Animisha Is this typical for Kawai? Or do all dp's / controllers have this problem?

There are no even piano keys. They are all uneven. But the longer the keystick, the less dramatic the effect with the increased force near the fallboard.

Details:
Everything is just physics. Lever principle to be precise. This happens if the pivot point of the key is relatively close to the fallboard. Most digitals of all brands have issue. Exceptions are only the expensive ones with the long keysticks like Kawai GF-III, Yamaha Grandtouch and Roland Hybrid Grand action.
Some upper midrange actions like Kawai GFC and Roland PHA-50 have the keystick at least long enough to reduce this lever effect to an acceptable result.

Both my Kawai RH-III action as well as your RM Grand II have only about 20 cm long keysticks (15 cm visible, 5 cm behind the fallboard to the pivot point). Let's do a little math. The cm-value is always the distance from the pivot point to the center of the chips pile. E.g. The gray pile is about 2 cm away from the fallboard plus additional 5 cm keystick behind the fallboard. This makes about 7 cm in total:

White keys (+5 cm behind fallboard):

  • B / gray: 7 cm * 329 g = 2303 g*cm
  • C / red: 13.5 cm * 164 g = 2214 g*cm
  • D / green: 18 cm * 131 g = 2358 g*cm

The black keys have the pivot point a bit further back (+ 6.5 cm behind the fallboard):

  • G# / blue: 8.5 cm * 231 g = 1964 g*cm
  • A# / black: 15 cm * 154 g = 2310 g*cm

You see, the g*cm value is quite consistent every time, with the blue pile as a minor exception. I guess this could be caused by a lighter escapement-notch on that specific key. I increased the weight until the escapement notch "clicked".

For the physicists:

If you don't like <g*cm> because it sounds unscientific, let's factor in our earth acceleration of 9.81 m/s2 and we get this:

2300 g*cm * 9.81 m/s2
= 2300 * 9.81 g*cm*m/s2
= 22500 g*cm*m/s2 ..........................| replace <g> with <kg/1000>
= 22.5 kg*cm*m/s2 ........................... | replace <cm> with <m/100>
= 0.225 kg*m*m/s2 ........................... | replace <kg*m/s2> with <N> (Newton)
= 0.225 N*m (Newtonmeter, the SI-unit for torque)

q.e.d. 😏

      keff ...is measured at the end of the key ... ... In my opinion there is no point in measuring down weight at any other position ...

      This is true if you are regulating an action. As a piano mechanic or tuner you can not change the geometry of the key, so it is best to ignore it. All you need to do, is compare one key with another. Adjacent keys of the same piano, or the same key of another piano.

      But it does tell nothing about how much more force a player needs when he plays near the fallboard. The topic I made up with my post is about the geometry of the action, and it's effect for playing near the fallboard.

        I don't know what post you are replying to but it's normal that you need to apply more force at the back of the key than at the front. It's not a "problem". It's how every piano works (not just digital).

        Also, the lower keys are heavier than the higher ones. On an accoustic piano this is necessary because the strings are thicker and the hammer requires more momentum. Digital pianos usually simulate this effect by adding more weight on the lower keys.

          WieWaldi

          Thank you! It is too late in the evening to try to understand what you wrote, but I'll do that tomorrow.

          BartK Digital pianos usually simulate this effect by adding more weight on the lower keys.

          On my VPC1, the lower keys, especially the ones below middle C, require less weight in order to fall below escapement. 😐

          *
          ... feeling like the pianist on the Titanic ...

              I looked up the uneven key response problem about Kawai VPC-1. There is a long thread on Modartt's Pianoteq forum. Apparently the issue was originally noticed by player feeling it. Then someone devised a way to measure the response of each key. It's done by first manually holding the key so that it remains at the up position, place a 140g weigh on top of the key, the release the hold and let the key drop. The resulting MIDI value is recorded. This is done for each key and plotted on a graph like this.

              As I understand, digital piano generate MIDI value based on key velocity, which is measured by time elapsed between the triggering of two sensors placed at fixed distance. As long as those sensors don't shift position, I don't see how velocity measurement can be inaccurate. If so, the variation in MIDI value generate by same weight points to variation in resistance in the key's moving parts. Still, the result won't tell which part of the key travel has more resistance, whether it's just the beginning of the travel, the middle of the travel, or both.

              For one, VPC-1 is know to have sticky key issue due to deterioration of poorly designed slip tape. There are YT video on how to fix it. It's also clear from the video the entire key mechanism is easily serviceable.

                iternabe Ah - this is a complete different thing I was talking about. But I can see at a glance the graph is suboptimal. He has all keys (black and white) on a single line, resulting in a zig-zag shape. Would have been much wiser to make a dedicated line for the whites and one for the blacks.

                Still I can see it has different sections of midi values. And the sections itself are zig-zagging instead of going upwards in a steady shape.

                Do you know a free software that outputs the midi-value?
                Or better, creates a text-file?
                I am curious what my CN-29 does...

                    WieWaldi Do you know a free software that outputs the midi-value?

                    I am on Mac so GarageBand can do it. Pianteq, too, of course. I have not researched other options.


                      White keys were easier to test. I put the weight of 145 gram on the key (taking the hand away), while holding the key itself up with the other hand. By taking away the other hand, I allowed the weight to press the key down.

                      For the black keys, I placed the weight on the key (touching) and dropped it. There might be more inconsistency with this method, because if the weight already presses the key down a little bit before, there is less velocity. And if the weight was not touching the key entirely, the velocity is higher.

                        WieWaldi Wow, great data! Were you consistent with how far away from the fallboard the weights were applied?

                            rsl12 yes. White keys at the very front, and black keys after the bevel part.