Moid's Random Tremolo

[moid]

I did some checking on the mouser phototransistors and they are all aimed at the infrared end of the spectrum so I don't think they'll work with the coloured LEDs I'm using. I did find TEPT5700 which are aimed at human visible light, so I'll buy a few of those in case they come in handy, but I'm hoping to use one of the other LDR types I'm ordering.

[PRR]

The IR peak is intrinsic in Silicon photosensors. Digital cameras mostly need IR filters so we don't see "ghosts". (Or they use IR response to face-check in near dark.)

LEDs are BRIGHT. You do not have to worry about color preference in most cases. I drove lots of "green peak" LDRs with red LEDs. Often the problem is "too much" response.

[amptramp]

We were trying to build a touch panel for avionics use a few years (decades?) back and we decided that there was no way to filter out bright sunlight at high altitudes with IR sources, but we used green LED's with other green LED's as photosensors.  LED's will work as photosensors, but they have a high leakage current and this leakage has a strong temperature coefficient (higher temp = more leakage) and the leakage current is on a par with the signal current (sort of like germanium).  We couldn't make it work for our purposes, but it is another possibility to consider.

[Elektrojänis]

I kind of dropped out of this conversation for a while and I'm not sure if someone else answered or you gathered the info somewhere somehow, but anyway... I'll continue with in some things I started with.

A few ideas:

1. You could try to arrange several light sources so they only need one LDR. Use some sort of small light tight container where you have an LDR in one spot. Then arrange all the flickering leds to point the brightest spot of their light towards the LDR.

Thanks! What would be the benefit of only having one LDR? Would that reduce noise then?

Benefit of one LDR only would mainly be saving parts. The result will quite possibly be very similar if yo can arrange the leds pointing to the single ldr. Maybe you could use some sor of small container for them and hot glue the parts in. I'd probably use an old black 35 mm film canister but I quess the availability of those is not as good as it used to be... And some are translucent instead of black, which does not help with stray light at all.

2. Play around wiht the value of your R2. It should affect the depth of the effect, especially if you put a buffer or another effect in front of the Moidulator. With a passive guitar straight in to the R2 it might not make as big difference.

Just tried changing R2 - 100k resistor kills all sound, a 1k makes everything much louder and clearer! Thanks for that, I have changed to 1k. You were right about passive guitar - I couldn't hear much difference between the original 10k and the new 1k, but if I stick a boos pedal in front of the circuit there is a huge volume difference - same thing happens if I use a boost pedal.

The thing with those LDR's and the R2 is that they act as a voltage divider. It's much like a volume pot. Maybe a simplified picture will help. I this doesn't go too simple.



Volume pot arranged like on the left is probably familliar to you. It works like the thing I drew on the right. The two resistors added is the total value of the volume pot and turning the pot makse one of the resistors smaller and the other bigger.

If you strip down the signal side of the Moidulator to bare minimum with only one LDR, you also get the circuit on the left side (I even wrote R2 and LDR there). The output signal will be the (LDR/(LDR+R2))*input.

When the Moidulator is working, the LDR is changing acording to the light from the LEDs. R2 is not changing. If R2 is too small compared to the range of resistances of the LDR (from dark to light) the signal throigh will be loud, but not changing much. If the R2 is too big the signal through will be too quiet (and not changing that much).

The key to maximizing the modulation depth, is to find a value for R2 that is similar to the average resistance of the LDR for the operating conditions (light brightness) you put it in. Or somewhere in that ballpark anyway. If it get's too quiet, you need a booster... You have that booster there now already, and it is really needed, as the core element can only make things quiter.

Oh... Back to circuit with 4 LDR's. The rest of those are parallel to the first one so you might need slightly lower R2 with all four of them. Only slightly smaller. Like if the optimum was x ohms with one, then the optimum with 4 LDR's is x/4 ohms at the very minimum. As the all LDR's are modulated differently, x/2 is likely closer to optimum, but it really depends on the modulation your flickering leds provide.

I hope that makes sense to you and maybe helps a bit.

Edit: Oh... If there is no buffer at the input, the guitar pickup itself is in series with R2. You can consider it as a part of R2 on my simplification. The pickups usually have a resistance from a few kilo ohms to maybe 20 kohms, but that's at DC (0 Hz). At audio frequencies the impedance (resistance when it is not same for all frequencies) is much higher. That's why changing the R2 without a buffer/booster didn't change much, but with buffer/booster the change was more noticeable.

[moid]

The IR peak is intrinsic in Silicon photosensors. Digital cameras mostly need IR filters so we don't see "ghosts". (Or they use IR response to face-check in near dark.)

LEDs are BRIGHT. You do not have to worry about color preference in most cases. I drove lots of "green peak" LDRs with red LEDs. Often the problem is "too much" response.

aah... well I already ordered the parts, now I have to wait... I will rememberthat for next time though. I just lookd at the diagram in the datasheet that showed responsiveness to light and assumed I should always look at the brightest as being the best... it never occurred to me that too bright was an option! I've ordered a bunch of LDRs so hopefully the circuit can be made into something good with those and I won't have to look into how to add phototransistors (I did buy some of those) because I have no idea how to integrate those into the circuit. Fingers crossed the LDRs are what I need! I also ordered some larger breadboard because the circuit is too big as it is to add more parts.

We were trying to build a touch panel for avionics use a few years (decades?) back and we decided that there was no way to filter out bright sunlight at high altitudes with IR sources, but we used green LED's with other green LED's as photosensors.  LED's will work as photosensors, but they have a high leakage current and this leakage has a strong temperature coefficient (higher temp = more leakage) and the leakage current is on a par with the signal current (sort of like germanium).  We couldn't make it work for our purposes, but it is another possibility to consider.

That's interesting amptramp - could you not use UV filters like you get for camera lenses, or are they not strong enough at high altitude? I remember being taught to use a UV filter as a standard precaution on any lens in case you dropped it (this was at art college; they were probably tring to stop us students from breaking expensive lenses!)... and it was just a good idea when shooting colour slide film of daylit landscapes anyway (removed a slight purple tint from the film). I wonder what digital cameras on phones do? Presumably they have some kind of built in filter on the lens?

[moid]

I kind of dropped out of this conversation for a while and I'm not sure if someone else answered or you gathered the info somewhere somehow, but anyway... I'll continue with in some things I started with.

Thanks very much!

Benefit of one LDR only would mainly be saving parts. The result will quite possibly be very similar if yo can arrange the leds pointing to the single ldr. Maybe you could use some sor of small container for them and hot glue the parts in. I'd probably use an old black 35 mm film canister but I quess the availability of those is not as good as it used to be... And some are translucent instead of black, which does not help with stray light at all.

I have seen something similar inside my Moen Shakey Jimi vibrato pedal - that has two LDRs both aimed at an LED (I think, it has been a while since I opened it). In the original pedal the LDRs are a certain distance from the LED, and an easy mod to make a stronger effect was just to bend them closer to the LED and angle them at it! In the original they point upwards, parallel with the LED! So tilting them by 90 degrees and putting them closer gives a much stronger vibrato effect. Probably the easiest mod I've ever done

Anyway, yours is an interesting idea - for now I will try the multiple LDRs because I'm wondering if the different LEDs I am using would each benefit from an LDR with different characteristics - maybe, maybe not, but seeing as I've ordered some different ones I am going to find out!

The thing with those LDR's and the R2 is that they act as a voltage divider. It's much like a volume pot. Maybe a simplified picture will help. I this doesn't go too simple.

I'm not sure if it is possible to go to simple for me! I have a tenuous grasp of electrical theory at the best of times, and when I do make / change things it's kind of like watching an orang utan that has been given some tools to play with... I know these things in front of me do stuff, but the order you have to put them together can be challenging and the outcome may be wild and experimental with more passion than understanding!

Volume pot arranged like on the left is probably familliar to you. It works like the thing I drew on the right. The two resistors added is the total value of the volume pot and turning the pot makse one of the resistors smaller and the other bigger.

If you strip down the signal side of the Moidulator to bare minimum with only one LDR, you also get the circuit on the left side (I even wrote R2 and LDR there). The output signal will be the (LDR/(LDR+R2))*input.

When the Moidulator is working, the LDR is changing acording to the light from the LEDs. R2 is not changing. If R2 is too small compared to the range of resistances of the LDR (from dark to light) the signal throigh will be loud, but not changing much. If the R2 is too big the signal through will be too quiet (and not changing that much).

The key to maximizing the modulation depth, is to find a value for R2 that is similar to the average resistance of the LDR for the operating conditions (light brightness) you put it in. Or somewhere in that ballpark anyway. If it get's too quiet, you need a booster... You have that booster there now already, and it is really needed, as the core element can only make things quiter.

Wow, thanks that's brilliant - so that's why I need R2 in the first place, and also why I shouldn't just leave it out! So this means if I change the LDRs I may need to change R2 to compensate. This also explains why I thought the 'off' part of the tremolo was a bit deeper when I was using 10K for R2; however the overall volume of the circuit was much quieter. A switch to 1K made the circuit a lot louder (with a boost and / or a Boss buffer) but I thought the quieter parts were not as quiet as they were before. OK I will keep that in mind when the new parts arrive, and will try them and also adjust R2 if I don't get the volume where I want it to be.

Oh... Back to circuit with 4 LDR's. The rest of those are parallel to the first one so you might need slightly lower R2 with all four of them. Only slightly smaller. Like if the optimum was x ohms with one, then the optimum with 4 LDR's is x/4 ohms at the very minimum. As the all LDR's are modulated differently, x/2 is likely closer to optimum, but it really depends on the modulation your flickering leds provide.

I hope that makes sense to you and maybe helps a bit.

It does thanks - I will bear this in mind when 'tuning' the circuit!

Edit: Oh... If there is no buffer at the input, the guitar pickup itself is in series with R2. You can consider it as a part of R2 on my simplification. The pickups usually have a resistance from a few kilo ohms to maybe 20 kohms, but that's at DC (0 Hz). At audio frequencies the impedance (resistance when it is not same for all frequencies) is much higher. That's why changing the R2 without a buffer/booster didn't change much, but with buffer/booster the change was more noticeable.

Aahh so the buffer acts as... well, a buffer! (now I know how it got its name!) it sits between the pickup and the circuit so the pickup doesn't become part of the equation for the voltage divider. Thanks again, I've learned a lot from this post.

[Matthew Sanford]

To piggyback on Electrojänis, for R2, you might try a trim pot with a high value with pin 2 connected to pin 3 - makes it a variable resistor, turning CW reduces it. Once it's what you want, check between pin 1 and 2 and that should be your R2 value

[duck_arse]

To piggyback on Electrojänis, for R2, you might try a trim pot with a high value with pin 2 connected to pin 3 - makes it a variable resistor, turning CW reduces it. Once it's what you want, check between pin 1 and 2 and that should be your R2 value

to piggybank on Sanford/Janis, if you made R2 one of the ldr's, you could add some random depth. but it will still be noisy if lit by a candle.



my camera teacher said always have a daylight on all your lenses, because replacing a filter covered in dirt and smudges and grime is cheaper than replacing a fingerprinted lens. I used rubber hoods/shades, and tried not to drop lenses.

[moid]

Surprise everyone! I finally got a couple of hours away from work to redraw the schematic with a buffer in front and after the circuit.

Moreover just make sure a midpoint voltage (4.5 for 9 Vcc) is present on the + in. I don't think resistor values matter much if you set it like Jack's example, however... I like to set up the divider with a couple 10ks, one to V+, one to ground, then where they meet (giving 4.5 volts as half the power supply) use a 1 Meg resistor to bring it to the op amp. You should also put a cap, say 10uF from that midpoint to ground to keep it stable-don't use the cap if it's going to an LFO generator (thanks to Duck for that). Reason for lower bias divider resistors and big ones to bring it in is that big one essentially is parallel to the one going to ground, so making it decades larger doesn't change the divider resistor value (much...)

Use a dual, it won't mess with it, so save space!

Sorry this is rather cramped (and probably confusing). I am fairly sure I got the input buffer part right according to Matthew's instructions, however I'm not sure if I got the output buffer part correct... I couldn't decide if the output buffer of the TL072 required the same voltage dividor components or whether I could take a cable from the input buffer components to pin 5 of the TL072 or not... in the end I added the same components to both the input and output buffers so please let me know if I just unnecessarily doubled up the components!

Another thought - should the power to the TL072 and the DOD 210 part of the circuit come after the 1N5187 diode that is part of the power supply for the  LEDs?

If you think the above circuit would work I will start ripping the breadboard mess apart and test it (it might be a few weeks before this happens, work is going to be relentless until near the end of November ) so dont be holding your breath...

Hope you're all having a great weekend!

[moid]

To piggyback on Electrojänis, for R2, you might try a trim pot with a high value with pin 2 connected to pin 3 - makes it a variable resistor, turning CW reduces it. Once it's what you want, check between pin 1 and 2 and that should be your R2 value

That's a good idea! Thanks very much, OK I'll try that once I get to breadboarding the new schematic.

[moid]

To piggyback on Electrojänis, for R2, you might try a trim pot with a high value with pin 2 connected to pin 3 - makes it a variable resistor, turning CW reduces it. Once it's what you want, check between pin 1 and 2 and that should be your R2 value

to piggybank on Sanford/Janis, if you made R2 one of the ldr's, you could add some random depth. but it will still be noisy if lit by a candle.



my camera teacher said always have a daylight on all your lenses, because replacing a filter covered in dirt and smudges and grime is cheaper than replacing a fingerprinted lens. I used rubber hoods/shades, and tried not to drop lenses.

Random depth is a cool idea - but I expect you mean an LDR on the outside of the enclosure? I suspect that could be highly irritating to control (a friend of mine once lent me a weird DIY fuzz pedal he had bought from ebay that had an LDR on the enclosure to add oscillating noise in some way - which it did (I think it was supposed to modulate the gain somehow), but it always made annoying noises unless the level of illumination hitting it was just right. He said it was cool on stage with flashing lights, but it was next to useless at home if daylight came through a window. This was before I got into building pedals so I never took any photos of the insides and he got sick of it and sold it, so I've no idea what it was... only that I don't think I'd want an LDR on the outside of the box!

Yes our teacher showed us in a catalogue how much we would have to repay if we broke a lens versus how cheap a UV filter was to replace... so I remembered that lesson (also I try not to drop lenses, and I've so far been good at that!)

[amptramp]

Random depth is a cool idea - but I expect you mean an LDR on the outside of the enclosure? I suspect that could be highly irritating to control (a friend of mine once lent me a weird DIY fuzz pedal he had bought from ebay that had an LDR on the enclosure to add oscillating noise in some way - which it did (I think it was supposed to modulate the gain somehow), but it always made annoying noises unless the level of illumination hitting it was just right. He said it was cool on stage with flashing lights, but it was next to useless at home if daylight came through a window. This was before I got into building pedals so I never took any photos of the insides and he got sick of it and sold it, so I've no idea what it was... only that I don't think I'd want an LDR on the outside of the box!

Yes our teacher showed us in a catalogue how much we would have to repay if we broke a lens versus how cheap a UV filter was to replace... so I remembered that lesson (also I try not to drop lenses, and I've so far been good at that!)

There is one reason to put a photoresistor on the outside of the box - to control the indicator LED current.  This allows you to use a high-brightness LED in full daylight or a dark club ambient and allows you to see the light under both conditions without being blinded.  You may need parallel and series resistors to "tune" the brightness response.

[Matthew Sanford]

Surprise everyone! I finally got a couple of hours away from work to redraw the schematic with a buffer in front and after the circuit.

Moreover just make sure a midpoint voltage (4.5 for 9 Vcc) is present on the + in. I don't think resistor values matter much if you set it like Jack's example, however... I like to set up the divider with a couple 10ks, one to V+, one to ground, then where they meet (giving 4.5 volts as half the power supply) use a 1 Meg resistor to bring it to the op amp. You should also put a cap, say 10uF from that midpoint to ground to keep it stable-don't use the cap if it's going to an LFO generator (thanks to Duck for that). Reason for lower bias divider resistors and big ones to bring it in is that big one essentially is parallel to the one going to ground, so making it decades larger doesn't change the divider resistor value (much...)

Use a dual, it won't mess with it, so save space!

Sorry this is rather cramped (and probably confusing). I am fairly sure I got the input buffer part right according to Matthew's instructions, however I'm not sure if I got the output buffer part correct... I couldn't decide if the output buffer of the TL072 required the same voltage dividor components or whether I could take a cable from the input buffer components to pin 5 of the TL072 or not... in the end I added the same components to both the input and output buffers so please let me know if I just unnecessarily doubled up the components!

Another thought - should the power to the TL072 and the DOD 210 part of the circuit come after the 1N5187 diode that is part of the power supply for the  LEDs?

If you think the above circuit would work I will start ripping the breadboard mess apart and test it (it might be a few weeks before this happens, work is going to be relentless until near the end of November ) so dont be holding your breath...

Hope you're all having a great weekend!

It looks pretty good. On the bias, take the 10ks divider off your input line. Have the divider set up, then use the 1M resistor from the midpoint of the 10ks to the non-inverting input. The signal doesn't need to go through it, it needs to bring that voltage to the signal line. You can use one 10k divider, and one 1M resistor to take it to each input. Also good to have a 10u capacitor from divider midpoint to ground.
Some examples cut from others