Hi everyone,
My firts post here but I have been reading and learning a great deal here for a while.
Mach3 is absolutely superb and it`s continued development is taking the software to amazing new heights,
so congrats for that.

My project is a bit unusual,
I have built a 3 axis system to control and freeze a very fine jet of filtered water in the hope of producing
3D mold cavities made of Ice. Like other rapid prototyping systems this is done by building up 2D layers.
At the moment my 0.005" nozzle orifice is giving far too much flow and this means there is too great a volume
to properly freeze on contact with an already frozen surface. So I`m still working on fixing this.

This leads me onto my call for help.
When the machine does an axis change of direction, for example imagine a simple square border, the nozzle keeps pumping for the duration of the axis change at each corner, resulting in an unwanted `blob`at that point.
I have tried the following code :  (Units are in mm)

G1X0Y0
Z0
M08
G1X50
Y50
X0
Y0
M09
G0Z10(END)


This was slightly improved by adding nozzle On/Off commands
G1X0Y0
Z0
X50M08
M09
Y50M08
M09
X0M08
M09
Y0M08
M09
G0Z10(END)

What I am hoping is that there is a way to manipulate the way Mach3 sends the signal.
Imagine if when Mach3 reads ahead it could control the timing of the signal. So lets say
if it sees an M9 command coming up it will activate the command a few milli-seconds sooner
than it would normally. In other words I could control the on/off flow of water during pauses
in direction change.
If anyone can fathom what I`m on about and offer any help it would be greatly appreciated.
I`m not sure if a VB script or perhaps Brains could do the job.

BTW I have Mach setup on `Exact Stop` as I`m not sure about using Constant Velocity.
CV Distance and CV Feedrate ?

Regards
Wil

Welcome Will,
    Have you tried it in CV mode? I would think that would be the best path to try first. It doesn't sound like you have any tool load. With motors strong enough to acheive High accel. rates, I think it will do good. You could try it and see. Try it in CV mode. Don't worry about any of the settings, leave them all at the default values.

Brett

thanks Brett I`ll give it more of a try tonight.
My initial attempt showed very rounded corners under CV mode, as I said I`m not too sure
about changing the settings, my default is CV distance 180 and feedrate +1.0 which I`m taking
to mean just that, ie feed is the same as program feed along straight paths and moving around corners.

Any chance you could explain the cv distance please.
At the moment I`m running my motors at full speed (I think so anyway) and getting around 60 IPM
With a bit of luck, once I iron out all the problems this will be my build speed which would be awesome.

Lovin Mach3 I even managed a screen change with the Mach3screen prog. Screen 4 is great also
but at times I had some problems being able to edit the buttons etc.

here`s a pic of my latest effort, you`ll notice I have added the MDI line to the main program page
and also done away with tool offsets and anything else relating to spindle work.

Good job on the screen Will,
    You can do a search on cv and come up with lots of post explaining it. If your corners are rounding bad, I assum you need to be able to accel at a much higher rate. For what you are doing, you need it to be almost instant, nothing to wide open. You might want to try some test of tuning your max speed down and your accel up in motortuning to help with the rounded corners. For you set up you will need insane accel rates I think to pull it off. Great idea though and attainable I think. Might require stronger motors though. Look at teh rounded corners like this, a Corvette VS a cube van in the slolum. A big motor and good suspension, in this case a rigid frame, mounts, etc. will get around the corners much faster. I'm going to retrofit another mill soon I hope. I will be using 1KW med. inertia servos. In the Visual sizer, I should be able to achieve 750IPM rapids with accels of .05 seconds.

Brett

Ok great, I`ll look at the tuning specs also, I calibrated my axes using mounted digital vernier scales and then by trial and
error I ramped up the velocity as high as I could and adjusted the Acceleration to eliminate any stalling. I`ll post a pic to show what I have later on. ( at work atm)
That Visual sizer program looks interesting. Am I right in thinking that is a seperate `bench` test and you are not able to use your mach3 motor tuning settings ?

WoW - 750IPM rapids with accels of .05 seconds that`d be soming else

Yup, it is a stand alone program. It is just to help pick the right servo for the application. It is a free download. Cooperhill technologies are the arthurs. You can find it at automation direct, about half way down on the home page, servos link. It will take you to a sure servo page with a link to software or downloads, can't remember.

Brett

WIll I beieve a part of your problem is that when changing directions you are covering part of the area twice. We worked with some painting apps that had the same problem. The solution was to reroute your paths. I don't believe that mach cannot  control the m8/m9 that precisely.

SO in your path try moving past the intended turn point and then make a loop outside of the spray path after you have cleared the endpoint. Then come at the cross direction head on turning the spray off as you clear the endpoint and back on at the point where you have cross back over the previous endpoint. That will help prevent overlapping the spray.

You will need to go to a "fine"  mister nozzle to be able to quick freeze the water on the form . Once you have the correct mist pattern the overlapping is NOT that much of a problem  as there is no extra water flow to create the buildup or start to thaw the previos layer.

ALSO ART has just added a time delay to the M8 function. I know you don't think that would work BUT it will give you something that allows you to adjust the timing a bit. If you can adjust the delay you can plan ahead from when it would normally start then adjust the delay tt restart as needed. You just have to plan for the delay in planning your path.

Just a thought,TP

Thanks very much for your great comments.
I must say I hadn`t thought on trying to create a `mist spray` I have been hell bent on tryin to have the smallest possible flow from the nozzle, I can regulate the pressure and when going down to virtually zero pressure the flow takes on a life of it`s own , it comes out of the nozzle at any angle it chooses which is obviously no good.
And thanks for the info on going past the endpoint to prevent overlapping.

I have been looking at buying a Rapid prototyping software called VisCam (http://www.marcam.de/Eng/default.htm)  which has user configurable
post processors to suit most makes of RP machines and techniques used.
I`ll be able to import an STL solid model and set the slice parameters etc
and post it as G Code straight into MACH3, how cool is that.
When I think I am ready I will download their 30 day trial and give it a go.

Ok many tnx again, I`ll lookout for the m8/m9 delay function in mach


regards
wil

To get a uniform application you really need the water pump to be variable rate. I would then tie this to the feedrate. If the feedrate slows, the pump slows, if it stops, it stops etc. It is clear from Mach3 that the feedrate is constantly calculated already - and is therefore probably readily available on a DRO.

You could easily output this to the control for the pump.

Such an application would also be applicable to the paint spraying problem.

Jim

Jim HUM you could tie the pump output into a servo control valve and run the servo valvet from an axis that way you could  sync the two together.

BUT the problem is when you vary pressure/flow  the MIST pattern becomes unpredictable at times.

Will you can find the FINE misters from a good garden or greenhouse supply company. I have used them in the greenhouse and they can put out a very fine MIST and i believe that is what you are going to need.

(;-) TP