[keen]

Hi Folks. In this video I am discussing tool and work offset errors, and ways around them. Including: Discussing why errors such as referencing errors are removed when resetting work offsets using a probe, but are not removed when resetting tool offsets with a ETS. Also some thoughts on mechanical toolsetters.

Please post your experiences and thoughts on the subject in this thread.

Cliff

[metalmayhem]

your tool setter should be sitting on a fixed surface , be it the fixed jaw , the table , etc , rather than the moving jaw .
A safer method to the piece of paper is to use a dowel pin , you can move below the top of the dowel , then move upwards until the dowel slides below the tool . Otherwise one simple error could push the tool into the part and damage the tool and or the part surface . Honestly I do both methods but the dowel method is faster and safer . Paper can also compress a certain amount whereas dowels won't

With the tts holder I prefer to pre-measure the tool height with a mitutoyo height gage . The main advantages are that a tool can be set up and measured while another is running . Also if a cutter breaks on one mill then I can quickly pull a matching cutter from another machine , and carry over it's tool height .

[jwatte]

I use a manual height gage (Starrertt) on the TTS granite plate with hole, and get quite good enough precision this way.
I think this is one of the best parts of using the Tormach system.
I know they have a USB option where you'd pipe the tool setter value straight into the offset table, but I've never felt the need for that.
I bought enough collets and holders that I mainly need to re-measure when I break/wear out a tool.

I also use a Hallmark ITTP for probing the workpiece, and that's very nice, too!
Before that I used paper (in fact, a $100 bill, for style points) and added 0.11 mm to the offset when it started to drag.
Also, if you do aluminum, you can simply jog and lower the head until it just scratches the surface, because the tool is sufficiently hard compared to the work (as long as you don't go too fast/deep.)

[popspipes]

One problem I have with the height gage is cold or hot temperature in the shop, the tool height will vary by .001 to .0015 depending on the temperature in the shop.

All the parts I make arent critical in Z , so it isnt a problem for me. I think the fix would be to measure tools at a constant temperature.

[jwatte]

Your machine likes a constant temperature, too. And even your machine operator might like a constant temperature -- long sleeves collect chips, and can easily get caught up in tools.
Treat yourself to proper insulation and a thermostat. Cheapest machine upgrade you'll ever make :-)

[metalmayhem]

a good height gage shouldn't fluctuate that much

[popspipes]

a good height gage shouldn't fluctuate that much

this is the tormach height gage, I will have to check it for the actual reading, my approximated reading may be way off......

[cncoperator]

Different steels have different coefficient of expansions but assuming the height gauge is steel, stainless 410 is one of the lowest at 0.0000055 in/(in °F) and with temp varying between 50 and 75 and an 8" long height gauge 0.0000055 * 25 * 8 = 0.0011 inch change in length just as popspipes suggests

[metalmayhem]

A decent quality precision instrument is not going to expand that much over a 25 degree difference . The same theory would mean that a .0005 tolerance bore which is at the bottom of the tolerance may be ok on a machine that is in a hot 100 degree sweaty shop , but may be out of tolerance once it is measured in a climate controlled qc office .

My shop varies a lot in temp over the winter and every few weeks I verify that some of my tools are still at the same height on the dual beam , and nothing changes unless the tool got bumped and pushed up into the collect (which is what I check for) .
The cheap Chinese gage that tormach sells isn't what I'd consider precision by any means , so if thats the type of height gage that is causing the problem then I wouldn't doubt it will vary under all circumstances .Even the speed of travel can affect those cheap scales

[mountaindew]

One problem I have with the height gage is cold or hot temperature in the shop, the tool height will vary by .001 to .0015 depending on the temperature in the shop.

All the parts I make arent critical in Z , so it isnt a problem for me. I think the fix would be to measure tools at a constant temperature.

Around 74-78 deg f. I see some precision improve with my own eyes. I assumed the machine lubricates and moves with little less stiction resulting in more accurate touch off using a hameir probe. If I use an electronic probe the constant movements from probe routines provide better accuracy for offsets at a much wider range of temperatures . As for offline height gauge setting or presetting can be fiddly and can effect z accuracy at any temp imho. If i set all tool heights for each day i use machine i find decent accuracy. If i run code and just pull tools off the rack i can expect less accuracy. One reason a while back i was asking about using an ets on mill to set each tool height before you run code. This would use the mill itself to set tool heights with code and machine movements imho making the process more repeatable and accurate.
Anyway I dont often need to get this fussy so this is not always a concern for me

[metalmayhem]

As for offline height gauge setting or presetting can be fiddly and can effect z accuracy at any temp imho. If i set all tool heights for each day i use machine i find decent accuracy. If i run code and just pull tools off the rack i can expect less accuracy.

Is this due to home switch inaccuracy , tts vs taper fitted holder ?? The switches are definitely something that float on these mills and is one of the reasons I rarely power down my mills . A 123 block measured on a height gage should measure perfectly every time , a tool in a holder should follow the same way .

[mountaindew]

Is this due to home switch inaccuracy , tts vs taper fitted holder ?? The switches are definitely something that float on these mills and is one of the reasons I rarely power down my mills . A 123 block measured on a height gage should measure perfectly every time , a tool in a holder should follow the same way .

I dont work from home switches. X,y,z are about 0.0015 to 0.0008 off after power up. Sometimes more depending on a number of factors.
I always use a probe and set new offsets after any restart. And I use tormach height gauge and set all tool z offsets in tool table. I have noticed this can vary for whatever the reason. Its not much in my world but in Uber precision work this would take time to get consistent. If you even can. The reason for the comment about using a ets on the machine is for more consistent setting by removing the human, plate, tool, to machine variance factor.
But Not many choices, from what I see in methods the only step up in offline tool measure is optical.

Power is far to expensive to keep machine running. Or anything else for that matter.
Not clear but I'm guessing it won't be long before many more people are under the same thumb!

[metalmayhem]

Power is far to expensive to keep machine running. Or anything else for that matter.
Not clear but I'm guessing it won't be long before many more people are under the same thumb!

it's not cheap here either but my mills don't sit idle for long , I weigh out time is money so I allow myself to justify the balance of the power cost I don't work from home position either but powering up the mills do require homing and I find the y switches to be inaccurate and I need to probe my mills on power up . The y switch inaccuracy is mostly due to the filthiness of my machines . The z is good enough and my parts require a skim pass with the facemill , so any inaccuracy in the z switch isn't a concern .

[jwatte]

A decent quality precision instrument is not going to expand that much over a 25 degree difference . The same theory would mean that a .0005 tolerance bore which is at the bottom of the tolerance may be ok on a machine that is in a hot 100 degree sweaty shop , but may be out of tolerance once it is measured in a climate controlled qc office .

Your part will absolutely expand by the coefficient of expansion times the delta in temperature.
This is why industry standard tolerance for machine shops is 0.005" and anything better starts costing more (and the better-er, the more-er!)
Very high precision parts come with QC schedules that specify inspection climate.
Very high precision mills and lathes come with built-in heaters/chillers for all the important parts. (Check out John Saunders' factory visit with the Kern Pyramid for some over-the-top precision stuff!)

Physics is physics. Might as well try to fight death, or taxes!

[jwatte]

Power is far to expensive to keep machine running.

Let's assume the average $0.11 per kWh. Let's assume an idling mill (spindle still, axles still, screen in power save) draws 50W for the controller and 50W for the cabinetry. Let's say you use it 8 hours a day, leaving 16 hours a day unused. That comes out to 18 cents per day, which is dumb to waste if you get nothing for it, but seems like reasonable cost of business if it means you get better precision.

That being said, I turn off my mill, too. Not only do I pay $0.34 per kWh, but I want to save the wear on the capacitors/power supplies and the fans, which will probably cost more than that per hour by the time they blow and need replacement!

Regarding limit switches, I got some high-quality Omrons with the same form factor as the Tormach Chinese switches, and will as a future project replace them to see if it makes any difference. (Well that, and the X axis switch casing caught on the Y bellows and cracked, so that needs replacement no matter what. The existing clearance there is not enough IMO!)

[mountaindew]

Jwatte
Assume 0.21 cents per kw +fees residential rate. And or about 2.00 a day to just let it set. On plus side It would keep Shop warm in winter months but requires ac in summer.
commercial power is much higher sliding scale and your on a demand read power meter system. Highest use day times 30days in month is your bill.

Also thunder storms for next 5 months will let the magic smoke out of your tools if left running and a strike hits

[kstrauss]

Regarding limit switches, I got some high-quality Omrons with the same form factor as the Tormach Chinese switches, and will as a future project replace them to see if it makes any difference. (Well that, and the X axis switch casing caught on the Y bellows and cracked, so that needs replacement no matter what. The existing clearance there is not enough IMO!)

Could you post the part number for your quality limit switches?

[metalmayhem]

[QUOTE=jwatte;2448734]Your part will absolutely expand by the coefficient of expansion times the delta in temperature.
This is why industry standard tolerance for machine shops is 0.005" and anything better starts costing more (and the better-er, the more-er!)
Very high precision parts come with QC schedules that specify inspection climate.
Very high precision mills and lathes come with built-in heaters/chillers for all the important parts. (Check out John Saunders' factory visit with the Kern Pyramid for some over-the-top precision stuff!)

Physics is physics. Might as well try to fight death, or taxes![/QUOTE


Yes it is physics and of coarse temperature has an affect on materials , I've had 2 stroke engines seized at -65C and were good again once warmed up .
Many of the high end machines will have temperature compensation but that isn't bullet proof . Typically chillers are for the spindles and I've never seen or worked on one with a heater . I've worked on high end and lower end equipment in extreme heat conditions (livable) on extreme tolerance parts , I've always had qc who would verify my measurements in climate controlled offices on their cmm's . A height gage is not going to stretch or shrink .0015 over a 25 degree change , but at the same time I'm not looking to argue the point . I know what works for me and it isn't necessarily going to work for others

[metalmayhem]

The power thing on 4 mills sitting idle is pretty insignificant and the computers suck most of that power . The main reason for me to leave them running is the ability to walk into the shop , load the mills and be running in minutes . VS the time it takes to fire them up , probe etc . One it's laziness , two I'm making money sooner while trying to gain time on a backlog of orders . The money spent on power while I sleep is recouped within minutes
Aside from the torus which is hooked direct - the mills are on surge protection

The omron switches which match up to the tormach kent switches are about the same quality . I've got some and replaced a couple broken x switches with them

[MichaelHenry]

(Check out John Saunders' factory visit with the Kern Pyramid for some over-the-top precision stuff!)

Physics is physics. Might as well try to fight death, or taxes!

John had a Youtube in the last day or so talking about the importance of temperature.