New at runoffgroove.com: Tri-Vibe

[John Lyons]

Forgive me if I missed it in this thread but is there a way to get more depth form the
vibe and trem modes? The swirl mode has enough depth but the other two are on the edge of
being unusable except for the higher speeds.

[Scruffie]

Forgive me if I missed it in this thread but is there a way to get more depth form the
vibe and trem modes? The swirl mode has enough depth but the other two are on the edge of
being unusable except for the higher speeds.

I would say the best bet would be to add another 2 stages with an extra LM13600... using the uni-vibe cap values.

What voltage do you get at the LFO and after the resistor to pins 1 & 16, perhaps the V.Ref resistor values could be reduced without killing the chips.

[DrAlx]

I don't know if the Tri-Vibe designer(s) still read this thread, but I am intrigued to know how the 15:1 cap ratio for the 2 all-pass sections was determined.  
I can appreciate that it is desirable to have the phase shift varying linearly with frequency over as broad a range of frequencies as possible.
However when I look at the maths, I come to the conclusion that you would need to make the all-pass stages identical in order to achieve that.
To see what I mean go to

http://www.wolframalpha.com

and paste in the following command

plot (180/pi) arg(((1-ix*1)/(1+ix*1))*(1-ix*15)/(1+ix*15)) from x  = 0 to 0.125

That's a phase-shift plot for 2 all-pass stages with caps in the ratio 1:15 as in the Tri-Vibe.
The horizontal axis represents frequency (normalised), and you can see that between x = 0 and 0.025 there is almost linear variation in phase from 0 to 45 degrees.

To get similar variation in phase between x = 0 and 0.025 using equal valued caps, you need to make each all-pass stage contribute 8 "units of phase shift"
rather than have one stage contribute 1 unit and the other contribute 15.  The graph for that case is given by this.

plot (180/pi) arg(((1-ix*8 )/(1+ix*8 ))*(1-ix*8 )/(1+ix*8 )) from x  = 0 to 0.125

You still get a phase shift of 45 degrees at around x = 0.025 but the graph is straighter over a much larger range of frequencies.
In fact when you use caps in that ratio 15:1, practically all the phase shift comes from the all-pass stage with the larger cap, so you've almost got 1 all-pass stage rather than 2.

So to my understanding, replacing the 1.5nF and 22nF caps by a pair of 12nF caps should give linear variation in phase over a wider range of frequencies..  
Am I missing something here, or is this too simplistic an understanding ?

[snk]

Hello,
I am reviving this thread, because i have successfully built a Tri-Vibe (it's a great circuit ! Thank you so much, B Tremblay !), and I would have a couple questions for a better understanding of the circuit :

1- components :
I have used a NE5532 and a TL074 (instead of a TL062/TL064). It works fine so far, but I am wondering if it would be any better using a TL062/64 (I have also seen advices to use a LM324) ?
Any Pro/cons ?
Also, would a 13700 be any better than a 13600 (i have both in stock) ?


2- rate pot :
I didn't have a 500k with a C taper. So i used a 1M pot (with a C taper), with a 470K resistor (instead of a 820K).
Am I right assuming that the 500k pot is wired in parallel with the 820K resistor, and that it is the same to use a 1M pot with a 470K resistor?
My tests give me some good sounds and a nice rate on the pot, but i am wondering if my maths are correct ?


3- lfo thump :
Last, I noticed some audible LFO thumping with faster rates. I know that some tremolo and phasers circuits are prone to this, but this Tri-Vibe sounds so good that i am wondering if there is any way to reduce/remove it ? What can be the cause : the chips used ? Some components values ? I have read that some users experienced the same issue, but didn't find yet a way to fix this (I can live without, it is not a dealbreaker, but i prefer to tweak my circuit a bit as it is still unboxed)...

[Mark Hammer]

Hello,
I am reviving this thread, because i have successfully built a Tri-Vibe (it's a great circuit ! Thank you so much, B Tremblay !), and I would have a couple questions for a better understanding of the circuit :

1- components :
I have used a NE5532 and a TL074 (instead of a TL062/TL064). It works fine so far, but I am wondering if it would be any better using a TL062/64 (I have also seen advices to use a LM324) ?
Any Pro/cons ?
Also, would a 13700 be any better than a 13600 (i have both in stock) ?

I built one.  I doubt any of the chip differences you suggest would make an audible difference.  If you plan on using it with batteries, the TL06x chips would draw less current, but that's it.

2- rate pot :
I didn't have a 500k with a C taper. So i used a 1M pot (with a C taper), with a 470K resistor (instead of a 820K).
Am I right assuming that the 500k pot is wired in parallel with the 820K resistor, and that it is the same to use a 1M pot with a 470K resistor?
My tests give me some good sounds and a nice rate on the pot, but i am wondering if my maths are correct ?

For the full range of speeds in the stock unit, you might try your 1meg C-taper pot with a 1meg parallel resistor.  You can also use the 1M pot as is.  The only difference is that there will be some slower speeds available than the stock circuit.  Those will not be useful for the vibrato setting.

3- lfo thump :
Last, I noticed some audible LFO thumping with faster rates. I know that some tremolo and phasers circuits are prone to this, but this Tri-Vibe sounds so good that i am wondering if there is any way to reduce/remove it ? What can be the cause : the chips used ? Some components values ? I have read that some users experienced the same issue, but didn't find yet a way to fix this (I can live without, it is not a dealbreaker, but i prefer to tweak my circuit a bit as it is still unboxed)...

The seems to be control current `breakthrough`.  Maybe that`s a reason to use a TL062 for U2.

I find the circuit provides a nice swirl.  I packaged mine up in an enclosure with a custom-tweaked overdrive (TriVibe first) and named it "Box O`Robin" because it gives me my instant Robin Trower sound.

[snk]

Thank you, Mark.

I find the circuit provides a nice swirl.

I'm loving slow modulations, phasers and vibe-ish effects, so this Tri-Vibe is really enjoyed here too

I built one.  I doubt any of the chip differences you suggest would make an audible difference.  If you plan on using it with batteries, the TL06x chips would draw less current, but that's it.

Ok. The TL06x was my main interrogation (I had read on the official page that the 5532 and the 13600 were preferred over other alternatives, so this is what i used, but the only quad opamp i had in stock is the TL074). I was wondering if it had any drawbacks soundwise (i am not using batteries), so i understand the reply is "no"

For the full range of speeds in the stock unit, you might try your 1meg C-taper pot with a 1meg parallel resistor.  You can also use the 1M pot as is.  The only difference is that there will be some slower speeds available than the stock circuit.  Those will not be useful for the vibrato setting.

Ok. So my calculation was wrong ? Slower speeds are fine (for the phaser mode), so i might try that.

The seems to be control current `breakthrough`.  Maybe that`s a reason to use a TL062 for U2.

Ok, I will order a couple TL062 (as well as some LM324, as some users reported a clean, thump-free operation with them), and see what happens.
I see that the circuit already features a good circuit protection (with the 470µF cap and the small resistor).
I used a 100R resistor instead of a 22R : could it come from that ?
Would adding a very small ceramic cap across 9V & Gnd improve things ?

[snk]

Out of curiosity, since there is a phase mode, is there any way to tweak the resonance ?
(I like it like that, but i'm curious)

[Mark Hammer]

The thing about  Vibe-like circuits is that, while they may be phasers, in principle, the phase shift is so broadly distributed that adding feedback/resonance does nothing of any use or even audibility.  You will note that you NEVER see a feedback control on commercial Uni-Vibe units.

[snk]

  You will note that you NEVER see a feedback control on commercial Uni-Vibe units.

Indeed 
... and it sounds good as it is.

The strange thing is that now i don't hear any more ticking/thump !?
It's a good thing, but somewhat curious...

[jonny.reckless]

I really like the Tri-Vibe. It was the inspiration for the Reckless Vibe design. I like the way they do the allpass filters with OTAs. They're cheap, consistent from unit to unit, and easy to use. I did try to improve on the design a bit by using 4 stages, generating a true sinusoidal LF with a quadrature oscillator, and using bipolar junction transistors for the frequency exponentiation, but these were all ideas derived from or inspired by the Tri-Vibe I built a few years previously. I do use some feedback around the allpass filters which gives the effect a bit more grit and intensity.
https://www.diystompboxes.com/smfforum/index.php?topic=119211.msg1146780#msg1146780
https://postimg.cc/5HzsFfYy

[snk]

Thank you for the hint, Jonny.
I'm listening to the video right now. What do the 4 pots do in your design [edit : found at the end of the video !]? From what I see, last year you wrote in the thread that you may have new, smaller and enhanced PCBs for sale ?

[jonny.reckless]

Sadly I haven't yet finished that new PCB. I'm still working on simplifying and shrinking the design as much as possible, and I haven't had the time to work on it as much as I would have liked. I'll post here when I do finally get it ready though.

[tubegeek]

Schematic question:

Caps C24 C18 C27 C26 C7: are they bypasses "at the socket" for IC 3A, 3B, 5A, 5B? If so, then why C14, C19, C17? And what does the 5th one do? Bypass near the Depth control?

[jonny.reckless]

They bypass VMID to GND locally around the allpass filters to ensure good HF decoupling of VMID (which is a virtual signal ground) to the actual ground. They complement the power supply bypass capacitors. It might be a touch overkill but I've always been wary with split rail "virtual ground" designs because you can never get the signal ground to be a low impedance at high frequencies unless you use a 4 layer board, and transient currents cause the virtual ground to bounce which then gets injected in to the signal path when it's dereferenced to real ground. I generally put the vibe in front of the distortion pedal so any little transients get amplified up by a lot, so I want the whole thing to be as quiet and glitch free as possible. Clicks, pops and glitches should be 90dB down compared with the guitar so even with a high gain amp you don't get annoying ticks or pops. I normally prefer twin voltage rails using either a switched cap or inverting boost converter. I always use the top layer of the PCB as a ground plane.

[tubegeek]

I normally prefer twin voltage rails using either a switched cap or inverting boost converter.

I thought your preferred method was using an opamp as a rail splitter, especially for small power amplifiers?

No wait, that was me.

[chemosis]

how does one increase the depth because my vibrato setting is weak just like in the various demos. just need a little more depth. not much

[chemosis]

you guys don't find the vibrato setting a little weak?? like in the demos. mine definetly is. how can one have a little more depth??

[Mark Hammer]

Vibrato needs to be faster, at least medium fast, in order to be audible.  It's a basic perceptual principle as espoused by Weber, Fechner, and later Stevens.

The slow speeds that will work just fine for phasing and uni-vibe do diddley-squat for vibrato when you lift the dry signal.  generally only the upper half of the speed range does much useful.  The amount of pitch-shift that could be audible at very slow speeds would send you running for the depth control to turn it way down if you increased the speed.

Blame your nervous system for that, not the design.