4: Key Levers and Parallelograms

[@mikhailkaryagin2509]

Love the video! Very informative

[@DIYJankoPiano]

Thanks! Hopefully I dont come off too dry and technical...

[@Abakhan]

It's my favorite suspense series!

[@DIYJankoPiano]

I'll try to incorporate more cliff hangers then :)

[@JoakimMonsen]

Love this!

[@DIYJankoPiano]

thanks!

[@rico-raven]

I really admire your work, already the parts of the prototype you've shown so far look like a masterpiece! I'm looking forward to your videos demonstrating it and hope that one day I'll also get the chance to try it out 🙂

[@DIYJankoPiano]

Thanks! Ive been following your channel as well - think you and Glitchnator are the two channels posting Janko piano content at the moment, and I keep checking for new videos almost daily!

[@s90210h]

this is going to start a janko revival

[@DIYJankoPiano]

Not sure about a revival, but I certainly hope the Janko layout becomes more popular (or at least more known). There have been several new youtube channels that have been posting instructional and musical videos, so check them out if interested! www.youtube.com/@Glitchnator www.youtube.com/@rico-raven

[@JonFairhurst]

The parallelogram design is brilliant. I’d love to play a standard keyboard with this geometry to experience how it feels. It might feel perfect, or odd and synthetic. I’m not sure. I bought the Kawai VPC1 specifically to get an acceptably long pivot arm. I’m very happy with it, as it feels natural and I adapt to the remaining imperfections. With multiple pads, rather than long keys, I would guess that we would be more discerning on a Janko keyboard. The brain would probably ask why this pad feels different from that one, as compared to accepting the physics of a lever. For my top keyboard, I recently bought a Hammond SK1 for its small footprint and relatively long keys. It’s hard to find long synth keys in a simple controller. The long keys seem to be reserved for the large workstations. Having the high trigger point for non-velocity sensitive sounds (like organ and classic synths) and deeper trigger with velocity when needed is ideal. One question on the parallelogram design: does this risk adding friction and stiction? A simple lever minimizes the contact points and is insensitive to angle of force. Does the design risk binding, and is it slowed by friction? I’ve wondered why piano action digital keyboards don’t do this for good feel in a small, light package. Is it just due to cost, or is it due to practical issues like friction, binding, maintenance, and longevity? Again, best wishes on your project!

[@DIYJankoPiano]

The parallelogram design feels quite good to play, though for classical piano I think sticking to regular levers is the better option - not because it 'feels better' or anything, more because you dont want to get used to it and find it difficult to play on a regular one afterwards. Funnily enough there are a couple projects I read about which try to use a similar parallelogram style connection just with different length joints in order to 'simulate' a key with a longer pivot arm in a shorter piano body - think having a 25cm deep digital piano with an effectively 30cm pivot arm for the keys. In terms of friction & striction, its actually not a problem - I use plastic 3D printed parts (so no need to worry about humidity effects on wood) and felt bushings (yes - same as you would in an acoustic piano) that I bush very loosely (so there is barely any friction) for the key levers - so even with the 6 pins necessary (plus the 3 guidepins) as opposed to classical piano's centerpin+guidepin I end up with very little friction. I also bush the inside connection instead of the outside, which further decreases the effective area. All of that would make one think that there would be some wiggle room due to the looseness of the bushings, but felt bushings are quite good at preventing that so I also end up with no looseness in the key. Suffice it to say that I spent quite a bit of time trying out different key lever designs to minimize the effect of friction (and yes, the first prototype had almost 30g of difference between up and down weights... which is down to ~10g in the final version!) As for why digital pianos you see on the market dont do this... I would have to chalk it up to complexity (and thus cost). In a cheaper digital piano its assumed that 'hey - a shorter key lever is acceptable at this price range', so why bother making things complicated; while in a premium digital piano they just add longer key levers and call it a day. In the end upright pianos have rather short key levers by design and people have grown accustomed to it - so there isnt any incentive to do anything more complicated.

[@JonFairhurst]

Thanks for the detailed reply. It’s great that your implementation doesn’t bind or have friction issues. It’s an area where prototyping is crucial. I think one reason that the keyboard industry hasn’t done more is that many synth users just play pads or one finger beats. They’re not playing fast runs. Also, we can see a low action on a guitar and easily hear buzzing notes, so quality is important. On keyboards, the mechanism is hidden, and salespeople talk about sounds, menus, knobs, and buttons. The keybed discussion ends at “weight”. I think Kawai is the only company that markets their mechanism as an advantage. The problem really hit me when learning Fitrth of Fifth by Genesis. The intro is on piano. In the middle, it’s restated on the ARP Pro Soloist. I found that I couldn’t play it on the Arturia KeyLab mkII, which is ridiculous. How can I play something fast on piano keys and not be able to match it on synth keys? It should be the other way around. The key length was one factor. The second issue is that the synth line is staccato. This is where we want the high trigger point of a non-velocity synth. It allows us to quickly tap a key for a quick sound, rather than needing to get to the bottom of the travel. Modern controllers have completely lost the recipe. I wonder how the mass will affect playing speed on your project. A synth with plastic diving board keys, a simple lever and spring, and high trigger is probably the fastest possible architecture. My Kawai has a triple sensor that attempts to enable quick repeats, but it’s still limited. Hopefully, the plastic components are light enough to play quickly on piano.

[@DIYJankoPiano]

@@JonFairhurst I completely agree with your summary of salespeople and digital pianos... All too often the listed 'features' of the piano focus way too much on the '600 different instruments' stuff related to that instead of the design and feel of the action. Even finding information on how long the key lever is (as the longer it is the easier it is to play along the backs of the keys) usually has to be done by googling the piano and hoping some technical documents will list it somewhere. Funnily enough when I visited a local piano shop and started testing the various digital pianos without looking at any of the price&feature listings I ended up with 2 that I felt were truly a cut above the rest: the novus NV10S (basically a hybrid digital piano with an acoustic action inside) and... a regular upright piano... I didnt even realize it was an acoustic piano until I started looking for where to plug in the headphones after I decided 'wow - the action is quite good on this one!' So yea - that was why I decided on incorporating a proper action into my DIY piano instead of going with a simple 'weighted hammer' approach of the vast majority of digital pianos. As for my design, I can consistently repeat the same note 3 times at around a 12/sec speed. 14/sec is also doable, but at that point I run into issues with my own skill. Mass wise its not actually an issue - FDM printing means that the parts are all mostly hollow on the inside so most of the 'inertia' comes from the weighted hammers as its supposed to. In the 8 key prototype I have I can play both FF notes by slamming down on the keys as well as fast pp notes by not pressing completely down (just to the point of let-off, letting the inertia of the hammer finish the action) As for sensors, I took a completely different approach. I actually have a double sensor right at the end of the hammer's swing that records the velocity instead of measuring the key's movement. This means that if the hammer moved enough to (theoretically) hit the strings, a note will be played. If the hammer didnt manage to reach the strings, then even if the key was fully pressed a note will not be played. I did have to add a second (magnetic) sensor to measure the key's movement as well though - as sensing the hammer's movement isnt enough to send the 'note-off' signal... Going to have to talk about all that in depth during the electronics & sensors video.

[@JonFairhurst]

Perfect sensor implementation for Note On for piano. Nice. I really like my VPC1, except for the system weight. One limitation is that I can’t “throw the hammer” at the strings. On a real piano, one can play sharply, but not hit the bottom of the keybed and still get a note. The Kawai doesn’t support this. My piano teacher asked why I would sometimes not create a note on his digital Kawai upright, and this is why. Growing up, I didn’t need to follow through the entire travel to get a sound. Here is an area where a real piano action can be easier to play quickly. This isn’t unlike a quick organ action, where you can dance on keys. Since you can “throw the hammer”, your keyboard could be quick to play runs and arpeggios. The return time isn’t critical here. You can just play a quick impulse and move to the next note immediately. I’m excited to watch your project continue. I’d love to have a standard keyboard with your underlying design.

[@sphiarno-88]

私の知っているホールトーンレボリューションの鍵盤では、距離の差を減らすために、前後の列をアーチ状に、すなわち手前側は水平位置に奥側は上方位置に、構えることである程度解消しているそうです。 私も試しましたが、支点からの運動が直線的ではなくて曲線的であることを意識しないと、交差している半音違いの鍵盤どうしが干渉しやすい構造でもありました。三次元で考えないといけないのが難しいところです。

[@DIYJankoPiano]

(hopefully google translate works well here): Yes, the whole tone revolution keyboards feature a basic lever for their keys, meaning that the buttons along the back row move less than those along the front row. In order to combat this the back rows are arranged to be slightly angled. In my design I use the parallelogram linkages to ensure that all buttons have the same movement distance (10mm), so every key is parallel to each other with the movement being linear instead of curved. Additionally the keys are more rectangular as in the classical Janko piano as opposed to whole tone revolution's circular/hexagonal keys - so hopefully that will make it harder to misplay. You still need to think in 3 dimensions (kind of) as you have to decide which of the 3 rows to use for a given note, but with a few rules (first row for the thumb only, 6th row is only for transposition so ignore it, etc.) there is usually only one or two options for the button to use.

[@GameLoopLounge]

The keys you are showing in this video are 3D printed? They look so smooth!

[@DIYJankoPiano]

Yes, they are most likely the most 'processed' part of the entire build. All other parts? Just print them out, minor cleanup with an exacto knife, drill out the necessary holes to size and they are good to go. 5 minutes tops per part, with some being under a minute (then again, there are 15 parts per key, not counting the framework). Keys? Printed in ASA (instead of PETG), cleaned up with an exacto knife, vapor smoothed with acetone, then sanded (by hand) to give them the smooth (and yet textured) look. Takes around 30 minutes per key... I tried using acrylic paint, spray paint, and varnish; but all of them just didnt feel good to play on...