Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards DIY Acoustical Digital Piano - algorithm

A few of us are working on these inspired by CyberGene's DIY hybrid midi controller. Because what he built is so cool!

And I've been documenting my progress (or.... lack of progress!) on youtube. Its fun to get something working and then make a short video of it, and try to explain my thinking (haha, or... well, lack of thinking)

Just now I posted a video on the algorithm and thought maybe others would enjoy since my test system is able to capture and plot piano hammer dynamics. I find it interesting to see what hammers are doing.

That's very cool and very interesting. It shows how the interaction between hammer and string is very hard to model and to convert into a single velocity number. Kudos for this work!!

Osho

Thanks!

Next up, I am building the printed circuit boards.

What are you doing to get your "hammer position" measurements in those graphs? I'm pretty sure there's some important nonlinearity between the phototransistor current and the distance. And depending on how you convert that current to a voltage, there might be more issues. All of that could easily make the velocity look more varying than it is.

(I keep meaning to set up a CNY70 and a reflective surface on some sort of linear stage and take a series of current readings at varying distance, to duplicate figure 9 of the data sheet.)

Those adapter boards like you show in video 2 make me nervous for digital parts. Getting decoupling caps in there nice and close ends up being more of a pain than soldering down the TQFP part.

Good point, I should have stated measurement of hammer plus measurement system!

I've debated including a calibration step. And, regarding the boards, yes, I have a custom PCB in progress.


We need a forum for - piano DIY / Maker / building your own electronics, software, hardware, instruments, etc

Hi,

Great video, but your Damper jack should be activated at Half of the Pressed key (1/2):

the Damper is up to early :

https://www.youtube.com/watch?v=sLuIYDJClsk

And the Hammer is to low :

https://www.youtube.com/watch?v=8hOnMcsUh_c&list=PLyhFrvLWlRkAhfIf0PaJ1l5fH7Qky-ZS9&index=22




your Hammer sample is not well regulated like a Grand Piano

here better:

https://www.youtube.com/watch?v=hhkntn8PBYM



With a good regulated sample : simulation and measurements will be more like real Grand Piano behaviour

And Damper action distance :

https://youtu.be/bDgpLYn9774?t=359

On the Cybrid I measure the velocity in the last 1.5mm of the hammer travel because the hammer flies freely there, the escapement is set at 1.5mm, so this guarantees that there won't be any unpredictable movement caused by your finger pushing the key up to the string and staying there which you show in your video. So, problem solved Make sure you regulate your action exactly according to standards: escapement 1-3mm, blow distance of 47mm, hammer drop same as escapement (measured further below the escapement point), check distance 15-17mm. And damper usually lifts off the string exactly at halfway through the key travel, hence also the hammer travel, so measuring if the hammer is above/below 24mm can provide indications of damper position too which is how Cybrid works, as well as NV10.

I vote for this!. A Telecommunications Engineer here with a 22 years on electronics R&D career (and programming since 1982). On 2012 I left R&D but no electronics, and in 2018 I began playing synths and piano, and also started fixing synths (for myself initially). So, yes, we need a DIY sub-forum!

In case people here haven't seen Jay is way, way ahead with a project like this one. See https://github.com/jkominek/piano-conversion

Haha..wait.. what? Is it a contest? This was supposed to be a fun little side hobby, now I am starting to feel stressed out.

Juuust joking The projects I have seen all are taking different approaches, including comparing my approach to the project you mention. Jay's project is cool! I am building just something for me.

And I think there are even more possible ways to approach this kind of a project too. And, other DIY style piano projects that are possible.

I feel really lucky that I created my project without having competition 😀 If there was even the slightest one, it would mean I abandon everything immediately. Especially in a field where I’m a complete noob and especially if it involves multiple fields where I’m a complete noob, such as in a DIY hybrid piano: electronics, machinery, piano regulation, woodworking, design, low level programming, etc. Looking at how pathetic Cybrid is compared to Jay’s project now, I’m certain I wouldn’t even have thought of starting such a project 🤣 And I worked for almost two years on it just because nobody else was interested, so I basically just did nothing for two years and most of my work was about resurrecting the sh1tty old action whereas everything else was probably worth it just two weeks of work... So lucky to be the first one 🤣 (and no, I’m not the first one either, but the other projects I’ve seen weren’t published or demoed extensively).

Don't sell yourself short like that, Cybergene.

You're kind of like the George Dantzig of homebrew hybrids What you accomplished and built is beyond impressive!

I finally got a CNY70 sensor hooked up to my action, and recorded some video of the oscilloscope displaying the output from the sensor.



Seems like a very nice signal to me.

(The fuzz is at ~120Hz, so it's probably related to power lines some how. Not too worried, I've got a bunch of long cables involved in this little experiment.)

This weekend I got my first printed circuit board working (the "easy" one). Now to the second.

Once both boards are running, I think that eliminates most technical risk and then its just assembly. Although, I have many more ideas for the real-time software side.

Been trying to document every step and detail on youtube. I got behind and recent caught them all up.
Apologies in advance to anyone who attempts to suffer through any, as they are a bit nerdy

So here is today's accomplishment -



Also, Jay, thanks for sharing the interesting video!

Glad you liked it. There are some more pictures and video over on my github project.

Soldering tip(s): Hold the solder itself, not the solder container. You'll have much more control. And if the solder is getting too hot to hold directly, something else is going wrong. Get a flux pen and go absolutely nuts with it. If you're worried about flux smoke, get a solder station with a pump that puts an extraction nozzle right next to the soldering iron tip. It's common on inexpensive rework stations. (Switching to SMT will also use a lot less solder.)

tl;dr goals?

Thanks for the feedback.

And I got the end-to-end data path running!


This time I put a block diagram as video preview. So, no need to watch the video to see what is now working.

backcheck on your action model is adjusted completely wrong