Mike,

I had a lot of fun, I can tell you! I haven’t made my little window yet, but I can just peer through the pilot inspection hole with the help of an extendible plumbers mirror; it’s about 2” diameter and one of the most amazingly handy things to have. By fiddling it about in that area, I can just see the burner plate, both sides of the pilot. The view is distorted by a small amount of flame escaping from the inspection hole, but it’s good enough to get an idea of what’s going on. I just have to move it out of the way quickly before it melts or cracks with the heat…

I completely concur about the pressure / flow changes around the venturis and stack exit due to movement – but it does seem to happen immediately the car’s moving. Perhaps I should try blowing across the venturis and stack exit with a desk fan to simulate movement. I forgot to mention that on my return home with the car, thinking similar to you, I made up a simple extension to the front lip of the stack exit (about 2”), hoping that this may disrupt the flow and put off the onset of howl. I couldn’t have been more wrong – the howl reappeared with a vengeance, not as bad as with the 0.049” jets, but hideous nevertheless. I think it’s fair to say that flow conditions around these areas greatly influence burner howl.

Re. the superheater / vaporiser - yes, I did mean the superheater – my thinking was that when the car is parked with the burner running, the superheater can get very hot, then, when the driver opens the throttle it is quite rapidly quenched by the relatively cool steam entering it. Maybe this drop in superheater temperature can have a small but significant effect on the acoustic conditions in the combustion chamber. After all, temperature is by far the largest influence on the speed of sound in a gas.

If what we actually have is good old ‘combustion driven thermoacoustic oscillation’ (which has dogged gas turbine designers for decades) then the geometry of the entire intake / combustor / exhaust tract deserves scrutiny. Maybe there are some unintentional ‘tuned lengths’ buried in there that, when coupled with the much higher speed of sound at elevated temperature, provide sufficient positive feedback to sustain howl.

Check out the Rayleigh Criterion with reference to combustion on the internet – and don’t expect to get much sleep after reading about it…

Can’t do much more with the car at present as it goes to the upholsterers soon, but I’ll be back in the ‘driving seat’ just as soon as I can!

Mark

Mark,

Googled for Rayleigh Criterion in combustion - as you say there's a lot of info.

This link [home.earthlink.net] sums it up pretty well.

The various steamcar forums and (going back 40 years Light Steam Power) have been speculating about howling when there was already a scientific explanation well known to combustion engineers. Several point made in the article and in others fit very well with our Stanley experience. Good for you for pointing it out.

To summarise, the combustion chamber, boiler flues and smokehood/exhaust flue form an oscillation chamber just like an organ pipe while the burner flames have a natural tendency to high frequency pulses as the flame front velocity and gas flow interact (and here no doubt the length of the venturi, the fuel flow and vaporisation all have their effect). When the two frequencies more or less coincide the flame oscillations reinforce the natural frequency of the combustion chamber in a sustained resonance.

Your comment about the influence of temperature on the speed of sound - maybe this explains why we get a howl in situations where vaporisation is incomplete, if the jet is too big, or when forcing the burner during firing up from cold, or after a short stop - perhaps the temperature of the fuel/air mix in the venturi is lower than when the fuel is fully vaporised. This would change the speed of sound in the venturi and hence alter the resonance frequency. If the fuel is fully vaporised it is a hot gas mixing with the air to arrive at a combined temperature. If it is hot but still partially in droplet form presumably these are evaporating frantically as they pass through the venturi and the temperature of the mixture may be lower.

Our messing about with a shield at the venturi or deflactors at the flue obviously alter the natural frequencies of the inward and outward ends of the system or possibly by changing the pressure difference slow down the gas flow and change the frequency of the flame oscillation.

My maths is not up to using the formulae in the Seebold paper but it really gives us food for thought.

Mike

I have been around hundreds of operating steam cars in the last 25 years and I have noticed that the small Locomobile has a high pitched squeel. The 10 hp Stanleys have a high C pitched melodious howl, the 23" diameter boilers have a middle C howl and the large 30 hp boilers have a lower note that they howl at. Something like the brass horns in the band. The bigger the horn, the lower the note that it plays. The point I am trying to make here is: The larger the boiler, the lower the note of the howl that it makes. No big discovery, just interesting. All of these tunes that they play can be varried with a change in fuel pressure, jet size, and the temperature of the air. I still think that it is the oscillating flame front over the burner that is the mouth piece for the howl. The size of the boiler helps determine what note that it plays. Please prove me wrong.

Mike, SSsssteamer,

Thanks for the link, there sure is lots of information on the topic.

I think that the observation by SSsssteamer is bang on. After much thinking, I’ve decided on my next course of action (when I get my car back!).

It’s a bit rough and ready, but see what you think…

The fundamental note generated by my 735 is just under 400 Hz (I’ve estimated this using a middle C tuning fork, so I may be well out). That’s somewhere between G and G sharp.

I’ve guesstimated the mean temperature between the flame front and the smokebox top inside surface to be 400°C (I might be well out with this, too).

The speed of sound in air at 400°C is about 520 metres per second. I’m assuming that the speed of sound in the products of combustion is similar to that in air.

The wavelength of this note at this temperature will be around 520/400 = 1.3 metres.

If a pressure pulse is created at the flame front with sufficient energy, is then reflected off a surface half a wavelength away and returns to the flame front just in time to reinforce the next pulse (i.e. in phase) a tone could be generated. This line of thought led me to try to find a flatish surface, half a wavelength away from the flame front – that’s about 650mm. Guess what – on my 735 the smokebox inner surface is about 630 mm away.

As I’ve said before, I’m sure that there’s lots of other mechanisms that are involved with howl, but if I can find one which is the major player then it may be possible to much reduce the effect.

My intention is to install a sound damping material in the smokebox inner surface; my hope is that this material will remove sufficient energy from each pulse to break the feedback loop. I’ll try to obtain some kind of very low density ceramic fibre board – anyone got any ideas?

If anyone has done something like this before, I would really like to hear about it.

Another thought – many people have installed pancake type smokebox feedwater economisers. I would be interested to know if this affected howl in any way, it’s possible that the shape is a relatively poor directional reflector of sound and may reduce howl?

Finally, I know that many of us have witnessed howl appearance or disappearance when closing/opening the smokebox flap – perhaps this is having a greater effect on mean temperature (and thus speed of sound) than one thinks – I can feel another experiment coming on….

I’ve just heard from Don Bourdon that my 30hp burner is nearly ready, going to try to have it airfreighted before Easter, so no prizes for guessing what I’m going to be doing!

Mark

Dear Mark, My guesses at the note values for the different sizes of boilers are not to be taken as accurate note measurements but as an approximate width of sound range between each size of boiler. Have fun with your new burner. I have done the same 30 hp installation in my 1922 735 B and it is a tight fit. To do it over again, I would leave it with a 20 hp tall boiler. Running the 30 hp boiler at full throttle, the condenser cannot keep up and much water is wasted. It will climb a hill very well though.

Mark,

I think your 400C is about right. My smoke hood gas temperature is 280C on full fire and the temperature by the superheater fluctuated between 600 and 1000C when I had a thermocouple in there. The lower temperature was when pootling along with the burner modulating and the higher when climbing a good hill. The thermocouple didln't last long - cooked to a frazzle.

I have a smoke hood pancake economiser which did not change the howll. I think any economiser spaced out enough to let the gas flow would be a lousy reflector of sound.

I had another read of the Seebold paper which I linked yesterday. His story of the of the oil flare burner tells us a lot; it occurred to me that we also have two accoustic resonators in our setup, the combustion chamber with the boiler tubes, smokehood and exhaust flue on the downstream side of the burner, and the venturi upstream of it. The venturi is like the gas supply pipe in the flare burner which comes into Seebold's calculations. I can't get my head around Seebold's maths but from your posts here I suspect you can.

If I have understood it the pulsing flame causes vibrations in the combustion chamber and in the venturi. If the venturi is less in length than one quarter of the wavelength of the vibration in the combustion chamber then it is likely that the standing wave in the venturi will kick back the pulses and sustain the vibration. May be this rather than the top of the smoke hood is where the reflecting is going on. The wavelength is determined by the frequency of the pulses and the speed of sound in the gas in the venturi which changes with temperature and gas density. In my experiment using a double venturi which howled a lot I had obviously shortened the effective length by a critical amount.

I think we may also find an explanation here of why Stanleys developed much more howl when they changed from burning petrol to kerosene. Seebold explains that the flame pulsation is a result of a battle between the velocity of the gas through the burner holes and the speed of the flame. The gas flow is pushing the flame up and the speed of the burn is pulling the flame back down to the burner holes. It flip flops up and down, pulsating down as the flame consumes gas down to the hole and up as more gas emerges. Fuel of a higher specific gravity like kerosene (having more heat in it per pound) flows more slowly than lighter grades like petrol to produce any given heat output. Presumably when Stanleys switched to kerosene this interfered with the balance of gas flow and flame speed and changed the flame pulsation rate. Swapping from the slotted burner to the drilled type, presumably by reason of the smaller dimension of the holes compared with the previous slots, pushed up the gas velocity and recovered the balance. I wonder if the Stanleys knew of Rayleigh's work?

There is rather a lot to think about in Seebold's short paper!

Mike

Another thought. How is it when there are 7000 holes in the burner plate that the pulses do not just overlap and cause "white noise"?

It must be a feed back of pressure into the venturi which merges into a single pulse in the mixing chamber and makes all the holes pulse in a synchronised fashion.

Mike

Mike,

I too have read the Seebold paper – I completely agree that the resonance(s) that we experience could be anywhere in the entire system and not just downstream.. If I had the time I would dearly like to go into the mathematical detail of his paper, but time does not allow at present! Seebolds explanation of flame pulsation with reference to flame velocity is exactly in line with my first thoughts on burner howl.

Like you, I have been turning my thoughts towards the intake setup on a Stanley – I will continue to pursue the smokebox top resonance possibility, mainly because it’s quite an easy experiment to do. When looking at the intake resonance possibilities the speed of sound becomes an important issue again as the intake temperature conditions are so different to the rest of the system – no problem, just needs to be taken into account.

SSsssteamers observation about the tone of the howl being directly related to boiler size is an important pointer and could well blow my ‘smokebox reflector’ resonance idea out of the water! This is because I assume that the note that he hears is more related to boiler diameter – and thus venturi length – than boiler size, mass or height. The height of various Stanley boiler sizes do not necessarily change in proportion with boiler diameter; for example my 20hp boiler is 1” longer (taller) than my 30hp boiler.

The fact that the tone becomes lower as the boiler diameter becomes larger suggests that it is a geometric feature whose dimensions are directly related to boiler diameter that creates favourable conditions for thermoacoustic feedback. The intake venturis are a rather obvious starting point.

On your last comment on the synchronisation of the pulses – my mental model of the combustion process sees the flame as a single, if complicated, flame – the reactants provided by 7000 little holes; and not 7000 individual flames. As the burn rate is increased, if any part of the flame burns slightly faster than another part (due to mixture variation etc) than a small pulse is created. This pulse propagates rapidly across the flame (hot – high speed of sound) creating a secondary pulse. Both the primary and secondary pulses almost merge as they are created in the hottest place, supplying energy to the resonance mechanism – wherever it is – which in turn provides the return acoustic pulse to start the process over again. After just a few cycles it develops into a stable howl. Lower the burn rate and there is insufficient acoustic energy to overcome the systems natural damping and the effect dies away.

Well, that’s how I see it anyway!

If I carry on like this, I’m going to have a mountain of experiments to do when I get my car back…

Mark

Mark,

I've done a bit of experimenting with my car and a few sums relating to howling which are too long to put on the forum. Please can you contact me by email through the forum - go to profile and use the send an email function. I'd like to discuss it with you to see if we can write something useful on howling for the Steam Car magazine.

Thanks

Mike

So is Howling good or bad for efficiency of the burner.

My gut feel is bad, as resonance and possibly turbulence, is usually a bad thing.

Some say "when a burner Howls it is giving the most it possibly can" but I think it has got to the point that it is using more fuel and producing less heat.

Any more thoughts ?

Regards
Ian Vinton