Jetty's Wifi Web Based Laminar Jet Project

Code Dependencies

Here are the libraries and changes that are required for the Arduino Mega 2560 / CupperShield Wifi and the circuits in the above schematics:


- WiShield Wifi Support:

Here's my WiShield fork that adds support for the Arduino Mega 2560: https://github.com/jetty840/WiShield_user_contrib

(Note WiServer calls sendPageURL multiple times to save memory on regular arduinos. The Mega2560 has more memory and multiple calls
can result in occasional corruption due to different sensor values since the last call. A simple page reload repairs the page, but a permanent fix would
be to increase the buffer to be larger than the size of the page, so only one call occurs. The above repository doesn't contain this fix. I'll detail it at some point later.)



- Accelerated Stepper Motor Support:

AccelStepper 1.8: http://www.open.com.au/mikem/arduino/AccelStepper/



- Dallas 1-Wire Temperature Support:

371 Beta (labelled as TCL 3.7.1): http://milesburton.com/Dallas_Temperature_Control_Library



- Free memory monitoring (useful whilst developing):

http://www.arduino.cc/playground/Code/AvailableMemory



- One Wire Support (for Temperature Sensors):

(Version 2.0)
http://www.pjrc.com/teensy/td_libs_OneWire.html



- BlinkM MaxM Support:

http://thingm.com/products/blinkm/quick-start-guide.html (download "Example Code"). Pull out BlinkM_funcs.h from the Arduino examples and put in your project folder.



- Statistics Support (used for Wind Sensor):

http://www.arduino.cc/playground/Main/Statistics (Note file names are incorrect. Save as Statistic.cpp and Statistic.h (not Statistics.cpp and Statistics.h))



- Time Support and Time Alarm Support (needed for various polling functions):

http://www.arduino.cc/playground/Code/Time
Change dtNBR_ALARMS in TimeAlarms.h from 6 to 20.

Here's the software for my Laminar Jet as demo'd in the videos. This works with the Arudino Mega 2560 and the LinkSprite Cupperhead Wifi Shield.

Note that although it supports 2 nozzles, I haven't built the 2nd nozzle yet, so it will need some minor changes at a higher logic level to work with 2 nozzles. I'll update it when I've built the 2nd nozzle.

Additionally the code needs some cleanup and the wind sensor is currently reporting stats until I calibrate it.

Otherwise it works fine with 1 nozzle. Enjoy.

https://github.com/jetty840/Jetty-Laminar-Jet

Jetty,

That is an amazing project and thanks for sharing SO much. I'm going to be adding more to mine someday and will definitely need some electronic support! Thanks for sharing everything!

Hi John. Thanks. Back at you. I've learnt a lot from everyone on this forum (and saved some cash too). I started off with looking at your pages (mad lab) and Marios (and others) and based a lot of my nozzle design and cutter design on the information contained on the detailed descriptions you have both given. It's only fair to contribute back. Thanks

wow !

Have not been here for a while
But your work is the most impressive I have seen for a long time!!!
I have been to busy with other projects the last 1-2 years :\
But I promise to be back next summer to make my 2 jets up and running !

They work OK, the biggest problem is to have a working cutter.
So the use of a stepper motor is maybe the way to go.

Sadly the only motor I know how to use is a DC motor and my honda VTEC motor lol :-)
So I think when the times come I need help to use a stepper motor.


but this is for sure the best place to start.

Thanx again for all the incredible updates
please continue your work and share your progress


cya later

Filip
Norway

Sadly the only motor I know how to use is a DC motor and my honda VTEC motor lol :-)
So I think when the times come I need help to use a stepper motor.

It was my first time using a stepper and it's was a lot less scary than I thought. Using the Darlington array, I found was the simplest / cheapest solution for driving it. From a programming perspective, you just send binary values to it in sequence, i.e. 00, 01, 10, 11, 00, 01, 10, 11 for example, if you want it to go the other direction, then 00, 11, 10, 01, 00, 11, 10, 01.

Each binary number is the next step, you just hit em in sequence.

However I just grabbed an arduino and a software library for steppers, much simpler then.

For a robust stepper driver use the Easy Stepper (open source) but available from Sparkfun.com. I used it and it is fast. I did it with the darlington array and I did it with the Easy Stepper and the Easy Stepper was a lot faster. Basically you just tell it step by 1, 0, 1, etc. and then direction. It's a little easier to wire IMHO.

Which reminds me of something I tested along the way that I forgot to mention. I tried the following motor driver with the same library I used in the final solution: http://www.adafruit.com/products/81

Unfortunately it couldn't keep up with the speed needed and ended up needing to run much slower to avoid missing steps, which is why I went back to the darlington array.

Note, this has nothing to do with Easy Stepper (I haven't tried it), it just jogged my memory when John mentioned it.

Hey Jetty,

Excellent post. Not sure if you are still monitoring this.

I notice that you tried some of my ideas from https://laminar.forumotion.com/t125-designing-the-perfect-cutter but say they didn't make much difference.

I am curious as to how closely you attempted to follow them. I am not criticizing you or promoting my ideas, which are purely theoretical. I am just interested in whether or not they are worth experimenting with further.

1. Speed of the cut.
As I pointed out in my calculations, to achieve a square cut of the stream a rotating cutter slicing through a 1/2" stream at a 60 degree angle has to be rotating at 2000 RPM at the time it starts cutting. If the cutter moves slower than the optimum speed then the entire exercise will be pointless as it will cause the water to spray as it cuts.

If you are using a standard stepper motor starting from a stationary position I believe achieving such a speed (within quarter of a revolution) is impossible. Were you able to do anything to achieve the required speed? I suggested that it would probably require a spring-loaded mechanism, as shown in my YouTube video. I believe that even a normal linear solenoid would struggle to reach the necessary speed.

(The other alternative of course is compressed air. It would achieve the required cutting speed but greatly complicates the logistics.)

2. Razor edge.
I proposed a razor-edged cutter, similar to the razor-edged orifice, not just the blunt edge of a cut sheet of metal. Did you try that? (But pointless if the cutting speed was not high enough.)

3. Cutter angle. (Not the angle that the stream is cut but the angle of the cutter itself with respect to the plane of rotation of the stepper motor).
The diagrams I used explain this point better. See the very first diagram. It shows a long arrow representing the plane of rotation of the blade (the stepper motor shaft would be at right angles to that) together with the blade itself.

The non-leading edge of the blade is tilted back with respect to the plane of rotation, to avoid any surface tension caused by the back end of the cut stream sticking to the surface of the blade.

4. Cutter mounting.
I recommended that the cutter be mounted completely separately from the nozzle to avoid jitter. Did you try that?

All valid questions. First off, I think the theory is sound, but I gave it a reasonable try and couldn't get it to work. I also could
have goofed somewhere.

From memory (and it's a while ago now), what I did was the following.

1. Cut Speed. First off I measured the speed of the laminar jet by interrupting it and timing the distance for it
to get to the end taking into account the arc. This gave me a meter/sec speed for the jet which I had to match with the cutter,
To achieve the speed in the stepper I used acceleration so that it would be at the correct speed when the cutter hit the jet.
The cutter formed an arc on a long arm, so that by the time the cutter hit the jet, the speed was matched, therefore the stepper didn't need to reach 2000 rpm, as the longer the arm, the faster the end of it moves when outlining a circle.

The verification of this is that you can also do the same thing water powered (which I did), where a bit like some crazy sprinkler cutter.

Also, because of the long arm, and the small diameter of the jet, the circle cut approaches straight wrt to the jet.

Compressed air I would think would work.

2. Razor edge. I sharpened the cutter to a razor edge. Also tried testing with a semi circle shaped cutter (matching the diameter of the stream).

3. Not sure, it was a while ago, I'll have to think about that one some more.

4. Cutter mounting. I tried both approaches. Holding the cutter manually (i.e. not connected to the tube), didn't solve it.

The above design was simple to prove the concept, mainly on the cut as there was no handling of back cut, and it was one huge cutter and impractical. However if I saw the cutter work effectively on the cut, then I'd have put the work in to implement something smaller.

From memory I think the arm was something in the range of 30cm - 50cm in length. It was Aluminimum with an aluminum cutter and aluminum hub (i.e. lightweight). Not much torque, but awesome to watch flying around.

Out of interest, have you got it to work? Certainly the theory seems sound, and I'd like to use a better cutter implementation is possible.

After a 3 year break I pulled out the nozzle today that I built (https://laminar.forumotion.com/t123-how-to-make-a-pentair-nozzle) and turned it on. Worked like a dream, but only a relatively short arc about 4' high x 6' long due to water pressure in the hose a bit low today. Must have been a lot of people using water.

I then set about building a spring-loaded cutter pivoted at one end. Not intended to be the final version, just to test the theory. It was about 10" long with the spring at about 3" so the far end cut super-fast into the stream.

Unfortunately the cutter blade I used was a bit narrow (30mm x 3mm flat aluminium) so the water sprayed around it both sides. I put a video camera on it and then had a close look on my PC afterwards. In amongst the splashes I believe that I saw a clean end to the stream, but I will have to do some more work tomorrow to confirm. I will widen the cutter so I can see the end of the stream clearly to check if it is clean.

To get the angle on the blade I simply twisted the flat strip around its centreline by about 10 degrees, about 3" from the end. I then filed it on the bottom of the leading edge to create the razor edge.

I believe the theory is still sound. I just need to prove it.

The current setup is clearly not suitable for actual installation so I am still not sure that it will help with the final product.

I widened the cutter but the results are not very good. A lot of splash that follows the end of the stream.