Pulse generator for testing flow

[pezman]

I think that Bruce Simpson's DIY Gasflow Modeling Rig is pure genius, but I was interested in coming up with a way to test how parts will respond to a pulse -- things behave differently at hight tempertures, pressures and speeds. I came up with a trick that might be useful.

If you take a Bernz-o-matic MAPP torch, stick it's end in a tube and pull the trigger, you get a pretty sharp pulse. I did a little fiddling around with this and longer tubes seem to generate more powerful pulses with a lot more low-end kick to them. If you do one pulse at a time, even an aluminum or plastic tube will work well.

[Bruno Ogorelec]

This is not at all a bad idea!

[larry cottrill]

This is not at all a bad idea!

I agree completely. I've been convinced for a long time that measurements made on individual pulses could be one of the most effective design tools you could use. For one thing, you should be able to build models out of papier mache, etc. - i.e. flammable materials, or materials that would melt quickly under a continuous run. Even composite structures globbed together with modeling clay would be feasible.

There are a few problems. You would need to verify that the pulses you get are reasonably consistent and duplicable, or measurements would be meaningless. Then, it seems to me we need a cheap, super-simple design for an impulse gauge, so several of them could be mounted at stations in the tube before firing off our pulse. Ditto with pressure and velocity 'gauges', whatever they might consist of.

You'd probably never be able to duplicate what happens in an engine under fully heated conditions - it would be more like a cold start condition. I don't see any simple way around that, in a single pulse mode. Still, it would probably show us some things we don't know now, even with its imperfections. Being able to work with wood, plastic, papier mache, modeling clay, etc. would be a huge advantage for preliminary testing of design ideas!

L Cottrill

[pezman]

Thanks for the flowers!

As far as conditioning the pulses is concerned, the pipe geometry seems to be key. This is just a qualitative judgement based on the sound of the impulse -- maybe not the best way to go based on the fact that hearing is log-normal ;).

You can get a manifold pressure sensor at Pep-boys for about $100, but I think that these top out at 1khz -- probably not fast enough for a structure the size of your average pulse-jet.

Another dea that I was thinking about for flow visualization is a time-delay strobe and cross-polarization (i.e. polarize the flash and put a polarizing filter with different orientation over the lens) -- hopefully the pressure variations will polarize the light such that high-low pressure areas are revealed. With the ability to generate multiple pulses, you could make a series of these images to show how the flow evolves over time.

At any rate, I think that Larry's post that pulse consistency is key is well taken, because it allows you to substitute a series of measurements for a pile of instruments.

[Mike Everman]

pezman, did you run the flame continuously in the pipe? anywhere from 1/5 to 1/4L with the tube vertical will make it resonate. forgive me if the Rijke(sp?) tube is old news!

[pezman]

I did a little playing around with it. If the flame is held a few cm from the pipe, you get a shrill whistle. With the head a few inches into the tube, you can get a pulsing effect, kind of like what you are describing. I was thinking if it as a kind of quick and dirty Gluey.

I didn't experiment with holding the pipe vertically, but now it's on my to-do list.

[pezman]

On a whim I tested it with propane -- works every bit as well.

I think that the secret is the Bernzomatic TS4000 torch head. I plan to build a simple pulse-jet using the TS4000 as a component.

[Tom]

Thats what i was getting with my tube in the off topic (Howler)

Tom

[pezman]

I was talking about the incredible kick that happens when you charge up a tube with about 3 seconds of gas and then hit the spark. It's a two step process:

(1) Fit a 1' long x 1/2" diameter tube to a 3' long x 1" diameter tube, charge it with gas for a few seconds and then pull the trigger all the way to ignite the mixture
(2) Change pants