ideal circuit would be two JFET op-amps with a CD4066 switch and hold cap (or both sample and hold caps?) between them.
just my 2c

Thx a lot brett...but it's enough complicated for me now...maybe when i will have sometime

[brett]

Hi
after a quick look at the layout, it seems that the two chips are biased independently (if I'm interpreting the layout correctly - I can't find a schematic).  That can be a problem.  The components that are duplicated for each chip (e.g. the 2 x (10k and 1M) resistors) will need to be closely matched for low bleed-through.  Without thinking it through, I suspect that a small re-design of one or two parts of the circuit could eliminate most of these issues (e.g. single Vref using noiseless biasing).  In the meantime, try matching your resistors for each chip as closely as possible by taking a batch (of 6 or 10) and using the closest pair.  The others will be fine for other projects, but not this one.

cheers

[gigimarga]

Thx a lot brett!!!
I hope that's the problem...i will test your ideas tonight.
Anyway, the schematic is here: http://www.dedalofx.com.ar/bioroids/external/nyquist_aliaser.jpg