Leakage Testing

[Terry McGee]

Woah, it doesn't take long to come up with a list of agenda items that might well exceed my remaining life expectancy! But treat it like a wish list and see what I've forgotton or what can be ruled out as patently unnecessary.

 

Concertina pressure and airflow tests

Test protocols

Determine standardised test pressure (250mm, 10" water?)

Determine typical playing pressure

Determine minimum operating pressure

Other?

 

External leakage paths

Devise ways to test leakage to outside the concertina through these paths:

• through bellows

• around bellows (i.e. through gasket seals)

• through or around pads

• other paths?

Determine effect on leakage through leather when stretched

Determine typical leakage values for above

Determine threshold leakage values for above.

Devise informal ways to test above, and correlate with formal tests.

Correlate overall leakage rates with players' perceptions.

Other?

 

Gasket leakage

Investigate alternative materials for gasket applications, e.g. chamois, leather, closed cell foam, other?

Relationship of clamping pressure to sealing ability

Compare results of spring clamping and screw clamping to determine typical screw-clamping pressures employed.

Capacity of various materials to accommodate irritants of varying thickness. (Sorry about the pea, Princess, but we had to be sure...)

Determine by how much does roughing up the surface of chamois improve the seal?

Does gluing gasket material down reduce its ability to accommodate movement? Are there "better" glues?

Does doping have anything to offer? Neatsfoot oil, Dubbin, silicone rubber, other?

Consider methodologies for investigating Intra-chamber leakages

Other?

 

Pad leakage

Determine relationship between pad leakage and applied spring pressure

Determine influence of pad plate surface and surface treatments on leakage

Determine relationship between pad leakage and the polarity and magnitude of applied air pressure

Other?

 

Reed characteristics

Determine (for large, medium and small reeds):

• typical operating pressure and airflow

• minimum operating pressure and airflow

• maximum operating pressure and airflow

• reverse airflow (when used without valves)

• starting characteristics vs applied pressures

Other?

 

Valve (leather) characteristics

Determine (for an average size and typical material)

• minimum opening pressure

• minimum closing pressure

• closed airflow at playing and test pressures.

• effect of adding a Mylar spring

• effect of surface treatment to the reed pan timber

• effect on starting characteristics of reeds

• other?

Other?

Are there other areas of the concertina overlooked?

Edited May 1, 2014 by Terry McGee

[Terry McGee]

And Don, you'll be pleased to see that, rather than summarily dismissing your concern that the spring-clamp pressure I'm using might be too light, I've accommodated it by inserting:

 

Relationship of clamping pressure to sealing ability

 

in the gaskets section. Hmmm, which means we can actually draw some conclusions about how much pressure we normally apply with our six screws, so I'll go back and add that too!

[JimLucas]

Now, would I be right in saying this. If a concertina is really leaky, the relatively small difference in leakage one way or the other through the pads is unlikely to affect the overall speed of closing or opening significantly. Whereas if the concertina is really tight, that (hopefully) small difference might be more noticeable. So it's people like Geoff, afflicted as they are by tight concertinas, who might be in a better position to detect a faster hang speed compared to closing speed?

I would suggest also measuring the mechanical factors that affect the tightness of the pads: spring force, and possibly length of lever and relative lengths either side of the fulcrum.

 

In particular, I suspect that far more significant than the permeability of the gasketing material is the tightness of the seal between the reed pan and the bellows frame. I've known instruments where the seal seems "tight" even though the reed pans need almost no pressure to seat them, while others need significant pressure to seat them properly into the bellows frame. Do the former still have residual leakage compared to the latter, or is there a relatively "loose" threshold beyond which there's not much change? We all know what happens when there's a visible gap between the reed pan and the bellows frame, no?

Measuring leakage inside the concertina presents greater challenges than leakages from inside to outside, but we should have a look at it once we've mastered external leakage testing. I don't even know how concertina makers and repairers approach the internal leakages. I'd be inclined to put a bright (but cool!) lamp in behind the reed pan and clamp the reed pan down with a sheet of glass or perspex. Has that any merit, or are there better ways to find out what's going on in the dark?

No merit according to what I'm alluding to. My comparison would be between situations in which no light could possibly get through, but where compression would affect the size of the spaces between the fibers that comprise the leather.

 

A possible exception to my above argument would be the valves, where leakage through their thickness -- especially when multiplied by the number of valves, if many are open for chordal work -- might possibly be significant. But this also raises the spectre of what happens when valves don't fully close/seal. And then there's the fact that a few of the highest reeds (at least in a treble English) normally don't have valves, so the amount of leakage past those reeds has been judged not to affect playing. Just how much is that, and how does it compare to the parameters you're trying to measure?

I think they are matters we can look into and put some numbers against. I've always been intrigued by the reeds with no valves. I guess it's not just a matter of "we can get away without them", but also a matter of "on these tiny reeds, valves would present a problem"?

 

I believe it's been mentioned more than once that one (the?) reason is that valves on such small reeds adversely affect the speed at which they start to sound. Maybe that's another thing to measure?

[Terry McGee]

 

 

I would suggest also measuring the mechanical factors that affect the tightness of the pads: spring force, and possibly length of lever and relative lengths either side of the fulcrum.

 

I think I've covered that one under: Determine relationship between pad leakage and applied spring pressure.

 

 

 

No merit according to what I'm alluding to. My comparison would be between situations in which no light could possibly get through, but where compression would affect the size of the spaces between the fibers that comprise the leather.

 

I guess we'll learn a bit from: Relationship of clamping pressure to sealing ability, and Capacity of various materials to accommodate irritants of varying thickness (in Gaskets). I'll add a reminder to "Consider methodologies for investigating Intra-chamber leakages" so we don't overlook the issue.

 

 

 

I believe it's been mentioned more than once that one (the?) reason is that valves on such small reeds adversely affect the speed at which they start to sound. Maybe that's another thing to measure?

 

I've already got: starting characteristics vs applied pressures under Reed Characteristics. But I'll add "Effect on starting characteristics of reeds" under Valves.

 

Thanks, Jim.

 

Terry

Edited April 29, 2014 by Terry McGee

[Terry McGee]

Hmmm, I've also added:

 

Determine relationship between pad leakage and the polarity and magnitude of applied air pressure

 

which goes to the question of how suck and blow might treat pads differently. It might also reveal whether there is a threshold effect on sucking and a compression effect on blowing.

 

Terry

[JimLucas]

I think I've covered that one under....

Looks like I missed your list while I was composing my post.

 

One more potential factor, though: As the late Rich Morse mentioned more than once, leather has "grain". So permeability may be expected to vary among different planar directions.

 

I suspect the effect of grain direction is too small relative to other factors to matter, but if you're looking for yet another thing to test....

[Terry McGee]

Yes, I can imagine it does, and we could probably test the variability of leakage through it or across it, along and across the grain. But I wonder where in a concertina you could take advantage of any difference. Obviously not in a round pad! Perhaps in a gasket? Or do other considerations rule that out? We'd need to hear from the professional repairers and makers on that one.

 

Terry

[Terry McGee]

While we wait for the 20Litre per minute flowmeter to arrive, I thought we should try a test that can take advantage of the current 0.5LPM full-scale sensitivity. I turned up a "through-leakage" tester in acetal (Delrin):

 

 

The hole in the adapter is 11.3mm diameter, which gives the hole a cross sectional area of 1 square cm. I figured that makes the numbers easier to scale up or down as needed. Note that the backing plate (at right) also has a same size hole through it. We're interested in what flows through the material under test, not what squeezes out around the sides as in our earlier test.

 

Testing the same leather I used earlier, the leakage was almost zero - probably in the order of 0.01 litres per minute (at a pressure of 250mm of water). So, definitely up to the task of being a pad covering. Nice to find something that doesn't leak like a sieve for a change!

 

To put it into a more human context, that's a leakage of two teaspoons of air per minute per pad when squeezing the bellows really hard. And we're measuring what flows through the pad, not around it.

 

Terry

Edited April 30, 2014 by Terry McGee

[Chris Ghent]

One more potential factor, though: As the late Rich Morse mentioned more than once, leather has "grain". So permeability may be expected to vary among different planar directions.

I suspect the effect of grain direction is too small relative to other factors to matter, but if you're looking for yet another thing to test....

Grain might be a shorthand word to explain leather to others, it doesn't have grain as such, it has surface texture and non-linear stretch characteristics. Surface texture seems to be a factor only if the leather is stiff or if it is exaggerated, I doubt you would have much success with Moroccan leather. The stretch characteristics differ in different parts of the hide but the most basic one is a tendency to stretch more in one direction than the other; this must be taken into consideration when selecting leather for various parts of the bellows. This effect is vaguely grain like.

 

Leakage factors would probably be; stiffness of the leather, pores in the skin, and looser or tighter fibres in the underside of the skin.

 

Some time ago I finished a concertina that did not feel tight but I could find no leaks. Isolating the ends from the bellows (take the ends off the bellows, put a piece of printing paper between the ends and the bellows, put the ends back on) showed the issue was in the bellows. I could hear nor feel no wind from holes. Eventually I realised the leather I had used as gussets was slightly porous. It was so slight that no gusset showed it particularly with a soapy water test. However like most things the porosity varied and a couple of tiny could be bubbles led me to try a suck test on the leather. Unsurprisingly the tiny amount of leakage detected (using vaseline on the lips, Terry's gadget sounds better to use) was exacerbated when the leather was stretched.

 

I think we have all learned to accept a small leakage in our instruments as normal, I'm talking about a bellows drop of which you would say, I think it might be dropping, is it? Oh yes it is, very very slowly. I have played an instrument by a great maker which was so airtight for the first few years that it made all movements rigid, as soon as you took your finger from the button the bellows stopped dead. It was very disconcerting and exacting to play, but it was probably good for you, forcing you to be accurate. It was not lightly sprung.

[Terry McGee]

Very interesting Chris. So I've added "Determine effect on leakage through leather when stretched". Might be a harder one to test, but very relevant to bellows tightness as you've explained.

 

Terry

[Terry McGee]

Now, our through-leakage test fixture mentioned above permits us to start looking at the question: Determine relationship between pad leakage and applied spring pressure. Although there is one little element missing - I don't have any pads to test! While I look into remedying that deficiency, I thought I should at least do a system test to see if it looks like it would work. And the results are quite encouraging:

 

 

As we go from left to right across the graph, we are increasing the pressure on the pads from zero to 150 grammes To put that in context, 150 grammes is a pretty heavy action. As we'd expect, the amount of leakage through the pads drops as the pressure is increased, linearly at first, but we reach a point where a substantial increase in pressure is needed to reduce the leakage any more. I'd suggest the turning point comes around 60gms, which is a figure that has come up before. It's not heavy, so if it keeps the leakage under control, it would be a good value.

 

I should prick this happy bubble by reminding us that, having no concertina pads, I used a sax pad for the tests. I don't yet know if there is any substantial difference between the two. Sax pads have the same card backing, felt interior and leather cover, but the leather is stretched over the backing, not just a simple sandwich.

 

The second caveat is that the seat is nicely faced off acetal (Delrin), which should be smoother and less porous than the usual wood seat. A subsequent test needs to investigate how less good wood is than acetal.

 

Speaking of which - what timbers are typically used for action boards?

 

Terry

Edited May 1, 2014 by Terry McGee

[Terry McGee]

Some of the items I recently ordered for my pneumatics lab arrived, wrapped in that fine translucent closed-cell foam often used in packaging these days. Oooh, I thought, remembering Geoff's notion that closed-cell foam might make a modern replacement for chamois, I wonder what that stuff is like? So I sandwiched a sheet in my concertina-end sized leakage apparatus and applied the Magnehelic. At first nothing happened - the Magnehelic read as if open to air. Then the flow dropped, smoothly, over a period of some seconds, settling down to around 1 on the scale of 10. When I turned the Magnehelic off, the flow meter reversed, implying that air was now coming out of the test rig and back into the tester. Eh what?

 

A few seconds thought, and I realised I'd invented the airflow capacitor. You'll remember when a capacitor is connected across a battery, current flows into the capacitor until it is charged, whereupon flow stops. When you remove the battery and short the capacitor terminals, there is a spark, as the energy stored in the capacitor is suddenly dumped. The sheet of closed-cell foam acted exactly the same way. When pressure from the Magnehelic is first applied, air flows into the test chamber, building up pressure there. That pressure compresses the little bubbles that make up the sheet of foam, a process that takes some seconds. Once the little bubbles have all reached the new pressure, airflow into the chamber stops. When I turn off the Magnehelic pump, the bubbles expand back to their normal rest state, pushing air back down the line and backpinning the flow meter.

 

Now this is unlikely to prove an issue if the stuff was cut to form a gasket, clamped between wood around the concertina's edge, as not many bubbles would have access to the air pressure being applied. (I'll certainly test that statement!) But it does alert us to an interesting issue - are there any materials or constructions we do use that have a similar effect? So far, I have been only considering static issues, but particularly Anglo players should be concerned about dynamic effects too. We don't want, for instance, anything to be slowing down the effect of changing bellows direction.

 

Fascinating stuff, when you look into it. Bit by bit my items are coming in, and I think I've devised a way of putting them together to form a comprehensive and flexible test rig. We'll soon see....

 

Terry

[Terry McGee]

Now that I've got the pneumatics lab together, I can freely rearrange the components from the old flute leakage detector into the order I'd like them, rather than having to put up with the order they came in. There was always something niggling at me about pad leakage test I'd run earlier. I would have expected to see a sharper knee in the curve, above which the pad seals fairly well, and below which the pad seals badly. Rearranging the components so that I can monitor both pressure and flow at the same time, we get a different picture (in red) for the same pad being tested:

 

 

So, unless there is at least 45 gms force on the pad, it's going to leak. Somewhere around 60 gms looks good.

 

That's at a pressure of 254mm of water, which is higher than we're likely to encounter in real life. I'll probably revise that downward once I have a good think about it.

 

My test pad hole is 11.3mm diameter, which has an area of 1 square cm. 254mm WC is equivalent to a 25gm force on a 1 square cm item, so it's obvious that if only 25gms of force were applied to the pad, the air pressure would be enough to blow it off the hole. We can probably assume that the remaining force is needed to flatten out the pad surface to render it airtight.

 

Terry

[Ann-p]

Terry, just catching up on your recent activities! With regard to materials for gaskets and/or pads I have used thin compressed closed cell foam as a temporary pad for my tenor recorder until I could get the proper pad. This was from the screw on metal caps on wine bottles, dig it out with a thin, sharp implement and cut to pad size. I fixed it on with some blu-tack as it was only temporary, main problem with it is getting any glue to stick, so not a real long term solution to pads replacement. However, if one could get hold of this stuff in sheets to cut to shape it might be possible to use that for the gaskets. It seems to me that it ought to be pretty airtight as it's a polythene type plastic.

 

Ann

[Terry McGee]

Ha ha, that has to rate highly in the annals of desperate interim repairs! We flute-makers used to claim we drank so much wine only because we needed the corks for flute head stoppers. But then they changed to screw caps and blew our little scam out of the water. But now you've given us a whole new future! Six-key flute = six bottles of red! That's a 600% improvement on the one-cork-per-flute consumption rate!

 

Hmmm, wait a minute, how many pads are there in a concertina?

 

You may have also stolen some of my thunder. While playing with the new system earlier today, my eye fell on the whitish closed-cell foam sheet that had come wrapped around many of my pneumatic lab components. I think it's called cushion wrap. I punched a disc of it and tried it as a pad, and it seemed plausible. I had previously found that a full concertina-sized sheet of the stuff did seem to demonstrate memory, but whether that's a problem or just a scientific oddity remains to be determined. Anyway, more to come on that.

 

This actually raises a question. I appear ready to be able to test materials as potential pads. Obviously I'll try the regular traditional approach - the sandwich of card, felt and leather. I'll also cast around the workshop for other materials to try, such as the cushion wrap. Might even crack a bottle of wine (hmmm, possibly two - one as a "control") and try its cap gasket, if I can winkle it out alive. But what else should I try?

 

Green hats on, folks.

 

Terry

[Terry McGee]

Hmmm, just to add what I already have in mind:

• traditional clarinet pad (fine leather wrapped around felt on card)
• synthetic clarinet pad (Valentino)
• leather disc on card
• leather disc on felt on card
• cushion wrap disc on card
• neoprene rubber disk on card
• adhesive-backed neoprene foam on card
• wine bottle seal on card

Any other suggestions?

 

Terry

[alex_holden]

With the traditional leather/felt/card composite pads, I'd be interested in experimenting with different types and thicknesses of leather and felt, and different glues.

 

I bought some commercially made pads and the glue used to hold the layers together was very weak - several of them fell apart before or during installation. I tried gluing them back together with hide glue and found that although it stuck well, it soaked into the felt and made it very stiff, which caused those pads to not seal properly.

[alex_holden]

I've also read that new traditional pads have a 'bedding in' period (a few days?) over which the seal gradually improves as the leather and felt conform to the shape of the hole.