|
Post by videosplicer on Mar 9, 2024 22:48:50 GMT
Hello all, My first post and apologies that this post is not related to Ethan's book but rather one of his prior articles in "You can DIY!"
I have a few questions regarding Ethan’s circuit described in “Building a Guitar-Controlled Synthesizer.” Specifically, on the Input & Pick Detector section. I need both a guitar pick detector and a noise gate circuit for my experimental guitar gadget project. I’ve successfully built separate pick detectors (e.g., Craig Anderton’s AMS-100) and noise gates and while they have worked, they didn’t have the richer features that Ethan's design has. His design is optimal as it combines both pick detector and noise gate which saves space on the PCB too.
Has anyone built the circuit on a PCB and got it working well? The reason for the question is that I am designing a ‘test’ PCB for it and am wondering how (or if) the PCB layout and capacitance affects the triggering accuracy. To my understanding, the entire design was done in LTSpice and I do not know if that takes into consideration any of the physical aspects of a PCB. A resistor in the schematic (R35 5.1K) is notated, “Pick Sensitivity.” I’m wondering if using a 10K user adjustable pot in its place would add any advantage. If not, I won’t put a pot footprint on the PCB. I understand that the input op-amp (U1) is a JFET type, are the other op-amp the same type? I plan to use the TL084 types. Speaking of op-amps, can someone clarify if the two comparators (U7, U8) are simply op-amps (e.g. TL084) used in a comparator configuration or are they really comparator chips (e.g, LM339)? The article’s text states they are op-amp comparators but I do not see any feedback resistors being used which leaves me a little confused. Knowing this helps me plan the PCB by telling me if I can use all of the op-amps in a chip, or if I need a different chip all together. The circuit is designed to be driven with +/- 15V; however my main project is driven with +/- 12V. I don’t think that makes a difference as long as the input (V3) is still kept at 2Vpp – is that safe to assume? Thank you for your time and 'input', Paul
|
|
|
Post by Ethan Winer on Mar 11, 2024 21:18:09 GMT
Sorry for the delay Paul, but I'm here now. I never built any of the guitar synthesizer circuits for real. They all exist only as LTspice simulations. Do you have LTspice? The program accepts Wave files as input, so you can record your own guitar (direct from the pickup) and use that to tweak the pick detector and noise gate to be sure they work the way you need. I don't think a few pF of capacitance on your PCB will be a problem because these circuits are not super high impedance. Even R19 in the noise gate at 2M won't be an issue because it's connected almost directly to 15 nF. As for R35, you could try values between, say, 2k and 10k, or use a 10k trimpot. That's why I suggested using LTspice with a Wave file of your actual guitar. But adjust the Wave file's dB level so the voltage at the plus input to U1 is about the same as comes out of your guitar. All of the op-amps in these circuits use the "Universal Op-Amp 2" model that comes with LTspice. This is not an FET op-amp, though it works fine in my simulations. I just happen to be working on a new book about audio circuits with LTspice, so I have this list of parameters handy: U7 and U8 are the same op-amp. Comparators don't generally need feedback, and when they do it's positive for hysteresis as with U7, mentioned in the article text. So sure, just use the same op-amp for everything. The TL084 seems excellent. And finally yes, 12 volts should be fine. The Wave file input will always be 2V peak to peak. But you might need to change the resistors at U8.
|
|
|
Post by Ethan Winer on Mar 11, 2024 21:22:28 GMT
You didn't mention if you're using surface mounted parts or not, but if you are, and you're concerned about wasting valuable PCB space, this trimpot is only 3mm in diameter! www.bourns.com/pdfs/TC33.pdfI bought some for a project and they work well and are easy to adjust - using a small screwdriver of course!
|
|
|
Post by videosplicer on Mar 12, 2024 4:29:03 GMT
Hello Ethan,
Thanks for taking the time out of your schedule and clearing up some of my confusion, I really do appreciate it.
So, it looks like I will be a test case for a real build of the circuit – I love this stuff so I don’t mind trying it out.
The “Test” board I am making uses through hole components since I like clipping scope leads into circuits and viewing waveforms. Actually, I have already submitted the Gerber files for a quote to make the boards and am waiting for that. It’s rather sloppy, but it is just to test the operation of the noise gate and pick detector.
If the circuit works as expected, I will be using surface mount components in the ‘real’ project. I will have to take some measures to create some of the resistor values as they don’t seem to be ‘standard.’ For example, R18 is 147K so I’ll use 100K and 47K resistors in series. I’ve also added some level shifting transistors as I need 0 /5V TTL logic levels out to drive my main project plus a ‘Gate’ LED. Another thing about the ‘main’ project would be the use of digital pots instead of the mechanical ones. This allows me to store and recall settings for the input and pick sensitivity settings, allowing various ‘presets’ to be set and recalled via MIDI and front panel control …plus it reduces the chances of picking up noise since there’s no need to mount a mechanical pot away from the circuit. I have to admit that I already have an input preamp that will supply this circuit with 2Vpp, meaning, I may not need a pot here at all and can use a fixed resistor.
Here’s a digital pot chip I am considering using:
OK, Ethan, thank you again. I will keep you updated on the results of the test board. It’ll be a bit since the board will take about 2 weeks to create from the PCB fabrication house.
Best Regards, Paul
|
|
|
Post by Ethan Winer on Mar 12, 2024 17:35:24 GMT
Sounds like a plan. I'll point out that 147k is in fact a standard 1 percent value: Also, test points area available in a large variety of shapes and sizes. I used these tiny ones for my last surface mounted project: www.digikey.com/en/products/detail/keystone-electronics/5019/3907343And yes, please do keep us updated with your progress. Believe it or not, having you be the first to build this for real is exciting for me too. I bet audioXpress magazine would welcome an article by you as a follow-up to my original article.
|
|
|
Post by videosplicer on Mar 13, 2024 16:16:49 GMT
Hi Ethan,
Thank you!
Wow, after 40+ years of electronics hacking, I really did not know 147K is a standard 1 percent value! Actually, most of my projects didn’t need that kind of accuracy, but I can see this one does in certain places.
Thanks for the Test Point links also. Ha, at my age, my eyes aren't the best anymore - so sadly, I can't hand solder very small components as I used to …but I'll keep trying!
I have to tell you, the audioXpress article was a great article! I’ve spent a very long time looking for such circuits to no avail. There are also quite a few guitar DIY forums out there where people are looking for such circuits and have basically been given bad news. As I mentioned before, Craig Anderton’s AMS-100 pick detector was basically the only one available until yours, and that was published in 1979! I suppose software and plugins have taken over that aspect, but I’m old school and still prefer hardware.
So, I am looking forward to building and playing with this ‘test’ board. If it works, I can provide you with the modified schematic (containing the logic level shifters) and JPGs of the board layers – keep in mind that I’m a hobbyist, not a pro like you!
Best Regards, Paul
|
|
|
Post by Ethan Winer on Mar 13, 2024 17:43:30 GMT
|
|
|
Post by videosplicer on Mar 16, 2024 18:04:22 GMT
Hi Ethan, Those articles were very informative, with some great advice! Some of the topics brought back memories of being in school learning PCB layout ...using tape to draw the trace lines as you mentioned. Speaking of which, that was a great layout pictured of the project that fit into the mic housing; how does one keep all the traces in an orderly fashion!?! CAD has rally made things so much easier compared to the tape days, but still, layout and running traces are art forms in themselves.
I've considered using a stencil and squeegees to spread solder paste for my larger SMD projects, but if you took a look at my car's windows, you'd know I'm not good at using squeegees! My age difficulties are just limited to my eyes right now, fortunately. You're right, nothing that a good magnifying glass can't correct (or an updated prescription for the old spectacles.)
I'm also going to try some of the PCB vendor recommendations listed in your articles as the one I submitted to has not responded with a quote!
Best Regards, Paul
|
|
|
Post by Ethan Winer on Mar 16, 2024 18:51:43 GMT
Yes, PC board layout is very much an art. It's also a lot of fun! As I mentioned in the article, it's not unlike playing chess where you have to plan several movers ahead.
|
|
|
Post by videosplicer on Mar 26, 2024 22:36:02 GMT
Hello Ethan,
I have built the test board and it is kind of working. It seems to me that the Noise Gate section is not working as expected but the Pick Detector section is. I'll troubleshooting over the few days. Technically, the circuit is a success for my use, as I only need the pick detection but the gate would be nice too!
Attached is a picture of the board (picker-sm), a scope capture of the reset signal (pic_467_3), and a scope capture of the gate signal (pic_467_4). Also the schematic (picker.pdf)
The gate signal output is not what I was expecting and seems far too short to me. A capture of the scope at C5 of your schematic shows the peak follower cap (pic_468_2, blue trace) not being charged? I'll resolder it and or replace it ..who knows maybe the ground side of the cap is not really grounded (OK, according to my DVM, it is grounded.)
As you pointed out, all of the resistor values are standard, but Digikey did not have the 102.3K and 204.6K in stock (ETA was 11/24, too long for me to wait.) So I added resistors in series to make the required values - give or take 100 ohms.
There are some physical issues with the board - I used an old footprint for the transistors that I had made a long time ago, where the pinouts are backwards, so it may seem the transistors are not oriented correctly. I also added, in my haste, an input jack so I can connect a guitar to it ...completely forgetting that the input level required is 2V p-p.
Also, keep in mind that I added some other components that give me TTL outputs. I am using +/- 15V and 5V in this test board.
I also added a pick sensitivity trimmer pot instead of R35. It seemed to make no difference until it reached one extreme. So in the final design, I will use your recommended 5.1K resistor.
Anyway, I used a signal generator as an input. It outputted a 440Hz sine wave, but in bursts about 300ms apart. I haven't tried a guitar as a source yet.
Best Regards, Paul
Attachments:
picker.pdf (110.87 KB)
|
|
|
Post by Ethan Winer on Mar 27, 2024 17:26:18 GMT
Thanks for the update. My best suggestion is to get LTspice and compare its "oscilloscope view" with what you see on your own board with a real 'scope. Maybe first record your guitar so you have an identical signal to feed your board, and as a Wave file version to feed the LTspice simulation.
|
|
|
Post by videosplicer on Apr 14, 2024 20:43:01 GMT
Hello Ethan,
Just wanted to give you an update on this.
I found the issue with the noise gate section of my version. Seems like the reference of comparator 2 (U8, + input) in my board was measuring at 9.2V instead of 9V. Since I'm too lazy to figure out resistor voltage drop values, I just simply swapped out R23 with a 9V zener diode and that did the trick. I think it's safe thing to use in case the supply voltage varies slightly around 15V, such is the case with my crappy bench power supply.
In the main project, I will be using with + / - 12V supplies; LTSpice shows the zener usages works there too - although I have to change the other comparator's reference to compensate (R17 was 7.15K now is 9.1K) for 700mV.)
Thanks for the great article and the help with this. Also, for the introduction to LTSpice - it's not too bad!
Best Regards, Paul
|
|
|
Post by Ethan Winer on Apr 14, 2024 20:45:53 GMT
Thanks for the update, and I agree that a Zener diode makes sense.
|
|