Cardinal Tremolo (Vactrol-base Harmonic Tremolo) - schematic & discussion

[midwayfair]

Such a clever way to create a new sound from the old tremolo. My question is about the indicator LED and Vactrol LEDs. I recently built this on the 1776 pcb.

I found the main tremolo section worked perfectly but the high pass section hardly trembled at all. Multimeter could read the changing resistance of Vactrol 1, but Vactrol 2 seemed to not be changing at all. I disconnected the low pass side completely (lifted one leg of the LDR) so I could hear the high pass side by itself. Then I jumpered the current limiting resistor for the second vactrol and then I could hear the treble throb, subtly. The main channel provides complete silence on the wave troughs with depth set high, but even with the CLR jumpered the high pass side is just gentle.

I thought "with the three LEDs in series are those CLRs necessary?" I thought maybe the intention was to provide lower voltage to vactrol 2 for a more subtle effect. I breadboarded the three LEDs and CLRs as per the schematic (with normal LEDs replacing the vactrols), and they seem to be equal in brightness. So the designer had his reasons...

Don't you really only need one resistor on the end of the three series LEDs?  The 9 + volts already go through a diode and a 100ohm resistor first n the power section and each LED drops around 2v.

I am thinking of replacing vactrol 2. (I am using the recommended VTL5C1, and I bought extra.) Should have breadboarded first!

Also: Why diffused on the indicator? I am not aware of any reason why a diffused would behave differently than a clear. Or why a 3mm should have different fv than a 5mm. On the other hand, color would seem to be important. I first used a blue one and found a red seemed to work better, lower voltage drop, leaving a little more juice for Vactrol 2.

"Diffused" is usually red, orange, yellow, or green. I know there are super bright blue diffused (I have a bunch), but those are kind of rare. I guess we should have specified the color. Anyway, it can't be blue or white from what I've tried; the total voltage drop seems to mess things up a little, and I couldn't get any depth out of it. As you discovered, red works better. Green is still fine and is actually the color I tend to use.

It says 3mm in the schematic because in Eagle, the part has the measurements. In Josh's PCB, it says 3mm because that's the size of the pad. The size doesn't matter.

As far as the depth problem on Q3:
It's possible to have a vactrol out of spec enough that it won't have on resistance low enough to pull the gate of Q3 low enough to trem completely. It's unfortunate, because although vactrols are more consistent than most discrete LDRs, they are a heck of a lot more expensive if they have to be replaced. I have one VTL5C1 in my box of parts right now that works on Q2 but not on Q3, since the light resistance never seems to get below 30K and most of the time isn't below 100K (!).

But before you go pulling an expensive part off the PCB, set the depth to minimum and measure the resistance of the LDR side of vactrol 2 and compare it with Vactrol 1.

You can also try simply jumpering the indicator LED to see if that helps. (If it does, there's another way to wire a rate indicator LED: Just run some wire from pin 7 to an LED's anode, and connect the cathode to ground through a CLR. Ugly, I know, but still probably better than desoldering an $8 part.)

The extra resistors to ground increase the length of time the LED spends on the dark swing. You can see something similar in the Tremulus Lune, from which the LFO in this is derived. The cascading values were needed to keep the LEDs similar brightness. One thing that happens with them is that they tend to soften the on-off, so the sine wave side is a lot softer than the original lune's (the extra LED helped, too).

On the other hand, Vactrol 2 doesn't need dark resistance any where near as high as vactrol 1, so there are actually other optocouplers that will work. If you have some other stuff in your bin, measure a couple and see if you have one with light resistance below 5K and dark resistance anywhere above, say, 500K (but preferably above 1M) and a relatively far on/off time. That should provide a sufficient range. Just a heads up: I can tell you for certain that the NSL-32xxx ones DON'T work, though; they never get dark enough and seem to have a turn off time much slower than their datasheets claim.

[idy]

Thank you, you answered all my questions and got me to ask better ones. I really admire the patience and love that go into giving an idea away and teaching everybody and their dog to make it work. I'm now satisfied that both sides are treming in contrary motion.  I still don't hear the phasey sound I heard in the demos. I have a few other boards and will definitely breadboard before no 2... and read the whole thread again.

[midwayfair]

Thank you, you answered all my questions and got me to ask better ones. I really admire the patience and love that go into giving an idea away and teaching everybody and their dog to make it work. I'm now satisfied that both sides are treming in contrary motion.  I still don't hear the phasey sound I heard in the demos. I have a few other boards and will definitely breadboard before no 2... and read the whole thread again.

Try this:

Remove the 150pF cap from Q1.

Increase the 2n2 cap connected to the switch to 4n7.

[mwnovak]

Here's what I'd do: Use a 2N5457 for Q3 at the very least, and raise its bias to around 4-6V (just decrease the 4.7K a hair, 3.3K usually does it in mine).

Hi Jon.  Hi all.  I wanted to report back on my quest for a bit more clean headroom on my Cardinal build.  So I swapped Q1-Q3 to 2N5457s from Smallbear and rebiased as follows (w/ a 9.21v power supply):

Q1 7.90v w/ 1.2k R4
Q2 6.89v w/ 3.6k R7
Q3 5.57v w/ 1.2k R9

Do these values seem reasonable?  I'd expected to fiddle with R9 to get the Q3 voltage up as suggested, but I was surprised at how much I needed to tweak R4 to get Q1 back near 8v.  The end result appears good--much more clean headroom, an overall more useable circuit for my guitar/amp (with GORGEOUS harmonic tremolo )--but I wanted to sanity-check the numbers.

Thanks again for any help,

--Matt

[midwayfair]

Here's what I'd do: Use a 2N5457 for Q3 at the very least, and raise its bias to around 4-6V (just decrease the 4.7K a hair, 3.3K usually does it in mine).

Hi Jon.  Hi all.  I wanted to report back on my quest for a bit more clean headroom on my Cardinal build.  So I swapped Q1-Q3 to 2N5457s from Smallbear and rebiased as follows (w/ a 9.21v power supply):

Q1 7.90v w/ 1.2k R4
Q2 6.89v w/ 3.6k R7
Q3 5.57v w/ 1.2k R9

Do these values seem reasonable?  I'd expected to fiddle with R9 to get the Q3 voltage up as suggested, but I was surprised at how much I needed to tweak R4 to get Q1 back near 8v.  The end result appears good--much more clean headroom, an overall more useable circuit for my guitar/amp (with GORGEOUS harmonic tremolo )--but I wanted to sanity-check the numbers.

Thanks again for any help,

--Matt

Those all look fine. The biases aren't critical, and especially not on Q1 -- it's just there to provide high input impedance and bump up the gain to make sure it's above unity (really, a J201 was never the best call there, but I wanted to simplify the bill of materials). You could have left R4 the way it was (it probably would have settled near 6.5-7V). I gave Josh the voltages for the build doc so people could have some idea if they were even close to my original build.

I wonder if at this point, I should give Josh a new bill of materials that will run on 2N5457s instead of J201s. Everyone including me seems to like those better.

[midwayfair]

Folks, I'm going to say that the Macron clones can't he used in the cardinal with the values shown. I have a friend who is going to help fix the issue, since he has two built with both sets of vactrols (and he says that the Macron photocells sound completely different and have much less depth). If we discover a solution, I'll be sure to post it here.

[soggybag]

Such a clever way to create a new sound from the old tremolo. My question is about the indicator LED and Vactrol LEDs. I recently built this on the 1776 pcb.

I found the main tremolo section worked perfectly but the high pass section hardly trembled at all. Multimeter could read the changing resistance of Vactrol 1, but Vactrol 2 seemed to not be changing at all. I disconnected the low pass side completely (lifted one leg of the LDR) so I could hear the high pass side by itself. Then I jumpered the current limiting resistor for the second vactrol and then I could hear the treble throb, subtly. The main channel provides complete silence on the wave troughs with depth set high, but even with the CLR jumpered the high pass side is just gentle.

I thought "with the three LEDs in series are those CLRs necessary?" I thought maybe the intention was to provide lower voltage to vactrol 2 for a more subtle effect. I breadboarded the three LEDs and CLRs as per the schematic (with normal LEDs replacing the vactrols), and they seem to be equal in brightness. So the designer had his reasons...

Don't you really only need one resistor on the end of the three series LEDs?  The 9 + volts already go through a diode and a 100ohm resistor first n the power section and each LED drops around 2v.

I am thinking of replacing vactrol 2. (I am using the recommended VTL5C1, and I bought extra.) Should have breadboarded first!

Also: Why diffused on the indicator? I am not aware of any reason why a diffused would behave differently than a clear. Or why a 3mm should have different fv than a 5mm. On the other hand, color would seem to be important. I first used a blue one and found a red seemed to work better, lower voltage drop, leaving a little more juice for Vactrol 2.

"Diffused" is usually red, orange, yellow, or green. I know there are super bright blue diffused (I have a bunch), but those are kind of rare. I guess we should have specified the color. Anyway, it can't be blue or white from what I've tried; the total voltage drop seems to mess things up a little, and I couldn't get any depth out of it. As you discovered, red works better. Green is still fine and is actually the color I tend to use.

It says 3mm in the schematic because in Eagle, the part has the measurements. In Josh's PCB, it says 3mm because that's the size of the pad. The size doesn't matter.

As far as the depth problem on Q3:
It's possible to have a vactrol out of spec enough that it won't have on resistance low enough to pull the gate of Q3 low enough to trem completely. It's unfortunate, because although vactrols are more consistent than most discrete LDRs, they are a heck of a lot more expensive if they have to be replaced. I have one VTL5C1 in my box of parts right now that works on Q2 but not on Q3, since the light resistance never seems to get below 30K and most of the time isn't below 100K (!).

But before you go pulling an expensive part off the PCB, set the depth to minimum and measure the resistance of the LDR side of vactrol 2 and compare it with Vactrol 1.

You can also try simply jumpering the indicator LED to see if that helps. (If it does, there's another way to wire a rate indicator LED: Just run some wire from pin 7 to an LED's anode, and connect the cathode to ground through a CLR. Ugly, I know, but still probably better than desoldering an $8 part.)

The extra resistors to ground increase the length of time the LED spends on the dark swing. You can see something similar in the Tremulus Lune, from which the LFO in this is derived. The cascading values were needed to keep the LEDs similar brightness. One thing that happens with them is that they tend to soften the on-off, so the sine wave side is a lot softer than the original lune's (the extra LED helped, too).

On the other hand, Vactrol 2 doesn't need dark resistance any where near as high as vactrol 1, so there are actually other optocouplers that will work. If you have some other stuff in your bin, measure a couple and see if you have one with light resistance below 5K and dark resistance anywhere above, say, 500K (but preferably above 1M) and a relatively far on/off time. That should provide a sufficient range. Just a heads up: I can tell you for certain that the NSL-32xxx ones DON'T work, though; they never get dark enough and seem to have a turn off time much slower than their datasheets claim.

I had exactly this problem! I just built a Cardinal board from 1776. The tremolo effect works but, the phasey trem was not sounding correct. I have built a similar project before so I'm familiar with what this should sound like. After futzing around for a bit I realized the treble side was not changing in volume.

I removed the the Vactrol on the bass side and I only hear trebly guitar no changing in volume. I removed the vactrols, they were in sockets, and inserted two LEDs. The bass side, VAC-1 on the 1776 schematic, cycles from bright to dark correctly. VAC-2 on the treble side is dark most of the time, with a tiny blip of light for each cycle. Obviously this is the problem.

I'm trying to decide what to do. I'm using VTL-5C1 for the vactrols, these are the correct part. It seems the problem has to do with the current driving the LEDs, not the LDR. I measured R16, 17, and 18, these are all the correct values. Seems there is not enough current to drive the third LED in the chain.

I am using a 5mm LED for D2, the parts list mention this should be 3mm. This is a random red LED, not a super bright, from the bin. Now that I've gotten this far, I'm thinking D2 might be the culprit. Adding a jumper across D2 should be a simple test...

Sure enough, as soon as I jumper D2 VAC-2 lights up and cycles like it should. I inserted the VTL-5C1s again, turned the speed up to max and measured the voltage across each LED.

• D2 1.76V
• VAC-1 1.5V
• VAC-2 0.8V

I'm guess the voltage drop across D2 is too great. I can try other LEDs here. I'm wondering if there is anything else I can do to increase the current to VAC-2?

[midwayfair]

You can move the indicator LED and make just a speed indicator LED (no depth indication) by connecting the anode to pin 7, cathode to a resistor, and the other leg of the resistor to ground.

You might also try a different op amp, sometimes that's an issue in this oscillator.

You can try fiddling with the resistors near the LEDs. They change the "dark" cycle time for the LEDs.

[soggybag]

The more I look at the LEDs I feel like the two LEDs in the LDRs should be in parallel. In series the last LED is in a spot where the operating voltage is very narrow. There is a voltage drop across each LED. The last in line getting barely enough enough to turn it off and on. This also makes the depth control less effective.

With the LEDs in parallel they both get a better range and more current.

Another alternative might be to use a single LED driving two LDRs.

[midwayfair]

The more I look at the LEDs I feel like the two LEDs in the LDRs should be in parallel. In series the last LED is in a spot where the operating voltage is very narrow. There is a voltage drop across each LED. The last in line getting barely enough enough to turn it off and on. This also makes the depth control less effective.

With the LEDs in parallel they both get a better range and more current.

Another alternative might be to use a single LED driving two LDRs.

If you put them in parallel, you have to match the Fv in vactrols, which means you need a pile of vactrols around. 5C1s are really expensive, so that's just making an expensive project even more expensive. If they don't match, one will suck up most of the current and the other will barely turn on. I've tried this in the past.

If you want there to be no difference in voltage between the two vactrol LEDs, remove the 4.7K resistor between the indicator LED and the first vactrol. The voltage drop across the two remaining LEDs will be identical then, because they'll be in series. (Or R16 in the 1776 schematic, which is backwards from mine.) But then you have to deal with the fact that you don't WANT the same on/off time for both FETs, because they create the tremolo effect differently. But by all means, go ahead and try it, because it's your pedal that you're building. If you're using the PCB, you can just lift one leg of the vactrol and run a little wire to tie the cathodes together and add a jumper to tie the anodes together.

And I'm not trying to be rude, but your comments are making it sound like the circuit doesn't work, when there are dozens of properly working builds out there (just going by build reports) built on the schematic. The only issues I've ever run into with the vactrols is (a) the Smallbear Macron work-alikes don't work alike and (b) I had an actual dud vactrol once and (c) there's not enough current to use a superbright for the indicator.

[soggybag]

No offense taken. I'm just trying to discuss the subject. It seems like what you have here works but, it requires finicky part choices. I'm just suggesting there might be room for improvement. People have been trying to improve the Fuzz Face for 40 odd years now.

I'm also trying to discuss this to solve my issues. I have the same problems idy seemed to have. I'm using a 1776 PCB, and VTL-5C1 for the vactrol which is the correct part.

If the LEDs are in parallel but, each have their own current limiting resistor current should be divided between them and they will both light up without being matched.

I'm trying to puzzle out the math involved. I'm sure there will be a few mistakes here. I'm guessing we get about 4.5v to 9v at D2. Each LED has roughly a 2v voltage drop. After the first two LEDs there is very little left to drive the third LED.

[anotherjim]

Just a thought,
If the LED's are in series, but you want to tweak their relative brightness, and you know which one needs to be less bright - put a trimmer resistor across that one. A 5k trimmer should give enough range. This resistance will bypass some current around it, making it weaker.

Jim

[solderfumes]

I'm hoping to build a tap-tempo enabled version of this in the new year (I have a friend coming from the US who is gonna ferry back a Small Bear order for me).  I found the schematic a few pages back, but wanted to double-check that I can use a 78L05 in place of the 7805 voltage regulator.  The current draw on the TAPLFO and LEDs won't be higher than 100mA, right?

[drolo]

I'm hoping to build a tap-tempo enabled version of this in the new year (I have a friend coming from the US who is gonna ferry back a Small Bear order for me).  I found the schematic a few pages back, but wanted to double-check that I can use a 78L05 in place of the 7805 voltage regulator.  The current draw on the TAPLFO and LEDs won't be higher than 100mA, right?

That should not be a problem. I have been using the TAPLFO with 2 Optocouplers fed by a 78l05 in a different tremolo and have never had issues. (I never measured the current though, but it's been running every day for more than a year...)