I just built a univibe style effect over the weekend based on the materials offered by JC at http://www.lynx.bc.ca/~jc/

and from materials from the RGs Neovibe.  RG Keen has excellent information in both the Neovibe documentation and in his article on the Technology of the univibe.  Even if you do not like the effect, RGs article is an excellent read.  I think it is here http://www.geofex.com/

Thanks to both gentlemen by the way, this project would not have ended happily without the above mentioned documents sitting on my workbench for reference. :lol:

I used JCs modified driver with a Ratshack bulb, mu-stage input and per his idea, replaced the 4.7k resistor in each of the three main phases stages that is in series with the 1uf cap and LDR with a 10k pot.  I used trimmers on mine.

The unit sounds fantastic.  I bought a variety of LDRs that I am going to experiment with in the next few weeks.  To get it working, I used the LDRs suggested in the Neovibe documents.  I will post my LDR findings of my LDR experiments in the next week or so as I have time to experiment.

Here is my question.  (by the way, I did my own layout, and this may have affect on my only real build problem)

The first time I fired up my circuit, I got the same problem that I have read about a lot lately, that is, the LFO worked great at its max speed, and did not work at all with the 100k reverse taper pot (from Small Bear).

After checking and rechecking the layout, my soldering, and scanning the posts on this forum...

I first tried several different transistors thinking I was not getting enough gain.... but to no avail.  

I then replaced the three 1uf caps (electrolytic) with better quality metalized film caps thinking the problem was there.  Same problem with no slow lfo speed.

I desoldered just about the entire LFO pulling each resistor and diode one at a time to measure it and make sure it was ok.

I got (of course) to the last resistor.  The 2.2 meg.  My dmm only measures up to 1 meg.  The resistor was a metal film, and sometimes I do not trust my eyes reading color codes on blue metal film resistors.  So to be safe, I chucked the resistor and replaced it with a carbon 2.2 meg that I could clearly identify by the color bands.

Sure enough, that fixed my problem.  

After fixing the resistor, I found that the LFO was happy with 3904s, 5088s, I even tried an MPSA-18 and all worked fine to get the thing to oscillate.  All that work got me interested in this LFO architecture.

So, I am figuring that the 2.2 meg resistor is instrumental to the LFO generating enough gain to start oscillating.

Can someone point me to a reference on transistor oscillators?  I would like to understand how these things work.  For example, how do you compute the necessary values of components to get oscillation?  How do you tweak the frequency range?  Are there cookbook formulas for this stuff?  In RGs article, he indicates that the LFO gets stronger as the frequency increases.  I observed this to be true on my scope.  Is that an intentional design, or inherent in the topology?

Thanks in advance.

Did you notice that the wiring for the speed pot is different on the layout than it is on the schematic?  Switch pad C and E, and it would match the schematic.

I had the same problem you had, and switching the wiring solved it... after about 2 months and 2 seperate builds.. hah.

Yea, I did notice that.

On JC's web site, he has a schematic of the original univibe.  I looked at all the resources that I had and finally figured out the correct way to get everything wired up.

My problem was that 2.2 meg resistor.  I do not think I was getting enough gain at the lower frequency settings to get the circuit to oscillate.

It seems like everyone has a different problem with the LFO...Mine seemed to be the transistors. Mine only worked using 2N3904 transistors for Q11, Q12 and Q13. Although even then mine will still stop oscillating if I turn either the speed or depth below 1/4, then I have to turn it all the way up to get it to oscillate again.

Did you happen to do the blinking LED described on JC's site? If so how did you wire the LED in?

Also, does anyone's pedal sound muted when using the chorus mode?

Quote from: ChicoCan someone point me to a reference on transistor oscillators?  I would like to understand how these things work.  For example, how do you compute the necessary values of components to get oscillation?  How do you tweak the frequency range?  Are there cookbook formulas for this stuff?  In RGs article, he indicates that the LFO gets stronger as the frequency increases.  I observed this to be true on my scope.  Is that an intentional design, or inherent in the topology?

If you check out books on micro-electronic design you should find plenty ... turns out that oscillators belong to the same family as filters (from a mathematical classification viewpoint) - oscillators are nothing more than unstable filters ... a condition expressed in terms of the number of loop function poles lying in the R.H.S of the S-plane ...

The Univibe oscillator belongs to the phase-shift oscillator family - where a combination of greater than unity loop-gain and at least 180degree phase shift in the loop is required for the circuit to start and maintain itself ... each cap produces 90deg of shift so three caps gives a total maximum potential of 270deg ... this kind of principle is used in Fender amp Tremolo oscillators for example - except in the Vibe case the voltage gain of a Darligton follower is less than unity, so the oscillator loop must be operating in the current variable - with a (current) gain of Beta^2 - and capacitors providing phase shift in the same way they would in a voltage-mode oscillator ... in theory phase-shift oscillators exhibit exponentially increasing Amplitudes with time and so require amplitude limition to keep the output signal steady, hence the two diodes clamping the middle shifting cap ... using LED's or Germanium diodes will alter the output size indirectly ...

You'll find that the amplitude of the Vibe oscillator doesn't grow monotonically with frequency but rather peaks at some point and starts dying down slowly passed it, this is typical for these types of oscillator circuits ... this means that the Vibe will looses a little effect depth at slow and fast speeds which is why I recommend adjusting the bulb driver gain and offset levels at the peak frequency first and then readjusting at slow-speed settings so that maximum intensity will give a nice round sway to the filter at slow speeds ... to mod the oscillator you may want to try varying the size of the phase-shift caps if you want to alter the range specs of the oscillator but I haven't met with much success "stretching-out" the range of the vibe LFO circuit ... you can replace downside cap on the Intensity pot by a 470uF cap to squeeze slightly more signal to the bulb driver circuit ...

:idea:  

FYI ... incidentally I just spent the last two nites canabalizing one of my older and simpler vibe clones to the circuit of the Shin-Ei "Resly-Tone" schematic which I think is the precursor to the original Shin-Ei Univibe and also related (identical?) to the JAX Vibro-Chorus ... aside from some strange switching which, to me, didn't produce anything substantial in itself, the magic cap values are different than the Unicord Vibe cap values ... they are 0.0047u, 0.0033u, 0.0022u, and 0.001u ... I modified the last follower stage as in the RT-18 schematic and used a 100k pot to mix the front-end and back-end signals as well as a 100k volume pot fed by a 0.015u cap to cut down on the exaggerated low end (not sure why that came up so much) - and I'm running it on a 9volt Boss adapter since I noticed the Electrolytics caps are rated at 10v in the circuit ...  it seems to sound a bit more authentic than my Unicord based clones and I find the signal amplitude doesn't jump out as much with strong picking attack - it's a little more chewy and psychadelic sounding and the top end of the churn seems more characteristic ... earlier I was playing Bridge of Sighs and Dragonfly through it and sounded so nailed ~

... jcm

Quote from: fretbuzz2003It seems like everyone has a different problem with the LFO... Did you happen to do the blinking LED described on JC's site? If so how did you wire the LED in? ...Also, does anyone's pedal sound muted when using the chorus mode?

... the LFO circuit is very finnicky ...

In my early Vibe-clone days I tried building alternative versions using current mirrors fed off the oscillator collector circuit current to drive photo-diodes ... I got it to work but cancelling offsets and setting the quiescent current through individual cells was too tricky - mainly the resistance curves that my circuit was producing didn't mimmic a bulb and photocell transfer so I gave up on that and instead decided to focus on the getting the bulb driver to work properly with various available bulbs ...

I'm just saying this to mention that you can stick and LED in the unused collector circuit of the Vibe oscillator if you don't want to stick one in the driver circuit in series with the bulb - and doing this doesn't alter oscillator behavior much depending on diode voltage swing - the only difference in this case is that the LED will not respond to the intensity control as in my posted mod ...

btw, a somewhat muted sound in Chorus mode is normal because of the signal cancellation involved ... one nice thing with the Resly-Tone cap values is that there isn't as much body cancellation in this mode ...

J.C.

Thanks for the explanation and tips.  I think that I am going to have to digest this info for a while.

I plan on spending a little time tonight/tomorrow tweaking my vibe.  I still have a bunch of LDRs that I want to experiment with.

btw, I really like your design of the mu-stage input.  I have the 500k volumne mounted to an external pot and plan on panel mounting the control after I am done hacking at my board.  

I think I will also try playing with the cap values as well.

Thanks again.

JC,
Sorry but I'm having a hard time figuring out your schematic for the external LED... You show the LED in parallel with a 500 ohm pot- is that the trim pot that biases the lamp?

Thanks.

Quote from: fretbuzz2003JC,
Sorry but I'm having a hard time figuring out your schematic for the external LED... You show the LED in parallel with a 500 ohm pot- is that the trim pot that biases the lamp?

I don't think you're refering to the bias pot ... first of all, I put the LED in series with the bulb to make double use of the current that's flowing through the bulb and robbing the least amount of voltage in that circuit in the process - then a 500 or 100 ohm trimpot is wired across the LED as a current divider to bleed some of the current away from the LED and therefore controlling its overall brightness against bulb current  (and in the process minimizing voltage drop !)...

This is provided in the case some of you end up using an LED that needs much less current than what is flowing through your chosen bulb, you may not need to do this depending on how you adjust the bias on the bulb driver circuit and your choice of an LED ... since bulb adjustment is crucial and will be re-visited over time I assumed one might need to trim the LED indicator brightness accordingly ...

... thanx for the feedback on the mu-stage front-end - I like the bite that it gives the circuit ... you can put a 100pF-470pF shunting cap at the mu-stage output if the overall circuit sounds too bright ...

... jc