Kawai VPC1 faulty key weighting (Page 2 of 4)

Hallo
I´m not sure about it, but i gues the problem (in case of MY VPC1) ist the let-off function...!? Some keys need a little more pressure to be pressed completely down. And when i try to play very gently on the keyboard thease special keys does make some problems: they have a lack in velocity or there is no sound because i pressed to gentle.
I would gues there are about 8 keys which act this way, and so i´m thinking about returning the VPC1 - sadly.

Before i owned an Studiologic SL88 Grand. And now i think this was the better choice for me.

Hello,

the OP had removed the Let-off simulation of his VPC1 (remember picture). If the friction of the Let-off simulation would be the reason for your problem, then maybe two different sources for unevenness exist in a VPC1.

Normally the let-off-rubber is touched before the note-on contact is triggered, right? I'm not sure about this, but it would make sense to me. In a real piano first the mechanical let-off/escapement can be felt, then the hammer "flies" and then the string sounds (=Note-On). 
 

Because you didn't mention it: Are those "about 8 keys" the eight adjacent white keys in an octave?

I couldn't find something about let-off/escapement simulation on the StudioLogic website. Is the SL88 Grand free of it?

Interesting find there groovy! Just re-did my VPC1 action, I reduced the key dip distance and find it makes way better playing. The key dip I'm more convinced than ever is the most important part of the calibration, e.g if the key dip is uneven then chords will be played with unintentional grace notes, or at least that's how they will play/sound as the sensors will be hit at different times. Volume is less of a priority to me, the accuracy of evenly triggered (timing wise) is essential as this cannot be corrected properly during playing but volume can. The white and black key also should be adjusted so as to trigger together and not give the grace note effect. Soz going off point groovy, just very exited at my VPC1 action this week, I'll have a look again at your link to see if there is some comparison to my way of calibration

Hello,
sorry for my bad English. Google Translate helps me.
I bought KAWAI VPC1 and started testing the keyboard. Actually wooden keys, nice action but a big ... problem with midi speeds.
I did the test with a weight of 140g and 170g. The results in the drawings, for comparison I added the M-audio hammer88 140g test and Studiologic SL73 Studio - 170g. You can see that Studiologic is very hard. M-audio has very equal action.
In contrast, kawai vpc1 is very uneven.

The result surprised me very negatively. I expected a good quality product at this price.
Thanks to your descriptions I have seen that everyone has similar problems. Many thanks to the user "groovy" for The "Key Switch Adjustment".
I cleaned all contacts with IPA isopropanol. The capstan was unregulated ...
Two days of work gave the following result:

If anyone is interested I can put a short description of the regulation.

There is a total lack of quality control. Please see what the left key cover looks like. This is a scandal !!!

MDF board is badly made technologically, it will crack in all models. MDF is wastworn and must be rotated 90 degrees.
I myself do many CNC MDF works as a hobby. Some constructions are over 80kg and nothing breaks. Example:

In general, the keys are cool. But the workmanship is poor. Key speed sensors using rubber contacts is not the best idea - not durable over time.
I am surprised by the technology used, because I myself can develop a contactless sensor position sensor with automatic compensation.
I have been a programmer, constructor and electronics specialist for over 30 years.
It is a pity that Kawai does not apply innovative solutions. If they used them, they would be the best in the world.

Hello,
there was a small error, I did not notice.
M-audio had a curve shifted by 32, i.e. the curve has the same shape, but values between 45-50 midi.
All curves are linear and x = y.
Unfortunately, Studiologic is tough, even very hard. At 140g, the keys barely reach 20 midi.

Kawai:
Some Capstans were set too low and there were false double keystrokes, e.g. 72 and immediately 3.
I was afraid that the erasers are not working properly - but they are ok.
Capstain has a screw: adjusting to the left increases the capstain - and decreases the force, while adjusting to the right causes lowering and has little effect on reducing impact force / time.
There is another problem there! Sometimes capstain does not symmetrically lie on the felt with silicone at the end of the key and hits the rubber sensor not centrally - and an underestimated speed arises. In principle, the capstain was set to an equal height, but more or less. Detriment. The adjustment is very time consuming and must be repeated several times. Each time you have to remove the key and capstain.
I have a comfortable situation because I have a neighbor who once tuned pianos.

Unfortunately in my copy the conductive rubbers and PCBs were not smeared with anything.
They were dry, dusty and a little scratched as if the pcb were lying to each other. And this is a new keyboard !!!

Lachnit:
I saw this optical solution (Lachnit), based on the Studiologic Grand keyboards. Work unlimited in time, but a problem in calibration.
I have a different idea, but I will not describe it here.

If I have a moment of free time I will make a new control and connect it to Kawai matrix. Then all the problems described will disappear.

Regarding the service manual, the S1 sensor is activated by pressing the white key by approx. 4.5mm, S2 by approx. 8mm and S3 by approx. 10mm.

Would you also mind going into more detail? I'd like to perform the same procedure on mine.

Thanks for clarification, Rtykon.

The SL73 Studio could be such a nice Rhodes emulation with a normal and quick hammeraction - sad to hear it seems to be on the heavy (and sluggish?) side.

Sorry, in my question I meant the friction points:
hammer <-> rubber-dome
hammer <-> let-off-rubber

Are these points lubricated by factory or dry? I'm just interested in the design of those wooden hammer actions and really don't know. And you were so brave to open the monster ...

Hmm, the 3rd sensor (S3?) should be triggered at ~10 mm? At "fullstroke" distance, where a white key is stopped by the underlying felt?!

Good night
(oh yes, factor 88 is a pain!)

Hi groovy
The hammer-rubber friction point is dry.
The third sensor S3 shows the full release of the key. Therefore, the PCB with sensors must be perfectly mounted. If installed incorrectly, there will be no S3 closure - i.e. no sound!

I also regret that SL73 is hard.
I was happier if I didn't know it.

Ah, that's good news for me, thanks. One variable less and zero maintenance (no lubrication needed by design and no greasy dirt).

Strange, normally the last contact in a "rubber-bubble" triggers the note-on/velocity event before a key hits the ground.

... when something has to be perfect, then Murphy's Law comes into effect

It is 2020 now and building a fine balanced and full range (0-127) velocity keyboard still seems to be rocket science ...

Hello,
I'm starting to be happy with the VPC-1 keyboard. All the keys began to behave as they should. I have learned a lot and most importantly they are wooden ...

The description of improving the VPC-1 keyboard operation will be divided into two parts: cleaning of rubber sensors and setting mechanics.

All changes described are made only at the risk of the person performing them. Improper performance may result in damage to the device, loss of warranty, etc.

The user "gasparka" (19-11-2019) has placed beautiful photos of the rubber sensor plate. In fact, everything is visible there!
It consists of two PCBs. Do not wrap two at a time - only one at a time. It is difficult to fit the tiles correctly, as there is a lot of slack in the screw holes. First, disconnect the keyboard from the power supply, then disconnect the two straps fastening the plates (in the white connectors, pull the latches and remove the wires). The contacts are washed with IPA isopropanol but with a slightly damp swab. We do this on a pcb and a rubber sensor. We wait for evaporation, we turn the plates and check the action.
We do the test with Pianoteq and a standard weight of 140 grams. The weight must be glued with a silicone or felt pad. Not to scratch the keyboard. The weight must be gently applied to the tested key and released freely. If the results read in the 3-5 series are accurate to 1 midi unit, then the test is correct. You need to practice this task. We do not take the average as the final value, but the maximum value with repeated results.
I don't know any grease suitable for this type of rubber contact, please don't use WD40 (it may damage the rubber).
There is a high probability that your keys will work much better.

If this does not help, I will describe the mechanical adjustment tomorrow.

I think although credit there, reducing the key dip is good, when you put the cabinet back in place a gap between the name board red felt and the keys will be an issue here. I would take some time and adjust the capstan to achieve your desired key dip. The name board and felt help to control excessive bounce. Hope this will help you on your mods Edit..remember all these parameters are set for good reason, even the key dip, should still be close to original spec

Yes I was going to say your 'tape method' resembles the Mozart rail. When you say 'key down' is that the felt at the far end where you put the tape? If so then there's no need to change this even with a raised key dip when capstan adjustments are made

If you lower the key dip that will mean tightening the capstan, this will increase weight throw and volume but yes, if excessive then the key will hit the front felt before the note is triggered. With the wear over time to the pads on key ends, an increase to the factory setting key dip will, as the pads compress, increase the key dip. It would be more likely that to compensate for this, adjust the capstan anticlockwise ( I'm guessing here, trying to visualize?) to slacken (so to speak). Again tiny amounts here or the screw will either fall off or weaken the thread. Edit... just to clarify the key dip is the distance the key travels (at the playing end) from rest position to pressed down (with a gentle press) and not the distance from underneath the key to the green felt on an un-played key

I see white grease on the let-off-simulation mechanism (what you called "silicone bands") on your foto.  Have you lubricated it now?

Thank you. I asked that a few posts above and the answer from Rtykon had not been clear.

Thank you for your response.
For black keys (also for white) the capestan position for adjusting the impact strength. If you want to have a keyboard set from 50 to 60 for a weight of 140gr - you cannot set the cabestan differently.
The most interesting is that the attached coins do not change the pressure of the key. Maintains dynamics such as adjacent keys.
It looks like a heterogeneous wood (plywood) structure in the key.

Sorry, I was wrong - the screw weighs about 4.5 grams and the screws are under the keys 01-33.
adding the screw (mass) to the key will increase the pressure, but at the same time will slow down the return of the key.
It is better to regulate the weight of the hammer. I thought to fix the rivet and use a countersunk screw. You can use your own weights.
Only then will I lose the guarantee for 100%. Adjustments made to date are not visible.

Interesting site about weighing
https://www.kawai.co.uk/products/archive/ca51/

@groovy

Not completely.
200g causes the bottom position of the key, but 140g also causes the bottom position of the key. The pads are used to adjust the position of the keys when idle (white and black).
In contrast, the weight of the hammer regulates the strength of the impact and the height of the key when pressed. This is a compromise. To a small extent you can adjust the strength of the impact without noticeably changing the work of the key.
90 grams causes the key to swim about 75% of all movement - i.e. about 7-8 mm.
Please remember that the keys should also have a controllable return time. About 92-100 midi units in pianoteq.

For me it looks like just the grey background has been rendered.

The next generation VPC2 will have the GFC or the just released GF3 action, I'm sure. Hopefully these are better designed technically and do not just introduce new flaws :-/

Very interesting. It seems, that the VPC1 has been the only Kawai with a MIDI-offset compensation.

Hi bijlevel,

I don't have experience doing this myself otherwise I would offer help. (I realize you may have seen me just post to a few other threads and log out - wanted you to know I'm not ignoring you).

Give some time, and others (like maybe some from this old thread) might see you revived this thread and have some help for you. Since it's an old thread maybe it won't be immediate, esp. with time zones different for everyone.

One day I might have time to do this with my Kawai, so would also like to keep reading the help you will receive here - but I have never opened or cleaned my keyboards yet.

Good luck to you.

[Edited to change username bijlelel to correct bijlevel - apologies]

The latest Kawai hybrid piano have the wooden keyboared with longer pivot point, making it closer to the action of a concert grtand piano.
Kawai VPC1 it's very good, but the pivot point it's a bit s hort. compared to concert grand's action.

I wonder why Kawai use rubber strip contact sensors, instead of optical sensors like in the PNOscan.

Hi bijlevel. If your keyboard is triggering ok then I would leave it. Cleaning won't suddenly bring about a big change or improvement in the VPC1 performance.

Nick

PNOscan it's not mass produced, so it's costy. But the optical sensors little piece underneath each key itself, it's not expensive, I think.

By the way, how VPC1 and the 3 contact sensors per key works???
I know that for 2 sensors it use the time difference between the first and second contact to get a vewlocity value, since the higher tehe velocity the lower will be the time difference between each sensor.

But for 3 sensors... Uhhnnn. Is it the arevage of the time difference between the first to the second contact sensor, and between the second to the third contact, used to determine the key velocity?