Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards "DIY" DP possible?

So in general it it seems that the holy grail is grand action and expressive sound reproduction. I'm totally okay with being tethered to my computer for playing. And while I like (but have not yet bought) PianoTeq, the on-board samples in my PX130 are sufficient for me. I haven't even thought of trying Galaxy.

But I started thinking about about it the other day. If true grand keybeds are somewhat easily removable and you can retrofit a piano (PianoDisc, PNOScan, and Doepfer come to mind) for MIDI output, why not combine the two?

I'll admit to not having looked around for salvage pianos, but I can't imagine that it's terribly difficult to find a grand piano keybed for sale. If you were to retrofit one with midi sensors and paired it to your VST of choice, wouldn't you basically have a N3 without the fancy case, speakers, and haptic feedback?

I mean, the only thing I could see as being a problem is if any of the retrofitting kits are prohibitively expensive. I can't find any reference to the price of a PNOScan or PianoDisc's midi strip.

good luck, but you know life is too short?...

I have actually looked into this extensively and actually arranged with a local piano tech to get an action. Here's a bunch of relevant info, in no particular order:

1. It is indeed difficult to find an action. Getting a new one will run you thousands of dollars, believe it or not. The best solution is to find a piano retailer or repair person and locate a dead grand. That is, one with a working action but otherwise unusable. The guy I talked with was willing to sell me such an action for $750 (regulation included) but it was not from a concert grand.

2. PNOscan will run you around $1000. It can only be had through piano retailers. Mine offered it to me for $1100. I bet I could have talked it down. Though the factory where it is made is driving distance from me, you can't find a way to get it in the US without going through a retailer. Apparently they don't like to sell their product for some reason. Terrible idea in my opinion.

3. Remember that you need to figure out a way to house it and something for the hammers to hit that is not too loud.

4. Here's a video of someone who did this on youtube. The guy is completely nonresponsive, though, so good luck getting more info out of him.

http://www.youtube.com/watch?v=korhxsv7xhw

5. PNOscan must be calibrated and it's apparently a painful process. You have to do key by key and it apparently isn't easy to get even. And some people paint the bottom of their keys to get better response.

6. PNOscan uses key movement only. As discussed in various places here, the right place to measure velocity is on the hammers (there are cases where it makes a difference). Thus I believe PNOscan does not provide a better sensor system than the hammer-based sensors in digitals.

7. PNOscan sends repeated notes (the ones without lifing the keys completely) in a nonstandard format. It sends a sequence of note-on, note-off, note-on, note-off, rather than note-on, note-on, note-one, like VST's expect. I think some VST's, especially PianoTeq, which is somehow connected to PNOscan, have overcome this hurdle somehow.

8. PNOscan has a single set of sensors. It's not 100% clear how they work, but I have never been able to determine whether they emulate three-sensor behavior (i.e., note calculation at the end of the stroke, damper disengagement half way through). No one seems to know or is willing to say.

9. You have to rig something for a pedal.

I was very serious about building one of these for myself. I finally gave up not because it would be too difficult or too expensive, but because after researching it, I decided that the finished product would be inferior to a regular digital piano. It would feel amazing, but I fear it would not be as responsive or work as well.

The AvantGrand N1 is more expensive, but it looks a lot nicer and probably works better, and it comes with its own speakers and sounds, as well as support.

On the plus side, I probably could have set it up such that there was an actual set of dampers that was lifted. That would cause the difference between up and down weight that is present in acoustics but not in any digital (including the AvantGrands).

Supposedly Petrof will build you a piano like the one we are thinking of. But they never answered my attempts to contact them. Jerks.

A thread about petrof's midi controller

Yep, this is the deal breaker for me as well...I guess the only was to find out is to travel to a dealer that sells pianos with PNOscan already installed, bring in your laptop / midi interface and give it a shot (assuming that would be ok with them!). My biggest fear about the strip is having to release the key fully to "resting position" as calibrated before being able to strike another note, but perhaps lowering the "resting position" on the sensors would eliminate this problem? Man, so many mysteries about the strip...whoever decides to pull the trigger on one and go for the pianoworldforums dream setup (for those who can't afford an AvantGrand) will be a true hero to post meticulous details about this whole ordeal!

Other things I fear in the strip:

- Not having triple sensor like response

- Having notes sound without full key depression

- Having notes sound FF when pressing a key with a fast initial movement but a soft landing (difficult to explain what I mean)

- Never being able to find a good calibration

- Being stuck with a $1000 piece of hardware useless for anything else if I decide I don't like it / can't get it to do what I want it to do

- Wasting an incredible amount of my time!

Hi Pandamonium and all,
Is it possible to go the route that Infinite Response has, but cheaper? To be more precise, instead of retrofitting a grand action with hundreds of moving wooden parts that are susceptible to road wear-and-tear and humidity especially, would it be more satisfying, reliable, and less expensive to emulate the Vax77 midi controller somehow?

I have always dreamed of owning a light, portable, consistent, and most importantly - *accurate* action that I could carry under my arm and hook up to a laptop for gigs and home use. Could one start out with just the keys of a piano, or even the plastic keys of an old Yamaha keyboard or something, and somehow put sensors on those? The Vax77 has spring-loaded mechanisms I understand, and Stevie Wonder has called it the best action he has ever played in his life. The only caveat being that it doesn't have the escapement feel of a grand piano. But who needs that when you have much more midi resolution and accuracy, even more than a well-maintained concert grand action?

Then you would say, why not just buy the Vax77 - well there is only one problem with that - the price tag! $$$$ Is it possible to make something like the Vax77, but for around a 1000 dollars instead of 3-freakin'-thousand dollars?

Maybe there is a reason for that insane price tag for a midi controller - be it parts, labor, patented sensor/spring mechanisms - but I find it hard to justify putting a $3000 price tag on it.

Can it be done more cheaply, or is this a pipe dream?

Hmmm okay so let's say that finding action is the limiting factor. What about making your own? There are plenty of CNC cutting outfits that let you send them designs. I'm googling images of some actions, but the best I've been able to find is http://www.rennerusa.com/VirtualActionModel.asp - it looks like a ton of parts for each key, but looking at it from a perspective of "what would I need to do to make this CNC-friendly?" and "where could I build a laser sensor system to handle input to MIDI duty?" and "how could this accomodate a case design that we find palatable?"

To that end, I've got my hands tied right now because I don't have access to any acoustics at the moment.

A laser system looks like it's possible but would definitely need some more googling on my part.

But as for a design that would be acceptable, I have a thought. The whole mechanism could be turned into a rather handsome parsons style desk. You could lift up the edge (much like a Roland DP-990) in order to expose the keys and a built-in music stand. When not in use, it would look like a long console table. I'm not sure how the pedals could be worked into it, but I'm thinking something that telescopes would work. That way you could retract it close enough to the bottom of the piano that an unaware passerby would probably not be able to see them.

Make your own grand action? Now we really are going of into a land where unicorns roam freely...

Would have to agree. A grand action is such a complex beast it is very likely impossible to make for mere mortals.

As others have already pointed out, the trouble is that to come up with anything decent, you're going to have spend a great deal of your own time and probably more money than an Avant Grand, which will do that particular job far better, look nicer and have some resale value (I guarantee you, your home project will have a resale value of virtually nil).

Life is one long compromise. I'd love a Steinway D but I'm never going to be able to afford one. The important thing is to find something you can live with, at a price you can afford, and then get on with making some great music on it. Don't confuse the tools with the product.

+1

exactly where we are in this thread...
But i agree, musicians should dream.

I don't have any hands-on experience with PNOScan, but have looked into it a tiny bit. It obviously uses the changing optical reflectivity of the key underside due to angular movement to sense key position. It seems to be a clever way to retrofit, but I'm not sure it's the best for new construction. Positional sensing is ideal because it allows you to go way beyond triple sensor, but of course the devil is in the details of exactly how it is implemented.

The VAX uses magnets and linear Hall-effect sensors to sense key position, which is likely more reliable and less interference prone. Someone should take a VAX apart and measure critical distances, weights, and spring constants. My guess is they have the weight located near the end of the played side of the key, the spring could be almost anywhere and still feel the same. I applaud the direction they have taken with this product.

Oh come on, it's just a 3D printer away

(Kidding... but, maybe one day...)

Hi Dewster,

Thanks for clarifying the sensor types in the Vax77 by Infinite Response. I have enjoyed and appreciated your DPBSD thread and your time and effort into exposing DP's and giving us here at PW an honest and unbiased look at them.

Do you feel that the Vax77 price point is right? I realize they have to make a buck like everyone else, but the price seems high (like doubly high) for a midi controller without sounds.

In keeping with the spirit of the OP, would it be possible to go this route and make something with magnets and linear Hall-effect sensors, but for less money? Are these parts inherently expensive? Is the labor too hard? There must be a reason for Infinite Response's $3,000 price tag other than profit margin (at least I would hope so).

When we talk of starting from scratch and buying and assembling all the parts of a keyboard like this, are we indeed wandering into la-la land? Or do you think it can be done more cheaply, if one is willing to put in the hours?

Thanks for your input.
-Erich

The manufacturer puts in thousands of hours doing these things. Do you have the time? Even if you work for "free"? Do you think you could ferret out all of the design, build, and reliability problems?

Precisely MacMacMac. Is this why you think the price tag is $3,000? If it ain't the parts, it's the labor. I was looking at this problem like one would look at purchasing a desktop computer. Sure, you can order a pre-made, pre-configured Dell or Powerbook - at a certain cost. Or you can order the individual parts online for less money and put the computer together yourself.

Certainly people DO have the time and willingness to make their own computer. Is it the same for DP's, or is it too cost-prohibitive (price of Hall-effect sensors / keys), or lack of knowledge and too time consuming on piecing the different elements into a portable keyboard package?

There are plans on the internet to build electronics such as computers and radios, but I'm not sure there's plans on how to build your own DP.

Edit: Found some "plans" - I stand corrected
https://docs.google.com/viewer?url=http%3A%2F%2Fece.uprm.edu%2F~pol%2Fpdf%2Ficee04CruzPol.pdf

Building a computer is trivial. Pile of parts, plug them in, load the software. You're really not inventing anything. Nothing at all. The parts are MEANT to work together.

Building a piano is far from trivial. I presume we're not talking about buying a pile of parts and then assembling them, right? (That would be pointless, yes?)

So you have to invent lots of things. And test them. And repeat the cycle. Thousands of hours ... with no assurance that your result will be any better than (or even as good as) the pianos sold at your local dealer.

I would think the reason for the high price is the relatively low demand. They have to write off their development and manufacturing costs with something less than mass production.

Jack

I've never seen one so I can't vouch for the quality. But for something that they maybe sell 100 to 1000 of a year (total guess) they have to charge a lot to recoup all the up-front costs. With that $3k you're paying for several people sitting around and thinking.

I personally don't like the way it folds, and would prefer a unit that didn't do that.


I see linear Hall Effect sensors at DigiKey for around $1 each in quantity. You could multiplex the outputs into a high speed 10 or 12 bit A/D converter. The rest could be in a small FPGA (state machines) or processor (program). MIDI is fairly stone age at this point.


Certainly as a one-off for yourself it is doable. People obviously build their own harpsichords and clavichords, I don't think a MIDI keyboard controller could be any more difficult than that - particularly if you don't want to specifically emulate the "hammer feel" of a piano - though the skill set would obviously be somewhat different. For the base of the keys you could use aluminum channel or tubing, or basswood / poplar, or maybe even flakeboard on edge. If you go with aluminum U channel you could perhaps orient the channel opening down and use it to guide the key. All the black keys could be identical, but there may be ~6 different types of white keys. I'd go with a really long distance from the played end to the fulcrum (250mm white / 200mm black) and use a return spring rather than a counterweight.