Bi-color LED for LFO rate wiring?

[pavels07]

Apologies if this has been addressed elsewhere.

I am working on a project based off of a Phase 90 and I have successfully added a rate LED using Circuit D found on this writeup I came across on the forum:
https://postimg.cc/fSSNt8Jm

My question/idea: Is there a way to implement or modify this circuit to work with a Bi-color LED (common cathode) to get it to cycle between the two colors? I did some testing giving one half of the Red/Blue LED a constant voltage from my 9v jack (blue side) and I could see the red side blinking, but it was easily overpowered by the constant blue. I then tried changing the dropping resistor on the blue side to make it dimmer and give a better chance for the red side to been seen when flashing, but that was undewhelming to say the least.

I've seen something similar to what I am hoping for be done by Spaceman Effects, so I know its possible, its just over my head.

I did come across this Instructables Page https://www.instructables.com/Simple-Blinking-LED-Circuit/  which achieves what I want in a 2 LED setup, but I don't know how to implement it in the circuit.

Including the section of my schematic for reference.

[Ripthorn]

I did something like this with Aeronometron circuit. If I recall, I just balanced the red and blue using current limiting resistors, but you could use the LFO with a transistor to change the balance as well. https://scientificguitarist.wixsite.com/home/aeronometron

[MrStab]

Maybe. TL062, big resistors and ultra-bright LEDs. Not sure how disruptive it'd be re. LFO clicking.

If you wanted to get more hardcore, ie. Red, off, Blue, off, then maybe a CD4017 decade counter with outputs 0 and 2 set to either anode, 1 and 3 empty, and RESET wired to Output 5. LFO op-amp output connected to CLK. Not sure if op-amp LFO would hit logic low threshold though, especially with the TL0* series always being 2V off the negative rail. Rail-to-Rail op-amp might work, but ofc affect functionality on the phaser FETs with a wider voltage swing.

Expand that concept a couple more 4017 output pins and use an RGB LED for God Mode.

Actually nvm that wouldn't make sense unless CLK was double the LFO rate. I'll stfu and go to bed. Lol

[Rob Strand]

The original ckt modulates the level of the LED.  The way it modulates is a fine balance of the Vgs_on of the MOSFET the LED voltage and the voltage swing on the gate.

Trying to get a complementary RED/BLUE LED with a common cathode to do that will end-up with a finicky circuit.

Things can be greatly simplified by having the LEDs go on/off only.  That's pretty much what MrStab's ckt does.

An alternative to adding an opamp is something like this:

I'm sure you can get it to work but it might need some fine tuning.  Ideally you want the Red and Blue LEDs to have equal perceived brightness.    You might even have to swap the positions of the LEDs in the circuit to get that to happen.


In other circuit, having the LED switch off/on in a hard manner like this can induce clicks on the power rails which then get into the audio.

[MrStab]

Flip flop controlling anodes, one on Q and the other on -Q, while LFO output controls cathodes and CLK?

Ok I promise I am going to bed now

[MrStab]

https://tinyurl.com/2o5on5ru

[Rob Strand]

https://tinyurl.com/2o5on5ru

That's going to do red for a whole up and down sweep then blue for a whole up and down sweep.
(The previous solutions are level dependent, instead of clock-edge, so the LEDs follow the sweep.)

[MrStab]

So transistor on the cathode with triangle output on base or gate instead of square wave output, which may increase part count beyond practicality. Just didn't think analogue level was explicitly mentioned as a requirement to begin with, though perhaps implied.

A 2 op-amp LFO could remove need for transistor here.

[MrStab]

Alternating, analogue LED ramping (common cathode) with quad op-amp. Unused op-amp used for buffered Vref.

https://tinyurl.com/2rxdjtb7

Cathodes to Vref may be preferable to cathodes to ground.

[pavels07]

Awesome suggestions.
Special thanks to Rob Strand and MrStab for really diving deep on this project.

I think the closest thing to what I had envisioned/poorly described was achieved by MrStab's latest post

Alternating, analogue LED ramping (common cathode) with quad op-amp. Unused op-amp used for buffered Vref.

https://tinyurl.com/2rxdjtb7

Cathodes to Vref may be preferable to cathodes to ground.

Now my other, more important question is how do I shoehorn that into the schematic? I'm still relatively new at self-taught circuit designing (especially something this complex), and this is my first "throw in the deep end, see if I survive" project I'm challenging myself with. So while I have some understanding of the jargon, its still pretty rudimentary to me.

I appreciate all the help and suggestions!

[MrStab]

I have an idea how you could retrofit something similar into the existing LFO using just a dual op-amp, which I can draw out once I bath my terror of an offspring.

A good case for doing the whole quad op-amp shebang and ditching the existing LFO would be if you're currently using the audio Vref for LFO Vref, because ideally you'd want an independent Vref for the LFO anyway (as mentioned by someone in another recent thread about the Phase 90). But maybe better to start simpler.

[MrStab]

Using a dual op-amp (hopefully you can see the original LFO - I've substituted the pot with a 1k fixed resistor for now):

https://tinyurl.com/2e4v5bbu

What's going on here? The op-amp furthest left is a unity-gain buffer, intended to avoid loading down your phaser control voltage in any way and to hopefully provide some uniformity. This feeds LED 1. The next op-amp is a unity-gain inverter, driving LED 2 in the opposite direction.

4.5V is, ofc, your Vref.

Putting the cathode(s) to ground would get you brighter LEDs but they may not go fully off, whereas putting it to Vref would make them dimmer but make them go fully off. Why? Because what matters is the relative voltage potential. Say your LFO swings from 2 to 7V. If cathode is referenced to ground (0V), then at the peak of the swing, the anode sees 7-0=7V, and at the trough it sees 2-0=2V. If the cathode is referenced to 4.5V, then at peak it's 7-4.5=2.5V max, and reverse-biased (-2.5V) at the bottom. Hope I'm making sense in my sleep-deprived state.

I'm omitting diode drop and forward voltage from this explanation, which will further complicate matters. Experimentation is required. Because of forward voltage (which is typically around the same voltage as the bottom of your LFO swing), I would try cathodes to ground first.

Alternatively, maybe it'd be simpler to use an N-channel (left) and P-channel MOSFET (right), such as the 2n7000 and BS250 respectively, but I think working in the "ohmic region" of MOSFETS in such a way may be leaving things too much up to chance with the characteristics of individual transistors. I could be wrong. Plus it involves scoring different, less-common parts and all the fun of ESD mitigation.

https://tinyurl.com/2frcp96c

In either case, the series resistors on the anodes of each LED would be adjusted to get the desired brightness of either colour.

[pavels07]

MrStab,
If this is what you are capable of in a sleep deprived state, I'm impressed (and a little terrified of what you are capable of at full strength, haha).

This is exactly what I was looking for with the added bonus of an understandable explanation for a newcomer. I'll give these options a go and do whatever tweaks I need to to get the brightnesses and such how I want them.

Thanks for everything!

[MrStab]

Thank you, and you're welcome! I'm easily in the top smartest 5,000 users here.

If any complete maniac would like to try a decade counter IC to alternate between each colour on an RGB LED, with triangle wave rise and fall, here's how you'd probably do it (can definitely be condensed to a dual op-amp):

https://tinyurl.com/2nx99l2a

If anyone builds this, you must call it "God Mode Disco LEDs" and pay me royalties in Mesopotamian shekels.

Maybe an LM3914 instead of a 4017 could give you even more fun, with signal envelope instead of LFO pulses advancing LED illumination. Dot or bar graph mode for different flavours of silly.

[Kevin Mitchell]

Thank you, and you're welcome! I'm easily in the top smartest 5,000 users here.

If any complete maniac would like to try a decade counter IC to alternate between each colour on an RGB LED, with triangle wave rise and fall, here's how you'd probably do it (can definitely be condensed to a dual op-amp):

https://tinyurl.com/2nx99l2a

Pro tip: replace the decade counter with a binary counter and you get every color variation in between - rather than just red, green and blue 
https://tinyurl.com/2q29t2x6

Edit: Trimmed the fat
https://tinyurl.com/2g9kdxwg

[R.G.]

No, a complete maniac would use an 8-pin DIP PIC and use PWM modulation to generate three-phase sine intensities on the LEDs, and also use the internal NCO hardware to generate the LFO as well. 8-)

[ElectricDruid]

No, a complete maniac would use an 8-pin DIP PIC and use PWM modulation to generate three-phase sine intensities on the LEDs, and also use the internal NCO hardware to generate the LFO as well. 8-)

Lol, that'd be me then

[Matthew Sanford]

Mr. Stab! Wish I'd looked at your first prior to diving in to an MFB solution...but I'm on, I'll finish that way.

If you want to fade back and forth and have a couple vactrols, you could do two voltage dividers with a vactrol to ground on one, 9v the other, other resistors with a resistance a decade higher than your low R on vactrol and another paralleling the vactrol (same R or maybe higher) with the LFO feeding the led part of vactrols. That way it should pan back and forth between colors, right?

[R.G.]

Lol, that'd be me then

I was actually thinking more of me and my recent string of "just use a PIC" posts, but you're welcome to join the club.