The FM Tube Modulator Jimi Loved

His guitar slung across his back, his dusty boots is his cadillac.

Moderators: VelvetGeorge, BUG

Post Reply
User avatar
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

Post by Eb7+9 » Mon Nov 09, 2015 3:51 pm

wow, that's a really cool idea Dave ...
the more ways we have of emulating the system, the better !

excuse me if I'm being a little adamant about it
but, just for my own edification, I'd like to be clear on how the system is defined, .. even if ideally (!)
and see what every (other) emulation option might be ... even if it leads to a new invention ... ;)

---

some background ...

Wayne's patent (the basis for the Baldwin "Chora-Tone Projector") provides much in terms of technical details

http://www.google.com/patents/US3004460

and, as you pointed out Dave (thx!), he refers to a much touted paper by Sydney Darlington

https://ia601003.us.archive.org/33/item ... 9-1-94.pdf

I'm pointing this out as I think there's a major descrepency between theory presented in the patent and how the Lattice is designed ... a point I want to explore a bit

to this end, further background might be useful ...
so, here's a list of patents referenced at the end of Wayne's patent:

http://www.google.ca/patents/US1773116 (Potter) ... Single side band system

http://www.google.ca/patents/US1964522 (Lewis) ... Phase control system

http://www.google.ca/patents/US2045796 (Plebanski) ... Method for radio transmission

http://www.google.ca/patents/US2117739 (Miller) ... Signaling system

http://www.google.ca/patents/US2230836 (Hammond) ... Musical instrument

http://www.google.ca/patents/US2408692 (Shore) ... Signaling system

http://www.google.ca/patents/US2559644 (Landon) ... Pulse multiplex system

http://www.google.ca/patents/US2566876 (Dome) ... Phase shift system

http://www.google.ca/patents/US2624041 (Evans) ... Amplitude modulator of the outphasing type

http://www.google.ca/patents/US2627549 (Kell) ... Bandwidth reduction

http://www.google.ca/patents/US2635226 (Harris) ... Phase modulation apparatus

http://www.google.ca/patents/US2660708 (Nakken) ... Frequency division

http://www.google.ca/patents/US2661458 (Saraga) ... Phase splitting network

http://www.google.ca/patents/US2677721 (Bedford) ... Color TV // ... 3-phase modulation

http://www.google.ca/patents/US2680153 (Boothroyd) ... Multichannel communication system

http://www.google.ca/patents/US2758204 (Norby) ... Wide-band phase shifter network

http://www.google.ca/patents/US2817711 (Feldman) ... Band Compression ...

http://www.google.ca/patents/US2857465 (Schroeder) ... Vocoder ...

http://www.google.ca/patents/US2905040 (Hanert) ... Chorus Effect ... Fig.6 shows a 5-octave system

if anybody spots the origins or Wayne's All-Pass lattice, please point it out ...!
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/

User avatar
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

Post by Eb7+9 » Mon Nov 09, 2015 4:21 pm

I gave Wayne's patent a more careful read and I have several comments, and questions ...
maybe we can discuss here

in the patent Wayne talks about taking the band-limited output of a phase-splitter gain stage (same signals but 180 deg out of phase) and running them thru an "All-Pass Lattice" that will maintain its two output at 120 degrees out-of-phase over a certain bandwidth ... this is the basis for referencing Darligton's paper (line 10, column 7)

the only problem, the frequency response is anything but all-pass ...
besides, you need a purely L-C network to do that,

see here: Synthesis of Constant Phase Networks (Kell)

http://www.highfrequencyelectronics.com ... Itemid=189

unfortunately, I'd have to brush up on the Filter Theory ... as its been 20+ yrs since i last studied that stuff

anyway, something to study and explore ...

the bottom line is, the lattice circuit that is provided in the patent, and seemingly used in the Baldwin units, are in fact Low-Pass outputs ... now, of course, it's 2015 and we have the advantage of using SPICE on the circuit // so we can derive our own conclusions ...

what I'm seeing is something that is resembling other Chorus/Vibrato circuits ... where a Bright/Dark signal is "progressively" being summed in anti-phase manner // that's why I'd like to hear a sample of the ChoraTone on a straight (electric) guitar chord strum (as it dies down, etc ...) ... so, we can hear what the effect does

my SPICE sim shows that the two outputs on the Wayne lattice remain exactly at 180 degrees, same as the inputs to the lattice ... and as far as I know, no R-C could ever produce a "near constant" 120 deg phase shift as he describes in the text ... that's my first point

plus Darlington's paper refers to a purely L-C lattice to produce the "near-constant phase difference over a bandwidth" response ... and, which involves negative inductances and capacitors // so, a purely theoretical concept at this point

how/why Wayne makes the jump from a "pure" L-C lattice to a "lossy" one I'm not sure ...
well, because it makes for a pleasant effect anyhow ... (I'm guessing)

incidentally, I've seen this type of lattice in other "tremulant" type circuits ...
see the Ampeg 435-S schematic for example :

http://bmamps.com/Schematics/ampeg/Ampe ... S_635S.jpg

---

what i found is this ...

instead of producing three phasors operating 120 deg apart
and turning like a 3-phase motor

he's got two that are 180 apart, and a third one (their sum) running at about 90 deg over a 100 Hz to 2k Hz bandwidth
so, in that sense I can see Darlingon's principle being applied on the sum output

but not, the way he describes it in the text ...

if I'm right, then it means that (yes) we would have phase-shifting happening ... but also "some" AM cancellations, occurring when the modulated gain on the 180 outputs are crossing at the same level (in a 3-phase system, other two phases meet where third is at its peak ...) ... whereas the purely balanced system would not (!)

sorry, not trying to sound like a killjoy, just pointing something out here ...
but it would seem that the Choratone unit is NOT working as the evenly rotating phasor system described in the patent

that's all ...
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/

User avatar
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

Post by Eb7+9 » Mon Nov 09, 2015 5:27 pm

I was curious about the R-C lattice and wanted to see what it was doing ...
so I drew up a simple SPICE simulation to replicate the conditions shown in the Baldwin schematic
but without tube distortion, etc ...

in the first case I'm using an idealized Phase Inverter followed by source load resistances of 82k on each node
in theory the Cathode impedance is much lower ... (so I have to redo these sims in a second case - to follow))

so, first assuming equal driving point impedances this is what i get
when I look at the two main outputs alone ...

the Phase plot shows two outputs 180 deg apart the whole way (as the source port is)

http://www.lynx.net/~jc/BaldwinAllpassL ... se180s.jpg

the Amplitude plot shows that they are NOT-All-pass, but rather Low-Pass ...
with the top port exhibiting more Treble ... so we could characterize these outputs
as Bright anti-phase LoPass, and the other Dark in-phase LoPass ...

http://www.lynx.net/~jc/BaldwinAllpassL ... deOuts.jpg

next, let's probe the summing point as shown in the patent diagram, and Baldwin schematic ...

http://www.lynx.net/~jc/BaldwinCT-Latti ... sponse.jpg

http://www.lynx.net/~jc/BaldwinCT-Latti ... sponse.jpg

the Phase response is telling ... it shows a near 90 degree phase shift being held between 100 Hz and 2kHz ...

but, this is the first (unlikely) 82k/82k equal driving point impedance case
Last edited by Eb7+9 on Mon Dec 07, 2015 9:46 pm, edited 1 time in total.
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/

User avatar
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

Post by Eb7+9 » Mon Nov 09, 2015 7:11 pm

ok, apologies for my lapse of reasoning here ...

yeah, the impedance seen at the cathode of the phase splitter driving the All-Pass Lattice is way lower
then the 82k I used above ... this, as a result of internal circuit feedback

to have an idea of the exact "equivalent" value I simmed the driving circuit using a high-accuracy SPICE model I once developped for this kind of work (https://viva-analog.com/298-algebraic-t ... n-tina-ti/)

unfortunately I only have it for a 12ax7 triode at the moment, and not a 12au7 ...
but the number I get will serve for illustrating what's going on here (much better than the 82k assumed above)

first, I need to verify that my dual-source model is sitting right at DC (ie., yielding proper Operating Point convergence)

http://www.lynx.net/~jc/12ax7%2082k%20bias%20table.jpg

then, AC analysis with current monitoring is used to determine Zout at the driving point
(notice, as the circuit is designed not taken directly at the Cathode)

http://www.lynx.net/~jc/12ax7%2082k%20f ... 20Zout.jpg

leaving us with a driving point figure of 3.16k Ohm-AC (by taking the inverse of the sensing current)

I'll now use that number and redo the previous sims,
noting that it would be slightly different using a 12au7 triode instead ...

ok, let's see what we get for Amplitude and Phase response at the Lattice's output
this time, things are markedly different ... and make much more sense (!!)

http://www.lynx.net/~jc/3.16kAmplitude.jpg

http://www.lynx.net/~jc/3.16kPhase.jpg

ok, now we're seeing some All-Pass phase action (ie., that inversion occurring between 500Hz and 1kHz)
the mix output is nothing unusual either:

http://www.lynx.net/~jc/3p16threeOutputs.jpg

things start making even more sense once we move the traces closer to each other and we add 180 degrees to the MIX outut ...

http://www.lynx.net/~jc/3phaseDifferences.jpg

finally, ...

in this case we see more clearly that between 1kHz and 20kHz all three outputs are
in roughly constant phase relationship to one another ... in particular we see a "roughly calculated" separation of
140deg, 140deg and 90deg
... (as opposed to the ideal 120deg x3)

using cursors in TINA-TI it's actually more like 68 deg (as opposed to 90 deg) ... but anyway

still, this is more in line with the patent ...
my apologies for jumping the gun here

I needed to SEE this with my own eyes first ... (heh heh !)

OK, this is pretty cool ...

the picture tells us even more as well ...

namely that the constant phase region takes place just above 1kHz ...
which may then explain the use of a crosss-over network prior to this unit
(ie., avoiding the sub-1k region ... another guess !)

I should note that all three curves would have been obtainable in the day using
signal generator and scope ... and so there was never any real reason to doubt mr. Wayne

;)

I will thus work with these two driving-point impedance values in my (op-amp) emulation of the front end ...

let's leave open the possibility that someone could try to devise a better Constant Phase Difference lattice one day
it's all good for now, things jive (at least loosely) with the patent

onward ...
Last edited by Eb7+9 on Mon Dec 07, 2015 9:48 pm, edited 2 times in total.
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/

User avatar
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

Post by Eb7+9 » Mon Nov 09, 2015 7:28 pm

dang, I just noticed that the forum truncates large images ...
ok, will post URL's instead
Last edited by Eb7+9 on Tue Jan 05, 2016 7:41 am, edited 1 time in total.
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/

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

Post by daveweyer » Mon Nov 09, 2015 7:44 pm

JC, I don't know how you set up your Spice simulator, but it looks to me as though your generator inputs are in series with a couple of 82K resistors. These inputs should be low impedance--the driver in Wayne's circuit has about a 200 ohm source impedance on the in-phase signal, and maybe a 3 to 4K impedance on the inverted driving source.
Ideally it was meant to be zero impedance on both, although I'm not sure what effect that will have on your phase response. I believe the bandwidth of the thing was supposed to be 3KHZ or maybe a little beyond--this was all that was necessary for chorusing audio signals, the guitar falls in this range too so it's a pretty good fit.
In the mean time I'll get some samples of the device's effect on a guitar or very familiar sound that you can compare.
The output signals are inverted of course, so you might want to represent that in your simulation--if no alteration of the phase takes place in the lattice, then the signals should just cancel each other out except for what's left of the frequency spectrum that is actually different between the top side and the bottom side of the lattice.
It looks like the 470K resistors are used to determine the final load impedance on the lattice above the .01 3db point, supposedly they are above the ideal load on the lattice and therefore unimportant in the response calculation--they do bias the tubes which amplify and invert the signal. If you're looking to see what you've got on the summing point, be sure to experiment with the specifically chosen resistor to ground at the junction, it should only matter in setting the input level for the third stage of the inversion of the summed signal, and creating a load for the 470Ks to work into.

User avatar
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

Post by Eb7+9 » Mon Nov 09, 2015 8:03 pm

Dave,

btw, can you confirm that our All-Pass Lattice component values are correct ?!
(ie., as in the Baldwin schematic)

I guess I'm a little confused about the 1kHz bend in the phase response...
as some the Choratone's BP frequencies lie under it (at 277, 554, etc ...)

wondering now if that 1Khz phase bump is right or not

unfortunatley, Wayne leaves the lattice component values out of the patent

---

also, are you operating the Choratone (flat out) as a full range unit
I think you mentioned running off the speaker (thru an attenuator) of a small dirty amp ?!

... or is crossover-like pre filtering crucial for optimum results ?!

---

another thing ...

Wayne's patent states 5-octave BP frequencies centered at 277, 554, 1108, 2216, and 4432 Hz ...

while Baldwin lists the following four L-C values:

1.0h/18uF (~37 Hz), 0.56h/0.1uF (~672 Hz), 0.27h/0.047uF (1.4k Hz), 0.12h/0.022uF (3.1k)

are these later ones the same values you used in your unit Dave ?!
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/

User avatar
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

Post by Eb7+9 » Mon Nov 09, 2015 8:23 pm

you're right Dave, the driving point impedance at the inverting node of the pre-Lattice driver
is going to be less than the 82k plate load ...

but, I'm not sure if it's by that much, ...
my previous simulation rewired to stimulate the plate node tells me it's 76.1k Ohm AC
(which is in line w prev testing I've done)

ok, granted these are for a 12ax7 values with a 300v supply voltage ...
I'll try to muster up a 12au7 model and see what the numbers say there

thx much !!
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/

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

Post by daveweyer » Mon Nov 09, 2015 8:51 pm

The later values are the ones I used.

I don't know if your spice simulator has a built in 12AU7 model, but the cathode impedance is quite low, generally 1/Gm, but in the case of the bootstrapped input, even lower, along with the plate resistance (5 to 6K in parallel with the plate load resistor dropping to 3K to 4K). I wonder if the upper values of the matrix are specifically to compensate for the higher driving impedance of the 12AU7 plate; if so, then a cathode follower or equivalent could lower this value to equal the cathode impedance, perhaps allowing you to use similar values on the top and bottom sections of the matrix and getting a wider all-pass phase response.
Another option would be to drive the matrix with a 12B4 tube or the like, with a plate resistance almost the same as the cathode impedance. If you use OTAs for the first stages, you'll be able to make it essentially zero, at least from a design perspective.
I'd try that in my spice simulations to see what happens to the phase response.

User avatar
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

Post by Eb7+9 » Mon Nov 09, 2015 9:08 pm

btw, I'm wondering if there's a typo in the Baldwin schematic ...

I just noticed that the lattice is shown to take its in-phase feed from the 82k resistor in the phase splitter circuit
and ... NOT at the cathode directly, as it is done with all other buffer circuits in the unit

Q: was this done on purpose ...?!!
(or is it a typo ?)

again, this would change the driving point impedance going into the lattice // and affect overall response ...
(in fact, it would bring the DPI down to 1.7k Ohm AC in my 12ax7 sim ...)

just wondering how that lattice feed is wired in your unit Dave

... certainly, there will be many options to consider once the prototyping starts

thx
~jcm
Last edited by Eb7+9 on Mon Nov 09, 2015 9:22 pm, edited 1 time in total.
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/

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

Post by daveweyer » Mon Nov 09, 2015 9:19 pm

The cathode circuit drives the matrix from the junction of the 1.2K bias resistor and the cathode load resistance. The effect of this is to bootstrap the grid, also raising the input impedance. I can barely read the resistor value on my schematic, but it is 1.2K in the actual circuit. All this above action lowers the driving impedance.
Alternatively they could have used a capacitor directly from the cathode, or, biased the grid separately. At any rate the impedance is in the hundreds of ohms from that node.
A 12AU7 running enough current for a Gm of 3000 has a cathode impedance of 333 ohms.

User avatar
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

Post by Eb7+9 » Mon Nov 09, 2015 9:26 pm

alright, thx much for that Dave ...!!

yeah, I just checked w the old tube data ...

they list plate impedance on a 12au7 at about 5~8k Ohm AC
much lower than I thought

ok, will recheck everything again in the lattice using your DPI numbers
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/

User avatar
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

Post by Eb7+9 » Mon Nov 09, 2015 10:56 pm

lattice response looking way better after using 330 ohm and 7.5k DPI's in the phase inverter ...
(thx Dave!)

I took some "phase difference" measurements between points selected on the three curves

notice the MIX node is now followed by an inverter function block in the schematic
... yielding proper relative phase information for that output


http://www.lynx.net/~jc/330cathode101deg.jpg

http://www.lynx.net/~jc/330cathode121deg.jpg

http://www.lynx.net/~jc/330cathode123deg.jpg

http://www.lynx.net/~jc/330cathode213deg.jpg

http://www.lynx.net/~jc/330cathode248deg.jpg

although the inverted MIX curves moves up and down a little, for the most
part I'm seeing 100 to 120 degree "near constant" phase deviation between all three
for a good section of the useful audio band

and the output "difference" (unused in this invention) really shows the all-pass nature of the lattice
(bottom curve)

http://www.lynx.net/~jc/330cathode4outs.jpg

... theory and simulation are jivin' well

8) 8) 8)
Last edited by Eb7+9 on Mon Dec 07, 2015 9:49 pm, edited 2 times in total.
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/

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

Post by daveweyer » Mon Nov 09, 2015 11:18 pm

Nice, I think we're getting somewhere. I would take my driving impedance down to 50 ohms on each phase and plot the difference. This may help explain the value choices in the lattice.

User avatar
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

Post by Eb7+9 » Mon Nov 09, 2015 11:37 pm

ok, will do ...
modern VT circuit analysis and modeling: https://viva-analog.com/product/ifmta-book-pdf/

Post Reply