vibrotron clone

[maxkracht]

A few months ago someone got it in my head that I needed a tube vibrato. Here is my version, of Merlin's version, of the Dorf vibrato. I am pretty happy with it, but it probably still needs some tweaks. I didn't have any C taper pots in the right size, so the intensity control range could be better. It is pretty quiet on the tremolo and bypass setting, but there is some hum in the vibrato section. Not sure if that is a layout error of mine or inherent to the circuit. Also, the LFO doesn't always start if I power it up while on the vibrato setting. Briefly switching to tremolo then back to vibrato will make the light start flashing again... If anyone has any ideas I would really appreciate it.

IMG_0163.jpg

IMG_0165.jpg

Edit:
For anyone who would like to play along at home.

ValveVibrotron.jpg

valve vibrato layout.jpg

IMG_0167.jpg

Trap for the young gamers. I measured for clearance to get tubes out, I didn't account for the shield. I guess it doesn't need shields...

[maxkracht]

I missed a link in the history of this circuit. Merlin was working off of a vox design that was lifted from Dorf or Wurlizter. This article on vox vibrato history is pretty interesting. https://www.voxac30.org.uk/vox_ac30_vibrato_gl.html

I think my vibrato having trouble starting was due to the diode on the first 1/2 of v3. It seems to have trouble starting and/or takes much longer to start when using an LED with a forward voltage of less than around 1.8v. LED's around 2-2.4v work and start up faster. I also tried a couple white and blue LED's, that my meter wouldn't measure in diode mode, which didn't start until I turned the trem/vibe off and on. Once the LFO kicks on they would work in trem mode, but would only work for a bit in vibe mode before the LFO stopped. Adjusting the speed and intensity controls also change how quickly the LFO will turn off with the wrong diode. I don't really understand the logic behind any of this.

[R.G.]

Very interesting.
Being the original adherent to the MOSFET Heresies, I can't help but note that at least two of the tubes, and with minor value changes you could replace all of them with MOSFETs, perhaps even the TO-92 LND150.

[maxkracht]

Yes, this certainly isn't the most practical device out there. I made a vibrato pedal for someone not too long ago that sounded pretty good, and had no extraneous noise or thumping and had more options than just a sine wave for the oscillator, but the glowy glass things look cool... I have been meaning to play with adding a solid state LFO in a tube amp's tremolo get more wave shapes. I am sure it has been done.

[trobbins]

A few queries:
Is the bridge rectifier a large current device?
Does the heater have a CT and is it grounded?
The LFO may start more consistently if the gain can be tweaked a bit higher - do you know what your cathode, anode and B+ voltages are for that stage? Increasing the plate load from 47k, but still having mid-bias type operation, may help.

[sluckey]

I think my vibrato having trouble starting was due to the diode on the first 1/2 of v3.

I found that red or yellow LEDs worked best (at least in the typical Fender oscillator). That phase shift LFO really needs a lot of gain to start up and reliably oscillate over a 3 to 10Hz range. Start by using a 12AX7 (ECC83). Increase the plate load to about 220K. Give it as much B+ as you can squeeze outta that PS. I suggest a 100Ω dropper for that first B+ node. Maybe even power the oscillator directly from the bridge.

I don't see any means to stop the oscillator (ie, via footswitch or panel switch. IAW your schematic, that oscillator should be running continuously. Start/stop should be no issue. I think your oscillator drops out simply because it doesn't have enough gain.

[maxkracht]

Is the bridge rectifier a large current device?
Does the heater have a CT and is it grounded?
The LFO may start more consistently if the gain can be tweaked a bit higher - do you know what your cathode, anode and B+ voltages are for that stage? Increasing the plate load from 47k, but still having mid-bias type operation, may help.

I am not sure what you mean by a large current device. I used my last NOS NIB radio shack BR. Something like 3A @ 1kv, so way overkill for this project.
The posted schematic is Merlin's, not mine, so there are a few differences in the power supply. I used a 380v CT transformer that was collecting dust, and used half of the secondary to get close to the 150vac listed on the schematic, then added another dropping resistor to make up the difference. So, at the BR I have is 245v DC, which is dropped to 228v when it enters the board, then 220 and 199 on the next PS nodes. I think the B+ is 10-15v higher than the original, using the closest convenient resistor values I had.

The heaters are 6.3v instead of 12v on the schematic with an artificial CT grounded at the rectifier. I don't believe much if any of the hum I have left is from the heater. It is only apparent when the vibrato is on and it really isn't bad, just noticeable in relation to only the tremolo or bypass.

(measurements taken with LFO off)

oscillator:
P - 136v
C - 1.66v

LFO driver:
P - 190v
G - 19.5v
C - 33v

I think my vibrato having trouble starting was due to the diode on the first 1/2 of v3.

I found that red or yellow LEDs worked best (at least in the typical Fender oscillator). That phase shift LFO really needs a lot of gain to start up and reliably oscillate over a 3 to 10Hz range. Start by using a 12AX7 (ECC83). Increase the plate load to about 220K. Give it as much B+ as you can squeeze outta that PS. I suggest a 100Ω dropper for that first B+ node. Maybe even power the oscillator directly from the bridge.

I don't see any means to stop the oscillator (ie, via footswitch or panel switch. IAW your schematic, that oscillator should be running continuously. Start/stop should be no issue. I think your oscillator drops out simply because it doesn't have enough gain.

I wound up with a green LED in the LFO driver and red in the LFO. I only experimented with different LED's in the driver, mostly because It would have been more of a pain to change the one in the LFO used as an indicator. I did try a 12ax7 for V3 earlier on, but it doesn't work well in this circuit. You lose a lot of depth in the tremolo, and the vibrato doesn't vibrato very well, just sounds like a lousy tremolo. I haven't looked on my scope, but I am guessing the extra gain is making the LFO clip.

I think you are right that it didn't have enough gain, but the problem was in the driver section. No problems at all with the correct LED installed.

No means to stop the oscillator by foot switch. The effect is turned off by grounding V2. With V2 completely on you have vibrato, either half grounded you have tremolo, and both halves grounded it is just a preamp. I wish I thought about this more when designing the face plate. Instead of just the switch to go from trem to vibe, I added a 1M pot to the back of the unit to make it variable between all trem and all vibe and I think I like some of the in between settings a bit better.

[trobbins]

Merlin used red LEDs back in 2006 for his prototype, and indicated he used the ECC81 essentially because it was cheaper, but of course part variations could cause issues. Changing LED colour (voltage) will change the anode bias voltage-current operating point, so some checking needs to be done that a change in LED does not cause a voltage swing or gain restriction. The power supply voltage and cap levels could also be an influence for start-up as it is the disturbance of the circuit that effectively starts oscillation.

My earlier comment on diode bridge package was because the photo seemed to show a 4-terminal spade lug package, which usually indicates a high current diode rating (eg. 10-35A). And I can see the 100-100R humdinger now in your photo (although it is going to the noisy end of the 0V bus). Wrt hum/noise, the LFO intensity pot and ECC82 grid circuitry is a high impedance node, so quite prone to noise ingress, especially as the cathode is not heavily bypassed.

[maxkracht]

No spade lug package, just bent wires that looks like it. Is there an issue with using larger than necessary rectifier packages, aside from cost? I wasn't aware that artificial center tap ground location was so critical. Is there a preferred location?

[trobbins]

In general, ss diodes with larger current rating have larger off-state capacitance, and if relevant - larger reverse recovery charge to clear. If you were aiming to minimise any noise transfer from before the diodes to after the diodes, then that capacitance is a leakage path. It's unlikely to be noticeable, and could be difficult to measure, but like the many precautions that can be taken, it is worth at least appreciating imho.

Similar for other very subtle noise/hum causes - the humdinger connection is worth appreciating, but unlikely to be noticeable. Not sure why you read my comment as it being critical. One role of a humdinger is to provide a leakage path flow for parasitic heater-cathode noise/hum currents - minimising the chance for those current paths to pick up noise along the way is a reasonable precaution if its practical. The link between the rectifier bridge and the first filter capacitor terminal has probably the largest pulse currents in the amplifier, hence my side comment that it is the noisy end. Everything from and after that first filter capacitor is a lower noise point of reference.

[maxkracht]

Interesting. I knew diodes of different composition would have different switching speed and noise, but I wasn't aware that current rating would have any correlation. It's a bit outside of my, very, limited understanding of these things. With the heater, I assumed putting an already noisy high current ground with the others wouldn't matter as long as it is upstream from other sensitive circuitry. I am always happy to learn better practices.

[Kagliostro]

Very Nice Build

I'm interested on your project and I'll follow your experience with interest

EDIT:

I was missing this

No problems at all with the correct LED installed.

Does the problem is solved with the correct LED ?

THANKS

Franco

[maxkracht]

Thanks Franco. Yes, that issue was solved with an LED change. I am guessing the slight change in B+ voltage from the original and difference between forward voltage in LED's was my main issue. I will probably order the correct resistor value to get closer to the original circuit. If I were to do It over, I would make a number of layout changes, but I am pretty happy with it as is.

[Kagliostro]

If I were to do It over, I would make a number of layout changes

Which ones ? To me the layout seems good ....

--

I added a 1M pot to the back of the unit to make it variable between all trem and all vibe and I think I like some of the in between settings a bit better.

Please can you explain, I'm not able to understand how you did it

Thanks

Franco

p.s.: I think that if you place the sockets like here, you can solve the problem you have with shields and tubes extraction



no other mod necessary, well, may be a sheet of aluminium with small spacers to preserve the top of cabinet from excessive heat

[maxkracht]

Like most of these projects, this one was intended to be quick and dirty but got out of hand. My main design constraint was to fit everything on that predrilled board, which I mostly succeeded. This made it harder to put each stage’s filter cap close to that stage. It probably doesn’t matter much, but that is how Merlin did it, and is usually good practice. I didn’t spend too much time thinking about the grounding scheme, so that could be improved. I assumed if I kept the preamp quiet, everything else would be fine. Not exactly layout, but I wish I would have shielded the under board signal wires to eliminate a variable when looking for sources of noise. This likely isn’t necessary, but couldn’t hurt. I’m sure there are other improvements that more experienced people on this forum would be able to point out, but it is hard to change one thing on a layout without changing the whole thing.



I designed the faceplate without thinking of clearance between the board and the input and output jacks. Once I had the board built, I realized it needed to be pushed back farther to make room. This also pushed the tube sockets back, requiring them to be rotated to leave room for the mounting screws. My layout, on paper, put each triode half at a right angle to the board, allowing shorter wire lengths and easier heater routing to the corner of the chassis. I used floating heaters because I ran out of room.



I thought about placing the tube sockets where you suggested, but opted for what I thought would be a more compact cabinet design. I also find it oddly satisfying when the transformers in an amp are just taller than the tubes. This is purely aesthetic when there is only one transformer. Your way makes more sense in the cabinet I ended up with.



I did all of the electronics first, then looked around for scrap wood. I found a piece of decorative trim and a leaf from a beat up table that the previous occupants left in my house. Those dictated the cabinet dimensions. The width was about right, so I didn’t have to bother with ripping anything down to size, just cut to length, glued and screwed. I used a broken, wood, coat hanger as a dowel to cover the screws. I initially planned to cover it with vinyl, so I didn't bother sanding the interior before gluing together. This is why the outside finish doesn't match the inside.

The trem/vibe mix pot is just a variable resistor between the node marked “A” on my layout and ground. You could eliminate the selector switch entirely, if you used a pot with a switch connecting the node marked “B” to ground when it is in the off position. To put it more simply, when “A” or “B” are grounded you have tremolo. When “A” and “B” are not grounded you have vibrato. When “A” and “B” are grounded the effect is bypassed.