Gallery, Projects and General > Project Logs
Diy optical pickup for guitar -- is it possible?
sorveltaja:
I've been looking for a way to suppress the ~50Khz 'carrier' signal from the string's one, that modulates it. Resistor-capacitor low-pass filter could be one way, as there are calculators for that online.
Another way that comes to mind, could be to somehow extract only the difference between the higher carrier frequency, and again the lower one that modulates it.
But as they are already mixed together, it doesn't look that simple.
Simplified signal chain goes like this, where the red one represents the vibrating/modulating string:
The carrier square wave, when it goes through sender to receiver, has always more arbitrary form, which seems to be very sensitive to any and all kinds of adjustments.
PekkaNF:
Few random thought....not sure if this is usefull...
1: Signal to noise ratio looks bad. Is it good enough? Any way to improve with electro-optical or mechanical construction.
2: Modulation...I lost the plot somewhere why this modulation is used? I think that if modulation is need ideal would be sine wave, because it has only one frequency. Square wave needs to be really high compared to information signal and still simple low pass filter will strugle.
3: Modulation/demodulation reminds me of AM-stuff I used to read and even experiment a little when I were a kid....like one transistor mixers:
https://www.engineersgarage.com/circuit_design/circuit-design-how-to-demodulate-am-signal/
4: Still on this demodulation. If you want to get rid of the IR-led carrier signal you have several options. Have you tried to add pick up signal and inverted signal to subtract the unwanted carrier signal out (that effectively should be very close to IR modulation signal) Is't it.
I haven't been fiddling with audio devices much, this "information" may not be relevant.
BillTodd:
I'm not sure why you want a square wave o/p (the fundamental frequency only will be a sign wave) but..
one way is to synchronously sample the o/p with the illuminator's clock i.e. only test the level when the light is on (this should also improve s/n) but , you should really have a sample frequency (anti-alias) filter if you want to remove all sampling artefacts anyway.
WeldingRod:
I think the reason all your variation is on the trailing side is that you are triggering on the leading edge, which fixes that location on the screen. If your scope will do it, try triggering on the led drive signal.
Electrically subtracting the led drive signal should be easy, and would take out the bulk of your modulation.
Sent from my SAMSUNG-SM-G891A using Tapatalk
sorveltaja:
Pekka,
1: Yes, the noise level is quite high so far, when using the previously mentioned circuits. What comes to electro-optical components, I have a few different types of IR-receivers, which are not yet tested.
And the mechanical construction.. there is always room for improvement, one of the most important being the accurate height adjustment of the sender/receiver pairs. Likely by grub screws.
The current mechanical setup is too narrow for that, but it's just one of the first practising platforms.
2: I got the idea for the modulation from several sources on the net. It seems to be true, that by pulsing the leds, one can get a lot more 'power' out of them, when compared to using plain DC.
At first I thought: "more is better". But when used on a project like this, overly powerful output from the sender IR-led goes way above the receiver's range.
Also, when using something like 50-100KHz square wave for modulation, it attracts all kinds of noises, if not properly shielded. Those noises don't necessarily show on the scope, but they are audible.
As I'm on the testing phase, there are a lot of wires around to act as an antenna.
3: Thanks for the link. I'll check it out.
4: As I've looked info about demodulation, it seems to be a very complex subject. As I see it, to cancel out certain noises/frequencies, one should be able to provide exact, inverted copy of it/them.
Bill, I'm not sure either. When I got the resulting ~50Khz square wave output, the vibrating guitar string was able to modulate it, but with a lot of (audible)noise.
WeldingRod, I'll admit, there is options/settings on the scope, which I'm not familiar yet. I thought about subtracting/cancelling unwanted frequencies, but as mentioned above, it isn't necessarily an easy task to do.
Today, after reading these replies, I decided to test using plain DC for the sender IR-led. I tried it earlier, but it was just too easy to burn out the buggers. One way to see, if they are dead or alive, is to use camera.
Now, that I'm aware(sort of), that the ultimate brightness is to be avoided, perhaps more delicate approach is justified.
Result: the sending IR-led was fed with mild ~3,5mA DC, and the output was strong, not so much on the scope, but audible. And this time there was no noticeable noise at all.
But it wasn't enough for the schmitt trigger. Needs probably more amplification to produce string's frequencies in a square wave form.
I'm led to believe, that audio measurements on the scope could be very tricky. Don't know why, but that seems to be the case.
For audio probe, instead of laptop, I now use a guitar multieffect device, as I guess it's designed to handle better the incoming signals, without adding strain, or loading them down(impedance?).
The device is used without any effects, so only dry input signal appears at the output.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version