Whistle for Doris

[runner42]

Hi all,

LBSC appears to have overlooked the construction details for a suitable whistle for his 3 1/2" gauge Doris locomotive. I have tried to get details from the Internet and it appears that model locomotive whistles are a compromise, one cannot scale and expect the same note obtained from full size. The main design features are a brass tube closed at one end the other partially closed except for the steam input, with a preformed opening and inside the tube close to the preformed opening is a partition with a gap, this partition is a set distance from the steam input end. See the sketch.

What measurements of a suitable whistle, viz tube diameter and length, size and shape of aperture, gap size and distance from the steam input end?

Thanks in advance,

Brian

[nonort]

Someone on the forum will know better than me but LBSC did design a whistle that has a chamber attached to it. It has a small interconecting hole. I have carried out work on fairground organs that have a simular arrangement. Over to the rest of the forum to supply a drawing?

[GWR 101]

Hi Brian. Both LBSC and Martin Evans have devoted several pages in their books to whistles and show methods of attempting to get a deeper note. I say attempting only because I haven't made one yet and it is something I need to undertake shortly.

It would appear they both advocate a thin wall brass tube (treblet tube) LBSC specifies 7/16" dia where as ME says 3/4", both go for a length of about 4" with a cut out about 5/8" from one end. There is a disc fixed in the bore which is relieved around its od. for a distance, it appears that this relief needs to be a constant gap with the bore and hence not a flat. Both describe the addition of sound boxes to give a deeper note.

Apologies if I am telling you something you already know, I am sure there are contributors who have been down this route and can offer more detail and practical advice. Which I will also follow with interest, regards Paul 

[jma1009]

hi brian,

you need to have a look at LBSC's book 'Shop Shed and Road/'The Live Steam Book'. if you havent access to a copy i can send you the relevant pages for the LMS 'hooter'.

there are some very bad whistles for miniature locos both in published drawings and sold commercially. there are threads a few years ago on here where proportions and details of design have been discussed. it is one of my pet subjects!

cheers,
julian

[jma1009]

also have a look at the following

modeleng.proboards.com/thread/6583/poor-whistle-sound

[steam4ian]

Brian.

One mistake people make is to have an arch shaped hole. A whistle needs fipple which is the lip which the steam strikes as it issues from the steam slot. Have a look at an organ pipe, a special kind of whistle, to see what I mean about the fipple.
One book I have suggests buying a "police" type whistle and modifying the mouth piece so steam can flow into it: I got one years ago from "the Scout Shop".
The book Hayden suggests should be in the club library.

Ian

[ejparrott]

My 2.5" has a modified Scout whistle.

One of the biggest mistakes people make is making the opening where the steam escapes far to large. I saw a piece the other day, think it was actually an LBSC piece in the 1953 volume of ME I was reading, the drawing said 1/32" annulus...this will not work!!!!! Annular gap should be around the 8 thou mark.

By far the easiest whistle for a beginner to make is the bell whistle. Touch more work getting the stalk for the bell in place, but it allows adjustment of the bell up and down to get the sounding right, not the pitch but whether it blows properly - pitch is pre determined by the length of the bell. My Hunslet has a whistle all of 1.25" long, it's very high pitched! The MW I'm building has a 4" bell whistle, it's built and tested on air, but I've not yet finally fixed the bell as I want to sound it on steam to make sure it's correct before I do. The annualr gap on that is about 4thou. Details are probably in my build thread.

[jma1009]

as well as sending Brian the LBSC LMS 'Hooter' drawings and description ive also sent him the following link

www.fonema.se/whistle/hotwhiz.html

this is one of the best concise bits on the net ive found without any irrelevant rubbish and dates back a few years, but after my own experiments.

i have Groves Dictionary where i got most of my basic stuff from re the maths but has to be treated with some caution when transfering to steam whistles. i also knew a few Church organ builders/repairers who tuned organ pipes, and read a few dusty books on the subject.

most commercial plain tube type steam whistles in miniature are made of far too small diameter and too long, and the slot in the languid/disc too big hence 'overblowing' and lack of a stable note for a range of pressures on steam. if you get the proportions of the tube ok and use as big a diameter tube as possible (in proportion to length) and get everything else right a stable clear loud note will be produced. various other refinements add to the harmonic effect and clarity of note and tone. however a miniature 'low' note 'hooter' as fitted to LMS locos isnt easy especially in 3.5"g and resort has to be had to adding a box on the back of the whistle to lower the frequency by upsetting and lengthening the nodes.

cheers,
julian

[runner42]

Thank you all for your replies. It appears that a lot of science has gone into designing a reasonable sounding whistle and probably some experimentation is required to get the sound to what's required.

Brian

[jma1009]

unfortunately very little science (or experimentation) has gone into designing the whistles that appear in print on our published designs for miniature locos! that is why so many people complain about them and the badly made commercial products!

[Jim]

Thank you very much indeed Julian for that very interesting link, it's rekindled my interest in making a new hooter for my son's vintage launch along with a rework of Boadicea's three chime whistle.

Jim 

[jma1009]

hi jim,

very pleased that the link was of interest.

as a rough start for those fed up with whistles that dont sound ok and might overblow ie jump an octave and shriek, check the languid/disk gap. it is in all of martin evans' designs and LBSC shown as 1/32". it ought to be 1/64" and if you can get it less then great. very careful silver soldering required to avoid the gap being filled by capillary action. this is why a greater gap is found on drawings and commercial products as easier to make.

opening of top of mouth on tube as Ian says almost rectangle shape with curve in top corners so not an arch (and definitely not an inverted 'V', and ive had success with an opening without any curves at the top). the wider the 'mouth' is, the louder the sound but see the link quoted above. the languid or disc must be just a few thou above the bottom of the mouth. length of stopped tube can best be judged on air at say 60psi with an adjustable stopped end (exactly as LBSC described years ago to give a clear note). if you want a particular specific note as many of us do, then that's where a bit more work and maths comes into play plus testing on steam. however there are very few people who will say your GWR whistle is a semi tone out!

in all the above 3/8" dia tubes are out and i would start at 1/2" dia. when you get to 3/4" dia you can play around with all sorts of half nodes etc depending on length to get a clear note over a range of pressures with a pitch/note that doesnt get dogs howling or those with hearing aids switching off the batteries!

cheers,
julian

[ejparrott]

Steam and air also create different pitches for the same length whistle, IIRC steam blows a semitone higher than air

[oldnorton]

I have spent some time playing with small steam whistle designs. Initially I followed the making advice of a respected engineer writing in ME a few years back and the result shrieked like a wounded cat - I was not very impressed. After much reading around and some experimentation, I also took the view that people were making them with far too big a slot, but that these whistles would work if you restricted the steam flow from the valve, hence they were being blown at a pressure much lower than boiler pressure. Also, the thin wall brass tube being advocated would saturate with higher harmonics, and these would overwhelm the intrinsic frequency - this is over-blowing.

Even a bad steam whistle will work fine on compressed air at 70 psi, so all testing has to be on steam. Everything Julian says I agree with, and I go further on narrowing the slot. 1/64" (0.015") will work but my best whistles have slots between 0.008" and 0.012", just as EJ says above. They still work at 0.003" to 0.005" but are quieter. I use thick wall brass tubing, 1/2" diameter but 0.050" wall. Others say this causes slow heat up and condensation, but it avoids the higher harmonics. In use, it heats up in about two seconds, during which the whistle tone rises in quite a realistic way. A smaller slot means that there is a higher pressure and steam exits at a higher velocity - a higher velocity means higher volume and this rises according to the velocity square law. Large square apertures work as well as curved top ones for me. My whistles seem to be quite loud.

I agree with Julian that 1/2" might be the smallest reliable diameter. I have had success with a design that starts at 1/2" and then opens up to 3/4" and that produces a lovely deep, clean tone. I have yet to experiment with add-on resonator chambers, to produce the "hooter", and that will be my next bit of playing.

Norm.

[steam4ian]

More to add.

The whistle action relies on the Coanda Effect with the fluid (steam) action on the surface of the fipple (top of slot. The top of the fipple must be narrow but can be wedge shaped if thicker tube is used. The steam jet passing over the fipple stays on the outside edge of the fipple (Coanda Effect) and draws a vacuum inside the bell (tube) which ultimately draws the steam jet inside the bell so it adheres to the inside of the tube (Coanda Effect). Pressure now builds up inside the bell which then forces the stream jet out and so the process is repeated. The rate of repletion is dependent on the length of the bell (tube) and the speed of sound in the fluid; this depends on the fluid density and temperature. To get a deeper note the bell can be lengthened or it can be turned into a Helmholtz resonator by adding a box.

Making a whistle is making an oscillator and the key component is the fipple and its relationship to the steam jet. Look up fluidic logic to get the idea.

Brian I came across a simple whistle design the other day for which I shall try and bring a copy for you on Tuesday night.

Meanwhile I have to do something about the whistle on Sevenplex which is close to hopeless.

Ian

[steamcoal]

Funny that this topic should appear. I was just talking to my neighbour at work last week about this very subject. I think they can help me out as they know a bit about this in their line of work.

www.pipeorgans.co.nz

[runner42]

Thanks Ian,

I have made a rough and ready whistle from a 1/2" dia copper tube about 4" long and cut a rectangular hole 3/8" wide  x 11/32" long placed at 1/2" from the steam input end. The deflector or languid is 1/8" thick with an annular gap of 1/64". The languid is positioned at the front end of the opening. I don't know if the distance between the languid and steam input end is critical since I blew into the pipe with no end cover and again with it fitted with a gap of about 1/10" and the note appeared the same. I have compared the note from this whistle with steam whistle sounds downloaded from the Internet and it doesn't sound too bad, maybe I got lucky with my wet finger guess at what's required. I shall make the LMS Hooter that Julian kindly provided details of and compare the two, maybe this will put the rough and ready version in perspective.

I am struggling to understand what exactly the fipple is, is it a ramp on the trailing edge of the opening? If it is in a thin walled tube how is this engineered? Is the trailing edge depressed into the tube by some small amount?

Brian 

[jma1009]

hi brian,

in simple terms Ian's 'fipple' is what i call the slot in the languid or disc. by having a slot of only 1/64" you will already have overcome the main problem that defalls most miniature steam whistles.

i suggest you enlarge the top opening of the mouth higher say 7/16" from the top of the languid or disc.

if you havent fixed the stopped end yet, try adjusting the stopped end like a 'swanee whistle' to say 1 1/2" above the languid or disc on high and low air pressure and note the difference in note plus it's stability over a pressure range compared to a stopped end length of your original 3 1/2" approx.

something like 22swg copper pipe will be perfectly for the whistle tube.

cheers,
julian

[Kevin]

Brian

I have sent you an email with a PDF of Making Chime Whistles it was to big to post here

[runner42]

Hi Julian,

yes I already fixed the end stop because one thing I noticed that the smallest of air leaks around the end stop greatly degrades the tone and sounds something like me trying to whistle, LOL. Apparently, altering the length of the tube alters the tone as you indicated in the construction of a swanee whistle.

Kevin (thanks for the PDF file) has sent construction details of a 3 chime whistle, which essentially is 3 tubes 19/32" OD of varying length 6.45", 4.78" and 4.03 " providing notes C, F and Ab respectively, serviced by a single manifold. All other construction details are similar to details already provided, except the voice disc (languid) is a two part construction, see sketch.

I assume that all three notes play at once, since there is no mechanism for playing them sequentially or is there some fluidic principle that operates to separate out the notes to produce the chime effect?

Brian

PS I have listened to three chime whistles and obviously had the wrong impression of what a chiming whistle was all about. I thought it was like the ding dong bell chime that we have.