Whistles (again)

[doubletop]

I'm struggling to get a whistle to work and have tried the LBSC instructions in "Shop Shed and Road" and read through all the "whistle" threads on this forum. Amongst other things Julian's post of 2011 was useful in correcting the gap in the languid from LBSC's 1/32" to ~6thou.

Sep 7, 2011 at 9:21am
Quote
like
Post Options
Post by jma1009 on Sep 7, 2011 at 9:21am
i have spent some time investigating miniature steam loco whistles, and wrote a paper on the subject some years ago.

firstly, the steam line and size of valve is as important as with injectors. the 'speed' of the steam as it passes the 'languid' or disc is critical. some whistle valves have ridiculously small passages, and some pipes to the whistle are ridiculously small, and the pipe should be as short as possible with a downward fall in all places.

your whistle with the mouth facing downwards is correct. the whistle should have a slight upwards tilt so no water collects at the 'stopped' end.

if the above is all ok, then the problem is with the whistle itself. this is quite a lengthy subject, but in short the whistle mouth may not be wide enough, and too long. try the whistle on compressed air and slide a piece of card wrapped round the whistle from the stopped end towards the disc slowly reducing the height of the mouth. the languid or disc should be level with the bottom opening of the mouth if not a few thou above. a small error here in silver soldering the disc is critical, but can be easily remedied by filing the bottom edge of the mouth if the disc is too low. if the disc is too high scrap the whistle. there is a relationship between the diameter of the whistle tube and its length with a given operating pressure (or more correctly the speed of the steam). the width of the gap between whistle tube and the cut out of the disc also has an effect (i have found about 5 thou to be ok for our miniatures).

LBSC's 'Shop Shed and Road' has a chapter on whistles and how to 'sound' them, and whilst LBSC doesnt go into the theory, he is spot on at how to achieve the correct result!

btw if the mouth is too high (avoid a 'V' slot, and a 'square' slot is quite ok), soft solder a piece of shim around the tube to reduce the height of the mouth.

2 days ago i altered a bell mouth whistle which is for a 7 1/4"g loco and was a commercial item. the whistle produced a sound like you have described...awful! i raised the height of the disc and lowered the 'bell' (easy on this type of whistle) and the whistle is now loud and clear!!

getting the 'right' note for a certain prototype is another matter entirely!

hope this is of some help and happy experimenting!

I've now got 6 whistles on my bench, including a commercial 3 chime and they are all hopeless. The whistle is for my 7.25" Dart. The only location close to the whistle valve and with a fall in the pipework is under the footplate so I'm limited to 6.0625" long including the steam inlet. I'm using a 1" copper tube in anticipation of getting a better tone.

From Julian's post I have the languid dimensions and location wrt bottom edge of the mouth and the details of mouth shape (slot or curve but not V) The information I seem to be missing is the relationship of the mouth dimensions to tube size and working pressure. Julian's note above indicates that a common problem is the opening could be too long, but how long is too long? I'm loathe to keep filing away at the slot in the hope that there will be an improvement, at some point too much metal will have been removed and whistle #7 will be required.

Any clues please?

Pete

[steamer5]

Hi Pete,
I've got an article by Bob Banson on whistles, cant remember the ins & outs, i'll check it out & see if its got what you are after. Got a wedding this arvo so wont be until tomorrow.

Cheers Kerrin

[doubletop]

Thanks Kerrin

[doubletop]

I think I may well have answered my own question or at least gone some way to answering it. I registered on the Yahoo Steam Whistles forum Yahoo Steam Whistles. There is a lot of information there, but most of it relates to bell type of whistle and I couldn't see how that related to our tubular whistles (I was wrong). One useful piece of information was the spreadsheet "Slit width vs Pressure.xls". It confirms the information that the slit in the languid should be in the order of 15 to 30 thou for our pressures of 60-100psi. However, the size of the slit also is dependent with the ratio of the whistle length to its diameter (Scale). The spreadsheet only covers scale ratios of 1:2 to 1:4 but it can be deduced that if we were to use a tube of 1/2" dia and (say) 6" long a scale of 12 then the slit needs to be less than 15thou. Which concurs with the statements made on the other whistle threads here that widths of 6-20thou have been used successfully.


That of course didn't address my immediate question of how to determine the slot size. I then returned to the informative page on basic whistles Basic Whistles. Again this page deals with bell type whistles, but realising that chime bell whistles were just a cluster of tubular whistles the information should be relevant to my question. The penny dropped when I realised that the area of the mouth of the whistle should be equal to the cross-sectional area of the tube. To test this, I created this



To confirm I measured the mouth of a working whistle and it was there or thereabouts. I then took one of my failed whistles and opened out the mouth to be the same area as the cross-sectional area of the tube and bingo! it worked albeit at a lower pressure, but the languid slot width was known to be incorrect at 32thou. I then did the same to whistle #6 (languid slot 8thou) and that worked at the higher pressure. I think what had put me off extending the mouth was Julian’s suggestion that failures could be due to mouth being too long (high) and to try reducing them with some shim.

That then brings me to my next question on Julian’s point on the size of the steam feed plumbing. As I had been having problems in ‘overblowing' my whistles so I had opted for smaller bore pipes and even made a pressure relief valve in an attempt to get the pressure down. I think what Julian is saying is once you design a whistle for say 80psi with the correct languid slot width you need to ensure you give it 80psi for it to work optimally, otherwise you are throttling the thing. Maybe it’s a trade off, if you need to use narrow bore pipes just increase the slot width a tad.

Hopefully I’m on the right track here and I’m looking for feedback please

Pete

P.S. and thinking further, the possible reason V mouths don't work is the mouth height would need to be twice that for a conventional mouth to get the area the right size?

[houstonceng]

Some time back, I collated information on whistles from ME and rage SMEE Journal and wrote an article for my MES magazine. I then made two whistles for the club locos and found that both irked on the Simlex but not so well on the Maid of Kent. Increasing the size of the pipe on the Maid got its whistle working as per the Simplex. Basically, each was made from 15mm Cu water pipe with the slot formed by the languid the same width as the wall thickness of the pipe.

A fellow MES member, wanted to fit a longer whistle on his loco - longer gives a lower note - but couldn’t find anywhere to fit it. We discussed musical instruments - he used to play a trumpet semi-professionally - and decided that the tube didn’t need to be straight. So he made a whistle bent double using a couple of end feed fittings and it works a treat. The only thing is, it has a low note that sounds LMS ish rather than LNER ish which diesn’t really match the Enterprise that it’s fitted to.

[doubletop]

Apr 1, 2018 10:08:41 GMT houstonceng said:

...............

A fellow MES member, wanted to fit a longer whistle on his loco - longer gives a lower note - but couldn’t find anywhere to fit it. We discussed musical instruments - he used to play a trumpet semi-professionally - and decided that the tube didn’t need to be straight. So he made a whistle bent double using a couple of end feed fittings and it works a treat. The only thing is, it has a low note that sounds LMS ish rather than LNER ish which diesn’t really match the Enterprise that it’s fitted to.

I'd had a go at the LBSC whistle with the Helmholtz resonator but that didn't go too well either. Folding the tube using standard plunbing fittings did cross my mind. As length only determines frequency I decided to park the idea until I got a whistle to resonate properly at the design frequency rather that squeal at a harmonic. (I suppose you could fold the tube at three times the design length and you'd get the frequency you required when it squealed at the third harmonic... )

Pete

[oldnorton]

I spent some time time playing with whistles and can now make new ones that work; but it is just as easy to make one that doesn't :-)

I disagree with the writings of many in the magazines who say to use a thin walled tube since that resonates and produces the colourful tones. Well that's fine if you are a flute manufacturer and its going to be blown by an expert. But I do not think all the additional, high frequencies help when we have small devices to start with. Also, as soon as you over blow the high additional frequencies (harmonics) kick in and it shrieks.

So I use very thick wall 1/2" tube, on one occasion I drilled some 1/2" bar. This heavy whistle now needs to be blown at high energy to produce sound so the slit width needs to be small to keep the pressure up and the steam velocity high. Mine start to work with a 0.006" slot but get louder at 0.008" and work up to around 0.012". I mill the slit onto the edge of a cylindrical insert using a rotary table. This insert is then put into the tube and aligned with the opening. I don't think the opening is all that critical if the basics of the steam feed out of the slot and velocity are right.

I use the usual 1/8" feed pipe, but the whistle valve needs to open properly or it will sound strangled. If the valve leaks a little bit it helps keep the whistle hot. If it gets cold it is no problem and the sound picks up over a second or two in a realistic big whistle fashion. They must drain back to the slot. The pick up delay is the condensing water coming out.

I have tried a tube that starts at 1/2" diameter for 3" length, and then the last 3" length is 3/4" diameter. The connection has to be an eccentric to ensure it drains when horizontal. Small whistles discharging into bigger chambers is a known method for dropping the frequency of resonance. This produces a lovely tone. I am going to experiment with expansion into much larger, up to 3" diameter chambers, and see what types of sounds I can get.

Norm

[doubletop]

Apr 1, 2018 11:01:10 GMT oldnorton said:

............... I mill the slit onto the edge of a cylindrical insert using a rotary table. This insert is then put into the tube and aligned with the opening. I don't think the opening is all that critical if the basics of the steam feed out of the slot and velocity are right.



Norm

Norm

Thats what I have been doing and means you know what slit width you are working with. One of the threads on here asked the question of soft vs silver solder and the general view was to use silver solder. Faced with the possibility of the silver solder flashing over the very norrow slit I used a graphite pencil and rubbed ot over slit prior to soldering. It worked a treat.

I've been re-reading the "basic whistle" web page and the more times I read it things make sense and actualy answer the questions I've been asking. On the subject of valve and pipe size.

2.5 Inlet and valve

When the blowing medium passes a constriction its pressure is reduced in proportion to its linear speed squared, as described by the Bernoulli equation above. One issue may be that the medium should find no essential constriction on its way to the flue, because that will reduce the power available to the whistle in relation to what is supplied by the blower. If we assume that the passage has a single constriction area A while the flue has area B (slit width times circumference), then the fraction of pressure lost can be formulated as B2/(A2+B2). If A=B, then the loss is 50%, but increasing the constriction area A e.g. three times will reduce the loss into only 10%. One can adopt the latter as a rule of thumb: any constriction in the feeding path, like inlet or valve diameter, should have a cross sectional area that is at least three times larger than the flue area

Presumably, with steam driven train whistles and the like, the foot pressure is substantially lower than the boiler supply pressure, due to the throttling action of a narrow valve passage. As a counter example using the formula above, if valve area A is one third of flue area B, then foot pressure is only 10% of the supply pressure

Pete

[oldnorton]

Hi Pete

Yep, you and I agree on how to make the slit.

A couple of years ago I was a bit precious about my design and had a mind to offer an article to a magazine. But I'm more than happy to share it with friends here.

I gave up soft soldering and I use a close fitting insert with a small screw to locate and hold. A tiny wisp of steam leak really does not matter. The screw together design means it is so easy to swap inserts with different slots, and swap bodies, and test the results on a steam boiler (no point testing on air).



That's all I have to show as I 'lent' all my other whistle parts to a dear friend a couple of years ago and I have yet to see them again.

The 'nice tone' larger chamber has the same construction design, but I got the dimensions wrong in my last post as the 1/2" bit is less than 2" long. Here's an image of it.



Norm

[oldnorton]

Pete,

Forgot to agree with what I think you are saying about pressure being important.

To my mind if you want sound volume from a heavy wall whistle you need more energy from the steam, which means as much velocity as possible since energy is the product of velocity squared. This means a smaller slit.

Most all the magazine articles I have seen advocate thin wall whistles to resonate and use large slits of 0.020" or more. It's all wrong. Those large slits reduce the pressure and lessen the energy put into resonating thin bodies, in an attempt to stop them shrieking. Sometimes they work, but most of the time they sound like someone trod on the cat.

Norm

[jma1009]

What a most interesting post by Pete, and I was surprised to see me being quoted from some 7 years ago.

I have yet to make Stepney's whistle. Of course, it has to be near enough the correct note and octave for the Stroudley whistle.

If Norm has got good results from his methods from 'heavy' whistles of tubular type then I applaud him. Treblet brass tube is not easy to silver solder, especially in these non-cadmium Easyflo 2 days.

The harmonics side is interesting.

When I made Stepney's whistle valve, I was very concerned to make the valve as compact as possible. The steam inlet in effect goes into the inside of the spring from the manifold. If the spring is too compressed, then the steam is restricted and strangulated before it even gets to the valve itself.

Then consider the standard LBSC/Martin Evans ball seat and push rod for the ball. The push rod takes up most of the area of the passageway underneath the ball seat. So you do a bit of jiggery pokery, and fit a larger ball, larger passageway, and smaller push rod.

The 'slit'/ opening in the disc or languid is crucial. A 1/32" slot is simply far too big as per Martin Evans and LBSC in their drawings. But if you aim for a 1/64th slot width or less you seriously risk capillary action filling up same when silver soldering same. If you have a 1/32" slot width you will get overblowing and the whistle jumping an octave or two and shrieking.

Commercial whistles are made for ease of manufacture without any considerations of the correct sizing for optimum performance.

A most interesting thread by Pete with very interesting contributions by Norm.

Cheers,

Julian

[steamer5]

Hi Pete,
Dig out the article, it’s by Bob Bramson, ( not Banson as per my last post).
Sorry to say but he is at odds with Norm on tube thickness. By using thin section tube you avoid condensation of the steam & the velocity energy isn’t dissipated instantly. He also recommends using thin wall with comparatively large bore tube to feed the whistle, lagged pipe work can also be an advantageHe also talks about slot length being dependent on the pressure used, which then sets the diameter of the tube. The whistle to be mounted such that any condensate that forms can drain out of the voice slot. Mount the whistle with the minimum of support, prevents condensation & to can prevent resonating
He recommends tube of 0.75” with a wall thickness of 0.020”, slot length 0.75”, the slot gap equal to the tube thickness

Hope this is of some help, but it looks like you are well on the way of solving the issue!

Cheers Kerrin

[jma1009]

Hi Kerrin,

Bob Bramson referring to the disc/lanquid depth of slot being the thickness of the tube is surely a nonsense!

So you make 3 otherwise identical whistles for the same boiler pressure. One is made out of thin brass treblet tube, one is made out of commercial central heating boiler type copper tube, and one is machined out of solid as per 'Norm' with thick walls.

According to Bob, they will all have a considerable difference in the depth of disc/lanquid slot! But in each case the whistle is the same length, diameter, and boiler pressure.

What I presume Bob meant was to make the disc/languid depth of slot the same as thin brass treblet tube - which makes more sense - but is still too wide a depth of slot to prevent overblowing and jumping an octave and shrieking in 3.5 and 5"gauge locos.

In respect of the above, a small diameter long whistle tube is far more prone to overblowing and jumping an octave and shrieking, so it is quite a complex subject.

Incidentally, organ pipes work at something like 2 to 4 psi. They are made of lead or wood in most applications. The complex analysis of same has no application to miniature steam whistles.

Cheers,

Julian

[builder01]

Actually, organ pipes operate at pressures even lower than what you can measure in PSI. Organ pipes pressures are typically measured in millimeters or inches of water pressure. 1 PSI is about 25 inches of water pressure. Typical organ pipes operate on about 4 inches of water pressure. Fairground organs at about 8 to 10 inches. If a pipe is over blowing, you probably need to raise the cut up of the upper lip. Making the slot, or wind way smaller will work also, as a smaller wind way will "throttle" the steam, or, air. A higher cut up will allow a pipe to operate at a high pressure even with a rather large wind way.  Here's a photo of a few organ pipes I have made. These play on 7-1/2" water gauge air pressure. Many of the wind ways in the flue pipes are in the range of .025" to .030". The cut ups are very low to match the pressure. If the cut up is raised, these could play on much higher pressure, but, in this case, is not necessary. 

Steam whistles are organ pipes in every way. They can be tuned and voiced using the same methods that organ pipes use.

01849Adjusted40-40 by Builder16, on Flickr

[doubletop]

Norm
I do like your integrated feed/languid as it deals with a whole lot of issues and allows re-use. Thanks for sharing

Julian

Thanks for the feedback. It was your earlier posts that helped me on my way. Your info, the "basic whistle" web page and the 'Steam Whistle" Yahoo group enabled me to improve my understanding of this subject. I do like to know what is happening rather than get a result that you can't explain. I've always been of the view that you learn nothing if things work first time. In a way failures are a benefit rather than a hinderance as you gain more knowledge.

A bit more to the mix. I've just discovered that one of the files I downloaded from the Yahoo site was a spreadsheet "_Whistles Scale vs Pressure.xls" this is probably the most useful spreadsheet for us as it allows the languid slit width to be determined given the scale (length/diameter) = and the working pressure.

I've entered the parameters for a whistle with a typical scale of 12 (6" long with 1/2" tube) and pressure of 100psi and also a slit width of 0.008"




You'll see that such a whistle with slit of 1/16" would work at 0.9psi, a slit of 0.015" at 16psi. For 100psi the slit would need to be 0.006" and if the slit was 0.008" the operating pressure would be 57psi.

So if my assumptions on mouth dimensions are correct, armed with those two spreadsheets, and, as Norm and Julian have said, the material is actually immaterial there should be no excuses for any further whistle failures…..

(no doubt now that’s going to illicit some feedback)


Pete

BTW - I believe that the line in the spreadsheet "cutup=unity" means that the area of the mouth equals the cross sectional area of the tube you are using. e.g my earlier spreadsheet. (experts please confirm?) If so it does mean that we now have a direct relationship between length, diameter, pressure, slot width and mouth area. The whistle note is a function of length, the diameter is whatever you want it to be and the rest just falls out of those choices.

[doubletop]

Just found this in the "definitions" sheet from the Yahoo group

Unity-The point where the area of the radiating mouth is equal to the cross-sectional area of the bell

Actually you may as well have this

Definitions

Bell-The body of a whistle

Bowl-The part of the whistle below the bell that is comprised of the slit formed between its inner edge and the outer edge of the languid plate, as well as the inlet

Chime-The term given to a whistle or whistles designed to sound a chord

Cutup-The mouth height or distance between the bell and bowl

Dynawhistle-An inverted, horn loaded (Type 2) toroidal whistle

Fundamental-The lowest primary frequency at which a whistle is designed to blow

Harmonic-A multiple (usually an odd number) of a whistle’s fundamental frequency (Harmonics usually occur along with the fundamental, but in the case of overblowing, the fundamental itself can be weak or absent)

Hooter-A plain bell whistle designed to sound a single note

Inlet-The hole at the bottom of a whistle where air or steam enters

Languid Plate-The plate that together with the bowl forms the slit (a.k.a. spreader plate)

Mouth-The open radiating area of a whistle between the bowl and bell

Operating Pressure-The pressure at which a whistle is designed to operate (it is influenced by the factors of scale, cutup and slit width)

Organ Pipe Whistle-A whistle design that uses a mouth of substantially less than 360 degrees (for the purpose of conserving air or steam without reducing the output, as the cutup is made proportionally higher)

Output-The acoustical power (usually expressed in decibels at a specified distance)

Overblow Point-The pressure at which a whistle loses all or most of its fundamental to (usually) the third harmonic

Plain Bell Whistle-A whistle designed to sound a single note (a.k.a. hooter)

Scale-The ratio of a whistle’s working length to its diameter (the smaller the number, the larger the scale)

Slit-The thin opening between the languid plate and bowl for the passage of air or steam (a.k.a. slot)

Toroidal Whistle-A whistle designed to have about 30 times the radiating area (and output) of conventional cylindrical whistles of 3:1 scale

Ultrawhistle-A plain (Type 1) toroidal whistle

Unity-The point where the area of the radiating mouth is equal to the cross-sectional area of the bell

Working Length-The length from the inner top of the bell to the languid plate, which includes a cutup of unity

[oldnorton]

I am finding this whistle re-visit interesting as I put it all away a couple of years ago and it is nice to dig out again. I have strung together some sound recordings at the end of this posting. They were made on an i-phone at the time and it lets us hear what we are talking about.

Yes I am at odds with Bob Bramsom. I started making the thin wall whistle but found them unreliable and unpredictable. It was the creation of unwanted higher harmonics that was spoiling them. My desire for investigation had me trying thick wall tubes and then by experimentation I settled on slits of around 0.008". I did like to see your finding Pete that 0.008" is the theoretical size for a 6" long by 1/2" diameter whistle, thank you. But, I maintain that if you put all that energy into a thin wall whistle it will over resonate.

With the thick wall tube it then progressed to the single piece design of push fit and screw inset and languid which is so much easier to make, and no soldering.

Bramsom is right that the problem with a thick wall is that it takes more energy and time to heat up. But I find a leaking valve helps or just give the whistle a two second gentle blow to warm before the full push. My thick wall whistles do sound best at the full 90 psi, at 70 psi they are quieter, and by 40 psi it is just a polite noise to tell the station you have arrived!

Even though they are thick wall they still resonate through the body so do need suspending with a light, spring bracket screwed in at the far end, and supported by the inlet pipe at the other.

In the video following there are four whistles being tested. I am sorry that I have not spent time editing and making it look posh - it was a string together and upload jobbie and you will have to tolerate my mumbling which, at the time, was my labelling of each test. I have picked out four from about two dozen tests:

1. Whistle 1 is my manufacture of the thin wall design, soldered together, similar to those on sale commercially. This one sort of works, but was the final after a few failures.

2. Whistle 2 is a friend's reject commercial whistle, but sounds just like many that are in use.

3. Whistle 3 is my final, thick wall, single piece insert, 0.008" slot design. You can hear that it takes a couple of seconds to clear, and that the basic pitch is the same as the similar length thin wall whistle.

4. Whistle 4 is my large chamber extension, fitted to the same 0.008" languid insert. The tone is completely different and, to my ears, much better than the standard 1/2" diameter throughout.

Norm

[doubletop]

I'm continuing to read through the Steam Whistle Yahoo Group. It was set up by Richard Weisenberger and all the spreadsheets are his work based on this document

Flue Pipe Acoustics (this has audio samples)

or as a document

Flue Pipe Acoustics

Its worth a read and provides the credability to the documents on the Yahoo Group

I'm also now realising that although the area of the mouth is to be equal to the cross sectional area of the pipe the key is the height of the mouth (cutup) and not my basic mouth width = 75% of pipe diameter. (I think LBSC could have crept in here as he may have proposed a mouth width of .375" on a 1/2" pipe). Somewhere along the line the documentation proposes that the cutup should be 1/12 pipe length (L/12). I'm trying to establish how Weisenberger got to L/12 and have unlocked the spreadsheets to see where that came from, but it is just used "as is".

Pete

[steamer5]

Hi Norm,
That's just dandy!!
This is quickly turning into yet another great thread!
Think I'm a convert to thicker whistle after listening to these! Like Pete I tried to find info etc. My Dad wanted a 3 or 4 chime whistle for his loco.....failed miserably. How ever while searching for something else I came across the info from Bob, we swooped a couple of emails, sounds like he used to make whistle's for sale, but no longer & still has / or had tubing, his info made sense at the time, & hence my reply to Pete.

Julians reply re using heaver tube & hence using Bob's idea of making the slot gab the same as the wall thickness I think was incorrect as Bob only used thin wall tube hence 0.020 slot.

I can see a few changes to my loco coming up! The old gent who started me in this hobby commented that you have to have a whistle valve that leaks a bit....it keeps the whistle pipework warm, but note that all 4 of yours blow water at first & then is clear with a lovely sound. I do really like the large chamber one!

Cheers Kerrin

[mr swarf]

I'm curious to know how you define thick & thin wall tube.
Treblet tube may have been available in LBSC's day but I cannot any reference to it being available to buy today.
I want to make a 1/2" diameter whistle, but a quick check of suppliers show that only 18g or 20g tube is available in that size.