Killa, I have not used that technique, I was presenting it rather tounge in check as it seemed like overkill for a harp application. I found it trying to research flatness required for gasket-less air tight seals. What I found was that this type of seal is used in laboratory vacuum equipment and oxygen handling equipment (where a gasket could easily become fuel).
This paper from Edmund's describes the measuring process. Here is another good description. If you understand the physics of this you will see that it has a lot in common with how a tremolo harmonica works. In this case the tremolo is in light intensity rather than sound intensity. And just like a tremolo harp, if you are not really close to being in tune (or flat in this case) the effect will not be noticeable. Actully you are not measuring flatness, but variations in the spacing between the optical flat and the workpiece,
A source I look to on the rare occasions I need optics is Edmund Optices. I see they do have optical flats , but one big enough to check a reed plat is about $2k. They have float glass windows that would give you an interference pattern but these would only be as flat as window glass and the interference pattern would sum the waviness of the reed plate with that of the glass.
You could try is with a red led and a piece of window glass. Red has the longest wavelength in the visible spectrum at 700 nanometers (.0000275 inches), compared to violet the shortest at 400nm. You might get a useful pattern, but I expect the interference lines will be so close together they are hard to detect (just like a tremolo harp where the two reeds are not close enough to in tune). Actually, the same may be true if you used a $2K optical flat as I don't think most reed plates are flat to within a few wavelengths of light, even big red 400nm waves!
Last Edited by STME58 on Apr 11, 2016 8:09 AM
Alright Grey Owl, it was using a silly pun about taking something for granite that got me started down this rabbit hole! Now with a comment about smoke, someone will suggest that the tar from cigarette smoke will seal the gaps between the reed plate and the comb and that is why smokers harps sound better!(Actually, I think I remember a thread on this.)
By the way, a wood smoke particle is about 1 micron (1000 nano-meters) and is definitely big enough to impact an interference test if it gets between the work-piece and the optical flat.
Lol. Interesting stuff. I got the gist by your original post on the other thread. But i was wondering if it was practical or not. There has to be a relatively cheap and easy way of checking and attaining reedplate flatness.
Idk hopefully someone smarter than Me will come up with something. ----------
Try a glass plate with a thin film of light oil. The parts of the plate that do not come in contact with the oil will be visibly different from those that do. IT gives you the same information sanding the plate using a piece of sandpaper on a flat surface does, but without scuffing the plate.
STME The oil idea is excellent aswell. Good Stuff.
The problem with the sandpaper isnt the scuffing. Because im going to flatsand it anyways. But after you adjust it you cant tell if the parts that were in contact before, are Still in contact. Does that make sense?
So even thought i will still flatsand after, the oil way will be much more efficient.
Plus you could use it to straighten the blow plate. ----------
Last Edited by Killa_Hertz on Apr 11, 2016 3:29 PM
I am not a harmonica tech, just a mechanical engineer with some experience in sheet metal design but here is what I would use to flatten small pieces of this brass sheet in small quantities. You need to run it through several times turning it over each time and making adjustments to the roller to get it flat but it does work. This would require removing all of the reeds. This should have been done before the reed plate was stamped on a machine like this note in the picture that the curl switches direction and gets smaller with each set of rollers so all the work you would do with the little hand machine is done in one pass. Factory reed plates are not flat either because the leveling process was not set up correctly, or the plate was warped in the stamping or assembly process. Dialing in the machines to produce flat parts can take hours. This is economical if you run 10's of thousands of parts at a time but I have a feeling the harmonica industry is not that big.
Yea. I figured you for an engineer by how u talk. Lol. That meathod would surely work, but removing the reeds is more trouble than is worth.
Ive also delt with my fair share of process lines, setting up the instrumentation and wiring plants/factories
The plates are likely bent when stamped, like you say. But even if they weren't, they surely would end up that way after being banged around the line and handled by all the workers in assembly.
I like your glass oil method. It seems like it should work pretty well. ----------
I've found that a ~ 4" x 2" glass weld filter (the dark lens for eye protection) works well. They are fairly cheap and easily found at any weld supply house. They have larger sizes also. Flatness is not guaranteed on all, tho.
I know it's been mentioned but surface finish of the comb or plates also comes into play.
JustFuya, have you been able to get interference patterns with a monochrome light source using the welding filter? Or are you using it with oil or just a a convenient surface plate to compare to the reed plat to see if the plate is flat?
You are certainly right about the surface finish. In the little research I have one on metal to metal seals I am seeing recommendations of a micron or better finish, with a strong recommendation to use a gasket if at all possible.
One common metal to metal seal is the valves in an automotive cylinder head. Those are lapped to ensure the valve and the seat conform to each other even of they are not perfectly conical. I wonder if a similar idea could be applied to reed plate and comb. Perhaps using an ultrasonic vibrator with a head that contacts the reed plate at each of the mount screw holes and vibrates at a frequency and amplitude that is well below that required to weld so that the contact points between the two are worn down until the whole surfaces are in contact. This should work for wood metal or plastic combs as you don't really care whether the reed plate or the comb is worn down as long as the two surfaces end up in contact everywhere. The reed and comb would not be interchangeable without repeating the process.
Nate, I see you were paying attention in geology and have put the knowledge gained to a very noble use in taking on the mantle of the punster. Igneous!
Interesting. I assumed the better finish the better the seal. How far is to far? Or is there no such thing as to far?
Either way you can get amazing results usein sandpaper. Stepping it up to finer n finer grit.
Im mostly interested in getting it "flat" before sanding. That way you dont have to spend as much time at the lower grits, using the paper to make up for the uneven surface. And just use the lower grits to knock off the burrs and rivet heads ,etc.
That way you can spend more time polishing with the finer grits. Aswell as have a much quicker turn around.
Could also use the oil to check the finished product. Aswell as check your comb flatness and progress during flattening.
What would be a reasonable meathod of getting the plates even further polished?
Re Gaskets ... I still think gaskets could be a good idea. Someone just needs to figure out the logistics. I know the problem in the past has been microbes n such. I'm sure someone can make it work. ----------
STME58: How flat is flat enough? I find the flatness obtained by making both surfaces rest together without letting any light through is flat enough. I hold a straight reference along the axis I am measuring and look for light peeking through.
I like the edmundoptics article, although the link you posted is garbled (I fixed it here).
"If you can measure it, you can make it." That's so true. I insist that you can make a piece flatter than the lapping surface by measuring where the piece is unflat and applying a little extra pressure in strategic spots. You can make the components of a harmonica flat using a cheap kitchen cutting board from the dollar store. The trick is to use a good reference.
As for the colors of the spectrum, It's interesting that as I am measuring the flatness of a comb I hold up to the light the interface between my reference comb tool. As I approach flatness, the slot of light will show a bluish or reddish color.
It's not really helpful other than to let me know I am getting close.
And The F-Tool I use to flatten blow and draw reed plates uses the same principle as the tiny bending machine you linked.
Do you know of an optical flat that would fit in between the slots of a harmonica plate? The surface of the blow plate that needs to be flat has the reeds in the way!
Killa: As for gaskets, I think they are a tone killer. To me the goal is more than to just make the thing airtight. You get a boost when the pieces of the harmonica fit together so well it's almost as though they are one piece. Using a gasket will work fine, but you will probably be missing out on that extra sweet tone. That being said, I think the stock ABS Special 20 comb works so well because it behaves just a little bit like a gasket. (But not too much, it still has a great tone).
I still want to see someone try to design a comb that also acts as a reed plate. I know it would make things like embossing a pain, but it would fix the problem of getting the 'gap' between the reed plate and comb sealed once and for all!
@STME - First off I'd like to apologize for a smart-ass comment I made on another thread. It came from ignorance and frustration. You were in range of the flack and handled it like a gent.
I used the glass as a quick and easy surface plate, primarily for combs. I am not an engineer and am unfamiliar with the science of monochromatic light although my quick read on the subject was interesting. It sounds like something that harp manufacturers could use to spot check. I've been farming out my tweaks to the resident pro but will revisit my approach when I get around to working on my GMs.
Considering all the irregularities on a reed plate, I was using a thin Starrett 6" scale which also feature various and surprisingly accurate graduations. I check both sides of the plates diagonally and axially. [I've tried "General" brand 6" scales usually found in hardware stores and found them wanting for straightness.]
For comb straightness I use the blade of a machinist's square and sanding marks as was previously mentioned.
Using your cylinder head analogy, I have heard of backyard mechanics replacing blown gaskets for aluminum heads without resurfacing the head. The idea is that extra torque at assembly will correct the warpage. I see this a lot on harps (plate to comb). I've only opened a few Manji's so I can't speak of the entire line but they were the best examples of this.
Last Edited by JustFuya on Apr 12, 2016 2:47 PM
Andrew ... ah. The gasket acted like a sound dampener. That makes sence.
Flatness. Yes i quickly figured out that even a pretty noobie attempt at a custom harp can do wonders. The first time i went all out and bought a custom comb (one of yours actually) and flat sanded everything, etc. I was blown away. I got lucky and those plates were relatively flat already. Still needed work, but not warped or anything.
Now ive Done a few and im getting "decent" at it. The hardest part of the process is getting out a two or three slot leak. Even with the F Tool is tricky not to overbend and make thing worse.
That and really good shaping.
As far as the glass with oil or light i think it should basically make a topographical map. And tell you what you need to fix. I'm hoping for the stars on that one, but it might work ....
Matter of fact imma give it a try now n see.
BTW Nate the comb idea would be killer. The solid comb could have tapped holes for screw in replaceable reeds. Totally user servicable harp. Changing the Key would even be a breeze.
Idea is Copyrighted By KHz & Nate Inc.
(Ok ... its not, but if you steal it. Give us some free ones. ) ----------
Last Edited by Killa_Hertz on Apr 12, 2016 5:22 PM
So, has anyone actually succeeded in using the optical interference pattern to test reed plate flatness? It sounds like it may bear fruit but I'm not sure how practical it is. I also don't understand how to translate the interference pattern to actual mechanical improvement of flatness. But I am here to be educated.
One thing I do know is how machinists typically test the flatness of a surface and that is with a dial indicator on a surface plate. You can take the readings from the dial indicator to know where and how much material must be removed (or the object bent) to achieve a desired tolerance (+ or - x thousands).
But I've never heard of anyone using the method on harmonica reed plates.
Michelle
---------- SilverWing Leather - Custom leather creations for musicians and other eccentrics.
@mlefree - I saw a post about a flat surface which you posted somewhere else and it was a link to purchase a very flat heavy object for a reference of a flat surface. Is this good enough?
I've never liked dial indicators for checking flatness and I don't believe harp tweakers would have much use for them. The only times I've used DIs for flatness was for critical parts in order to quantify deviation for rejection reports. In these cases I would support the piece at 3 points.
For harp parts, I would do what the fella in the video does without the indicator using a known flat surface. Deal with the cause of the rocking and then fine tune with a straight edge.
Sitting here thinking of .. How Flat Is Flat Enough? ..
Lapping .. polishing .. etc.
Andrew whats the stages of sandpaper you go threw? The finest being? Do you do any final Polishing/Lapping?
@Michelle you said you don't know what you would do with the data from the glass. Well in my mind it would be like a typographical map. Then obviously you make the high spots match the low spots ... and this=Flat. Or damn close. Then you flat sand and get it really flat. But it takes out all the guessing and over tweaking and pain in the ass parts. Ofcourse if your plates had protruding rivets you would have to sand them first to get a good reading. Im going to mess with a few ideas and see if they are worth it.
Ide like to figure out a painless way to flatten combs aswell.
I also think this idea of mounting the reeds themselves ... on a comb could work. ----------
Last Edited by Killa_Hertz on Apr 13, 2016 5:44 PM
Good eye, Owen. The grey surface that the machinist has both the dial indicator and the object he is testing sitting on is a granite surface plate.
Surface plates are milled to very precise tolerances to be, for all intents and purposes, perfectly flat. They have many practical uses in a machine shop. They can be be made from various materials and come in lots of different sizes. Granite is popular because Chinese manufacturers churn them out very inexpensively. So inexpensively that one has to wonder about the carbon footprint of a 25 lb. chunk of rock that is manufactured in the first place and then shipped across the ocean, transported to Grizzly Tools and then finally delivered to your door for $36.94.
I have a couple. One is a 3" thick one residing on my indoor leather/harmonica workbench. Its main purpose is as a base for tooling leather. You need the mass to keep your surface from vibrating, causing your stamp to jump around. I also use it as a perfectly flat surface for sharpening and honing blades and polishing flat objects with a variety of wet/dry sandpapers and 3M polishing papers. A couple of harmonica-related applications are flat-sanding reed plates and combs. I also have one in my shop in the garage where my power tools are for similar purposes.
Both are integral parts of my workflows for a number of my activities. Once you have one you're constantly surprised by how handy they are.
Try one. You'll like it!
Michelle
---------- SilverWing Leather - Custom leather creations for musicians and other eccentrics.
Killa, I have used several kinds of interferometry. I know how useful it can be but I also know that it is a laboratory-level experiment, not something you pull off on your kitchen table.
I'm trying to visualize how you would use the concept on a reed plate.
You would be looking through the perfectly level glass tank from the bottom, not mean feat at the get-go. (You don't want the liquids spilling on your face while you lie on your back looking up).
Then you would see your topological map analogy of the intereference rings. You have, of course, calibrated what the space between interference rings translates to in terms of thickness based on the physical characteristics (temperature, altitude, etc.) of the two liquids you use and the ambient conditions in the room in which you use them.
OK, you've done all that and you see your topo map indicating where you need to make your bends or remove material and then somehow mark those locations. You take the reed plate from the bath, dry it off, bend/remove material. Then you clean the reed plate meticulously because the smallest amount of dust or detritus will totally foul the interferometry pattern (we're dealing microns here...). Rinse and repeat.
Alternatively, if you want to get anal about measuring flatness, there's the standard machinist way of doing things that is inexpensive by any comparison and doesn't require laboratory conditions.
You have your $36.94 granite surface plate and a $34.99 Harbor Freight dial indicator and you are, in your kitchen or garage, in the business of measuring flatness in the standard, widely accepted and applied way.
And, JustFuya, I agree that rocking a reed plate or comb on a flat surface is far more practical than interferometry. I only mentioned dial indicators as a method somewhere in between those two methods in complexity and expense. The goal of measuring the flatness or thickness of an object is different from rocking it on a flat surface.
Michelle
---------- SilverWing Leather - Custom leather creations for musicians and other eccentrics.
Last Edited by mlefree on Apr 14, 2016 5:14 AM
Michelle, you seem to have some experience with this. What are the chances one would be able to see any interference rings are all with on a reed plate using window glass and a red LED? I suspect slim to none, based on what you just posted above and the setup procedures I have read.
"I'm afraid this is one of those things that sounds great on theoretical grounds but is very difficult to pull off on a practical basis."
It seem a little like measuring the size of your living room with a micrometer. However, I think thinking about things in this kind of detail can sometimes give you insights you might not get otherwise.
Last Edited by STME58 on Apr 14, 2016 9:02 AM
thats why i was leaning toward the light coat of oil method. I think with a light coat of oil on a flat piece of glass you could be able to see where it touches and where it doesnt. I think thats good enough. I just want to see the high spots. I dont need to know how high. then once you make it to where the whole plate touches the glass, with no high spots in between slots or any of that, then you flat sand. n boom good enough for me.
Im not trying to figure out how to make them flat. Just an easier more fool proof method of doing so.
For now the sand n check meathod isnt bad. The main problem with it is .... if one of the spots that were previously making contact are now raised ... the sanding method wont show it. I counter this by using different directions to create new scuffs, but its far from fool proof.
----------
Last Edited by Killa_Hertz on Apr 14, 2016 9:16 AM
