freestompboxes.org

The scheme is good because it does not require identical jfets. But, as far as I know, it can have clipping problems, just like jfets. It seems to me that it is more reasonable to replace CMOS and Jfets with a photoresistor + LED. You could try using OTA, but not like in the EHX Small Stone circuit (I think OTA ruins the sound), but you can use OTA as a variable resistor (idea from the Engineers thumb compressor, not tested in practice).

Axldeziak wrote: ↑11 Feb 2023, 19:48 Here's a layout that allows switching between the 2 and the 4 stage versions.
If anyone finds an error let me know.

So many layouts! You're making me happy, guys. Be sure to also post the schematic for reference.

dylan159 wrote: ↑09 Feb 2023, 23:31 Thanks. How did it go? Series components may be swapped but that makes no difference. I've built that layout and it works.

Excelent! Still gathering components. I forgot to buy one of the 074 and the 2M2 res seems hard to find around here.
Will try to build it soon, and post with good news I hope.
Thanks!!

POTL wrote: ↑11 Feb 2023, 20:16 The scheme is good because it does not require identical jfets. But, as far as I know, it can have clipping problems, just like jfets. It seems to me that it is more reasonable to replace CMOS and Jfets with a photoresistor + LED. You could try using OTA, but not like in the EHX Small Stone circuit (I think OTA ruins the sound), but you can use OTA as a variable resistor (idea from the Engineers thumb compressor, not tested in practice).

Matching can be as much as a problem with LDRs as it is for FETs. OTA phasers definitely exist (DOD for example), and they work well, but they're awful in terms of noise and distortion isn't better than using FETs either. As I said in the "distortion" paragraph, I found it to be more than acceptable. I take the chance to link the mentioned app note since I forgot to in the original post https://www.onsemi.cn/pub/collateral/an-6603cn.pdf
Mosfets will always perform worse than JFETs in this terms, and JFETs fall short of LDR purely in terms of distortion even if they're good already. For guitar levels and this kind of effect, the benefit of not having to match more than offsets the distortion concerns.

Rodri_DiY wrote: ↑16 Feb 2023, 13:59 Excelent! Still gathering components. I forgot to buy one of the 074 and the 2M2 res seems hard to find around here.
Will try to build it soon, and post with good news I hope.
Thanks!!

I take care to not make things that require anything too specific. 2.2M is a standard value, but you can use anything higher you can find. the 074 can be two duals or another quad too.

Thought I would post my layout here as well. I'll post the Gerber file once I receive mine and it is verified. The extra footprints are for caps on each of the phase stages to experiment with different values.
The reason behind the dual op amps instead of quads is I have a full reel of LF353's. Also, I usually find it easier to route PCBs with duals. Quad packages get too crowded unless you go SMD. This would be a tough one to etch (double sided) though not impossible. Etch files, schematic, 3D renders are in the PDF below.
Enjoy!!!

Lani wrote: ↑16 Feb 2023, 16:55 Thought I would post my layout here as well. I'll post the Gerber file once I receive mine and it is verified. The extra footprints are for caps on each of the phase stages to experiment with different values.
The reason behind the dual op amps instead of quads is I have a full reel of LF353's. Also, I usually find it easier to route PCBs with duals. Quad packages get too crowded unless you go SMD. This would be a tough one to etch (double sided) though not impossible. Etch files, schematic, 3D renders are in the PDF below.
Enjoy!!!

Beautiful as usual. Don't challenge me with etching one of your boards, I've done it before even if it probably would be more painful than even the perfboard for this one
Using dual op amps is actually a good idea here, so you can use some 5532 for the allpass stages and keep the noise down. The Pup 6 had me wishing for a common low noise quad op-amp.
Cap or resistor, either one works to shift the minimum of the frequency sweep, while cap and bias set the upper limit.
About the lfo patch point: you need to provide the right bias from the external lfo! This isn't the end of the world and means the internal one with the triangle lfo can be left undisturbed, just something to keep in mind.
Let us know how this turns out. what kind of enclosure is it for?

dylan159 wrote: ↑16 Feb 2023, 15:22 I take care to not make things that require anything too specific. 2.2M is a standard value, but you can use anything higher you can find. the 074 can be two duals or another quad too.

Excelente! I think I've already bought everything I need to start. Is really useful to me to understand schematics in order to start building stuff. May be for example this difference between schematics and layout: schematics says SW3 (6-4) goes 1 to label 6 and 3 to label 4 in contrast with the layout, that seems to be 3 to 6 and 1 to 4 (backwards)?
I believe that this order really doesn't matter, cause the switch will work anyway, is just to check if I read it properly.

Thank you!

Rodri_DiY wrote: ↑17 Feb 2023, 21:35 Excelente! I think I've already bought everything I need to start. Is really useful to me to understand schematics in order to start building stuff. May be for example this difference between schematics and layout: schematics says SW3 (6-4) goes 1 to label 6 and 3 to label 4 in contrast with the layout, that seems to be 3 to 6 and 1 to 4 (backwards)?
I believe that this order really doesn't matter, cause the switch will work anyway, is just to check if I read it properly.

Thank you!

As long as both sides are consistent it's ok. I made the layout to have 6 when up and 4 when down.

Beautiful as usual

Thanks buddy!! Thank you for all the work you do and post here!! Keep up the great work!!!

About the lfo patch point: you need to provide the right bias from the external lfo! This isn't the end of the world and means the internal one with the triangle lfo can be left undisturbed, just something to keep in mind.

I'll probably only use the out anyways to drive other pedals in my chain, but you're right and I didn't even think of that.

what kind of enclosure is it for?

It "should" fit in a 1590BB, I have a Phase Tone here that uses the same board outline as this one (viewtopic.php?p=287069#p287069) will double check to make sure it fits and will report back only if it doesn't.... Just got my 1590BB's in yesterday.

Let us know how this turns out

Will do

Lani wrote: ↑18 Feb 2023, 15:10

About the lfo patch point: you need to provide the right bias from the external lfo! This isn't the end of the world and means the internal one with the triangle lfo can be left undisturbed, just something to keep in mind.

I'll probably only use the out anyways to drive other pedals in my chain, but you're right and I didn't even think of that.

Oh you want to use this LFO to control other circuits? I like the idea, but same considerations apply in reverse: the output will be high impedance and possibly loaded down by additional circuits, and it will be biased up by the trimmer. A possible solution could be to buffer the capacitor voltage before the 330k, send it through as before but taking the output from the buffer's output. Of course I wouldn't mention it if it wasn't possible by adding a little daughterboard and tacking a wire to the existing pads.

dylan159 wrote: ↑16 Feb 2023, 14:49

POTL wrote: ↑11 Feb 2023, 20:16 The scheme is good because it does not require identical jfets. But, as far as I know, it can have clipping problems, just like jfets. It seems to me that it is more reasonable to replace CMOS and Jfets with a photoresistor + LED. You could try using OTA, but not like in the EHX Small Stone circuit (I think OTA ruins the sound), but you can use OTA as a variable resistor (idea from the Engineers thumb compressor, not tested in practice).

Matching can be as much as a problem with LDRs as it is for FETs. OTA phasers definitely exist (DOD for example), and they work well, but they're awful in terms of noise and distortion isn't better than using FETs either. As I said in the "distortion" paragraph, I found it to be more than acceptable. I take the chance to link the mentioned app note since I forgot to in the original post https://www.onsemi.cn/pub/collateral/an-6603cn.pdf
Mosfets will always perform worse than JFETs in this terms, and JFETs fall short of LDR purely in terms of distortion even if they're good already. For guitar levels and this kind of effect, the benefit of not having to match more than offsets the distortion concerns.

Speaking of OTA, I mean a different approach. it definitely should not be approached like standard OTA phasers, EHX, DOD or EQD style. Imagine a phase 90 scheme in which jfet is replaced by OTA. the signal goes through conventional op-amps, OTA changes the filter.

POTL wrote: ↑18 Feb 2023, 20:24 Speaking of OTA, I mean a different approach. it definitely should not be approached like standard OTA phasers, EHX, DOD or EQD style. Imagine a phase 90 scheme in which jfet is replaced by OTA. the signal goes through conventional op-amps, OTA changes the filter.

I see how it could be possible, but I can't say if it will be more linear (since the same OTA input voltage considerations still apply) or quieter (since the OTA will output directly in the op-amp) than using the OTA traditionally. It sure won't be very compact.

Thanks @dylan159 I’m always impressed with your work! It’s always thorough and inspired.

Looking at the first two stage circuit I wondered about using the four left over inverts. Might be possible to use for an input or output buffer, maybe both, and two more for the LFO? Obviously is less than ideal but makes an interesting challenge of building a phaser with a single 4049.

@dylan159 I’m looking 2 stage pup, should the input or output of the unused inverters be tied to Vcc or vdd?

@dylan159, one more question, I notice on the two stage pup vss of the 4069 is connected to Vr instead of vcc? Maybe I missed this if you mentioned it in the description.

soggybag wrote: ↑23 Feb 2023, 22:20 Thanks @dylan159 I’m always impressed with your work! It’s always thorough and inspired.

Looking at the first two stage circuit I wondered about using the four left over inverts. Might be possible to use for an input or output buffer, maybe both, and two more for the LFO? Obviously is less than ideal but makes an interesting challenge of building a phaser with a single 4049.

This and the following messages all tie into the same answer.
The inverters aren't used in a conventional way, but as an array of matched N channel fets. That means the P channel is unused and not even connected to the supply. that's why you don't need to terminate the unused sections.
The sources are connected to Vref because the VCRs are DC coupled. AC coupling or buffering Vref are two things you could try to have them connected to something more useful for other purposes. I'll leave that to you, but inverter buffers? Oh no please no I can hear the noise from here.

good work, wondering if someone here could take the time to stick a 1khz tone into one of these phasors and state at what (input) amplitude the output signal starts to clip (at any time during the cycle with oscillator running) ?! some measure of headroom or an idea of non-linearity profile would answer an important question many have about these nMOS VCR's ...

Hotrats wrote: ↑28 Feb 2023, 21:56 good work, wondering if someone here could take the time to stick a 1khz tone into one of these phasors and state at what (input) amplitude the output signal starts to clip (at any time during the cycle with oscillator running) ?! some measure of headroom or an idea of non-linearity profile would answer an important question many have about these nMOS VCR's ...

I've done this kind of testing with an isolated VCA stage as detailed, which you can try to reproduce. I think it's more representative that way, if we accept that the in-circuit signal isn't much different (save from the PUP6 attenuation). I did try to get some numbers at certain amplitudes and gate voltages, then did the more useful test in this case, which is to hear if it sounds distorted or not.
The 6 featured input attenuation, which was a sacrifice aimed at reducing distortion which started to add up. In all cases my verdict is "Perfectly Usable", and if there's a bit left, it adds to the character of an effect in the strict sense. It might be different for other applications and I might be more rigorous there.

Hey Lani,

Thanks for the great work (and thank you Dylan159)!
I'm looking forward to your Gerber files.

Izhar

Lani wrote: ↑16 Feb 2023, 16:55 Thought I would post my layout here as well. I'll post the Gerber file once I receive mine and it is verified. The extra footprints are for caps on each of the phase stages to experiment with different values.
The reason behind the dual op amps instead of quads is I have a full reel of LF353's. Also, I usually find it easier to route PCBs with duals. Quad packages get too crowded unless you go SMD. This would be a tough one to etch (double sided) though not impossible. Etch files, schematic, 3D renders are in the PDF below.
Enjoy!!!
PUP - LITTER OF 6 - 3D RENDER - 1.22.23 -TOP (WH).png