Fail More: The Story Of [CNLohr]’s Clear Keytar

[CNLohr] is kinda famous round these parts; due to some very impressive and successful hacks. However, for his 20k subscriber video, he had a bit to say about failure.

Of course glass circuit boards are cool. Linux Minecraft things are also cool. Hacks on the ESP8266 that are impressive enough people thought they were an April Fool’s joke are, admittedly, very cool. (Though, we have to confess, posting on April 1 may have added to the confusion.)  For a guy who puts out so many successes you’d think he’d talk about the next ones planned; hyping up his growing subscriber base in order to reel in those sweet sweet Internet dollars.

Instead he shows us a spectacular failure. We do mean spectacular. It’s got beautiful intricate copper on glass key pads. He came up with clever ways to do the lighting. The circuit is nicely soldered and the acrylic case looks like a glowing crystal. It just never went anywhere and never worked. He got lots of people involved and completely failed to deliver.

However, in the end it was the failure that taught him what he needed to know. He’s since perfected the techniques and skills he lacked when he started this project a time ago. We’ve all had experiences like this, and enjoyed hearing about his. What failure taught you the most?

18 thoughts on “Fail More: The Story Of [CNLohr]’s Clear Keytar

  1. But wait: what didn’t work? Okay fine, you can write a thing about failure. Big deal – we’ve heard it before. So tell us what failed. What you tried to do that didn’t work. What did you intend for it to do, and what did it do instead? I see some beautiful copper on glass with chips soldered to it. Great. So what’s the problem? What didn’t it do????

      1. Hmm, this is the first anyone’s asked, so I don’t see any reason not to go into an in-depth explanation here.

        * The silicone PCBs were too fragile (though now one could use 144 LED/m WS2812B strips).
        * The connectorization and bussing for the signals from the processors on the boards was too awkward. (Though that could be done with ‘8266’s)
        * There was no good mechanism for schlepping the data on/off (though, this could also be done with ‘8266s)
        * No way to make it battery operated, would still require a dongle hanging off. (Though there’s a lot more options now)
        * Couldn’t make raspi have low enough latency for the sound synthesis (though the ESP32 would be plenty fast)
        * Messed up the insides, trying to freehand the router (though now, I could use the mill)
        * Never had any good synthesis stuff other than really basic stuff Axlecrusher wrote.
        * Touch sensitivity required the subject to be grounded to the keytar for it to read all 0-64 levels accurately.
        * It was hard to iterate over new designs of the PCBs since they took so long to produce and the yeild rate was so low.
        * Never figured out a good way of securing the glass PCBs to the body of the keytar.
        * Never found a good LCD or display that could give feedback to the user regarding keytar settings, etc.

        If I had as much free time as I did then, with the technology available today, I MIGHT be able to do it, but… with the ESP32 here, I’d need to not have a day job to do it.

        1. dislike.
          This looks like a beautiful thing to waste… so you’ve got 20k viewers, now, but yet you complain you’re too busy to finish something so many people helped you build… nor even turn it into something differently-useful… As though, somehow, this ode-to-fail is supposed to be enough to justify your having given up. All those issues seem solvable, and even if they weren’t, there’s certainly *something* that could be done with that work besides just throwing up a poorly-focussed video of it. The leaning tower’s still standing, even though it was under construction for adding new support when I visited.

          And, having just watched your series on building glass-PCBs and noticing your having done one that didn’t work… with the portion about roughly “this is why we make multiples… It’s only $x in parts, don’t waste time troubleshooting, just move on” … Nah, you don’t do that to things of this level of beauty, this level of community-effort.
          This is the first vid I’ve seen of yours, and throughout I had some strange feeling that you were *trying* to be all zen-insightful, but weren’t quite there yet, and I think the second vid I saw of yours kinda confirms it… You’re being *wasteful* my friend, not only of cheap electronics, but also of others’ hard work, and encouraging others to do-so, as well. Similarly with your dismissiveness of the AVR-linux-javascript thing. Dude put tremendous effort into making that possible, making it something worthy of being on your videos, and yet your commentary is like “why’d I waste my time with this?”
          I think you’re better than this… you took the time to USBify the ESP, etc… But I’m having a hard time liking you after seeing this and those other vids. Glad you’re getting 20k views, but question whether you sold your soul to do-so.

          1. I agree with “duh”.

            On the scale of denial, anger, bargaining, depression, depression, acceptance, you’re on the “bargaining” stage. “It’s not so bad that this project doesn’t work” etc.
            What I would do is use the AD7147 that allows for 12 capacitive touch screen sensors per device, and then hook them up to a common I2C bus so you can control 4 of them on one bus. Have two busses. Then you have 96 keys or sensors. That can be done by an attiny or MSP430 or whatever 8/16 bit processor. Arduino even.
            Most hurdles you’ve already given an explanation for.
            If you want more synthesis options, then just convert the whole lot to MIDI and you can even use your computer to generate the sounds.
            You can secure the glass to the keytar by using a glass drill and drill a mounting hole.

            Anyway, it would be a shame to quit now, and you’re also letting those people who helped you down.

          2. I’ve never seen the AD7147. That is a seriously cool chip and beats the socks off of everything else in digikey’s capacitive touch sensor section. And, @duh… I will admit you make some compelling arguments — though — I would argue the tower of pisa should be left to fall, get cleaned up and turned into a mall or something.

            There’s still SO MUCH that would need to be engineered, probably 400+ hours. There’s so much else that could be done in that time.

            There’s a principle of opportunity cost. If I were to take these 400 hours, and attempt to actually complete the keytar, there are several other projects that I will have signed the death warrant for. The ESP32 would keep getting pushed off into the distance – and there would be many more opportunities to hang out with friends, development of friendships and much else I’d have to fore-go.

            I admit after yours and mime’s responses it is tempting – and I will certainly name drop you guys if I am ever inspired to complete this project… But I want to stress that it’s OK. Not every idea needs to be completed.

            Sometimes the old needs to pass away to make way for the new.

          3. @cnlohr
            I’ve been working on a project for 5 years now and it’s frustrating from time to time that I still haven’t finished it. However the amount of failures and consequently redesigns I have done have been very educational and I don’t think I would have learned as much if I hadn’t been so persistent.

            You are looking at this project as a time sink, but I think it’s still ripe with learning experiences.
            If it wasn’t you wouldn’t need 400 hours to finish it, amirite?

            Sounds like you still have enough chances to fail in this project and by your stated logic that should be actively seeked instead of avoided.

            I would love to see that thing finished man, it’s okay if you don’t but admirable if you do :)

          4. I’ve known cnlohr for many years. I don’t think he’s sold his soul to get to 20k subscribers just because he decided to axe this project. Also, he’s not a full-time YouTuber so the views only matter in that it’s cool his hobby got this much recognition.

          5. I think anybody who wants can start off where cnlohr left, if he really wants to. Sometimes you quit projects, not because the problems are unsolvable, but because you feel solving them all is way to much effort in comparison to the personal gain (fun, enjoyment, even proud if you want) from completing this project.
            Of course that is sad for people who contributed to a project in any way. But on the other hand: Nothing is blocked with patents here, everything seems to be open so continue if you really want.

          6. Apologies, I was a bit harsh…
            And maybe I should rewatch your vid…
            I think the key is that failure is an OK thing, and sharing that is commendable.
            My step-dad used to say something like… when you meet a problem you can’t solve, put the project away and come back to it later with a fresh mind. There’s no reason one can’t have 10 or even 100 ongoing projects in various states of unsolved… Subconsciously you’re likely still working on ’em… unless you choose to consider it “the end.” And, upon realizing failure toward certain end-goals, it’s OK to revise its end-goals.

        2. Okay, thanks, that’s helpful. From the long list of things that went wrong, it’s pretty clear what happened: too many new things and new technologies in a single project. Silicone PCBs? Touch sensitivity with 64 levels on a capacitive sense board? Advanced pioneering all over the place.

          Just look at all the things you learned about ALL of these processes!

  2. I think its a waste to give up at this stage… If things arnt working, fix them… Whats stopping you now? I mean damn, my speedo board is onto revision 4… At least stage of failure im learning how to get around it. Improve on it…

    I realise that a large slab of perspex isnt cheap. So a new handle design might be needed to reenforce it… Why not add a metal frame that emphasises the design?

    All the work thats gone into your project, the copper designs the experimentation, it looks like the electronics are on there and should? work… Why stop now?

Leave a Reply to Mangy_DogCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.