Sound to light LED built into bass

[Mick Bailey]

An associate is including sound to light in a bass - an array of 12 high-brightness blue LEDs flashed in unison. The intention is to drive this off the pickup signal. He sent me details of where this was so far but I can see issues (low impedance input to driver circuit, no DC blocking, insufficient drive from directly connected opamp output etc...).

So, with a new design there would be no problem in flashing the LEDs and using a high-impedance buffered input with a decent MOSFET or BJT capable of easily handling the load current,. Peak consumption on pulsed mode is 420mA - so maybe a few hours with a LiPO rechargable battery of maybe 900mA rating.

But, the main problem I perceive is noise from switching the LEDs. Just the same issue as any digital/analogue system where the analogue ground gets polluted. Who has done this and had good results?

[dmeek]

I'm guessing the 12 LEDs, if they are regular T 1 3/4 size, would require far less than 100mA. Rather than switching maybe they could be driven with an amplified analog signal from an LM386.
Has decided which LEDs he wants to use?

[Enzo]

You can run LEDs in series, so the same 10ma can power a number of them. DOn't wire them up in parallel, that would be maximizing current need.

Control your LED strings with something analog like an op amp or tiny power amp or even a transistor. As opposed to straight ON/OFF. That way they throb instead of blink.

[Mick Bailey]

The LEDs are NSPB510AS. In pulsed mode they can be run harder (up to 110mA max depending on pulse width/duty cycle) than if continuously lit. This gives the possibility of a really bright display. With a nominal 9v supply and factoring in a reducing voltage I wasn't expecting more than 2 to be able to be run in series, so 6 parallel strings of 2 LEDs. There's the possibility of reducing the current consumption and reducing peak draw, though because the leds aren't continuously lit they need running harder then usual to give an impressive display on stage.

The main issue is potential switching noise just from the LEDs turning on and off. I mention digital/analogue because it presents the same problem of noise getting onto the audio side. The bass will also have an inbuilt Big Muff. A while back a customer wanted a transparent sound to light edge-lit base to a fuzz pedal. A simple BJT circuit and 4 green LEDS worked fine but I couldn't get rid of the noise when the LEDS switched. Maybe Enzo's idea is the best way to go - throbbing rather than blinking so they never turn right off. With high brightness leds I can see this working out because the difference between just about turned on and full brightness on peaks is huge. A little bias control maybe to just get them conducting.

[dmeek]

I was thinking of something like this. I didn't breadboard this but assuming the output sits at +4.5V with no signal, no LEDs will be lit. If you don't need the extra gain the capacitors can be left out

https://music-electronics-forum.com/...1&d=1588362152

[nickb]

Here's some inspiration for you https://bmamps.com/v01/2017/10/13/re...des-pixelator/

All runs off a serial data stream. Full 256 bit (IIRC) RGB pixels in a square grid. The body was milled out and PCB stuffed with LEDS and driver chips fitted in the cavity. The nice thing about the approach is the flexibility and all the heavy lifting is moved off the instrument.

If having it all on board is an essential requirement, I would use a PIC micro with FETs to drive the LEDS. A small bit of firmware could give you so much more flexibility than the LM386 proposal. Happy to help with coding if needed.

PS: Or maybe use I2C or SPI out of the micro to LED driver chips for more flexibility. Split the analogue inputs into low / mid / high bands to make it more interesting. Fun for all.

[Mick Bailey]

I'll give the LM386 idea a go today. It would need a FET input to increase the input impedance and I'm interested in seeing how well a 386 would drive the LEDs directly. I like the idea of using a PIC as a driver but just for a simple array it's maybe overkill at this stage. I do have programming kit and use Proteus as it does the circuit simulation including running C# code. I think it's about three years since I last used it to develop a programmable electronic ignition system so I'm very rusty by now.

That Pixelator guitar gives pretty much everything else an early bath and it reminds me of how a cuttlefish changes colour. I just watched the video a dozen times and it got no less impressive 12th time round

I sent a link to the guy building the bass for his consideration. Maybe that will shift the direction he was planning.