According to Pete - August Edition #1


After a one day delay, we're here the with August's first edition of "According to Pete." In today's episode, SparkFun Director of Engineering Pete Dokter attempts to make guitar pickups using the AMP401 MEMS Microphone. Check it out!


Vimeo version can be found here.

Please leave any comments or questions in the comments section below! We'll be back in two weeks with more "According to Pete." Cheers!


Comments 37 comments

  • Pete - have you looked at Ti’s http://www.ti.com/lit/an/slyt477/slyt477.pdf for your Piezo actuators? Take a look under “High-definition haptics: Feel the difference!” Figure 2. “Form factors for piezo actuators”- see “(a) Vibration of screen only”- couldn’t you use that for input for your sound? YES I know that’s output, but could you not do the inverse of that?

  • Hello Pete, nice project by the way,i noticed that you not use a good acoustic conductor betwenn the MEMS sensor and the body of gitarra, and because of this, you pick a lot of acoustic garbage, one possible option, is the similarity of echographs use a conductive gel to reduce the acoustic impedance between the two interfaces (gitarra / MEMS sensor), and eliminates a lot of garbage.

  • Could the reg be starving everything for power. I power my 3.3v parallax propellers with LIPO batteries with just a 3.3v zenor diode on the power line.

  • Pete-O: I am currently trying to interface with a similar transducer, home made, putting out around max 35mV, would you be thinking of using the AD620 like I am, If so just wondering if you would like to share your thoughts on PCB layout, right now I need 2 - AD620s because I have 2 sensors! Any thoughts or maybe even email me if you would like to pass around Ideas, I am far from genius but I have read through the AD620 and AD8222 datasheets plenty of times.

  • And not even a small sample of the sound you got from your setup? =(

    • Yeah, if you could listen with headphones, why not plug in a direct box, record and mix it into the video? You could make a direct box from stuff in the drawer.

      • I thought I’d take some flack for that. But really, my impression was better than what was actually coming out of the amp. At this point, I’ve scavenged the old circuit for the new, so I couldn’t make it happen, anyway. But I promise real live sound for the next one.

  • Why does she like it when you say “bust my hump”?

  • Hey Pete here’s a suggestion,

    I’ve figured out a sort of surface pickup using an inductor. It works very much like those guitar pickups, except instead of having the coil and magnet being attached to each other (then the vibrating metal above it disrupts the EM field, creating an AC wave), the magnet is attached to the surface you want to record. Then you just place the coil near the magnet. Downside is the coil has to be mechanically/acoustically insulated from the magnet. I usually place my inductor on a tripod with the magnet taped to the instrument, works pretty well.

    Here’s my blog post about it: http://www.orangebugstudios.com/blog/2012/2/3/inductors-pt1.html

  • The mixer circuit you describe (@3:30) looks like an adder (summing) circuit to me. If the input to the opamp is a virtual ground, then a 1V input signal will cause 100uA through each 10K resistor. The output will need to go to -3V to pull 300uA through the feedback resistor to compensate. Is there something that I’m missing?

    As always, awesome video. Your projects inspire me to start hacking stuff together every month. I can’t wait to see some LEDs on that guitar!

    • You know, that’s a good question. I don’t have specialty op amp circuits committed to memory, so I looked it up and basically verified through operation. But yeah, when you look up “op amp adder”, it’s virtually the same circuit. It could be that the addition of a pot at the input is a means by which to get the level out that you’re looking for (and less of an adder). Well, maybe that’s obvious. Somebody will chime in here and tell us where we’re screwed up…

  • Nice vid - Keep in mind that the LM358 output only swings from 0 to VCC-1.5V… If your divider gives about 1.67V bias the output will clip at 1.83V, limiting the AC peak amplitude to 0.16V.. If you change the bias to 0.9V the AC peak will be increased to, well, about 0.9V. If you have a rail-to-rail amp that would be even better.

    • The ADMP401 boards were clipping long before that. So I may have ultimately run into that, but not this iteration…

  • Pete- quick note: Glad you support Costco… But your Video today kept going green on me at the end of your “According to Pete”.

  • Great lesson in audio design. I have to ask, though, why you are trying to design a guitar pickup. The standard design is great: It has a pair of coils wired in opposite directions to cancel any hum, and it uses a magnet as the core, so that it does not pick up sound at all. It picks up variations in the magnetic field when the steel string moves. It is clever and it works.

    • Why? To bring the body shape and material more into the equation. A guitar with humbuckers is basically always going to sound like a guitar with humbuckers, regardless of shape an material. By picking up actual vibrations in the body, and in different places in the body, you should be able to muck with tone in much more interesting ways.

  • I for one totally got the gretsch comment. Reminds me of when I tried adding a strat pickup to my wife’s cello and fed it through my marshall. I thought it was awesome but something about the ensuing death glare told me she was less than enthused.

  • I’m not sure, but in the video you saw, they may have been using a MEMS accelerometer. I saw a video from Analog Devices where they were using an accelerometer attached the the body of an acoustic guitar.

  • piezo (pee-yeh-zo) sensors definitely have enough bandwidth - that’s what they use on acoustic guitar pickups under their bridge. I also have one on my (passive!) electric banjo.

    Pete, yer funny.

    RE: digital communication, maybe a MIDI project?

  • Was that my frikkin' head?

    You said “flexy” when you were talking about piezo sensors, that doesn’t seem right. Are you sure we’re talking about the same (SEN-10293) piezo element?

    • That was one of the ones I looked at, admittedly not “flexy”. If you’ve got one lying around, bring it over and I’ll try it out.

  • I think 300 Hz is not low enough for the low E string. Looks like you need to go down to around 82 Hz to get that fundamental. http://en.wikipedia.org/wiki/Guitar_tunings

    • Well, we shall see. I’ve got my gear in the office with me today to do some testing with the surface transducer. I tried it yesterday with Jeff Branson’s stand-up bass, and I was encouraged with the results.

      • Being that A-440 is yer A-string, yeah… early 300’s for low E. On guitar.

  • Why not pull in the data optically? Beam an LED onto each string and reflect onto a photo transistor the same way that optical mice work. This should eliminate all mechanical feedback…

    • I believe that the SynthAxe, played and toured extensively by Alan Holdsworth, used an optical system to trigger the untuned strings of the SynthAxe. This would then trigger a synthesizer designed specifically for this purpose. The advantages of optical triggering were that it was nearly instantaneous, as opposed to analog magnetic triggering which suffers from significant lag to this day in other guitar-triggered synths, mostly manufactured by Roland in Japan.

    • Could work, but how would you determine what is the string and what is ambient? I know a few guitar players, and they like to move around a lot, so you’d still end up getting some mechanical feedback anyways (because the person is moving around AND the ambient light).

      • I can think of three ways to possibly mitigate ambient light.

        1. I’m guessing a light cover would help block noisy ambient light. I think that’s how the commercial ones work. BTW, the beauty of optical is you can get much lower frequency coupling then electromagnetic – essentially DC, which is why I think the optical pickups were used on Bass instruments.

        2. You might measure the ambient light and subtract it from the string signals using DSP. That would be fun.

        3. You could call the ambient light a feature. Thinking about it, combine it as part of the show. Collaborative concerts. It could pick up movement from light reflecting off the audience.

  • Dear GOD…a pointy guitar…

    • It’s a Jackson Randy Rhoads model. Randy Rhoads, R.I.P. It’s made for heavy rock, as Randy Rhoads played with Ozzie Osbourne until his untimely death in a plane crash.

    • umm.. it’s called a “flying V”!

      • No. THAT is not a Flying V. THAT is an abomination of the Flying V.

        • Yeah, I’d tend to agree. I’ve got a Gibson faded V that’s WAY more satisfying to play.

          • If you’re gonna keep that transformer on there, you should wrap it in Spandex.

            • Then it would look too “finished”. I sorta dig the whole ghetto aspect to the project.

              • No doubt…however…if it’s gonna live on THAT guitar…there oughta be Spandex. ‘cause, really…there’s not much else that can be successfully paired with one of these beasts, than Spandex. And Aqua Net.


This Week

This Month

Heartbleed

Happy Arduino Day!

SXSW Create 2014