<Things keep changing. That's a bit different than the last time I looked at you program. I don't think you were using end of the stroke at that time.
This sums up the problem I am trying to address. If even an individual's private definitions of an action ratio are continually changing and evolving, there is no way that discussions between techs can not only avoid mis-communication, but effectively seriously avoid confusing and unintentionally misinforming the conversation. We, by definition, will always have to be talking about apples and oranges, or apples and orang-pples, or appl-ges and orang-ados.
By "misinform" I mean,for example, tech x reads about the ratio I declare to be 5.23. They then gather a bunch of info from any of the knowledgeably techs in this thread, or in Mario's book, or anywhere, it doesn't matter. Then they try to duplicate my results at 5.23 using a hodgepodge of other calculation assumptions, ending up god knows where, and taking a lot of time to get there.
One of the interesting things I find from my own post which started the "half-punching" thread, is that I Included a ratio number in order to communicate better ....well that worked dandy Jim, didn't it?...ah no...it just illustrated this problem well.
The interesting things about this is, much like Alexander above, my own process has evolved in a way that doesn't assign a ratio at all...none...nada..neitnte. I no longer find the number useful. I mention this, as it could be a way out of the communication morass, perhaps...ie, quit talking about AR's altogether, as they are not informative.
My own evolution in the action design arena, actually mirrors this "quit talking about AR" idea. Here is my own evolution in brief as an illustration. I was first mentored by Bruce Clark, and used both Bruce's proprietary program (which he has decided not to make public, at least for present), and CAD from my own design background. The procedure was: measure action cavity dimensional parameters. Based on the action cavity's givens, design to a given set of regulation parameters, using the Clark/CAD approach to design and draw it conceptually. Then model it in an action model, and then model it in the actual action. It was not until a number of years later that I realized, that since in the end, I was going to model it anyway on the real action, I really could just skip the CAD/computer work, as well as the action model work middlemen, and just go directly to empirically proving regulation parameters on the actual as built key frame/stack, by messing with the capstan and heel parameters...no computer necessary. One of the reasons this made sense to me was, that if one looks at the givens which are imposed by an as-built action cavity, the possible number of mind boggling variables in the work, can be strategically reduced to only a small number of variables...ie, there are less choices to be made than one might think.
For example
-String Height is a given.
-Action Cavity Entrance height (distance between keybed and stretcher/block) is given.
-This means Bore distance is a given and shank height/stack elevation are a given.
-Strike point is given, which means shank length is given (unless stack placement is screwed up).
-add to that, if one imposes a known set of regulation specs that you know empirically work with a range of hammer weights, there is another set of givens...ie hammer weight and regulation specs.
This reduces the variables to capstan position, whippen heel height/location, and knuckle distance, as the only variables (in an existing action). Balance rail position is a potential variable, especially if like Dean, you are making a new key set. Although on that score, even though I have done balance rail height and front back relocation on existing keysets, I frankly have gotten fine results by taking the balance rail position as a given and only considering capstan position, heel heights/placement, and knuckle location as a variable.
I think Nick said it well...
< In the final analysis, if what you are doing allows for a ─ "standard" regulation (close to the mfg. specs), and your hammer weights and key leads are under control (which means inertia is under control) ─ then your action ratios (whatever you think they are or however you measured them) are useful and relative to your work. However, they may not translate well over the internet. So why don't we refer to action ratio by not referring to it directly, but simply by referring to as-built regulation specs, relative amount of ballast required to hit the action's static DW? This would say something which tells you a whole lot about the given action, without unintentionally mis-informing anyone, including ourselves?
Another possibility, as in the pics in the OP in this thread, even though technically incorrect because it measures vertical linear dimensions instead of arcs, taking simple action ratio as shown in the pics, can in be informative in a statistical sense. Even though technically innacurate, it could indicate roughly, statistically what a given action would feel like in the powered portion of the stroke...a ballpark measurement, suitable for discussions, but not an absolute design value...just throwing out ideas. Problem with this is that the brain sees a number,and the bias is to inappropriately assign credibility of precision to that number , because its a number and because its published on the forum or journal ("But I saw it on the internets!" syndrome)
just thinking out loud...Nah, I take that back...I think any AR number is a realistically hopeless way to try and communicate about this stuff.
------------------------------
Jim Ialeggio
grandpianosolutions.com
Shirley, MA
978 425-9026
------------------------------
Original Message:
Sent: 01-17-2020 23:58
From: David Love
Subject: Action Ratio Measuring
Nick.
I see. It's kind of like following what you should and shouldn't eat. Things keep changing. That's a bit different than the last time I looked at you program. I don't think you were using end of the stroke at that time. Hopefully this helps to answer Jim's question as to why folks are arriving at different numbers.
thanks for taking the time.
------------------------------
David Love RPT
www.davidlovepianos.com
davidlovepianos@comcast.net
415 407 8320
Original Message:
Sent: 01-17-2020 18:41
From: Dean Reyburn
Subject: Action Ratio Measuring
Hi David and Nick,
Thanks to Nick for clearing that up. The end of stroke AR is I what I was referring to also as being about ~5.0. End stroke AR is especially convenient to use since both radial arm measurement and action regulation spec measurements come up with the same AR number, confirming each other within 1% or 2%.
I'm not saying anyone is right or wrong how they measure. However I'm attempting to objectively and simply measure action ratio and key ratio using the most standard and accepted methods. Gravagne and Baldassin's methods at least agree and confirm each other, and they agree with my ancient copy of Walter Pfeiffer's "Piano Key and Whippen" book, which is the old standard on the subject.
Nicks article in December 2018 is I think the best definition on how to measure radial arms, and his article in March 2019 covers the fact that the AR changes throughout the key stroke.
As Nick notes, most techs can just use their own method internal, and that's fine, I have to have a way to talk to my rebuilder customers so that we're both on the same page.
------------------------------
Dean Reyburn, RPT
Reyburn Pianoworks
Reyburn CyberTuner
1-616-498-9854
dean@reyburn.com
www.reyburnpianoworks.com
www.cybertuner.com
Facebook: www.facebook.com/dean.reyburn
Original Message:
Sent: 01-17-2020 18:00
From: Nicholas Gravagne
Subject: Action Ratio Measuring
Hi All,
One of the primary goals of my work and articles has been to establish a baseline for working out an AR, both physically (with the action under hand) and mathematically. From this baseline, I imagined that the technician could then depart to some extent and do his or her own thing. Having said that, ranging too far afield from the basics will land us in the Mutara Nebula (Star Trek fans understand that the MN contains high levels of static discharge).
In the final analysis, if what you are doing allows for a ─ "standard" regulation (close to the mfg. specs), and your hammer weights and key leads are under control (which means inertia is under control) ─ then your action ratios (whatever you think they are or however you measured them) are useful and relative to your work. However, they may not translate well over the internet. As to the ratios and measurements discussed in this thread, they can only mean so much as we are all not together in the same room working as a team.
The basic equation of distance remains as Dean has posted and we have published:
AR = (B-L) / (D-A) where B is hammer blow, L is let-off, D is key dip and A is aftertouch.
Making reasonably close measurements and running the simple math will always tell us what is there. The AR cannot be changed simply by changing the blow, or any of these variables. For example, shortening the blow will increase the aftertouch and vice versa. Correcting the AT will shorten the total dip, and so on. The AR can only be changed by altering a knuckle or capstan location, or else having a new keyboard designed.
The radial arms method should confirm the simple AR input/output method:
AR = KR * WR * SR
where KR is key ratio, WR is wippen ratio, SR is shank ratio. All the ratios must be based on radial arms dimensions. These dimensions are defined in my articles.
A particular point of confusion surrounds the ratio of weight and mass (what I call the MAR for Mass Action Ratio, Moment Arm Ratio or Mechanical Advantage Ratio). The piano action is a poor mechanical multiplier, since it takes a large input at the key end to lift and balance a small output at the SW. Thus, the MA is negative.
Technically, the weight ratios we see published or discussed should be a number less than 1. For example a 5.7 mass ratio should really be its inverse, or 1/5.7 = 0.175 to 1, meaning that the SW will only be 17.5% of the activating downforce resulting from the tandem efforts of the BW and FW*.
So then, if at E44 the downforce is 57 grams, then the maximum theoretical SW = 57 * 0.175 = ~10 gram SW (approx. 8.5 gram hammer). And so, equilibrium is attained. Should the actual SW exceed this maximum, then key leads begin to pile up in order to maintain static balance.
Since our tradition has set the weight ratio above the number 1, and since this ratio comes in as 5 "point something", it is often confused with the simpler AR of distance input to output.
*Strictly speaking, the BW in this sense does not include the weight of the wippen, which is an intermediary "gear". Its weight does not contribute to the ratio dimensions of its lever arms, which form the basis for the ratio calculations. This is why the wippen weight is subtracted from Stanwood's equations for AR and SW. The effective wippen weight at the key end (say 8 g), the BW (as used here, say 30 g) and friction (say 12 g) are always overcome by the pianist's touch (total 50 g DW). The FW contributes to the combined effort to rotate the key.
RE the Green Zone or Goldilocks zone: In my earlier work I tended to target an average AR. These many years now, I use the end of stroke AR, which comes is at ~5.0 (plus or minus a 10th). I like this better as it dovetails with the manufacturer's regulating specs. Any readers using my older programs should contact me for updates. If your current program includes a page called Compare, you are mostly up to date, but still may be lacking a small item or two.
Original Message------
′Dean
There are clearly different ways of measuring the in/out of each lever and I think that may account for differences. Last I checked, a 4.9-5.0 AR would be outside of the "green zone" for Gravagne's program. I use his system for measuring each lever and that would be low for me too. I can't speak to what Baldassin/Renner advocate now but know Nick has been involved there.
Interestingly, your numbers comport with what Rick Wheeler advocated but I believe they (Wheeler and Gravagne) used different methods. In the Stanwood system it seems that 4.9 - 5.0 would also be outside of his normal target, but I'll have to let him confirm that. Wheelers 4.9 yielded something quite different using mine or Gravagne's methodology
Product of levers is only as good as our agreement on how each lever is measured so we should verify that before going further.
------------------------------
David Love RPT
www.davidlovepianos.com
davidlovepianos@comcast.net
415 407 8320
Original Message:
Sent: 01-17-2020 12:22
From: Dean Reyburn
Subject: Action Ratio Measuring
Jim and David,
Thanks for your comments on action ratios. I've also been frustrated with the lack of standard methods to measure and express both the action ration and especially the key ratio. That crucial for us as we build keysets since we have to talk to our customer/rebuilders and have some common frame of reference. It appears both you guys use a different method for calculating or measuring action ratio than the Gravagne/Baldassin/Renner method which we use. Our action ratio targets end up close to 5.0 (4.9 to 5.1 is the goldilocks zone) since we subtract out both let-off and aftertouch when we measure the action:
AR = (B-L) / (D-A) where B is hammer blow, L is let-off, D is key dip and A is aftertouch.
Since we are usually proposing a change in key ratio, but don't have the key produced yet, we use a variable key ratio setup with two keys to test actual regulation. I've uploaded a picture of the two special keys we use. They have a moveable balance rail, adjustable capstan and front mortise. The keys will essentially adapt to most keyboards in a certain range. (we have three sets of these for small, medium and large pianos).
But we also measure and calculate the parts's radial arms and to calculate the AR:
AR = KR * WR * SR
where KR is key ratio, WR is wiped ratio, SR is shank ratio. All the ratios must be radial arms, and we define those the same as Nick G.
As long as everything is measured accurately, the two systems produce the same results, within a very small range of say 1%.
The stickler for us is the key ratio. We have to use a "linear" key ratio in keyset production, it's just not practical to measure or define the radial key ratio for the CNC machine. But for calculating KR and AR from parts measurements we have to use "radial" key ratio.
Sooo..... I've been working on a project for our customers to *objectively* take just a few (10) simple measurements, fill out a web form and submit it so we can analyze a action and keyboard. It's a free service to any rebuilder or technician. We find that very few techs out there really understand action or key ratios, so for our use, this defines our standard methodology. There's no question of what we, or the tech is talking about as far as ratios once they've done these measurements.
This has been through a number of iterations, but it's now in "public beta". Here's the web form:
https://www.reyburn.com/geometry.html
and the directions are linked from the form above, or here:
https://www.reyburn.com/geometry.pdf
I'd be happy for any constructive criticism on this!
Best regards,
------------------------------
Dean Reyburn, RPT
Reyburn Pianoworks
Reyburn CyberTuner
1-616-498-9854
dean@reyburn.com
www.reyburnpianoworks.com
www.cybertuner.com
Facebook: www.facebook.com/dean.reyburn
Original Message:
Sent: 01-16-2020 13:59
From: Jim Ialeggio
Subject: Action Ratio Measuring
I want to discuss measuring action ratios, following a confusing response on the recent "half-punching" thread. Distance measuring action ratios can be confusing, and we need to know we are talking about the same thing when trying to communicate about this stuff. Here's is David Love's comment regarding action ratio measurement, that makes me think we are not communicating precisely...David's comment below, says quite clearly that a 5.25 action ratio, at .390" dip, will result in a very shy blow.
David L< I agree that if you push the leverage down to 5.25 that you will want a higher static balance weight. However, I don't see you can possibly accomplish a 3.90" dip with a 5.25 leverage (if that's actually what it is--taking accurate measurements with dip/travel is tricky). I've seen plenty of 5.3 targeted leverages and the dip is nowhere near that low, not without a very short blow distance. Too short for anything I would want.
Contrary to David's comment, my own empirical evidence says something quite different. I see a 5.23 directly measured action ratio, providing a carefully measured set of pretty common manufacturer's regulation spec's: .390"dip/ .040" aftertouch/ 1.75" blow/ 2mm letoff. 1.75" blow is what I measure, produced by my measured 5.23 AR.
So, I don't think we are communicating precisely. As this action ratio confusion seems to come up often, at least to my mind, the confusion around measuring it can muck up the communication often, I think. So I want to suss out what the communication mismatch may be.
After reading David's comment, which I knew did not jive with my experience, I went out the shop to double check my measurements, to make sure I wasn't talking through my hat. There are two ways I measure action ratio. The two methods double check each other. One is direct measurement of the ratio, and the other is to confirm that the regulation specs I target are actually achieved. I measure both of these with the highest l level of precision the felts and other materials in the system will allow. As someone who set up production shop runs where the machinery had to cut 100's of water tight sash joints, in a life previous to life in piano land, I know where to look for measurement inaccuracy, and how to avoid the major problem of making dimensional assumptions. Here is how I take my measurements.
1- To start, before hanging hammers, I draw a line across the set hammer crowns indicating the strike point of each hammer, and refer to that line throughout the rest of the action process. In the installed actions, this line ends up being pretty darn close to the empirical strike point on the crown. In this case, I'm more interested in measuring to the exact same place on the crown for the two different measurements.
Using a 12" machinists height gauge, with a dial scale, Measuring from one point on the marked strike point on a given sample hammer, I take a measurement with hammer at rest, using the machinist's height gauge, and record the reading. Then with a separate machinists height gauge at the key front, set the gauge to just touch the top of the front of the key at rest. So I have my starting positions clearly indicated. Then, at the height gauge at the front of the key, I insert a shim of known thickness (.246), between the height gauge and the keytop, thus depressing it a precisely known dimension, ie ,246". Then I go around to the back of the action, and measure to the same point on the crown of the hammer I indicated to in the first hammer-at-rest rest reading, using the 12" machinist height gauge, at the new raised position of the hammer at .246" dip. Record that t measurement and do the math: (raised hammer reading - hammer at rest)/.246 . (7.695-6.413)/ .246 = 5.23
This is precise measurement, and does not depend on relative positions of other parts. So the action, at this note is according to this technique, 5.23, and is as precise as can be had in this kind of setup. Here are pics of this process...the last pic shows that the jack tender has not touched the letoff button, which is essential in taking this measurement;
------------------------------
Jim Ialeggio
grandpianosolutions.com
Shirley, MA
978 425-9026
------------------------------