CAUT

I'm starting to take measurements with Renner's new KMD - Key Measuring Device.  I'm finding that the results are generally higher by 3 or 4 grams than those taken with the traditional  Down Weight by finding the weight that causes the hammer to rise in a controlled motion after a few light raps to get it started and mirroring that motion with Up Weight.  I expect the consistency with the KMD will be much less subjective and way more consistent than the traditional method and it's way faster.   Any one else have input on this?

Following up on this question by David Stanwood after several rounds of testing on different actions.  

There are some inconsistencies to be aware of.  First, I would suggest multiple measurements per note, 2-3 is sufficient. 

Balance Weight:  I tfound that this measurement was fairly accurate when comparing with gram weights.  However, I found of range of up to 1 gram per measurement when the measurements were taken multiple times on the same note in succession. The range does seem to diminish with multiple measurements.  I would presume that this is due to the initial breaking of friction on the first measurement however it wasn't always the case that subsequent measurements were less in value so the cause remains unknown.

Downweight, Upweight and Friction:  When compared with measuring with gram weights I found the KMD values consistently different. The friction values tended to be much higher by a factor of sometimes 6-8 grams.  Thus, the DW and UW values also varied by the same amount.  Of course when you are getting a reading of 20+grams of friction the assumption is that you have a friction problem.  But when measured using gram weights, those levels, most of the time, registered in the normal range with DWs proportionately reduced and UWs proportionately increased.  It is my guess that since the device pushes the hammer all the way through let off, that skews the numbers as compared with the use of gram weights where we only measure until the jack/let-off button contact stops the rise of the hammer.  

My conclusion is that the device is still useful for measuring BW and since BW and FW have a 1:1 relationship, it's still useful as a quick reference for changes that need to be made to the FW to achieve your targeted BW.  That in itself is useful. 

As far as a guide for assessing friction levels (which would be nice for following up on what appear to be problem notes), it's not accurate or useful that way. Whether the relative values are consistent I couldn't determine.  If, for example, the normal friction levels through let off produce a reading which is consistently, say, 6-7 grams higher, then you could use it to detect problems where values fell either above or below those levels.  But I didn't find any consistency there.

Dip measurement seem to be accurate.  For me that's not really relevant as I set dip by aftertouch, not by measurement.  So, I tend to ignore that reading other than as a point of interest. For me it has no practical value. So this raises a bit of a conundrum; if you desire the dip function then it appears you will have to choose between that and a potentially, and yet realized, more accurate reading of DW, UW and F since an accurate measurement of dip requires depressing the note through let-off where the friction readings get a bit confused.  If I had to choose, I would rather have the device stop at jack tender/let-off button contact and give me an accurate reading of UW, DW and F.  

I did consider that maybe I could screw the let-off button up so the jack tender didn't contact it while using the device.  So I tried that but the KMD produces an "error" message when taking the measurement that way. Plus, not all let-off buttons have enough clearance to let you raise them that far and as a practical matter, if the piano has been regulated it would force you to "unregulate" it in order to take those measurements which probably defeats the time saving function of the device in the first place.

My general use, therefore, is just to be able to take a quick survey of BW which I do by averaging between two measurements (in my head) and rounded off to the nearest .5 gram.  That can be done for the whole keyboard in about 15-20 minutes which is faster than switching out and stacking weights, and trying to figure out how hard to pound on the bench in a consistent manner to get things rolling.  One half a gram seems more than adequate for these purposes.  I mean, do we really think we're balancing an action the tenth of a gram just because our scales read out at that level?  It's an illusion no matter how we do it.  None of our standard methods of taking BW, DW, UW or F measurements are that accurate.  

Since the F readings and thus DW and UW showed no real consistency, or accuracy, even with multiple measurements, I don't think it's useful for that, which is unfortunate for such an expensive device.

Perhaps a subsequent iteration can address those limitations.   

DL 

BTW after I wrote this it occurred to me that I would really want to check a complete survey of BW by KMD against a gram weights survey to see how consistent they were over an entire action note for note. Since I didn't do that, I can't be sure.  It appeared that the BW measurements were pretty good when I checked but I didn't check every note.  

It also occurred to me that a simple fix for this for the device would simply be to reduce the distance that the plunger depresses to the point of jack/let-off button contact.  That's an easy number to calculate and probably easy enough to restrict on the device.  It would give up the dip calculation but I think that's of no real benefit anyway.  The accuracy of the BW, DW, UW, F is much more critical.