Machsupport Forum

Mach Discussion => General Mach Discussion => Topic started by: Flipz01 on June 10, 2008, 05:17:29 PM

Title: Height Following
Post by: Flipz01 on June 10, 2008, 05:17:29 PM
Has anybody out there found a good way to do height following?  That is, to have the z-axis plunge, have a sensor read the material surface and set the height, then maintain that same height above the material as the part is being cut?
Title: Re: Height Following
Post by: jimpinder on June 12, 2008, 04:34:48 AM
What are we talking about here -
Are you asking a simple device to come down, read where the top of the work is, then move z down and cut level -

Or are you asking for a device that comes down, and continuously reads an undulating surface and keeps the cutter a set distance above (or below) it - i.e. cutting a groove of a set depth.
Title: Re: Height Following
Post by: Hood on June 12, 2008, 05:49:07 AM
If its the latter of what Jim said then it will not be an easy thiug to do. Just thinking aloud here but it may be ppossible to use a plasma style torch height control. I think they work off a voltage and adjust to keep the torch height constant when materilas warp due to heat. It might be possible to set up some sort of sensor that alters voltage as the material lowers or raises in  comparison to the axis height then feed this to a THC and it will vary the axis.
 As I said just thinking aloud and not sure even if thats how the THC work :)

Hood
Title: Re: Height Following
Post by: Flipz01 on June 12, 2008, 09:40:07 AM
#2 is the correct answer, almost.  We wouldn't be cutting a groove.  Think plasma torch, laser, or in our case waterjet.  This would be especially nice when cutting underwater where you couldn't see the material or the cutting tip (also done with plasma torches).

In general, plasma set-ups do exactly what you suggested.  They plunge to a set voltage and then maintain that voltage.  With plasma, this can be read in the torch itself as the voltage increases as the cutting tip gets farther from the material.

What we'd like to do is use a linear resistor that passes voltage from 0-10 vdc.  The setting for height would be at 5 vdc.  This could be attenuated using a potentiometer to adjust the perceived voltage and to set the start point (that is, plunge to position and let the basic setup pick a height - then turn the knob to adjust up/down slightly to a proper height).

So, the z-axis would plunge to a start point - the point where the feedback voltage is 5 vdc.  The height could be altered manually with a potentiometer, if required.  From there, the control would move the z-axis up/down to maintain the voltage at 5 vdc.
Title: Re: Height Following
Post by: da21 on June 12, 2008, 11:33:41 AM
i am working on a plugin at the moment to do exactly that for a Laser ,
let me know directly what voltages u need etc , or exact requirements

i hope to have it tested shortly

Dave 

Title: Re: Height Following
Post by: Flipz01 on June 12, 2008, 01:24:56 PM
That's great news!

I really don't care what voltages you pick - we'll work with whatever.  However, we prefer to work with or inside of a 0-24 vdc range.  The sensor we are planning to use is 10 vdc to 24 vdc.

What are you planning on using for the laser?
Title: Re: Height Following
Post by: da21 on June 12, 2008, 02:25:32 PM
here's a pic of the type of sensor we use , although it's set in it's holder , it's a linear sensor gives us + - volts out ,dc swing either side of 0v of around 8 - 10v
you can see the shoe that rests on the material being cut , this is spring loaded to keep the tension on the sensor shaft , it has a calibrated range of  + - 1.0mm
with a range or + - 3mm each side of 0v point. we are using 15v + & -  dc  so as to get a negative voltage swing from the sensor , but if your sensors only positive output , then we should be able to set it for a specific voltage for your zero reference point , this then depends on what voltage drift you may have etc .   

we are looking at either a remote triming encoder for fine tuning or just setting the points in mach 3 , i am hoping to get some machine time to start testing it all shortly
to test for voltage drift and repeatability of height control

email me off list
what's  required ? , the full electronics or just a plugin for mach3 , if so what type of connections do you need if any ,  parrallel port , serial etc


Dave
   
Title: Re: Height Following
Post by: Flipz01 on June 12, 2008, 03:14:41 PM
We've only just started thinking about this.  If we can buy a package that works (is the one in the picture a production thing or custom made?), we will use what's available.

Otherwise, we'll use a linear encoder and make something similar to your picture.  We don't care if it's +/- 10 vdc or 0-20 vdc, or whatever.  However, 0 vdc to 24 vdc is REALLY easy since every system we've ever done has this already.

Currently, we use two LPT ports for Mach3.  We'll use the Smooth Stepper when it's ready.  For this device, we'll use whatever you recommend that works.

Sure would be nice to be able to set the high, center, and bottom voltages in Mach.  That would eliminate the need for a manual adjustment to get the height right.  Actually, just the middle might work - with a choice of polarity for what raises or lowers the z-axis.  That is, depending on the resistor and wiring, an increasing voltage might mean you are either too close or too far away.  A simple check box could be used to set polarity.

Another possibility would be a variable for response speed if this is to be truly dynamic, or alternately a response distance.  For example, if the sensor reads "too high", the plunge is 1/20" and recheck.  That could prove easier by eliminating the possibility of "hunting" like a badly tuned servo.


BTW - forgot to ask if this could be used for twin heads - nothing's ever simple!!
Title: Re: Height Following
Post by: da21 on June 12, 2008, 03:55:47 PM
well one thing is , if it dont work then i won't sell it !.

the height unit shown is a one off , but we will be making similar units , it all depends on whats needed etc to suit the particular machine ,
the unit it's self is fully dynamic in that once you have set the parameters it looks after it's self , but i was looking at incorporating a mini mpg rotary control that could move the zero point over a small range , more as a final trimming pot , but actualy i hope we won't need it , this of course depends on the accuracy of height control required  .

we are in too minds to incorporate the controller electronics via USB  , or via an addon unit to the smoothstepper as done by warp9 etc , by which we hope to have a similar product , available in a few months but suiting servo & encoder system users .

Dave
Title: Re: Height Following
Post by: Hood on June 12, 2008, 03:58:39 PM
but suiting servo & encoder system users .

Dave

Just a bit curious about what you are meaning?

Hood
Title: Re: Height Following
Post by: da21 on June 12, 2008, 04:08:37 PM
Hood ,
We do more servo motor systems i.e closed loop  , as against stepper based systems .
and retrofits etc , thats all

nothing meant by it

Dave 
Title: Re: Height Following
Post by: Hood on June 12, 2008, 04:12:11 PM
As the smoothstepper can be used on servos systems as well, I was however wondering whether you were thinking of closing the loop to Mach with your product?
Hood
Title: Re: Height Following
Post by: da21 on June 12, 2008, 04:28:54 PM
the short answer is yes we hope to


Dave
Title: Re: Height Following
Post by: Hood on June 12, 2008, 04:34:38 PM
Sounds interesting, looks like there may be quite a few products that will be attempting this which can only be good news for Mach :)

Hood
Title: Re: Height Following
Post by: klmark on June 12, 2008, 05:33:45 PM
Hi guys
Jumping in here if that is to be used on a plasma cutter . It would be nice nothing is out there that is simply as of yet
                                             klmark
                                     
Title: Re: Height Following
Post by: Flipz01 on June 12, 2008, 06:03:06 PM
Wow - lots of interest here.  As Hood said "Good for Mach" - and us too!!

A good solution should work through Mach so that current position is always known - or "closed loop to Mach" .  I think that's the goal everybody is talking about.

Should work well on anything - plasma, laser, waterjet, etc.

Once the basics are worked out, I'm expecting that several entries will pop up for consumer sales.  The trick will be making one that is low priced for the masses.
Title: Re: Height Following
Post by: Chaoticone on June 12, 2008, 08:50:42 PM
I wonder if and how well something like this would work for sensing distance? There are other types avaliable form other manufactures. Could sense the distance and out put an analog signal depending on varience.

http://web5.automationdirect.com/adc/Shopping/Catalog/Sensors_-z-_Encoders/Ultrasonic_Sensors/30mm_Round_(TU_Series)/TU1-C0-0E

Brett
Title: Re: Height Following
Post by: Flipz01 on June 14, 2008, 12:20:31 PM
As far as I'm concerned, an ultrasonic device would be perfect.  It outputs a signal that varies from 0-10 vdc.

Now, if there was a way to read in this analog signal to the Mach interface that would plunge until the feedback was 5vdc, and then maintain that 5vdc by moving the z-axis up/down as the voltage varies - you're there.

Hopefully, something like that is possible.

Title: Re: Height Following
Post by: Hood on June 14, 2008, 12:29:00 PM
Would the 1Hz switching frequency be adequate for such a device?
Hood
Title: Re: Height Following
Post by: Flipz01 on June 14, 2008, 03:16:43 PM
For almost anything I do that would work great.  I think the toughest application would be laser where height is critical and cutting speeds can be fairly high.
Title: Re: Height Following
Post by: fer_mayrl on June 16, 2008, 01:40:52 PM
There are some high accuracy ultrasonic sensors if that 2mm repetitablility affects your process.
Regards
Fernando
Title: Re: Height Following
Post by: Chaoticone on June 16, 2008, 03:11:17 PM
Also, what I posted was just a quick example. I'm pretty sure you can get some with much faster switching freq.


Brett
Title: Re: Height Following
Post by: fer_mayrl on October 14, 2008, 01:37:13 PM
Hello,

Where are all of you standing regarding this subject?

It would be interesting to see that feature implemented into mach3

Also, for Flipz, i guess cutting underwater would throw off the ultrasonic height sensor, what kind of sensor are you thinking of using?

Regards
Fernando
Title: Re: Height Following
Post by: da21 on October 14, 2008, 02:41:04 PM
we use Linear Displacement Transducers ( LVDT )
made by solartron meterology

one model although an old one which is still available from RS
which with a 12v dc input gives  a + - 9v dc output with 0v at centre of travel
RS 646-482  , but be aware they are expensive !



Title: Re: Height Following
Post by: fer_mayrl on October 14, 2008, 02:57:32 PM
were you able to get the whole thing to work?

How did you implement it?
Regards
Fernando
Title: Re: Height Following
Post by: alfaalfa on October 17, 2008, 04:03:02 AM
Hi.

For laser cutting you need to use ring sensor around your nozzle and the position can be measured with linear potentiometer or LVDT sensor. The measurement can be translated to a analog voltage which can control servo drive directly - or if possible - through Mach.

The better way for metallic materials is capacitive height sensing where you have oscillator that gives frequency (in range of 200 kHz - 1 MHz) to your insulated nozzle. By monitoring the amplitude or phase locked loop comparison you can get out signal that depends of the nozzle distance ie. capacitance between nozzle and plate. This signal can then control your analog servo drive or given to control.

I have some data of capacitive sensors but no time to produce devices at the moment. If someone with electronics skills is willing to continue - I'm willing to co-operate.


BTW. I'm still waiting for faster M03 M05 handling to Mach as this is slowing my applications significantly!! Any news??

BR. Arto

Title: Re: Height Following
Post by: fer_mayrl on October 17, 2008, 08:13:51 AM
I guess there are quite a few types of sensors that would work. I have been looking into linear pots and LVDTs, and plan to have mach red the analog voltage and control the position of the z axis.

For some applications we are left with contact sensing, for example waterjet cutting where you may cut different types of materials that would behave differently on some sensors, and then there is underwater cutting where welll... it just makes things go out of whack.

Regards
Fernando
Title: Re: Height Following
Post by: alfaalfa on October 17, 2008, 10:12:47 AM
I think the electromechanical sensor (ring around the nozzle, linear slide and distance sensor) and capasitive one are those which can work in real life. You can think of optical and ultrasonic ones but how many you need of those as you must sense around the nozzle - just one spot is not sufficient?

I agree the sensor ring is the right one for WJ and AWJ. If you need to minimize scrathes to your workpiece - the sensing can happen when piercing and then only occasionally say once per 200 mm or something.

With thin materials some vibrations may occur. With mechanical sensing these can be damped too.

BR. Arto

Title: Re: Height Following
Post by: da21 on October 17, 2008, 02:17:55 PM
we only use one sensor as you can see in the picture in this thread
it works fine on our lasers , we can hold the height position to within 6 - 8um
Title: Re: Height Following
Post by: fer_mayrl on October 17, 2008, 02:22:35 PM
Hey Dave,
One question... does your foot sensor tend to get stuck when passing over already cut material? seems that if it touches a slot prom the side it might tend to get stuck.

Title: Re: Height Following
Post by: Sage on October 17, 2008, 02:47:45 PM
I'd be sceptical that an ultrasonic system would work, especially in the vicinity of the weldng/cutting process. Welding, cutting etc apparently produces a lot of sonic (and probably ultrasonic harmonics) "noise". Evidenced by the fact that hearing protection is recommended / required for welders. Never would have thought so, but apparently true.

I guess it really doesn't matter what you decide to use as long as you standardize on the voltage levels the software expects to receive. Output level shifting would be simple enough to implement for whatever sensor you choose.

Sage
Title: Re: Height Following
Post by: da21 on October 17, 2008, 03:03:20 PM
no not had any issues over it catching as you say , however the hoop (4mm dia ) around the nozzle is slightly bow shaped to help in that case and the hoop is spring loaded with a light weight spring just enough to keep the LVDT touching at the rear ( mounted inside the tube mounting ) , i think you can see it but it may not be obvious , i have had thoughts of making it more of a bowl or dish shape

i'd certianly be game to look at capacitive sensors and see if we can come up with a better solution

works fine on my 600w CO2 Laser i dont get any interference pickup etc 
 
Title: Re: Height Following
Post by: Flipz01 on October 28, 2008, 01:53:27 PM
fer_mayrl

Any progress on your end here?  You stated "I have been looking into linear pots and LVDTs, and plan to have mach red the analog voltage and control the position of the z axis".

That's the correct solution - or so it seems to me.

For our part, we are planning to use one of the PoKeys input units for the analogue in (they will handle up to 4 analogues).  That gives us a nice expansion of useable I/O, and allows the analogue inputs and ability to use encoders (for mapping as an example).
Title: Re: Height Following
Post by: fer_mayrl on October 28, 2008, 01:59:28 PM
Hey Flipz,

Not really, im still in the designing stages of the machine, and was just investigating the way i could (in the future) implement this.

Long way ahead since i think i am starting with plasma, and then go with the waterjet which is for what i need it.

Regards
Fernando
Title: Re: Height Following
Post by: fourxfour2 on February 22, 2009, 12:51:59 AM
Fernando,
Would really like to know how this project is going ?????
Dave
Title: Re: Height Following
Post by: fer_mayrl on February 23, 2009, 10:52:40 AM
I strayed from that proyect... and I have not implemented it yet.
It will be a while...
Regards
Fernando