The FM Tube Modulator Jimi Loved
Moderators: VelvetGeorge, BUG
-
daveweyer
- Senior Member
- Posts: 713
- Joined: Wed Oct 29, 2014 9:36 pm
- Just the numbers in order: 13492
Re: The FM Tube Modulator Jimi Loved
Just a side note, the added capacitor under the circuit board on the wah was a .003. You need that to get the right range with a 530mh inductor. I added that cap across the .01 Paktron film cap. Believe it or not, a ceramic cap worked best. The sound was so juicy.
-
Tek465b
- Senior Member
- Posts: 210
- Joined: Sat Jan 10, 2015 9:43 am
- Just the numbers in order: 13492
- Contact:
Re: The FM Tube Modulator Jimi Loved
Ok a little bit off topic here but since we are talkinga bout the wah.
I did some math....
(my goal here is to achieve the same resonnant frequancy and Q/bandwidth and damping/attenuation with the 500mH inductor and a 15nF capacitor)
If somebody want to try a close equivalent by changing the capacitor and resistor (instead of inductor/capacitor)
here is my result.
with the stock 500mh / 10n /33k
We get and angular frequency = 14 142
in hz = 2.250 khz
Quality factor = 4.666
angular bandwidth = 3 030
in Hz = 482.287
Damping = 0.107
Attenuation = 1 515
Now with your 530mH inductor and 13n, we get a sharper peak(more Q) and a lower resonant frequency.
530mH / 13n / 33k
angular frq = 12 047
in hz = 1.917 khz
Q = 5.168
Bandwidth in hZ = 370.990hz
Damping = 0.096
Attenuation = 1 165
And now with a setting that yield similar result as your 530/13n
500mH / 15n / 30k
angular frq = 11 547
in Hz = 1.837khz
Q = 5.196
Bandwidth in Hz = 353.677
Damping = 0.096
Attenuation = 1 111
That look darn close to the 13nF and 530mH
Resonant frequency is a bit lower (80hz)
damping is the same, Q and bandwidth and attenuation are darn close.
Now if we compare with the mod everyone usually do
500mH / 15n / 33k
angular FRQ = 11 547
in Hz = 1.837kHz
Q = 5.715
Bandwidth in Hz = 321.525
Damping = 0.087
Attenuation = 1 010
We can see that with the standard mod of 500mH / 15n / 33k
we get the same resonant frequency that my mod and close to your mod (80Hz lower)
But the Q(quality factor) is higher wich in turn make for a narrower Bandwidth, less damping and attenuation = sharp peak.
So replacing the 33k with a 30k and a 15nF/500mH get those last 3 parameter closer to the 530mH / 13n / 33k value
In fact if we use a potentiometer/trimmer instead of the 30k resistor and tune it around 29kOhm we will get pretty much the same exact value for all parameters except for the resonant frequency. -80Hz (if we factor 10% component tolerance that is next to nothing)
I did some math....
(my goal here is to achieve the same resonnant frequancy and Q/bandwidth and damping/attenuation with the 500mH inductor and a 15nF capacitor)
If somebody want to try a close equivalent by changing the capacitor and resistor (instead of inductor/capacitor)
here is my result.
with the stock 500mh / 10n /33k
We get and angular frequency = 14 142
in hz = 2.250 khz
Quality factor = 4.666
angular bandwidth = 3 030
in Hz = 482.287
Damping = 0.107
Attenuation = 1 515
Now with your 530mH inductor and 13n, we get a sharper peak(more Q) and a lower resonant frequency.
530mH / 13n / 33k
angular frq = 12 047
in hz = 1.917 khz
Q = 5.168
Bandwidth in hZ = 370.990hz
Damping = 0.096
Attenuation = 1 165
And now with a setting that yield similar result as your 530/13n
500mH / 15n / 30k
angular frq = 11 547
in Hz = 1.837khz
Q = 5.196
Bandwidth in Hz = 353.677
Damping = 0.096
Attenuation = 1 111
That look darn close to the 13nF and 530mH
Resonant frequency is a bit lower (80hz)
damping is the same, Q and bandwidth and attenuation are darn close.
Now if we compare with the mod everyone usually do
500mH / 15n / 33k
angular FRQ = 11 547
in Hz = 1.837kHz
Q = 5.715
Bandwidth in Hz = 321.525
Damping = 0.087
Attenuation = 1 010
We can see that with the standard mod of 500mH / 15n / 33k
we get the same resonant frequency that my mod and close to your mod (80Hz lower)
But the Q(quality factor) is higher wich in turn make for a narrower Bandwidth, less damping and attenuation = sharp peak.
So replacing the 33k with a 30k and a 15nF/500mH get those last 3 parameter closer to the 530mH / 13n / 33k value
In fact if we use a potentiometer/trimmer instead of the 30k resistor and tune it around 29kOhm we will get pretty much the same exact value for all parameters except for the resonant frequency. -80Hz (if we factor 10% component tolerance that is next to nothing)
-
daveweyer
- Senior Member
- Posts: 713
- Joined: Wed Oct 29, 2014 9:36 pm
- Just the numbers in order: 13492
Re: The FM Tube Modulator Jimi Loved
Tek, nice analysis.
Just keep in mind that components make a huge difference in the sound; if you you use Mullard Tropical fish caps, it won't sound exactly like the circuit with film and foil Paktron caps. The inductor ferrite makes a big difference too. So if you match the Q and bandwidth exactly, you may not have the same pedal.
As I have tried to explain here, Jimi liked a darker band, namely because he played a lot of tunes with the circuit engaged but only moved the pedal to hover around the bandwidth center of the strings he was playing. He didn't want the peak too sharp, obviously, and some wahs have a very sharp peak which makes them sound honky. I first used a Triad 500mh toroid, but its peak was too broad; the cup inductors had higher Q and the TDK was just about right.
Geoffrey Teese, the ultimate wah builder, ended up with almost the identical pedal I made Jimi, all through his own research many years later. For Jimi stuff, those parameters are pretty close.
But you have made a nice analysis for all to see, and I think it helps to explain the sound of the pedal I made Jimi. All those interested should copy and paste these figures into their notebooks for future reference.
The bandwidth and Q of tone shaping circuits was a big part of the engineering at Thomas Organ, because that form of subtractive synthesis was central to the voicing of their organs; the wah circuit originally was lifted out of a Thomas Organ, and then made variable with the feedback circuit devised by Brad Plunkett. Thomas tried all kinds of inductors but found the Q of the cup inductors most suitable for their organ voicing circuits.
In your next analysis, check the load resistance on the collector of Q1. If you make that collector provide much ac current to the load, the variable sweep bandwidth collapses. When I plugged the circuit into the three transistor fuzz, and took out the wah output resistor between the circuits, I could adjust the wah sweep with the input control on the fuzz circuit. If I turned it down to zero ohms, the wah would sweep to its highest frequency and remain there no matter which position the pedal pot was in. (The three transistor fuzz had to be at the low gain setting, i.e. maximum negative feedback.)
That's because the three transistor fuzz has a lot of negative feedback at the base of Q1, meaning the input impedance is very low. The wah circuit does not have the current to drive a low impedance and maintain its sweep range. That is one reason I used an output resistor on the wah.
The greatest sweep range is had by adding a source follower FET on the output of Q1, something that might be done on a revised model, something I would have done if Jimi had lived.
To Thomas it didn't matter because they expected the wah to be plugged into amps, high impedance inputs.
Just keep in mind that components make a huge difference in the sound; if you you use Mullard Tropical fish caps, it won't sound exactly like the circuit with film and foil Paktron caps. The inductor ferrite makes a big difference too. So if you match the Q and bandwidth exactly, you may not have the same pedal.
As I have tried to explain here, Jimi liked a darker band, namely because he played a lot of tunes with the circuit engaged but only moved the pedal to hover around the bandwidth center of the strings he was playing. He didn't want the peak too sharp, obviously, and some wahs have a very sharp peak which makes them sound honky. I first used a Triad 500mh toroid, but its peak was too broad; the cup inductors had higher Q and the TDK was just about right.
Geoffrey Teese, the ultimate wah builder, ended up with almost the identical pedal I made Jimi, all through his own research many years later. For Jimi stuff, those parameters are pretty close.
But you have made a nice analysis for all to see, and I think it helps to explain the sound of the pedal I made Jimi. All those interested should copy and paste these figures into their notebooks for future reference.
The bandwidth and Q of tone shaping circuits was a big part of the engineering at Thomas Organ, because that form of subtractive synthesis was central to the voicing of their organs; the wah circuit originally was lifted out of a Thomas Organ, and then made variable with the feedback circuit devised by Brad Plunkett. Thomas tried all kinds of inductors but found the Q of the cup inductors most suitable for their organ voicing circuits.
In your next analysis, check the load resistance on the collector of Q1. If you make that collector provide much ac current to the load, the variable sweep bandwidth collapses. When I plugged the circuit into the three transistor fuzz, and took out the wah output resistor between the circuits, I could adjust the wah sweep with the input control on the fuzz circuit. If I turned it down to zero ohms, the wah would sweep to its highest frequency and remain there no matter which position the pedal pot was in. (The three transistor fuzz had to be at the low gain setting, i.e. maximum negative feedback.)
That's because the three transistor fuzz has a lot of negative feedback at the base of Q1, meaning the input impedance is very low. The wah circuit does not have the current to drive a low impedance and maintain its sweep range. That is one reason I used an output resistor on the wah.
The greatest sweep range is had by adding a source follower FET on the output of Q1, something that might be done on a revised model, something I would have done if Jimi had lived.
To Thomas it didn't matter because they expected the wah to be plugged into amps, high impedance inputs.
-
Tek465b
- Senior Member
- Posts: 210
- Joined: Sat Jan 10, 2015 9:43 am
- Just the numbers in order: 13492
- Contact:
Re: The FM Tube Modulator Jimi Loved
Yes of course the component selection will make a difference.
Just the coil winding resistance alone has a big effect on the Q of the coil itself. If that was a LC circuit (without the 33k R) the Q and attenuation could be heavily affected by just 0.1 ohm difference from the coil. since we use RLC the effect is minimum. There are other parameter, but inductor always been quite a mystery for me.
Darker band and less sharp peak totally confirm what i expected and i think its a really good idea
. Now its time to warmup my soldering iron 
.
About the fuzz/wah.
The way i see it is simple, the input impedance of the Fuzz get in parallel with the Wah's Collector resistor.
This turn the Voltage gain wayy down thus cancelling the miller effect (can't make the capacitor look bigger, no more wah sweep).
We still get sound going thru because of the current drive.
This is why i made a Ge input buffer at the input of my fuzz (on a bypass switch of course, with a 12k between Buff and fuzz input to control gain), also it seem counter intuitive, but for BOG sound, it to do the trick for me.. strangelly. (BOG was never my goal, it was accidental, woodstock always been my favority tone)
Just the coil winding resistance alone has a big effect on the Q of the coil itself. If that was a LC circuit (without the 33k R) the Q and attenuation could be heavily affected by just 0.1 ohm difference from the coil. since we use RLC the effect is minimum. There are other parameter, but inductor always been quite a mystery for me.
Darker band and less sharp peak totally confirm what i expected and i think its a really good idea
. Now its time to warmup my soldering iron .About the fuzz/wah.
The way i see it is simple, the input impedance of the Fuzz get in parallel with the Wah's Collector resistor.
This turn the Voltage gain wayy down thus cancelling the miller effect (can't make the capacitor look bigger, no more wah sweep).
We still get sound going thru because of the current drive.
This is why i made a Ge input buffer at the input of my fuzz (on a bypass switch of course, with a 12k between Buff and fuzz input to control gain), also it seem counter intuitive, but for BOG sound, it to do the trick for me.. strangelly. (BOG was never my goal, it was accidental, woodstock always been my favority tone)
-
daveweyer
- Senior Member
- Posts: 713
- Joined: Wed Oct 29, 2014 9:36 pm
- Just the numbers in order: 13492
Re: The FM Tube Modulator Jimi Loved
Roger made a fuzz with an extra input transistor, I saw it while working on Jimi's fuzz boxes at West Coast. It's logical considering how these effects interact. Personally I liked the interaction on the base of Q1 of the fuzz, because it made things sound different depending on the amount of negative feedback and the input source. It allowed the guitar to control the gain of the fuzz like a remote volume control.
This only works if the guitar is plugged directly into the fuzz; the current in the pickup coils can overcome the current in the negative feedback loop causing the gain to increase, R1/R2.
Putting an input resistor or device on the fuzz can eliminate the interaction, good if using a wah first.
This only works if the guitar is plugged directly into the fuzz; the current in the pickup coils can overcome the current in the negative feedback loop causing the gain to increase, R1/R2.
Putting an input resistor or device on the fuzz can eliminate the interaction, good if using a wah first.
-
Eb7+9
- Senior Member
- Posts: 105
- Joined: Wed Mar 09, 2011 1:30 pm
- Just the numbers in order: 13492
- Contact:
Re: The FM Tube Modulator Jimi Loved
I got around the classic WAH/Fuz-Face "boundary" issue by designing an op-amp simulation of the original two-transistor circuit, which works with same value components, incl. the inductor and pot ... this is done simply by using the basic small-signal parameters that can be derived from SPICE sims on the original gain stages (1st year analogue circuit theory stuff)
I avoided making the mistake I saw in a couple of other (DIY) op-amp wah designs ... works perfectly and it can be "tuned" via trimpots, ie., instantly tweak-able with a small screw driver ... moreover, I can tweak parametric circuit values without altering bias (which NFB keeps at mid-rail or ground) ... in the original WAH's altering bias meant altering headroom // it's a juggling act as everybody knows ... my op-amp WAH runs at +/- 7.5 volts, so there's all the headroom in the world (if you like that) ... that aspect is esp. useful on Bass and Keys, or the output of a preamp ... which, in theory, can help obtain better signal/noise
in my Ge FF I use a slight variant of the classic circuit, one that avoids bias issues and blather/gating ... and with op-amps I can take advantage of the high drive capability of this circuit to run the input of a fuzz-face without issue, as long as I come off the op-amp directly - and not the WAH pot
what's interesting in all this for me is that the same mods done to the classic WAH classic circuit can be made in my op-amp "simulation" WAH as well, and obtain analogous results ...
I'll have an update on the 4-channel CTP shortly ...
~jcm
I avoided making the mistake I saw in a couple of other (DIY) op-amp wah designs ... works perfectly and it can be "tuned" via trimpots, ie., instantly tweak-able with a small screw driver ... moreover, I can tweak parametric circuit values without altering bias (which NFB keeps at mid-rail or ground) ... in the original WAH's altering bias meant altering headroom // it's a juggling act as everybody knows ... my op-amp WAH runs at +/- 7.5 volts, so there's all the headroom in the world (if you like that) ... that aspect is esp. useful on Bass and Keys, or the output of a preamp ... which, in theory, can help obtain better signal/noise
in my Ge FF I use a slight variant of the classic circuit, one that avoids bias issues and blather/gating ... and with op-amps I can take advantage of the high drive capability of this circuit to run the input of a fuzz-face without issue, as long as I come off the op-amp directly - and not the WAH pot
what's interesting in all this for me is that the same mods done to the classic WAH classic circuit can be made in my op-amp "simulation" WAH as well, and obtain analogous results ...
I'll have an update on the 4-channel CTP shortly ...
~jcm
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/
-
Eb7+9
- Senior Member
- Posts: 105
- Joined: Wed Mar 09, 2011 1:30 pm
- Just the numbers in order: 13492
- Contact:
Re: The FM Tube Modulator Jimi Loved
took the plunge, ... I'm about half way across now
biggest and most expensive PCB I've had made so far
http://www.lynx.net/~jc/CTP4pcb.jpg
http://www.lynx.net/~jc/CTP4corner.jpg
following the Baldwin CTP structure very closely
with a few modernizations:
as it stands,
* 4 independently-variable channels ...
incl. OFFSET, DEPTH and SPEED control on each 3-phase oscillator board
so, four sets of oscillator controls
* each channel also has a tank-cut switch, so the Band-Pass function can be inserted, or not ...
recall, the CTP was designed for synchronous "tone-wheel derived" signals
and so we (guitarists) may want to see what happens when we run each channel full-band instead
* each channel will also have its own GAIN control and "kill" switch
* the whole works has a Master Volume control
same tank frequencies as the Baldwin schematic
less lossy summing at the output (should be interesting ...)
picture shows results after two solid days of soldering
http://www.lynx.net/~jc/CTP4partway1.jpg
so far, I've got temporary LED's on the under side, to show PWM action
later, these LED's will be moved to the front panel
* each channel will have it's own trio of rotating "LFO" LED's to show status
http://www.lynx.net/~jc/CTP4traces.jpg
that side of it now works ... on all four Duino's
checked all the voltages, +/-9 volt everywhere required, etc ...
5volts regulators obviously doing their job
checked out one of the converter circuit without OTA's in their sockets
by adjusting the master bias trimpot I was able to get
a 0.06mA to 0.88mA spread in OTA current
by sweeping the OFFSET control end to end for that LFO
that's roughly 5% to 90% volume variation in the OTA at (close to) full depth
I'll take it ...!!
should provide ample separation between modulated signals A, B, and C
---
next, I need to snake the A, B and C signals from the lattice to the modulators
same order as they are shown in the Baldwin schematic (ie., orig. patent)
hook up, the kill switches, GAIN pots and MV pot
couple of wires ...
and I should be able to hear it in action
hopefully I can get around to doing that this weekend
although, I have a couple of preamps I have to finish first
should have confirmation soon enough ...
can't wait
---
in the mean time if anybody can recommend a source of 1U and 2U rack cases
I'd appreciate it ... I can't seem to find the last guys I used
I'm looking for an outfit that can supply brushed Alu front panels
~jcm
biggest and most expensive PCB I've had made so far
http://www.lynx.net/~jc/CTP4pcb.jpg
http://www.lynx.net/~jc/CTP4corner.jpg
following the Baldwin CTP structure very closely
with a few modernizations:
as it stands,
* 4 independently-variable channels ...
incl. OFFSET, DEPTH and SPEED control on each 3-phase oscillator board
so, four sets of oscillator controls
* each channel also has a tank-cut switch, so the Band-Pass function can be inserted, or not ...
recall, the CTP was designed for synchronous "tone-wheel derived" signals
and so we (guitarists) may want to see what happens when we run each channel full-band instead
* each channel will also have its own GAIN control and "kill" switch
* the whole works has a Master Volume control
same tank frequencies as the Baldwin schematic
less lossy summing at the output (should be interesting ...)
picture shows results after two solid days of soldering
http://www.lynx.net/~jc/CTP4partway1.jpg
so far, I've got temporary LED's on the under side, to show PWM action
later, these LED's will be moved to the front panel
* each channel will have it's own trio of rotating "LFO" LED's to show status
http://www.lynx.net/~jc/CTP4traces.jpg
that side of it now works ... on all four Duino's
checked all the voltages, +/-9 volt everywhere required, etc ...
5volts regulators obviously doing their job
checked out one of the converter circuit without OTA's in their sockets
by adjusting the master bias trimpot I was able to get
a 0.06mA to 0.88mA spread in OTA current
by sweeping the OFFSET control end to end for that LFO
that's roughly 5% to 90% volume variation in the OTA at (close to) full depth
I'll take it ...!!
should provide ample separation between modulated signals A, B, and C
---
next, I need to snake the A, B and C signals from the lattice to the modulators
same order as they are shown in the Baldwin schematic (ie., orig. patent)
hook up, the kill switches, GAIN pots and MV pot
couple of wires ...
and I should be able to hear it in action
hopefully I can get around to doing that this weekend
although, I have a couple of preamps I have to finish first
should have confirmation soon enough ...
can't wait
---
in the mean time if anybody can recommend a source of 1U and 2U rack cases
I'd appreciate it ... I can't seem to find the last guys I used
I'm looking for an outfit that can supply brushed Alu front panels
~jcm
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/
-
Xplorer
- Senior Member
- Posts: 2477
- Joined: Sun Apr 19, 2009 5:27 pm
- Just the numbers in order: 7
Re: The FM Tube Modulator Jimi Loved
WOW !!! 
what a work !!! that's some serious pcb here !
what a work !!! that's some serious pcb here !-
daveweyer
- Senior Member
- Posts: 713
- Joined: Wed Oct 29, 2014 9:36 pm
- Just the numbers in order: 13492
Re: The FM Tube Modulator Jimi Loved
Just cracks me up; I mean, to what lengths will we go for achieving altered guitar sounds? I love it.
Everybody cheer for JC.
Everybody cheer for JC.
-
Eb7-9
- New Member
- Posts: 4
- Joined: Sun Feb 25, 2018 9:41 pm
- Just the numbers in order: 13492
Re: The FM Tube Modulator Jimi Loved
hey y'all ...
it's been an interesting yr // I had to put down the CTP4 project after finding an un-foreseen bottleneck
in my last go-for-gold PCB ... haha, doesn't always turn out as planned
I tried fixing stuff ...
turns out I hadn't considered my current distribution ckt not working right when it came time to trim OTA's for offsets
yada yada ...
I've learned a few thing since then, and in fact busy applying pwm ideas elsewhere in the mean time ...
in some cases I'm avoiding OTA's in the process ... not that they're a totally poor choice for the job
but for the sake of operating at pro signal levels it's easier to go the op-amp route
op-amps also interfaces well with CMOS bi-directional (analogue) gates ... so do OTA's (just to be clear)
which makes op-amps equally well suited in the mixer section //
this means we can have a +/- 9volt rail pair accommodate up to 15 vpp signals throughout
targetting use in un-buffered FX loops and sends (either mono or stereo settings)
since the last time I've up with a more suitable comparator arrangement
or rather, a simple way of dealing with saturation drop-out
converting 0-5v to +/- 9v at high speed for example
when I get the chance I'm gonna shoe-horn a simple single-channel AVR based unit in a 1590BB box ...
will work the same as a chorus/vibrato, but take 15vpp signals ... so, providing low noise chorusing yada etc
the CTP4 is a wonderful dream ... but I think someone else can tackle that one day
ciao for now
take care y'all ...
~jcm
it's been an interesting yr // I had to put down the CTP4 project after finding an un-foreseen bottleneck
in my last go-for-gold PCB ... haha, doesn't always turn out as planned
I tried fixing stuff ...
turns out I hadn't considered my current distribution ckt not working right when it came time to trim OTA's for offsets
yada yada ...
I've learned a few thing since then, and in fact busy applying pwm ideas elsewhere in the mean time ...
in some cases I'm avoiding OTA's in the process ... not that they're a totally poor choice for the job
but for the sake of operating at pro signal levels it's easier to go the op-amp route
op-amps also interfaces well with CMOS bi-directional (analogue) gates ... so do OTA's (just to be clear)
which makes op-amps equally well suited in the mixer section //
this means we can have a +/- 9volt rail pair accommodate up to 15 vpp signals throughout
targetting use in un-buffered FX loops and sends (either mono or stereo settings)
since the last time I've up with a more suitable comparator arrangement
or rather, a simple way of dealing with saturation drop-out
converting 0-5v to +/- 9v at high speed for example
when I get the chance I'm gonna shoe-horn a simple single-channel AVR based unit in a 1590BB box ...
will work the same as a chorus/vibrato, but take 15vpp signals ... so, providing low noise chorusing yada etc
the CTP4 is a wonderful dream ... but I think someone else can tackle that one day
ciao for now
take care y'all ...
~jcm
-
Eb7-9
- New Member
- Posts: 4
- Joined: Sun Feb 25, 2018 9:41 pm
- Just the numbers in order: 13492
Re: The FM Tube Modulator Jimi Loved
quick update,
it's been an hectic 18 months since we last blabbed
I've been developing a bunch of new FX circuits ... incl. for example a couple of new op-amp based low-distortion / high-drive WAH's ... bunch of Limiter circuits, incl a modded 1176 clone for bass/gtr ... and a bunch of weird original sh*t
the CTP project is one of those things I just can't stop thinking about ...
Indeed, my last attempt tried to embody the original 4-channel unit using AVR's as source of 3-phase clocks
thinking I could just walk up to the plate and hit a grand-slam
turns out, the CTP-4 that was pretty much it as far as my going thru a phase of exploring digital-PWM designs
State-Variable WAH Pedal Filter, Mono-Stereo Phasors and the CTP efforts where mainly all I did with custom AVR technology with code generously provided by Tek456 ... I also used an Electric-Druide Tap-Tempo IC on one project ... wanting to see if I could in any way get away from the stale feel of purely digital controlled analogue effecs
though not entirely futile, my few digital PWM efforts it gave me a chance to estimate the effective range of resistance derived from cmos analogue gates (ie., CD4016 and CD4066 gates) ... yielding a range of about 440ohms to 440kohms AC when running on an 8-bit AVR platform (ie., with 0.39% resolution)
but shortly after my CTP4 mess, I abandoned digital-PWM altogether (not to mention Tek456 wasn't feeling well and couldn't code for me anymore) so, I started exploring Analogue derived PWM signals for a while - for example, encoding Univibe LFO signals to PWM in order to bring in some life into the PWM picture ... I'm glad I did it, but unfortunately my first effort to operate floating analogue gates in a Univibe signal path proved to be problematic ... it worked, but the noise (I suspect from charge injecting thru the gate capacitance) made the effect un-usable ... still, I had to try
I have now have moved on from PWM-gate resistor techniques altogether since ... was fun exploring
I took another detour to do a reversing/cloning of all the EMS-VCS3 synthesizer circuitry ...
more stuff sitting on the back burner
suddenly began obsessing over getting a variable speed 3-phase triangle output oscillator working, pref. of some sort not digital (a stringent goal) ... with absolutely nothing in the literature ... and around Feb of last year stumbled on the idea of generating a current-controlled ring-oscillator using three OTA's to produce triangle wave output ... musta spent two months chasing my tail on that one before quitting ... the thing worked, got lots of beautiful plots (on the simulator) but I couldn't figure out how to stabilize the signal swing limits over long periods of time // interesting yet frustrating ...
didn't get around to bread-boarding any of it // no cigar ...
then, earlier this summer - amidst a bunch of new design stuff - I stumbled on the idea of using digital counters running of an analogue CCO to produce 3-phase overlapping outputs ... my original aim was to produce triangle-like output but after remembering what Dave said about the step-like nature of the original tube circuit I decided to let go of my "triangle output" goal (maybe) and settle with plain pulses yielding a potentially simple and compact solution to a simple single-channel unit ... similar in size to my earlier opto-AVR version ... I do have a way of converting to triangle-lie output but adds a bit of complexity to the design ... something possibly worth exploring if I'm not happy with stepper signals
the idea I'm toying with these days is a 6-state ring circuit, continuously speed controlled, all in a single-channel circuit, with variable intensity control ... both versions fitting in a 1590BB style box ... because the choice came up, I have one version doing the mixing via opto-couplers and the other thru analogue gates ... I need to build both and compare performance ... otherwise I have a circuit for converting pulses to triangle that would require three more OTA's, making the circuit a little larger and a bit more finicky to adjust ... besides, multiple OTA's is what sunk the CTP4
it's still free exploration at this point
my next step puts the "stepper" aspect of the theory to the test ...
board files have been sent to the pcb makers for both versions,
just a matter of scoring shop/study space again
thanks for watching
yO!
it's been an hectic 18 months since we last blabbed
I've been developing a bunch of new FX circuits ... incl. for example a couple of new op-amp based low-distortion / high-drive WAH's ... bunch of Limiter circuits, incl a modded 1176 clone for bass/gtr ... and a bunch of weird original sh*t
the CTP project is one of those things I just can't stop thinking about ...
Indeed, my last attempt tried to embody the original 4-channel unit using AVR's as source of 3-phase clocks
thinking I could just walk up to the plate and hit a grand-slam
turns out, the CTP-4 that was pretty much it as far as my going thru a phase of exploring digital-PWM designs
State-Variable WAH Pedal Filter, Mono-Stereo Phasors and the CTP efforts where mainly all I did with custom AVR technology with code generously provided by Tek456 ... I also used an Electric-Druide Tap-Tempo IC on one project ... wanting to see if I could in any way get away from the stale feel of purely digital controlled analogue effecs
though not entirely futile, my few digital PWM efforts it gave me a chance to estimate the effective range of resistance derived from cmos analogue gates (ie., CD4016 and CD4066 gates) ... yielding a range of about 440ohms to 440kohms AC when running on an 8-bit AVR platform (ie., with 0.39% resolution)
but shortly after my CTP4 mess, I abandoned digital-PWM altogether (not to mention Tek456 wasn't feeling well and couldn't code for me anymore) so, I started exploring Analogue derived PWM signals for a while - for example, encoding Univibe LFO signals to PWM in order to bring in some life into the PWM picture ... I'm glad I did it, but unfortunately my first effort to operate floating analogue gates in a Univibe signal path proved to be problematic ... it worked, but the noise (I suspect from charge injecting thru the gate capacitance) made the effect un-usable ... still, I had to try
I have now have moved on from PWM-gate resistor techniques altogether since ... was fun exploring
I took another detour to do a reversing/cloning of all the EMS-VCS3 synthesizer circuitry ...
more stuff sitting on the back burner
suddenly began obsessing over getting a variable speed 3-phase triangle output oscillator working, pref. of some sort not digital (a stringent goal) ... with absolutely nothing in the literature ... and around Feb of last year stumbled on the idea of generating a current-controlled ring-oscillator using three OTA's to produce triangle wave output ... musta spent two months chasing my tail on that one before quitting ... the thing worked, got lots of beautiful plots (on the simulator) but I couldn't figure out how to stabilize the signal swing limits over long periods of time // interesting yet frustrating ...
didn't get around to bread-boarding any of it // no cigar ...
then, earlier this summer - amidst a bunch of new design stuff - I stumbled on the idea of using digital counters running of an analogue CCO to produce 3-phase overlapping outputs ... my original aim was to produce triangle-like output but after remembering what Dave said about the step-like nature of the original tube circuit I decided to let go of my "triangle output" goal (maybe) and settle with plain pulses yielding a potentially simple and compact solution to a simple single-channel unit ... similar in size to my earlier opto-AVR version ... I do have a way of converting to triangle-lie output but adds a bit of complexity to the design ... something possibly worth exploring if I'm not happy with stepper signals
the idea I'm toying with these days is a 6-state ring circuit, continuously speed controlled, all in a single-channel circuit, with variable intensity control ... both versions fitting in a 1590BB style box ... because the choice came up, I have one version doing the mixing via opto-couplers and the other thru analogue gates ... I need to build both and compare performance ... otherwise I have a circuit for converting pulses to triangle that would require three more OTA's, making the circuit a little larger and a bit more finicky to adjust ... besides, multiple OTA's is what sunk the CTP4
it's still free exploration at this point
my next step puts the "stepper" aspect of the theory to the test ...
board files have been sent to the pcb makers for both versions,
just a matter of scoring shop/study space again
thanks for watching
yO!
-
Xplorer
- Senior Member
- Posts: 2477
- Joined: Sun Apr 19, 2009 5:27 pm
- Just the numbers in order: 7
Re: The FM Tube Modulator Jimi Loved
Hey ! great to read some news !
-
daveweyer
- Senior Member
- Posts: 713
- Joined: Wed Oct 29, 2014 9:36 pm
- Just the numbers in order: 13492
Re: The FM Tube Modulator Jimi Loved
Yeah, I'm glad to see this topic rescued from the bottom. This is an amazing thread. Oh, and while I'm at it, many blessings to Tek, for his wonderful contributions to this thread and the forum, and for his health.
The Hendrix Foundation is still interested in these Jimi projects, and intends to do something with them as circumstances allow. The phase shifter circuits are an area of great interest for them, as is the Wah and fuzz, and a few others we'll talk about if the forum doesn't croak in the mean time.
The Hendrix Foundation is still interested in these Jimi projects, and intends to do something with them as circumstances allow. The phase shifter circuits are an area of great interest for them, as is the Wah and fuzz, and a few others we'll talk about if the forum doesn't croak in the mean time.
-
capacitor
- New Member
- Posts: 1
- Joined: Wed May 20, 2020 3:28 am
- Just the numbers in order: 13492
Re: The FM Tube Modulator Jimi Loved
Hi, all! I hope this finds everyone healthy and safe.
I have one of the original Choratone units, rescued a bunch of years back. It's in good shape, and was probably fully operational at the time.
I'm interested in getting it running. I've done enough tube amp & electronics tech stuff to feel good about tackling it.
Is this a good thread to ask questions? There's a ton of great theory in here, and I don't want to clutter it up if not appropriate.
I have one of the original Choratone units, rescued a bunch of years back. It's in good shape, and was probably fully operational at the time.
I'm interested in getting it running. I've done enough tube amp & electronics tech stuff to feel good about tackling it.
Is this a good thread to ask questions? There's a ton of great theory in here, and I don't want to clutter it up if not appropriate.
-
Eb7+9
- Senior Member
- Posts: 105
- Joined: Wed Mar 09, 2011 1:30 pm
- Just the numbers in order: 13492
- Contact:
Re: The FM Tube Modulator Jimi Loved
Hi George, Dave and all ...
hope y'all hanging in there and doing ok
myself, I've been locked in the tool shed since early Dec working on my things ...
thought I'd give an update in the Wayne CTP dept. since I do think about it once in a while
---
I think I mentioned this here before, I took a sharp left turn in late 2017 to go back to the EMS VCS3 synth project
that I had started w Juergen Haible in 1995 - dunno, some strange urge came over me out of the blue ... in the course of about a year I had the whole synth sussed out including SPICE simulations of every module ... in the process of doing all this reversing I learned a few neat circuit tricks - some of which ended up as nucleus for a few of the designs shown later down below ... I needed to get away from the stuff I had been working on, including lots of Digital and some Analogue PWM circuits // ... for example I did a Univibe work-alike that runs in class-D at 140khz driving transmission gates as replacements for photo-cells ... that sorta stuff ... the last two CTP units were based on Digitally PWM'd gates as well, and felt I had milked that cow enough and needed to move on - but not before doing one last weird trial with them gates ...
as you might recall in the second phase of my CTP development (the first being the discrete jFET based analogues of the original tube modulators) I had some moderately good success using TEK465's custom Duino code to produce perfect three-phase triangle wave clock signals, embodied as digital PWM and driven into opto-couplers to perform signal mixing ... a second version doing the same but driving a bunch of bi-directional CD4066 gates worked as well, in the second case eliminating the need to match opto-couplers ... the opto-coupler version had a better feel in the end and was more compact and less wasteful from a hardware point-of-view // ... both of those Duino versions worked well
I still thought that an all-analog solution with no component matching should be possible
for about a month and a half in 2018 I toyed with the idea of wring up three OTA's as a ring oscillator, analogously to the original triode (and later jFET) circuit topology ... I started out by drawing out the 3-phase waveforms against each other and deriving conclusions about the waves' crossing points ... and went from there ... things worked but I had stability issues
... so, close !
but I had to give up as my intuition wasn't taking me to a solid place // I still think it might be a viable solution (I'm letting it rest back there for now) ...
then, sometime last year I thought of taking a radically different approach to producing three-phase triangle waveforms but starting out with a binary clock signal and running it through a Finite State Machine (FSM) to produce a three-phase binary signal set and then doing some kind of RC cheat and current-controlled amplitude stabilization to get sorta-triangle three phase signals ... but all that quickly started to look complicated with several sets of OTA's etc ...
In the process it gave me an idea of only going part-way with that transfer chain, providing a test platform so-to-speak to help show how well tuned or balanced the constant=phase shift lattice was (without relying on BODE transfer plots and subsequent analysis) ... rather, by taking the outputs of the FSM directly and driving transmission gates as either 0 or 1 multiplier we can produce the phasors sums A, A+B, B, B+C, C, and C+A (using original Choratone schematic nomenclature) ... if A and B are phasors at 120 and 240 then A+B is a phasor lying at 180, etc etc ... so we get six points on the full 360 degree rotation of the WAYNE shifter
notice this is NOT like a BBD (chorus/flanger) with fixed delay across frequencies of interest, NOR either like a phasor/vibrato, whose phase flips 180 on either side of a phase-transition area ... instead, the WAYNE shifter does a full 360 shift on ALL frequencies across the audio band and keeps going in the same advancing direction ... I put together a cheesy video clip that shows the circuit in action fed from a signal-generator :
https://vimeo.com/427820298
definitely an exciting vid eh ... what it shows is that the 120/240 phasors are where they should be, otherwise summed together to produce the 180 phasor component and that signal mixed/summed with the input signal wouldn't cancel so well against it if the 120/240 were off ... the test does show perfect cancellation - at least at that frequency ... I still need to do more testing to confirm how far across the audio band this holds
ok, so this turns out to be useful for tuning the lattice in producing an even output across all 360 degrees of modulation ... it's fairly easy to show (by invoking independence) that if the resultant 6 amplitudes are equal then they should be continuously for the full cycle when used in a linear scenario ... that was useful and surprising in itself // I was forced to design a three-phase FSM circuit and that in itself brought me back to my University (digital design torture) days ...at least now I have optimal values for the WAYNE Lattice circuit (*!*)
with this it seemed like that might well be the end of my CTP explorations ...
that was three of four weeks ago
and then late one night, looking back at everything I had done with all this, just suddenly realized that the jFET circuit I had played with in the beginning could be made speed variable after all - and in a very simple manner ... and NOT ONLY THAT !!
and so, if there's any light at the end of the tunnel I think it might be from here on ... // clearly, at some point I will be modding my two channel jFET CTP unit and mount the controller circuit in a wah shell and see how it goes ... I will post about it once I have confirmation
---
in the meantime, thanks for keeping the forum open George
I'm keeping as busy as I can until work gives the green light again ... as we speak I'm going through a phase of building some novel dirt box designs - something I've avoided in the past, but came up with a cool idea I just had to try (and panning out I'm pleased to report) ... I also got a bunch of original compressors on the go and a bunch of new stuff Flanger/Chorus/ and Univibe related ... I'm also coming up with my third high-headroom low noise parametric WAH circuit ...
a year later ... for those interested:
https://viva-analog.com/viva-analog-nw- ... -jcmc2019/
https://viva-analog.com/precision-shifter-jcmc2018/
https://viva-analog.com/viva-analog-sub-mux-jcmc-2019/
https://viva-analog.com/viva-analog-hard-mod-jcmc2019/
and more to come
as for amp work, well ... (heh heh) there's my new 60w Analog Tube Amp Killer ...
https://viva-analog.com/non-linear-anal ... amplifier/
I've jammed it with a drummer ... I have to say - I've played worse !!
cheers ...
~jcm
hope y'all hanging in there and doing ok
myself, I've been locked in the tool shed since early Dec working on my things ...
thought I'd give an update in the Wayne CTP dept. since I do think about it once in a while
---
I think I mentioned this here before, I took a sharp left turn in late 2017 to go back to the EMS VCS3 synth project
that I had started w Juergen Haible in 1995 - dunno, some strange urge came over me out of the blue ... in the course of about a year I had the whole synth sussed out including SPICE simulations of every module ... in the process of doing all this reversing I learned a few neat circuit tricks - some of which ended up as nucleus for a few of the designs shown later down below ... I needed to get away from the stuff I had been working on, including lots of Digital and some Analogue PWM circuits // ... for example I did a Univibe work-alike that runs in class-D at 140khz driving transmission gates as replacements for photo-cells ... that sorta stuff ... the last two CTP units were based on Digitally PWM'd gates as well, and felt I had milked that cow enough and needed to move on - but not before doing one last weird trial with them gates ...
as you might recall in the second phase of my CTP development (the first being the discrete jFET based analogues of the original tube modulators) I had some moderately good success using TEK465's custom Duino code to produce perfect three-phase triangle wave clock signals, embodied as digital PWM and driven into opto-couplers to perform signal mixing ... a second version doing the same but driving a bunch of bi-directional CD4066 gates worked as well, in the second case eliminating the need to match opto-couplers ... the opto-coupler version had a better feel in the end and was more compact and less wasteful from a hardware point-of-view // ... both of those Duino versions worked well
I still thought that an all-analog solution with no component matching should be possible
for about a month and a half in 2018 I toyed with the idea of wring up three OTA's as a ring oscillator, analogously to the original triode (and later jFET) circuit topology ... I started out by drawing out the 3-phase waveforms against each other and deriving conclusions about the waves' crossing points ... and went from there ... things worked but I had stability issues
... so, close !
but I had to give up as my intuition wasn't taking me to a solid place // I still think it might be a viable solution (I'm letting it rest back there for now) ...
then, sometime last year I thought of taking a radically different approach to producing three-phase triangle waveforms but starting out with a binary clock signal and running it through a Finite State Machine (FSM) to produce a three-phase binary signal set and then doing some kind of RC cheat and current-controlled amplitude stabilization to get sorta-triangle three phase signals ... but all that quickly started to look complicated with several sets of OTA's etc ...
In the process it gave me an idea of only going part-way with that transfer chain, providing a test platform so-to-speak to help show how well tuned or balanced the constant=phase shift lattice was (without relying on BODE transfer plots and subsequent analysis) ... rather, by taking the outputs of the FSM directly and driving transmission gates as either 0 or 1 multiplier we can produce the phasors sums A, A+B, B, B+C, C, and C+A (using original Choratone schematic nomenclature) ... if A and B are phasors at 120 and 240 then A+B is a phasor lying at 180, etc etc ... so we get six points on the full 360 degree rotation of the WAYNE shifter
notice this is NOT like a BBD (chorus/flanger) with fixed delay across frequencies of interest, NOR either like a phasor/vibrato, whose phase flips 180 on either side of a phase-transition area ... instead, the WAYNE shifter does a full 360 shift on ALL frequencies across the audio band and keeps going in the same advancing direction ... I put together a cheesy video clip that shows the circuit in action fed from a signal-generator :
https://vimeo.com/427820298
definitely an exciting vid eh ... what it shows is that the 120/240 phasors are where they should be, otherwise summed together to produce the 180 phasor component and that signal mixed/summed with the input signal wouldn't cancel so well against it if the 120/240 were off ... the test does show perfect cancellation - at least at that frequency ... I still need to do more testing to confirm how far across the audio band this holds
ok, so this turns out to be useful for tuning the lattice in producing an even output across all 360 degrees of modulation ... it's fairly easy to show (by invoking independence) that if the resultant 6 amplitudes are equal then they should be continuously for the full cycle when used in a linear scenario ... that was useful and surprising in itself // I was forced to design a three-phase FSM circuit and that in itself brought me back to my University (digital design torture) days ...at least now I have optimal values for the WAYNE Lattice circuit (*!*)
with this it seemed like that might well be the end of my CTP explorations ...
that was three of four weeks ago
and then late one night, looking back at everything I had done with all this, just suddenly realized that the jFET circuit I had played with in the beginning could be made speed variable after all - and in a very simple manner ... and NOT ONLY THAT !!
it was staring at us the whole time, and not only that but the same technique can be used to make the original tube circuit speed variable as well
and so, if there's any light at the end of the tunnel I think it might be from here on ... // clearly, at some point I will be modding my two channel jFET CTP unit and mount the controller circuit in a wah shell and see how it goes ... I will post about it once I have confirmation
---
in the meantime, thanks for keeping the forum open George
I'm keeping as busy as I can until work gives the green light again ... as we speak I'm going through a phase of building some novel dirt box designs - something I've avoided in the past, but came up with a cool idea I just had to try (and panning out I'm pleased to report) ... I also got a bunch of original compressors on the go and a bunch of new stuff Flanger/Chorus/ and Univibe related ... I'm also coming up with my third high-headroom low noise parametric WAH circuit ...
a year later ... for those interested:
https://viva-analog.com/viva-analog-nw- ... -jcmc2019/
https://viva-analog.com/precision-shifter-jcmc2018/
https://viva-analog.com/viva-analog-sub-mux-jcmc-2019/
https://viva-analog.com/viva-analog-hard-mod-jcmc2019/
and more to come
as for amp work, well ... (heh heh) there's my new 60w Analog Tube Amp Killer ...
https://viva-analog.com/non-linear-anal ... amplifier/
I've jammed it with a drummer ... I have to say - I've played worse !!
cheers ...
~jcm
Last edited by Eb7+9 on Sun Jun 14, 2020 10:02 am, edited 7 times in total.
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/