Rotary Valve Pulse-Jets

Moderator: Mike Everman

SR71Fan
Posts: 22
Joined: Sat Oct 17, 2009 1:32 am
Antipspambot question: 125

Re: Rotary Valve Pulse-Jets

Post by SR71Fan » Mon Nov 02, 2009 3:40 pm

Mark:

I'm still in the "visionary" stage, so nothing is built to take photos of.
But simply imagine a round butterfly valve on a car throttle body,
except the butterfly can rotate through 360 degrees instead of having
a "stop" when closed. The butterfly plate would also be thicker and
heavier, to withstand the pulse pressure. The butterfly is not intended
to "perfectly" seal, as a small percentage of blow-by is acceptable.

Because of rotation rates and other limits, the shortest PJ engine
I would envision is about 5 feet long, resulting in a pulse rate of
about 100 Hz, and a valve rotation rate of 50 revs/sec, or 3000 RPM.

ST71

Mark
Posts: 10931
Joined: Sat Oct 11, 2003 10:14 pm

Re: Rotary Valve Pulse-Jets

Post by Mark » Sun Nov 29, 2009 5:36 pm

Happened to see this thought on rotary valves as I was looking through some papers.
Attachments
Gosslalu's scheme.JPG

tufty
Posts: 887
Joined: Wed Dec 24, 2003 12:12 pm
Antipspambot question: 0
Location: France
Contact:

Re: Rotary Valve Pulse-Jets

Post by tufty » Sun Nov 29, 2009 8:53 pm

I know you were thinking of using HC08 µCs, but you might want to consider something a bit more "beefy" - I've been considering getting one of these for playing with, looks like it might have a certain appeal, particularly WRT the number of analogue inputs (6) that you can use at once. Plus it's a 100Mhz, 32 bit ARM machine. I like ARM.

They are pretty cheap, $40 or so for the DIP40 mounted devkit; the only downside is that the toolkit is web-based, although the libraries can apparently be cajoled into working with GCC offline.

Anyway, thought you might be interested.

Viv
Posts: 2158
Joined: Sat Oct 04, 2003 2:35 pm
Antipspambot question: 125
Location: Normandy, France, Wales, Europe
Contact:

Re: Rotary Valve Pulse-Jets

Post by Viv » Sun Nov 29, 2009 9:22 pm

Hi Simon

I remember when the HC08 and its ilk first came out, we were pretty excited to get our hands on the first parts as I remember, I do however like that new part you linked to and its performance

Viv
"Sometimes the lies you tell are less frightening than the loneliness you might feel if you stopped telling them" Brock Clarke

Viv's blog

Monsieur le commentaire

SR71Fan
Posts: 22
Joined: Sat Oct 17, 2009 1:32 am
Antipspambot question: 125

Re: Rotary Valve Pulse-Jets

Post by SR71Fan » Tue Dec 01, 2009 2:11 am

Hello All:

It's interesting to see that someone [Mark] has found what appears to be a
rotary valve proposal for a pulse-jet. Too bad the text is unreadable. I didn't
think that I would have been the first person to suggest this type of valve.

I have been working on the software architecture for the ECU, and analyzing
the performance requirements. It may come as a surprise, but the HC08
microcontroller actually has more than enough computational horsepower
to handle this application. The HC08 version I am using costs under $5.00, so
it beats the ARM on cost. Software simulations indicate that the basic valve
speed control loop will use less that 15% of the CPU compute time when the
pulse-jet is running at 200 pulses per second. This is by using optimized
integer arithmetic instead of floating-point, which is totally not necessary for
this application.

The largest CPU loading occurs when the ECU is capturing real-time pressure
data while the PJ engine is running. Capturing 64 samples of pressure data per
engine pulse pushes the CPU software loading to almost 50%, but capturing only
32 pressure samples per enging pulse reduces this to less than 30%. Note that
these percentages also include the time needed to control the rotary valve speed.

My plan is to capture 64 pressure samples only at the normal 105 Hz operating
pulse rate; the 200 Hz pulse rate occurs only at high flight speeds. Pressure
profile data is mainly useful for analyzing PJ engine design variations; a mechanical
pressure switch is the normal mechanism for detecting combustion pulse events.

Of course, there is also a bunch of analog hardware that is required for either
the HC08 or the ARM. On-chip peripherals such as A/D converter and timer
hardware are provided by both devices. Since I happen to have a full
development tool set for the HC08, and I have a nuimber of years experience
in programming the HC08, it is the logical first choice for me. Having
determined that I will not run out of computational horsepower, there is no
need to jump to a faster CPU.

Regards,

SR71

Mark
Posts: 10931
Joined: Sat Oct 11, 2003 10:14 pm

Re: Rotary Valve Pulse-Jets

Post by Mark » Tue Dec 01, 2009 4:41 am

"Too bad the text is unreadable."

Click on the image to enlarge it. It then becomes readable.
Presentation is Everything

Suns
Posts: 23
Joined: Sun Sep 19, 2010 2:55 pm
Antipspambot question: 125

Re: Rotary Valve Pulse-Jets

Post by Suns » Thu Jul 14, 2011 7:27 pm

SR Hello

I have been thinking in something identical you did, do google some material and find your post in the forum.

I´m graduated in automation and control and know a lot of ECU and other control systems.
In this forum my only contribution is a 2d pressure simulation software.
viewtopic.php?f=7&t=5910

My idea is very similar of your.

put servo motor or a DC motor with an index or any position sensor to feedback a microcontroler to control the speed of rotation of a rotary valve.

I would put too a knock or pressure sensor in the engine to know the phase of the cycle of burn in the engine.

Close the loops with a PLL like you said to and control the speed of the motor. I think t is nice to send a pulse to the spark plug to detonate the mix anyway.
Possible open a fuel injector too with a pwm control.

I dont think it will not work. have nothing that can not work.

with a pressure sensor, fuel injector and spark control a rotative valve can possible turn any tube in any configuration in a working pulsejet.

If i synchronize the point where soon after the valve when i have vacuum my valves open, it will aways aspire air and controlled dose of fuel. close the valve pressure start increase by itself and pulse the spark, detonate the air, air start going to the pulsejet exit. When all air exists the inertia will make a little of vacuum and the cycle repeats.

the idle fuel will be probably very small.

The pulsejet exit tube will probably work better with a long tube.

Robinson

annodomini2
Posts: 5
Joined: Thu Jul 14, 2011 8:27 pm
Antipspambot question: 125

Re: Rotary Valve Pulse-Jets

Post by annodomini2 » Thu Jul 14, 2011 9:21 pm

Hey all, very new to this, how about:

Two discs with holes, one fixed the other can rotate, actuated by some form of motor?

Stepper motor may allow for greater control over phase timing.

Plenty of ECU solutions out there, a concept like this could run quite easily on a 4Mhz PIC.

Use an automotive 'knock' sensor to measure the pulse and regulate the timing?

Suns
Posts: 23
Joined: Sun Sep 19, 2010 2:55 pm
Antipspambot question: 125

Re: Rotary Valve Pulse-Jets

Post by Suns » Fri Jul 15, 2011 12:54 pm

Annodomini,

Yeah, the valve can be 2 discs with holes like rotating a strong petal valve in its own position.
or use a very non restricting valve like the paper Mark posted. But motor controled.

Its is probably true that it can be controlled by a 4mhz controller.

if you use a pressure sensor with some more advanced math you probably need a better controller.
For alternative sensor it is possible to use a microphone very muffled attached to the engine you can get the phase from this.

Anyway, for developing i think it is best to use a powerful controller, give more space and time to joke around with the control without great organization and everything. after you have the knowledge you can probably run easy on a hc08, an pic or an AVR/atmega.

I think with control like this we can put any pulsejet work on harmonic frequencies, and give power to archive more continuous stress conditions and thinking greater permit PDE operation.

Stepper motor might not work fast enough to actuate the valves. but is is possible and give other type of control. if we need to fast close and fast open to archive operation a stepper or i think a solenoid activated valve is better. could use cars engine valves.

The cycle is something like this:
Sensor= pressure zero going to Vaccum.
Open Valves and inject fuel
Sensor= pressure zero and rizing (or time delay) We need to wait gases revert it flow?
Close Valves and spark the ignition
Sensor= pressure rise from burn -> pressure peak -> pressure start dropping

Cycle Restart. Sensor position after the valves inside the CC.

I think a simple tube long enough for the sensors to work with the valves on one side will work with this.


For PDE little modification on the control software and a very long tube??
Valve with very little restriction needed.

Sensor= pressure zero going to Vacuum.
Open Valves (air need to isolate the previous burn and do not burn the new mix)
Time delay
Inject Fuel
Time Delay
Close Valves
Sensor= pressure high enough to detonate the mix
spark the ignition
Sensor= pressure rise from burn -> pressure peak -> pressure start dropping
Restarts

This software as described don't even use a PLL or something complex, use just the resonant frequency of the engine for the frequency operation and the value of the pressure. A fuzzy controler, dont need a stable frequency too.

You get the first POP and it start working.

Robinson

annodomini2
Posts: 5
Joined: Thu Jul 14, 2011 8:27 pm
Antipspambot question: 125

Re: Rotary Valve Pulse-Jets

Post by annodomini2 » Fri Jul 15, 2011 5:37 pm

Suns wrote:Annodomini,

Yeah, the valve can be 2 discs with holes like rotating a strong petal valve in its own position.
or use a very non restricting valve like the paper Mark posted. But motor controled.
Good to hear! :)
Suns wrote: Its is probably true that it can be controlled by a 4mhz controller.

if you use a pressure sensor with some more advanced math you probably need a better controller.
For alternative sensor it is possible to use a microphone very muffled attached to the engine you can get the phase from this.

Anyway, for developing i think it is best to use a powerful controller, give more space and time to joke around with the control without great organization and everything. after you have the knowledge you can probably run easy on a hc08, an pic or an AVR/atmega.

I think with control like this we can put any pulsejet work on harmonic frequencies, and give power to archive more continuous stress conditions and thinking greater permit PDE operation.
HC08 is old tech and pretty much dead.

PIC's are still alive, widely available and cheap, but you tend to need to resort to assembler and loading the binary can be a pig.

Atmega's are a good bet the arduino platform may offer something useful.

I would also suggest ARM Cortex M0 or M3, which are also cheap and very powerful.
Suns wrote: Stepper motor might not work fast enough to actuate the valves. but is is possible and give other type of control. if we need to fast close and fast open to archive operation a stepper or i think a solenoid activated valve is better. could use cars engine valves.
Varies greatly depending on the stepper motor, it's actuation and power source. The main issue will be power as steppers run constant current.
Suns wrote: The cycle is something like this:
Sensor= pressure zero going to Vaccum.
Open Valves and inject fuel
Sensor= pressure zero and rizing (or time delay) We need to wait gases revert it flow?
Close Valves and spark the ignition
Sensor= pressure rise from burn -> pressure peak -> pressure start dropping

Cycle Restart. Sensor position after the valves inside the CC.

I think a simple tube long enough for the sensors to work with the valves on one side will work with this.


For PDE little modification on the control software and a very long tube??
Valve with very little restriction needed.

Sensor= pressure zero going to Vacuum.
Open Valves (air need to isolate the previous burn and do not burn the new mix)
Time delay
Inject Fuel
Time Delay
Close Valves
Sensor= pressure high enough to detonate the mix
spark the ignition
Sensor= pressure rise from burn -> pressure peak -> pressure start dropping
Restarts

This software as described don't even use a PLL or something complex, use just the resonant frequency of the engine for the frequency operation and the value of the pressure. A fuzzy controler, dont need a stable frequency too.

You get the first POP and it start working.

Robinson

Finding a pressure sensor that can withstand this environment will be the challenge, hence why I mentioned knock sensor as it sits outside the combustion chamber.

The controller would be needed to calculate the delay time as the engine doesn't start at full frequency.

If you control the motor speed with PWM Voltage control for a brushed DC Motor this and the delay time should be proportional.

A rotary encoder on the motor shaft should help with fuelling and ignition timing.

How about using a stainless round vent for the valve?

http://www.seascrew.com/browse.cfm?STAI ... 0000000323

Suns
Posts: 23
Joined: Sun Sep 19, 2010 2:55 pm
Antipspambot question: 125

Re: Rotary Valve Pulse-Jets

Post by Suns » Fri Jul 15, 2011 8:14 pm

Annodo

SR has thinking in using the hc08, i like the arm7 LPC series and have used it a lot before.
For a test rig a microphone is like same-thing as a knock sensor of a vehicle.
As for the pressure sensor. there are a lot of sensors that can withstand this application. only need to stay away from the extreme heat.

About the knock/microphone or a pressure sensor the pressure is ideal, you know what is happening inside the engine and how actuate directly.

As for the delay time it is truly a phase relation like cycle 0% and 100% then resets, something happen at 22% of the cycle not a fixed time delay.

I think for a test maybe a simple solenoid valve and a microphone can give some test results.
I´m start to like the idea to use solenoid actuate valve. can open and close very fast.

for the rotary motor, the probably best is use a tree phase servo motor.

The rotary petal valve i has thinking in something like the link you posted, put the motor axis in the middle and finished hehe

Robinson

annodomini2
Posts: 5
Joined: Thu Jul 14, 2011 8:27 pm
Antipspambot question: 125

Re: Rotary Valve Pulse-Jets

Post by annodomini2 » Sat Jul 16, 2011 8:58 am

Suns wrote:Annodo

SR has thinking in using the hc08, i like the arm7 LPC series and have used it a lot before.
For a test rig a microphone is like same-thing as a knock sensor of a vehicle.
As for the pressure sensor. there are a lot of sensors that can withstand this application. only need to stay away from the extreme heat.

About the knock/microphone or a pressure sensor the pressure is ideal, you know what is happening inside the engine and how actuate directly.

As for the delay time it is truly a phase relation like cycle 0% and 100% then resets, something happen at 22% of the cycle not a fixed time delay.

I think for a test maybe a simple solenoid valve and a microphone can give some test results.
I´m start to like the idea to use solenoid actuate valve. can open and close very fast.

for the rotary motor, the probably best is use a tree phase servo motor.

The rotary petal valve i has thinking in something like the link you posted, put the motor axis in the middle and finished hehe

Robinson
A knock sensor is a 3 Axis accelerometer, a microphone would be susceptible to other noises within a car engine bay.

As the waveform ripples up and down the surface of the tube will cause noise in a microphone sensor.

You don't want the peak of vacuum, you want the ignition point as you need time to ensure that the valve is as open at right point to cause the appropriate peak of vacuum, this is the equivalent of the spring rate on a petal valve.

The benefit being that this can be dynamically changed and adapted by the control system, so the pulse rate limitation becomes the speed of the motor.

This may also make the motor easier to start as the resonant frequency can be adjusted. (Speculation!)

An Atmega option http://www.skpang.co.uk/catalog/arduino ... p-485.html, this can use the free and widely available Arduino development environment.

My recommendation http://uk.farnell.com/nxp/om11049/kit-d ... dp/1786279, comes with it's own C development tools and a JTAG Debugger.

I suggest this as it is small and very low power.

Another suggestion may be an expansion switch, basically a pin that moves out of the tube to actuate either a mechanical microswitch or an opto-switch, this is simple and cheap, but there would be a delay and this would need to be tuned.

Another quirky idea I had was having a small model propeller on the valve actuation shaft, removing the need to maintain the valve actuation with a motor when the engine is in forward motion.

When the system is up to speed using the motor(s) to brake and regulate the speed of the valve.

With the motor driving and the engine static this would also induce some airflow into the system, possibly helping with starting.

Rocket Man
Posts: 144
Joined: Sat Sep 26, 2009 9:59 pm
Antipspambot question: 125

Re: Rotary Valve Pulse-Jets

Post by Rocket Man » Fri Jul 22, 2011 12:39 pm

Have you done any research on Stirling Hot Air Engines. There are several basic hot air engine designs. The Beta Design Stirling Engine is the closest thing to a pulse jet engine. A large pulse jet engine will most likely work with an automatic valve because the engine has a low pulse frequency and it has the power to move the mass of all the valve parts. I think it will work but I think it is unnecessary unless the engine can open larger, stronger heaver valves that will allow the engine to run for many years with no damage.

I have several pulse jet engines in my work shop that will run for many hours with no damage to the reed valves. I have made some very interesting discoveries building and experementing to see what works best and why. I combined all the best ideas into a new improved engine. I have been building pulse jet engines for 40 years I have lots of research data done starting from 1950 when the USA experemented for several year up to date where many colleges and university students have done research on pulse jets. Each study learned something new that gave me ideas for my own experements.

I have redesigned my Dyna Jet Model engine it will run static thrust in my work shop test stand for 2 hours and the reed valves show no signs of damage. The Dyna Jet Engine has 4 design problems you change them all and the engine will run for many hours with no damage to the valves. I have larger engines 10, 20, 50 lbs thrust with the same results.

Tim36
Posts: 31
Joined: Fri Apr 25, 2008 5:11 am
Antipspambot question: 125
Location: Australia

Re: Rotary Valve Pulse-Jets

Post by Tim36 » Wed Mar 04, 2015 11:31 am

Hi Guys,

I haven't been active on the forum for ages, but I have been working on a project like this over the past little while.

I'm just wondering if this project ever got off the ground or if you know of any similar projects that have been successfully built and run?

I have most of the electronics built, I just need to calibrate the pressure sensing input and write an interface program so that I can datalog through the computer.
My goal is to have a working control system then carry out a study using a few different jet sizes and designs, to do a bit of a comparison and pretend I'm clever for a final year project at uni. (It's a year or so off but I want to get the ball rolling soonish).
I want to play around with running pressurised air and fuel inputs rather than relying on the engine to draw the fuel in naturally as well.

Let me know what you think!

Tim

Post Reply