CNC Touch Probe
HTML-код
- Опубликовано: 26 сен 2024
- I wanted to add a CNC Touch Probe for a long time to my CNC Router but never could find one that had a low height dimension in the Z-direction ... until now.
This is here a find on Amazon but you will find it on Ebay and Aliexpress as well. For the price of $76 it is quite amazing. The built quality is very good and my first impression if the accuracy of the probe itself is also really good. I test in this set up +-0.05mm however if you would turn the probe and also measure from left and right I think it is reasonable to expect an accuracy of +-0.025mm.
Getting spare parts is an additional bonus as you can easily bend the stylus of the Touch Probe and the M2.5 connection thread is restricting a available selection.
If you contemplating to get a 3D CNC Touch Probe I think this is a good place to start, you can't go wrong for the money.
Here is an updated link to a Normally Closed Contact Probe that I recommend.
As an Amazon associate, I make money on qualifying purchases.
Touch Probe Link (Amazon):
amzn.to/3Rivcu2
That has an amazingly short Z dimension to the tip. Makes it very usable on our non-industrial machines. thank you for posting this video.
Thank you for your comment Gerrit - hope to see pics soon of your finished JBWORX Cool Mister set up with your awesome pump control
@@JBWorx Just working on the mounting for the control box. And 2 tiny fluid leaks at the coolant solenoid :-(
I have bought pricy stainless steel fittings for the cool mister - however using alcohol they will leak after 6 month. The O-Ring inside will deteriorate.
@@JBWorx If I have to I can make hose barbs to avoid those couplers on the fluid line. Also this pump controller has a mode to empty the fluid hose, that might help.
Yes that sounds like a very good option
Thanks for all of the content and tips. I'm picking one up to give it a go.
I have tinkered with mine now and really enjoy having it.
Really a nice video and great series overall. I've been using the same exact probe for over a year now without any issues, the wired USB-C connection still works fine. Actually it is my second one, I did crash the first one and I managed to bend the main shaft of the probe, not just the stylus, so fixing it I think would be rather difficult. But lucklily at this price point it is no problem to get a new one. Got mine from Aliexpress. And I also found some sellers offering M2.5 stylus tips with ruby balls. They are difficult to find as most more expensive probes use M3 thread but they do exist. Have not ordered it yet though as I find the original stylus good enough for my application, well, atleast till I manage to crash it again :)
Thank you for your feedback, I have been quite impressed with mine so far. I wanted to build a ceramic stylus with a Ruby ball. I got the precision ball but I could not find a small diameter ceramic rod. If you don’t mind can you share the seller of the M2.5 Ruby stylus that would be awesome. Usually these will break and prevent further damage if the probe … ok in your crash it probably would not have mattered. Lol 😆
Mine arrived today, what a lovely device. No repeatability testing yet but x,y,z all work as expected on PathPilot. I have the 6mm shank, NC version, a bit more money for some reason but it is what I need for my controller. Thank you for finding this, I had given up hope of an affordable probe that fits my machine. Touch off is noticeably stiffer than my Renishaw TP2, but at least I can use this new one :-) , the TP2 takes a lot of Z height space.
Awesome!!! Now can you calibrate it in PathPilot?
@@JBWorx yes, they have a whole screen and process to go through including testing with a ring gage.
Nice, I really like the PathPilot for my lathe, the conversational programming is really good. I hope you can use the probe in your workflow and It will make alignment easier.
excellent video, Thank you. I have just ordered one of these for my CNC router.
Glad it was helpful! I updated the Amazon link for the NC version of the probe.
@@JBWorx my probe arrived today. It’s so tiny.
Yes, I thought so too when I unpacked it. I really like it.
I purchased one of these probes for use on my manual mill with Electronica EL700 DRO, which has probe functions, but requires only a passive input (no 5-24v input and it requires NC contacts). To use it I needed to disassemble and re-wire it to bypass the internal circuitry. I just removed all the components from the circuit board and located the pad that connected directly to the sensor ring, then connected that to the USB connector center pins. The common is already connected to the other side of the ring via the outside pins. I am in the process of testing now: inside / outside features, inside / outside diameters, steps, angles, etc. I don’t expect the advertised accuracy without the active circuitry but generally if I am within 0.001” for any of those functions I am close enough. If you have not had a look inside your probe you will be pleased to know that it is well-made with gold-plated contact balls and pins, nicely-machined internal parts, O-ring seals, etc. Mine was very clean. As you correctly point out, the USB connector is a weak link. Although the supplier’s advert states that they have replacement parts, in the US we must go through Amazon to contact them and I suspect this might not be straightforward. Thank you for posting your experience.
Good work around !! So your board expects a “dry contact” NC. Yes it’s very nicely made I opens mine hoping to change it from NO to NC as the NC probe was not available on Amazon when I bought mine. Unfortunately I was unable to change it. Also getting it back centered took me quite a while.
Let me know how the accuracy turns out.
In that case, you don't have to modify the sensor itself at all; you just run its output into an optoisolator chip, which's secondary (output) side you connect to your machine inputs.
Depending on whether you want to invert the input logic or not, you connect your signal to the anode or cathode side of the chip input and Vcc or GND to the other. Dead simple, costs you only a marginal amount of money, and saves you the hassle of disassembling the sensor and modifying it. That also solves JBWorx inversions problem in the same way.
It the optoisolator adds to the timing loop and some of them are quite slow. This is why I was avoiding it. However in reading up on this right now I see it’s only 5 micro seconds. So I have to change my opinion about using them in the future. Thanks for your post.
@@JBWorx The probe should be calibrated - distance between probe contact and trigger = effective ball diameter. Adding a relay had no change in my effective diameter(1.88mm). I use "Probe it" for calibration.
My guess is that during probing everything moves so slow that it does not matter if a relay is in-between. A transistor however would be a better choice. But if it works it works !
Hi Jörg, thanks so much for your excellent contributions so far. Very much appreciated! Regarding the wireless approach, I imagine the pro stuff would have some heartbeat going on to avoid accidents? For size and low-power I’d look at ATTINY85 and a rechargeable LIR2450 cell plus 433Mhz TX module. Thanks again!
Thank you so much for that recommendation. There is a new protocol using a direct link 2.4Ghz with a latency of 2.4ms that I have in mind but will check out the components that you mentioned definitely looking into that battery.
I made a circuit a couple of years ago using this cell and was quite happy with it. You can also find a TP4056 USB charge module for about $3 and change a resistor to set charge current to a more appropriate 130mA. Have you determined the power needs of this probe?
I measured the current connected / triggered and it is on 5V 11/16 mA. So rather low that is why I was able to connect it directly to the output of my EdingCNC board. The new S3 chip has 4 power modes active it should be 78mW so on 3.6V we should see 21mA sleep is in the microA range. The board comes with single cell lithium charge capability, probably for an 18650 so yes a charge resistor will be necessary. So about 2-3h of probing for every 100mAh battery capacity. Doable I think.
Great informative video like always. An ESP32 MCU has a wifi chip and is arduino compatible. A small lipo can be attached to drive it all. I’ve used them for smart home sensors and communicated wireless to a home assistant server over the MQTT protocol. Not no anything time sensitive though. There might be a better protocol that is simpler with less latency to do the job. You only need one way communication from the device. Other possible ways to send the signal could be radio, trigger off the rise of a beep sound, or a ring of IR LEDs that light up to trigger 1 or more IR detectors stationed in or around the cnc machine. I'm kinda thinking an IR based setup might be the way to go.
Nice thanks for your ideas. So I thought to pick the signal right of the USB C connection also for the Lipo I will need a charging board. Probably will have to print a small housing for the side. Lately using ESP32 for everything but there are even smaller Arm chips available … not to worried about the latency as the travel speed during sensing and the 2,4 GHz signal will make the over travel virtually no issue … I think 😂
Hehe great review :D way less "chaotic" than mine xD after my crash (that you apparently saw - naming the source would have been nice btw ;) ) I also thought about making it wireless... I don't have much time but might tackle this if there is an interest in it. Keep up the great work 👍
Now I have you, I took screen shots from you video !!! It was on my PC instead of my iPad where I do my editing and was not able to find your channel back. I usually always give credit if I use someone’s info so I will mention your channel when I make another vid about it. Thanks for reaching out !
@@JBWorx Hehe thx, no worrys. I know how easy it is to miss something like that. Anyways, should I get it to wireless I'll give you an update. Thank you for your content und weiter viel Erfolg ;)
Please do !!! I ordered some some components for it already … also I like to test if the probe has both outputs NO but also NC and only the provided wire will make it one or the other way. That would be an easy change.
As a dude thats has crashed one... er uh.. .two. They are 50/50 on rebuilding. If you bend the probe from the tip, youll likely be able to rebuild and recalibrate. I've only been able to get the rebuilt one to repeat to .003".. that said, I have about .001" of sticktion/backlash on my unit, so that would be .002 of the .003".
But yeahhhh.... I have 3 bodies now due to non-related CPU interrupt issues at VERY VERY unfortunate times.
Oh wow that’s not good I hope this happened over a long period of use so you got some life out of them. Luckily I have never had a controller issue in 5 years of use with EdingCNC nock on wood I hope it will stay that way.
I think "thou" is reserved for thousandths of an inch rather than thousandths of a millimetre. Really excellent and informative video as always!
Yea I am pretty much messed up when it comes to units … still prefer the metric system and most of my measurement tools are in metric. We would always say 1/10th, 2/100 or 3/1000 where I am from. Thank you for reaching out.
Americans and their fractions... always gotta make things more complicated than they really are.
Yes my big problem is when I try to explain that square feet is an area and has no equal to a volume of cubic yards to a concrete guy 🙈🙈🙈
To confuse the issue further, I have heard Americans using 'mils' to mean thou.
Yea it’s hilarious- however I do think the we give ourselves here in the US a huge problem dealing with decimals of an inch. I see so many engineers that have no “feeling” of a dimension and its tolerance. I also think that we miss out on not using enough standards. Every bolt, nut pin, clip tap and what not has a DIN standard in Germany. And that makes everything so much easier. Here it’s a bit wilde West where every supplier makes their own version. Add wire gauges, drill bits and dimensional lumber into the pot and you just have to ask who the hack came up with all that nonsense. Ok I get of my sop box now ….
Another great video. I have the same probe and love it. I am planning on adding a tool setter and using the touch probe for zeroing my z height work coordinates. Like you, I don’t have an atc. Do you have an accurate way to measure the length of the probe with your tool setter for subsequent tool changes. Both the probe and tool setter have springs. Thanks for the great content.
Thanks for your positive feedback. So measuring the length of the 3D Probe on the tool setter I can not recommend. My probe triggers first and it does not even deflect the tool setter plunger however I think it’s better to create a touch of block in a certain position on your table. Let’s say you use a 1-2-3 block and touch off on that one. At leased that was my plan so far, have not tested that.
@@JBWorx Surprisingly difficult to find information on how people are handling this in their workflow. Since the touch probe is not deflecting the tool setter, do you think the travel of your tool setter is measurable, and repeatable so that you can factor that into subsequent tool length offsets? I have a tool setter on the way, and will experiment once when it arrives. Keep up the great videos
Like I mentioned above I don’t think it’s a good solution this way. Having a firm surface with a known height will work much better. There is a video from Klaus Michel in this but it’s in German. He did it like I mentioned and it seems to work fine. Also a question how you can incorporate that into your controller settings
Why did you go for NO instead of NC for the two probes?
I'd also be quite scared that maybe a cable is not properly connected and BAM.
Because the NC version is not available in NPN … at leased when I bought it 6 -8 month ago.
In my macro for the z-probe in Mach3 I simply added a check for the max travelled distance allowed.
I think the same simple check can be programmed for the Erding controller too.
Yes it is part of the Marco. I think at the moment i have it set to 15mm. If it does not make contact during that travel distance it will stop
I touch of the material as I struggle getting the basics working.
Yes this is a good way to start. At one point you should investigate how to get a touch Probe going. Also for the tool length.
@@JBWorx I ordered it. But I ignore my tool length as I touch the whole lot of the stock. What I struggle with is centring so this would help
@@fior2373 yes definitely. You will need to put an offset into the controller so a value for x and y this is done only one time. After that x and y 0 is set automatically. Important is that you have a probing routine in your current software
@@JBWorx linuxcnc has a routine for touching of a part. I believe it has others but I never checked it. I have not used offsets for the machine as they confuse me and I broke some bits. Now I zero in the part and all is well. I have not made much. Today I made a fusion 360 design for a watch case. So it's starting...
Linux CNC is Good 👍 and congratulations to your fusion project sounds complex.
Your description tells that you probed accuracy at +-0.05 mm (which is not really high precision 😮) but what I've seen in the video it was actually +- 0.005mm, which would be really a good value, especially for the price.
So, did you do further tests and it was 0.05mm or is that a mistake?
I went with the Japanese approach of underpromise and over deliver. So I actually measured 0.005mm but I figured I would place more divination on top of that just in case anybody buying this and not be able to get my results they would sure get something 10x less than what I measured. And if ok with that anybody will be happy to find out the unit is more accurate.
@@JBWorx Ah, ok. Wise move 😉
I think you are one of the few people that actually read my description! Nice 👍and good catch on the measurements !
Z probing: standard macro in eding F11 > F2 (macro sub zero_z) test dry... :)
Yes the standard macros are working and also the one from Sorotec. I also found the length offset for the 3D Probe in the ini file. Thx for reaching out
Great video, I was looking at them but worried about spare parts, looks like that seller as you said has spare parts like the tip adapter that commonly breaks..
Yes they do and also the plastic part in the inside. The tip has an odd thread (M2.5) so it’s not easy to find a replacement from a different vendor. But it’s available on Amazon now
I see plenty of M2.5 tips on ebay, some with a jewel tip also. That shouldn't be a problem.
Excellent, yes I can find them now as well … wonder why I had such problems finding any at all before.
One may think that with modern tech no probe may be necessary. A concept may work as following: a laser diode illuminator is attached near the spindle projecting a laser line down the working area. The fan out angle of the illuminator is known, and so is the z machine coordinate. A decent monochrome camera with an appropriate filter for the laser color, looking at the working area from the side at some shallow angle should be able to distinguish features and distances down to micrometer ranges. Laser illuminators are cheap (circa $10), digital cameras are also cheap, only the software is lacking.
That is a very interesting concept. I guess one would need this system to be able to look in X,Y and Z to establish the work coordinate system zero points.
I have the same probe, only I use a mach. Do I need to set anything in the control software to measure with a 3D probe?
I don’t know the specifics in Mach3 however it has a probing routine. However the probe needs to be calibrated in the software so you can use it
So does the signal cables go into the existing Z probe connects? I have a Sainsmart 4030 ProverXL V2 and there is a 2 connector Z probe on the box.
Yes, both probes are wired in parallel so in the same input. Either one can make that input on the board and the software knows from the selected routine wich one is being used.
@@JBWorx absolute magic thanks.
@Festivejelly anytime
Nice content! I bought mine but can't find the instructions about how to install it on the mach3 board.
I don’t know what board you are using so I cannot tell you how to connect it. Look for the probe input terminals and make sure you connect the signal correctly to be either + or - so that nothing gets damaged.
I like the probe but the electronics are weak.. I wired mine up wrong and that was that... I ended up removing the electronics and use it as a simple contact. works great now....
Well good that you got some use out of it.
Agreed. See my earlier reply. My DRO expected NO and I deleted the electronics intentionally. It seems to be as accurate as the original poster states. It's especially good to have all the probe functions available.
👍👍
The wire coming out of the uccnc controller is +5volts and the spindle is grounded. Is this NO or NC condition?
I would think it is looking for a NO contact. It will trigger the input when the switch is closed or the circuit loop is closed. But best is to test it.
Will the 6mm shaft work on a 1/4" collet? Will the collet squeeze down enough?
I recommend using a 6mm collet. If you use the 1/4” the runout starts to get really bad. Also fir a cutting bit. So best is to use the correct diameter collet.
I’m having a npn no 3d probe coming in. I was thinking about to connect it to the input of the tool probe. But my tool probe does only use two wires. How do I connect it then ? 😂 the tool probe is using input and gnd. Do I just need to connect it like 5 volt. Then gnd and input on same wire from the tool probe?
So if both are NO probes you can connect them in parallel so to the same input to your board. Either one of them can close the circuit path and get the required signal. This is for the signal wire the power you pull from your control board but check if it supports the current. I remember I measured the current and it was really really low but don’t recall the value from top of my head. So my Eding controller supports it directly.
I got the same probe. It was $45 on China taobao site.. The accuracy is good enough for me. However, I spent some time to make it work. On my machine, the probe has to be grounded on the shaft or it will not work properly. If you have similar problem, try to ground it.
Thanks for the tip, mine worked perfectly out of the box.
@@JBWorxWell, It wasn't the probe. It was due to my spindle not grounded. The probe will get intefered somehow and behaved strangely.
Yes proper grounding of the spindle is important. Best is to open the connector at the spindle and check using a Ohm meter to see that everything is ok.
@@JBWorx My machine is from China. They preferred to use an alligator clip on the drill bit and pass the voltage signal for zero detector. I guess that's why the spindle cannot be grounded.
A proper ground is really important especially if you have a spindle that runs on line voltage let’s say 110C or 220v. It can be a life saver and protect you as the user. If it is a DC motor that is used as a spindle then it is less important.
I use Eding software and controller and I have ordered a touch probe same as yours. This may sound like a dumb question but how do you make the axis move toward the stylus? I am sure it wont be the jog function.
Cool ! Not many use Eding, and I think it’s such a good controller / software. So you have built in functions for probing. Go into the user menu (it’s the icon with the head) you find them there. There is also a bit of set up to do so the dimensions are correct. Too much to explain here but it’s covered in the manual.
@@JBWorx The only soft button I see here is for the tool setter, this will only move the Z axis, I need a soft button to command the Y and X axis.
A wireless probe? Hm. I am wondering if a simple setup is good enough or whether that might need some latency compensation. Interesting. But I sure would want a NC setup then :)
NC was also my choice so far for my DIY Z Touch Probe Lucky it provides both so for this setup I was able to change it. Most industrial 3D Probes are wireless these days. I think the rather low travel speed at the moment of contact make the latency a no issue and I think 2.4GHz will be fine. However … let’s find out together … thx for reaching out !
@@JBWorx Well, it's 2.4Ghz, but if you think wifi there are plenty of protocols on top. I am not saying it's impossible (see the industrial) but it would be good to calculate and check the requirements before starting to build. I just assume the industrial ones will have a proper protocol to model the equivalent of NC probe.
Thank you , all a concept as of right now but if I don’t start I will not learn if it works
@@tcurdt ok the ESP NOW protocol provides a latency of 2.5ms at 40mm/min that would be 0.0017mm travel I think that is totally doable.
@@joergbalzer7430 ESP NOW is a good idea as it strips quite a few layers from the OSI model. But are there smaller ESP32 boards? The ones I know feel too big for this use case. Maybe create a custom board?
Is the probing done via the MDI tab or is it done by a file within the cnc macro file?
I use a Marco. Eding CNC provides a very basic macro however if you get involved how the probing works you can extend on the existing one.
How do you bring up a command to start the probing?
Eding provides a user menu. Here you can assign a button to a Macro / or basically call up routine in the file called macro.cnc
@@JBWorx I tried to probe a work piece using MDI such as F30
G38.5 X30 when the probe touched the part there is no retraction of the probe and the accuracy is down to 0.25 which I am not happy about.
@SmokeRings-eg7md you do need to calibrate the probe. I use a calibration ring but if you have a ball bearing that will work as well. I do not use the MDI command for probing but it is possible.
A NO Probe would not be an option for me. If there is a broken cable, the controller does not notice this and there is a crash. I built my own Probe and in order not to have to install it all the time, it is mounted on a linear guide and can be extended using a compressed air cylinder.
Yes NC would also be my choice, however this is a really good deal I think. Yes the permanent mounting is next this will provide a fixed offset for Z.
@@JBWorx Fixed Z Offset? that you needed only when you have a toolchanger. Otherwise you doenst need it or maybe i be wrong in my thinking? I just move my probe in the left X Corner about 20mm over the Part. Then i start my Probe routine. First the Probe Zero the Z Axis (only for Probing) then he probe the Part in X and Y Axis. I thinks there is video about this function on my Chanel
You can probe Z using the 3D Probe like you just said and set Z - 0 my controller can use either one to set Z0. This way the tool length sensor can be permanently mounted.
I'm courios on how you built the linear guide side of the spindle (I suppose).
Any photo available somewhere?
You mean the Z axis ? Or my project for the linear guide for the probe ?
Do you have a macro (script) that will work with this probe.
It is not really the probe it is the controller that will need the function. Yes my Eding CNC comes with a rudimentary probing function. usually the vendor of the probe will provide the required macro for different types of controllers. Not in this case.
@@JBWorx that is what I'm asking about, is a script for my software, to be assigned to my probe button.
It’s not furnished by this vendor. What software do you use ?
the link takes me to this and not the probe. :( ER11-6mm
Thank you for pointing that out I will correct it right now !
Fixed it AND now the Normally closed contact probe is available as well on Amazon that is the updated link.
Could you please share the macro?
I am using the standard Eding CNC Macro at this time. It’s part of the Eding Software.
Hey nice video. I bought that 3D touch probe and I want to connect my Z touch probe paralel like you show in the drawing. But it won't work. I'm using UCCNC software the probes are both NO contacts . Separately they work fine. What can be the problem.
Ok so NO contacts will need to be wired in parallel, so either one can close the circuit. I had to change the expected contact in my software to make it work at first.
@@JBWorx Hey, thanks for the quick repley. Did you had to change the "activ low check". Do you also use UCCNC?
I use EdingCNC and yes in my controller I had to change this setting … I don’t remember at the moment if it was from high to low but it worked once I changed it.