[Bob La Londe]

How good are the homing switches on the 1100. Reliability I would guess it pretty good since I have not been reading about related crashes and failures. I want to know how repeatable. A few thousandths is good enough I am sure to prevent limit crashes and ATC related positioning crashes. Is it goo-d enough to make repeatable dimensioned parts? One of my machines uses covered and shielded prox switches and repeats within about .001-.002. Depending on the part that can sometimes be good enough. 3 of my other machines use un protected micro switches with cams. If I remember to blow off the swarf first they are good to about .005... if a resistant chip doesn't stick to a surface. LOL. Not really good enough to just home and cut a proven program for a lot of work. Even if I remembered to set a work stop to position the part before I shut the machine down.

[IMT]

I've checked it a few times and never seen more than .003.
Checked it "accidentally" on Saturday. Probed my part and set XYZ0, went to jog the probe up and hit the Home button which is next to Page Up on my keyboard. After it finished homing, I re-probed the part and all 3 axes were <.002 off.

[CadRhino]

I use to use a homing switch in my 4th axis and the repeatability was excellent. Unfortunately path pilot no longer supports homing very well (or as good as Mach 3)



QUOTE=Bob La Londe;2157158]How good are the homing switches on the 1100. Reliability I would guess it pretty good since I have not been reading about related crashes and failures. I want to know how repeatable. A few thousandths is good enough I am sure to prevent limit crashes and ATC related positioning crashes. Is it goo-d enough to make repeatable dimensioned parts? One of my machines uses covered and shielded prox switches and repeats within about .001-.002. Depending on the part that can sometimes be good enough. 3 of my other machines use un protected micro switches with cams. If I remember to blow off the swarf first they are good to about .005... if a resistant chip doesn't stick to a surface. LOL. Not really good enough to just home and cut a proven program for a lot of work. Even if I remembered to set a work stop to position the part before I shut the machine down.[/QUOTE]

[mountaindew]

I've checked it a few times and never seen more than .003.
Checked it "accidentally" on Saturday. Probed my part and set XYZ0, went to jog the probe up and hit the Home button which is next to Page Up on my keyboard. After it finished homing, I re-probed the part and all 3 axes were <.002 off.

I had a problem for a day or so but all I needed to do was clean machine . As mentioned above I get around that. It varies a little from day to day You can setup indicators and home mill very precisely manually if the home switch system is not accurate enough. I tried it for a while even designed mounts for 3 height indicators, but went back to using the machine as it was designed.

[Bob La Londe]

I use to use a homing switch in my 4th axis and the repeatability was excellent. Unfortunately path pilot no longer supports homing very well (or as good as Mach 3)

Can you explain what you mean? Not homing very well would be a huge issue.

[mountaindew]

Can you explain what you mean? Not homing very well would be a huge issue.

If you don't have a 4th axis this is not important!

I have the 4th axis optional homing HAL sensor. I use a auxiliary operation in sprutcam to command a return to 0 and not unwind 5x360 or some annoying amount of turns at break neck rotary 4th axis speeds . Other then that, when my 4th axis is set to 0 it returns there +- the backlash of the worm drive and I see no difference when used with path pilot.

[MichaelHenry]

I checked the repeatability on my 2007 PCNC 1100 (Series 1 upgraded to Series III) a year or two ago and found repeatability to be 0.0005" in X, 0.0023" in Y, and 0.0012" in Z, averaged over 10 sets of measurements in each. By comparison, the errors on my 15L lathe measured the same way were 0.0162" in X and 0.0410" in Z. The lathe results are clearly inferior to the mill results.

I expect that I could get better Y measurements on the mill with a bit of tuneup, but the parts made so far have been good enough that I haven't bothered. I suspect that the only way to get better results on the lathe will be to replace the limit switches. If your mill is much worse than what has been reported so far, a swap out of the limit switch(es) might be in order. That's neither difficult nor very expensive, as you probably already know.

[Bob La Londe]

So as long as we are on the topic: What type of limit switches does Tormach use on the linear axis? I know I'll find out soon enough, but...

I probably under estimated the accuracy of the switches on my current machines, but I don't like to be one of those guys who always says, "well my ____ never has that problem" or "my ____ is better than that." The thing is they are probably good enough on all the machines to make a second part... most of the time... but if they are off once on a part that takes half a day to machine it can be hell.

As to the rotary: If its just a matter of "unwinding" it may be addressable in code. Just insert a line of code that changes the offset to match what you want it to be. Change it from 362 degrees to 2 degrees or as needed.


From: https://machmotion.com/cnc-info/g-co...System_Offset_

G52 Coordinate System Offset

To offset the current point by a given positive or negative distance (without motion), program
G52 X~ Y~ Z~ A~ B~ C~ , where the axis words contain the offsets you want to provide. All axis words are optional, except that at least one must be used. If an axis word is not used for a given axis, the coordinate on that axis of the current point is not changed. It is an error if:


¨ all axis words are omitted.


G52 and G92 use common internal mechanisms in Mach3 and may not be used together.


When G52 is executed, the origin of the currently active coordinate system moves by the values given.


The effect of G52 is cancelled by programming G52 X0 Y0 etc.


Here is an example. Suppose the current point is at X=4 in the currently specified coordinate system, then G52 X7 sets the X-axis offset to 7, and so causes the X-coordinate of the current point to be -3.


The axis offsets are always used when motion is specified in absolute distance mode using any of the fixture coordinate systems. Thus all fixture coordinate systems are affected by G52.

I would have to experiment to see what happens if its used multiple times,

[mountaindew]

I'm lost now!
I use 54-59x in path pilot. With all the probing options both electronic and manual Hamier a user can set offsets in less then a minute and your good for the day if the 0.001 homing average is not accurate enough. Your mill comes with fusion360 cad/cam I think. Draw what you want to make. Then setup your production parts in fixtures create operations and generate code. Go to mill and set it up the same way and make stuff, no drama, no hand code, all fun. As I mentioned above PP allows a manual homing system that can setup to home machine to hyper accuracy of the super expensive indicators used . This would be a far better solution imho then any roller switch system no matter what brand!

[Bob La Londe]

I'm lost now!
I use 54-59x in path pilot. With all the probing options both electronic and manual Hamier a user can set offsets in less then a minute and your good for the day if the 0.001 homing average is not accurate enough. Your mill comes with fusion360 cad/cam I think. Draw what you want to make. Then setup your production parts in fixtures create operations and generate code. Go to mill and set it up the same way and make stuff, no drama, no hand code, all fun. As I mentioned above PP allows a manual homing system that can setup to home machine to hyper accuracy of the super expensive indicators used . This would be a far better solution imho then any roller switch system no matter what brand!

You are lost. I was answering the problem about homing on a rotary. Instead of unwinding just change the offset in CODE. You could do it with the offset system (G54-GXX) as you sort of imply, but that would be over complicated. Just change the current offset. You can store and retrieve, or you can just set it. I do both either by macro or in code, although not for that particular application. I'm not guessing.

I am a little familiar with Fusion360 and use it for things like 3D adaptive clearing which it does very well and rest which it does fair. I also use CamBam (its much faster for some things) and ViaCAD. I have used a handful of other CAD and CAM programs as well. I have also written some fairly complex hand coding solutions to problems and manually edited to do things Fusion CAN'T or doesn't do easily. I'm not a newbie, and I have been following Tormach for many years. One thing I am not and never will be is a fanboy. I want the reality not the rose colored glasses of the situation. I don't mean that to sound excessively harsh, but its in response my perceived tone of your post.

I must not understand your definition of "hyper accuracy" given what others have said about the accuracy and repeatability of their Tormach homing switches. Either that or your machine is the exception. Perhaps you mean hyper accuracy using a tool that cost $500 (+/-) and that's fine, but it wasn't my question, nor was your post clear enough to definitively separate one from the other. (I know what you meant) I do that if its really important. Sometimes I have to make parts to go with parts I sold a customer a year ago, or modify a part after its been off the machine. That's when I always touch off. Sometimes in multiple locations to verify position and orientation. (I sell parts all over the world.)

There is absolutely no need to defend the BRAND as you seem to conclude. Tormach doesn't make switches. They buy them. AND as I said early on in this thread one of my machines has shielded and covered proximity switches that never see a chip. Its a nearly 40 year old machine I retrofit and it will throw chips <slight exaggeration> the size of small rodents </exaggeration> across the room with enough force to dent the walls if I didn't have shields on it. It just a "different" machine. Does a different job.

Which brings me back to: "What "TYPE" of homing and/or limit switches does Tormach use?" which was the very first part of my previous post.

I'm just trying to learn more about the machine so I know what to expect. GEEZ!!!!!! You guys brought up a potential issue and I tried to help. Sorry. I'll try not to do that again. LOL.

[mountaindew]

Sorry! not defending NOBODY or NOBRAND................................... EVER.!!!!!!!

I'm confused you were asking about homing accuracy and it was explained very well by a number of people. Then you post some code using g52+ some number. ok I'm lost what does that have to do with Homing a Tormach.
Just stating my experience with the product and I understand homing a 4th axis and the reason I wonder why you didn't, with your many years of experience, why you asked that question above. Figured you were confuse but it was me.
Anyway I said enough back to my boring cam simulations only reason I'm on forum.

They take time to compile and review so i set here and read cnc post lol. not very good at answering or staying on topic I guess

Draw this up in about 20 minutes this morning

Attachment 388020


Then move parts one by one to cam
Attachment 388022

Then Friday night in the shop

[Bob La Londe]

Sorry! not defending NOBODY or NOBRAND................................... EVER.!!!!!!!

Good. It gives your replies a lot more credibility.

I'm confused you were asking about homing accuracy and it was explained very well by a number of people.

I know how homing works and I asked very specific questions. One of which has still not been answered. You were confused.

Then you post some code using g52+ some number. ok I'm lost what does that have to do with Homing a Tormach.

You said, "If you don't have a 4th axis this is not important!" in direct reply when I asked "Can you explain what you mean? Not homing very well would be a huge issue." about "I use to use a homing switch in my 4th axis and the repeatability was excellent. Unfortunately path pilot no longer supports homing very well (or as good as Mach 3)" As somebody who has direct experience with trying to save machine time and dealing with odd situations and who tries to think out side the box I tried to help with a perceived problem expressed by one person and then seemingly confirmed by another. I do that. I don't just come to take.

Just stating my experience with the product and I understand homing a 4th axis and the reason I wonder why you didn't, with your many years of experience, why you asked that question above. Figured you were confuse but it was me.
Anyway I said enough back to my boring cam simulations only reason I'm on forum.

They take time to compile and review so i set here and read cnc post lol. not very good at answering or staying on topic I guess

<WARNING. TANGENT.> Yeah I get that. One of my previous recent investments was in a faster multi (quad core 8 thread) processor PC for CAD CAM work. From what I can tell neither CamBam nor Fusion seem to take full advantage of the extra computing power. Even when I manually set affinity and priority. Interestingly though I can run CB and Fusion at the same time and neither seems to slow down.

[mountaindew]

Sorry! not defending NOBODY or NOBRAND................................... EVER.!!!!!!!

I'm confused you were asking about homing accuracy and it was explained very well by a number of people. Then you post some code using g52+ some number. ok I'm lost what does that have to do with Homing a Tormach.
Just stating my experience with the product and I understand homing a 4th axis and the reason I wonder why you didn't, with your many years of experience, why you asked that question above. Figured you were confuse but it was me.
Anyway I said enough back to my boring cam simulations only reason I'm on forum.

They take time to compile and review so i set here and read cnc post lol. not very good at answering or staying on topic I guess

Draw this up in about 20 minutes this morning

Attachment 388020


Then move parts one by one to cam
Attachment 388022

Then Friday night in the shop

Heck my big break thru last week was variable reduced depth peck drilling cycle. Its been a setting in cam I never understood how to use right. Researched and found a formula published in HSM advisor and now im happy with my deep drilling all fun and just trying to help. If your stuff don't work. You here about it from me

[Jim Dawson]

Bob, I have my own theory about machine homing. As you say, home switches are close enough for tool changes and to keep the machine from crashing, but IMHO, are never good enough to locate parts on the table. There are some really accurate switches available that repeat to a few microns, not that expensive either. But that is only useful if you have truly zero backlash in the axes or are using linear encoders on the axes.

If the software supports it, machine homing should be a 4 step process:
1) Move to the home switch, look for a state change
2) Move off of the home switch, look for a state change
3) Slowly move back to the home switch, look for a state change
4) Then move to the index pulse of the encoder.

This gives you a position accuracy of one encoder pulse.

I personally do not use any machine homing for locating parts, I always set 0,0,0 off of the part or the fixture using a pin that is installed for that purpose. If I have a setup on the table, I always leave the machine on and always park it at a known location when done for the day. That way, I know where it's at in the event of a power fail, which happens frequently around here.

Just my $0.02

[Bob La Londe]

I personally do not use any machine homing for locating parts, I always set 0,0,0 off of the part or the fixture using a pin that is installed for that purpose. If I have a setup on the table, I always leave the machine on and always park it at a known location when done for the day. That way, I know where it's at in the event of a power fail, which happens frequently around here.

Just my $0.02

Yeah. I'll sometimes park a machine at 0,0,5 at the end of the day before I shut it down if I know I'll be cutting an identical part from an identical or nearly identical blank the next day. My current 4 running (well 3 and 2/3 at them moment) machines are all manual tool change. My big concern was with going to an ATC machine (what I ordered). I guess the probe will get a bit more use now.

[mountaindew]

Good. It gives your replies a lot more credibility.

What is credibility?
I try to help only! I could care less about who does what and how fast or how accurate. Experienced users of these machines understand there are hundreds of variables that effect results. Example: I mentioned above getting good accuracy on my super spacer. Well I do small diameter parts and any problem would be very hard to see. If I was working on a 8" or larger diameter part with 4th axis in horizontal position any problem would be very easy to see way out there. And I would think homing a 4th axis using a magnet HAL censor would be difficult to get accuracy that way. I never could get 4th axis automatic homing to ever work. I just set it and hit home button on path pilot. Same goes for xyz homing accuracy I stated! It is based on measurements taken at vise work envelope. Z accuracy will vary over x axis table travel because of hooks law, so I read anyway.

[Bob La Londe]

You know I think we all need to use more emoticons.

[hanermo]

Most posts seemed to me to be great -
as details are determinant.

Very good japanese mechanical switches are accurate and repeatable to 1 micron.
About 100$ each.

Fairly good optical switches are accurate to about 2 microns.
5-30$ each.

Almost-all screws will *repeat* to a few microns, 1-2-3 micron max error on the same part/run/material + similar environmental conditions.

Typical ballscrews here on cnczone are c7/c5 on accuracy, meaning about 0.1 mm max positional error over 300-500 mm.
Almost-all screw mounts, yokes, machine designs/kinematics easily exhibit 0.02-0.05-0.10 mm errors due to flex, bias, sticktion, etc.

I have seen all above on my own stuff .. where I strive to work to much higher accuracies than 99% of people here, and spend more money and time at it than the typical average user here.
My comments and observations are meant to help others, that´s all.

For example;
On my 12x lathe I changed the 19 mm acme screw, z axis, to a 32/4 mm ballscrew.
With massive mounts, massive yokes, etc.

Yoke:
70 mm thick yoke, 80x100 mm, tool steel.
40 hours of fab:

Screw mount:
Ballscrew baseplate is 50 mm thick, 160x200 mm tool steel, iirc.
Fixed to the lathe HS with 6 x 8 mm shcs of 12.9 quality.
End of slab turned flat to 0.01 mm or less error, assuring excellent rigidity/fit for the BS mount.
40 hours of fab:

Modern machine tools are very good in design.
They tend to be thin-skinned, stressed-skin designs.
Typical strength vs load is 2%,
Ie any modern machine tool can take 100 tons load if it can cut 2 tons at kinematics. 50:1.
Typical 32 mm screws == 1400 kg, kgf, and max loads on z axis at typical 200-400 kgf.
Rigidity 54 kgf/um on 32 mm ballscrews.

Moral:
!!!
My C7 very thick ballscrew with Very Rigid mounts is not too accurate, at all.
If I try to make a workpiece 100.000 mm long, pre-calibration, it might be off 0.05 mm in length, +/-.
About 0.1 mm max error in length, in other words.
But..
after making one similar piece, and adjusting for the screw error and motion-control train error..
all the next pieces will be about 0.004 mm accurate in length.

[Bob La Londe]

Moral:
!!!
My C7 very thick ballscrew with Very Rigid mounts is not too accurate, at all.
If I try to make a workpiece 100.000 mm long, pre-calibration, it might be off 0.05 mm in length, +/-.
About 0.1 mm max error in length, in other words.
But..
after making one similar piece, and adjusting for the screw error and motion-control train error..
all the next pieces will be about 0.004 mm accurate in length.

Great post. I've found similar results although perhaps not quite as precise. In a recent project I had to make sets of steel plates to make a box that fit together with a tapered tab and slot. The tapered side of the slot forced the resulting box to pull together to a fixed dimensions as the plates were forced against the flat side of the slot. First the flat side had to be snuck up upon to get the dimensions, then the taper to get the proper fit. For each size the first piece had to be found by progressive light finish cuts, but its mirror piece was done simple by incorporating the finish cuts exactly into the code. I found this simply could not be done identically from one size box to the next. The results were close but measurably different when going from making a 25mmm x 35mm box to making a 135mm x 200mm box. (Fortunately it was a modestly well paid job. Unfortunately it took me a long time to come to grips with the solution. One of those jobs that turned out to be more difficult than it looked. Seriously, how hard is it to make a box. LOL.)

Fortunately allowable tolerances for the application were not particularly critical. They were all within easily measurable ranges. Only fit up was somewhat critical. It had to fit up and require some screw tension to pull the flat surfaces tight to each other. I found this rather frustrating. I had to use machine to spec methods to achieve a machine to fit result. I am accustomed to working either way, but it was some what difficult for me to incorporate both ways of thinking into a single project.

Thank you for your post. Its good to remind myself that while I sometimes struggle with light weight machines to achieve consistent repeatable thousandths or occasionally half thousandth others are repeating to ten thousandths every day.

[syscore]

I have found mine (linear) to be within a thou. I was actually surprised they worked that well, but I still touch off. As one poster stated, if you are doing parts start-to-finish from raw stock using a setup from Fusion-360, a thou or two (or even 5) is close enough. Still, touching off just once after powerup takes so little time and gives some piece of mind. Especially if your setups change often.