What's new
What's new

Positining error in new hmc

Nope. I actually have the bestest words, and knows how to useses them!

equilibrize - Google Search
I got nothing but learnt something new … lol

(third-person singular simple present equilibrizes... , present participle equilibrizing, simple past and past participle equilibrized) 1. To balance, or bring into equilibrium.

It sounded right ?

Always have to be an equilibrist and not an equiliberalist on this forum... shhhhh.
 
Seems I can't edit the original post to add, (D' oh),

@9106872762

Doosan HP 4000 M19 Issue

(from the forum);
^^^ That seems to be going in a positive direction in spite of difficulties.

And

Who makes or can make this type of machine

Who makes this type of machine discussion ? (also from the forum), Kinda scales versus temperature problems vs. on shop floor metrology versus full metrology department + corrections and various machines.

I'm sure there are many other good links that are relevant too.

@OP Machines with various types of scales (modern ones, not all , but some ) have temperature sensors in the scales/ scale assemblies and variously about the machine, but the compensations at best with a ten degree Celsius change* have a hard time keeping things under 8 micron shift over several hours with active automated thermal compensation on a machine from the control, at the best of times. larger machine, bigger distances, and much greater temperature differentials 30 micron could be your worst case "ball park" scenario, un gauged and un corrected. Two tenths over a thou ? 0.0012" / 30 micron ?

Some machines use a design approach where the scales expand at the same rate as the machine and if you are machining ferrous alloys then things can hang together more easily. For ferrous based scale, cast iron frame and bed and mostly ferrous based alloy parts to be machined.

Nickel alloy part, granite epoxy castings and glass scales ? Aluminum part , invar based scale, Concrete / aggregate base ? Quartz nano positioning scale … ceramic... Blah blah blah, The point is DOOSAN should be able to make or engineer a machine that thermally compensates fairly well ? They don't build rubbish ? , something seems wrong with "New Machine" maybe install / control settings etc. ? (I understand your environment is very challenging .). So does that mean you have thermal changes as high as 30 degrees Celsius I guess in the machine, as other wise if that was in the shop I think most people would die wouldn't they ? I know for example there have been terrible heat waves in Rajisthan but strangely the daily maximum and daily minimum (for example in Jaipur) rarely exceed 10 or 12 degrees Celsius, biggest difference being in the winter ? Not sure how that compares to California maxima and minima , or South Korea where the machines are built ?

The way you are cutting the parts are you putting a ton of heat into your parts ?


@9106872762
what is your biggest temperature swing ? Through a day or particular day of the year ?

Or is it literally your inspection department is at 21 degrees Celsius and your shop floor is at 40 degrees Celsius ? And your parts heat up by ten degrees through cutting ?

So comes back to the repeatability of the machine and the inspection department essentially indirectly calibrating your machine ?

I haven't done the research on shop floor based CMMs yet, but allegedly they are designed to take into account shop floor thermal conditions ? Up to a point ?



@9106872762 On one of threads cited above there was reasonably heavy finger pointing towards shaft wind-up on very long ball screws ? Stored motion and torsional springiness on a new machine versus a machine really well run-in and warmed up over a week at a time for a more normal run ?

_______________________________________________________________________________


* It's kinda Odd as John Harrison sort of solved this core problem nearly 300 years ago with a "grid iron" bi-metallic pendulum that uses dissimilar metals with different rates of thermal expansion to create a pendulum of constant length regardless of what the temperature is , so time keeping was accurate to within seconds a year ~ Eventually leading him to the development of marine chronometers and solving the problem of "Longitude" and accurate navigation at sea.

Gridiron pendulum - Wikipedia

John Harrison - Wikipedia

https://demonstrations.wolfram.com/GridironPendulum/

Here's one ^^^ with knobs on you can twiddle (imbedded app in web page) ^^^ (it's an interactive Harrison grid iron pendulum of constant length … ).


cameraman,

In india in summer temperature variation is highest. In summer at noon max tem. is 43' C and at night it will be around 30'C in Gujarat where i live and in qinter tem. variation is in small amount.

I only saw temperature control big machining workshop in MNC compny only here like SIEMENS,ALSTOM etc. But you have your own machine shop with 5-10 machines then it is not economical beacuse machining rates are very low here compare to other countries so it is a main problem. You can put small vmc in seperate room with A.C unit but machine big like doosan HM1250W it is not possible so even there is some problem in machine which cause error in accuracy, manufacturer first tell to control temperature and escape😂

we have one makino which is 25 year old and we bought it in second 13 years ago and we are running it now 24/7 since we bought it.Machine is box way and travels in x:1000mm , Y-650mm, Z - 600mm whithout scale. In this machine we bore two holes at 500 mm then we get readings in within 20 micron without temp. control . Thats why makinos are beast ...!!!

I have one more question,
when we bought this doosan hmc in report it said positining accuracy is 5 moicron / fulltravel and in accuracy test positining accuracy by boring

ISO standard is 50micron/320mm and
DOOSAN standard is 50micron/560mm

So what is the difference in position accuracy 5micron/fulltravel and and positining accuracy by boring 50micron/560mm? It shouldn't be same?

thanks
 
It seems to me that this thread and Can anyone explain how probing works?? deal with same machine and same problem, so I will refer to both of them here.
First of all some remarks to few statements.
1. “On machine probing is not applicable for quality assurance tasks.”

It is quite obvious, that if investment has been made and the probe is installed on the machine, it is at least used for part setting task – measure center of the bore, measure corner and set your WCS there, measure angle and rotate the coordinates system accordingly. All these activities (including not mentioned yet but widely used on machine tool setting ), are ON MACHINE PROBING, and importance of the reliability of the results of these measurements and their impact on parts accuracy is (I believe) clear to all.

2. “Probes are not accurate”.

Touch probes do not measure. The output from the probe itself, of from it’s interface, to machine control is normally closed or normally open contact, which is activated once the probe tip is deflected. At this moment the current machine coordinates are stored in register. The measuring means are machine axes, probe is responsible for generation of the trigger. Probe repeatability is important to assure, that the trigger is generated always under same circumstances (travel from the moment of physical touch until axes position noticed.

3. “CMM must be used in order to assure the quality of parts.”

This statement is based on theory, that CMMs are more accurate then machining centers. But they are not. Let’s compare the accuracy data taken in randomly taken from web catalog of CMM working horse Mitutoyo Crystal Apex S574 travel 500X700X400: 1.7+3L/1000, 3.9 micron for 700 mm axis, and from web catalog of working horse class machining center like YCM FP66A travel 510X660X400: 7 microns over whole travel. Both demand stable ambient temperature of 18-22 with gradient not exceeding 2 degrees Celsius per 8 hours.

4. “CMMs are periodically calibrated”.

This is really amazing argument. It is clear to all, that CMMs must be frequently calibrated. Why don’t you periodically calibrate production machines too. It seems just trivial to put all efforts to assure, that the part at the end of the production process is IN TOLERANCE (Process Control), rather than produce on uncertain machine, then measure it on CMM and reject it . In both cases only IN TOLERANCE parts will leave the factory, but the differences are obvious.
PROCESS CONTROL is the name of the game. Probing is a major player in this process. As shown in my previous post, as long as machine repeatability is acceptable (same results of probing same feature in short period of time), even not calibrated machine working in extreme environmental conditions will be able to produce IN TOLERANCE parts.

To the authors of the threads:
1. Perform the laser calibration of the axes of the machine.
2. Install the probe. Either contact DOOSAN, RENISHAW INDIA, and if not successful – me on PM.
 
It seems to me that this thread and Can anyone explain how probing works?? deal with same machine and same problem, so I will refer to both of them here.
First of all some remarks to few statements.
1. “On machine probing is not applicable for quality assurance tasks.”

It is quite obvious, that if investment has been made and the probe is installed on the machine, it is at least used for part setting task – measure center of the bore, measure corner and set your WCS there, measure angle and rotate the coordinates system accordingly. All these activities (including not mentioned yet but widely used on machine tool setting ), are ON MACHINE PROBING, and importance of the reliability of the results of these measurements and their impact on parts accuracy is (I believe) clear to all.

2. “Probes are not accurate”.

Touch probes do not measure. The output from the probe itself, of from it’s interface, to machine control is normally closed or normally open contact, which is activated once the probe tip is deflected. At this moment the current machine coordinates are stored in register. The measuring means are machine axes, probe is responsible for generation of the trigger. Probe repeatability is important to assure, that the trigger is generated always under same circumstances (travel from the moment of physical touch until axes position noticed.

3. “CMM must be used in order to assure the quality of parts.”

This statement is based on theory, that CMMs are more accurate then machining centers. But they are not. Let’s compare the accuracy data taken in randomly taken from web catalog of CMM working horse Mitutoyo Crystal Apex S574 travel 500X700X400: 1.7+3L/1000, 3.9 micron for 700 mm axis, and from web catalog of working horse class machining center like YCM FP66A travel 510X660X400: 7 microns over whole travel. Both demand stable ambient temperature of 18-22 with gradient not exceeding 2 degrees Celsius per 8 hours.

4. “CMMs are periodically calibrated”.

This is really amazing argument. It is clear to all, that CMMs must be frequently calibrated. Why don’t you periodically calibrate production machines too. It seems just trivial to put all efforts to assure, that the part at the end of the production process is IN TOLERANCE (Process Control), rather than produce on uncertain machine, then measure it on CMM and reject it . In both cases only IN TOLERANCE parts will leave the factory, but the differences are obvious.
PROCESS CONTROL is the name of the game. Probing is a major player in this process. As shown in my previous post, as long as machine repeatability is acceptable (same results of probing same feature in short period of time), even not calibrated machine working in extreme environmental conditions will be able to produce IN TOLERANCE parts.

To the authors of the threads:
1. Perform the laser calibration of the axes of the machine.
2. Install the probe. Either contact DOOSAN, RENISHAW INDIA, and if not successful – me on PM.

Cato the Elder this is great ! :-)

But what happens if the minimum sensitivity of movement of the behemoth machine is total pants ? (Raw friction and inertia, Stiction, shaft wind up, micro-lurches etc. ).

The machine in question weighs 66,000 lbs, and has box ways from Doosan.

AND does not have core cooled ball screws i.e. the ball screws on this machine are over 1.5 meter … Lot of friction + heat generated and the spindle on this is a monster thousands of pound's of torque. Most of the other Doosan horizontals have core cooled ball screws and half of them are using linear rolling element trucks and rails.

If something is loose then that needs to be fixed. (pointing the finger at double ended pre-load springs on both ends of the of pre tensioned ball screws on the casting that raises and lowers the spindle assembly and similar.). @45 degrees Celsius.

30 micron over 500 mm on a multimillion dollar machine that claims it can work in uncontrolled environments ? That's not great/ right …

As you mention, the Renishaw type ball bar plots and Laser sequences can diagnose a lot of problems with a machine.

But I would go one step further in this case and recommend that OP get the factory runoff inspection report for the machine to at least see what DOOSAN themselves consider to be "In tolerance".

That seems to be the key problem especially as the machine advertises itself as being really accurate and able to operate in hostile uncontrolled climates.
 
cameraman,

In india in summer temperature variation is highest. In summer at noon max tem. is 43' C and at night it will be around 30'C in Gujarat where i live and in qinter tem. variation is in small amount.

I only saw temperature control big machining workshop in MNC compny only here like SIEMENS,ALSTOM etc. But you have your own machine shop with 5-10 machines then it is not economical beacuse machining rates are very low here compare to other countries so it is a main problem. You can put small vmc in seperate room with A.C unit but machine big like doosan HM1250W it is not possible so even there is some problem in machine which cause error in accuracy, manufacturer first tell to control temperature and escape��

we have one makino which is 25 year old and we bought it in second 13 years ago and we are running it now 24/7 since we bought it.Machine is box way and travels in x:1000mm , Y-650mm, Z - 600mm whithout scale. In this machine we bore two holes at 500 mm then we get readings in within 20 micron without temp. control . Thats why makinos are beast ...!!!

I have one more question,
when we bought this doosan hmc in report it said positining accuracy is 5 moicron / fulltravel and in accuracy test positining accuracy by boring

ISO standard is 50micron/320mm and
DOOSAN standard is 50micron/560mm

So what is the difference in position accuracy 5micron/fulltravel and and positining accuracy by boring 50micron/560mm? It shouldn't be same?

thanks

@9106872762, You need to get hold of the inspection report for your machine - i.e. a specimen runoff /inspection report [accuracy book ] for the machine and then dig up your inspection report (factory runoff) for your machine.

That way you get to see how DOOSAN has exceeded whatever standards DIN, JIS, ISO etc.

Typically they will state the international standard (for that type of machine, which is probably a bit out of date ) and then show their internal standard and then show how the machine actually tests tighter for that.

Trying to interpret abstract occasional figures without seeing a whole inspection is tricky.

So you say 5 micron over full travel - That's a metric for straightness (which is largely independent of positioning accuracy - there is a cofactor but it's minor.).


So your laser tests will probably confirm that being the case or not, or at least get the measure of.

But 5 micron full travel or +/- 5 micron full travel, or +/- 5 micron unidirectional moves, +/- 5 micron bidirectional moves , or standard deviation aggregate of a sequence of moves +/- 5 micron and so on.

That's why you have to see an inspection report.

Positioning accuracy being different than straightness, sounds like from what you say they claim to position the center of the bore to within 5 micron basically two tenths and what you are experiencing is 3 to 6 times out of tolerance or out of tolerance maybe by a factor 5 ?

As PROBE and myself kinda bang on about establishing repeatability first.


Eventually sounds like DOOSAN India (like what PROBE is saying) may have to carry out a second systematic inspection report of your machine in situ to verify the install. Think of it like bringing in the "Millwright".

Maybe upload a video to show the procedures you are using to establish a 30 micron discrepancy - maybe at least rule out simple blunders like drooping indicators etc ? or systematic snafoooos on the control.

@9106872762 I'm sure you guys will be able to figure this out and get the machine doing what it says it does on the "Tin" / brochure.

It's a huge machine and is going take quite a bit of time to iron out the wrinkles, rather than just plonk it down and push the green button and it all goes "Lovely" (I know you know that) but wonder what DOOSAN INDIA do to make the machine right in terms of proper / full installation ?

66,000 lb machine... Even on a three point footed machine that could settle … IS the machine level ?

Bottom line is don't accept a 30 micron discrepancy in this case as "Normal" / OK. So it's not a Makino (as you say) but it's still a DOOSAN … i.e. Don't let your vendors BS you into the "Ohhh if you wanted higher tolerances you should have bought a YASDA or a MAKINO ". ~ argument.

It is possible to have an initial tolerance stack up of the order of 30 micron , but not for raw positioning.

And some discrepancies between dynamic tolerances versus static tolerances.

____________________________________________________________________________________________________


And YES I'm a huge Makino fan, it's pretty astonishing what some of their machines can do - even their most humble offerings. The Makino quoted tolerances and their margins to stay inside of, are pretty wide and very conservative, but essentially are intended for a 15 year + life span to hold those stated tollerances.
 
thanks all for your reply
yes it takes time to figure out problem

yes machine is big with x travel 2100mm, y and z 1400 mm and spindle with 50hp and main advantage of this machine is 300 mm quill so we can machining from all side without extra long tool and can make even small job easily.
we just completed leveleling and geometry setting second time about month ago for 7 days.

we are now talking to doosan and doing laser calibration in 1 or 2 week and lets see its effect on positining accuracy.
 
we are now talking to doosan and doing laser calibration in 1 or 2 week and lets see its effect on positining accuracy.

Considering all the money spent on the machine, and the delays due to inaccuracy and remediation attempts, I can't understand someone at least throwing a large tarp/tent over the thing and plugging in a couple room air conditioners with the heat exchangers ducted to outside the building.

Along with a dehumidifier it would cut down the humidity (a big deal in India), so lowering risk to electronics and reduce rust attack of the structure, along with improving consistency.

I'd guess under $2K USD for two A/C's and ducting, and a big tent or tarp. Maybe some lights needed if the tarp is opaque, but it could be clear plastic (just keep open flames away from it).

Under 1% (0.2%?) of the machine cost, and allows you to use it correctly. Seems a easy decision to me...
 
thanks all for your reply
yes it takes time to figure out problem

yes machine is big with x travel 2100mm, y and z 1400 mm and spindle with 50hp and main advantage of this machine is 300 mm quill so we can machining from all side without extra long tool and can make even small job easily.
we just completed leveleling and geometry setting second time about month ago for 7 days.

we are now talking to doosan and doing laser calibration in 1 or 2 week and lets see its effect on positining accuracy.

Hello in India. I happen to work for Doosan in the US. I have been paying close attention to this thread but since Doosan in India is a totally different company than Doosan America, we are not supposed to get involved. Looks bad on Doosan India I suppose? But, I decided to ask about this and was told about a similar issue with another Doosan HMC. This is something I would definitely look into when you have the machine looked at. It turns out that on a lot of machines, the spindle chiller turns off soon after the spindle is stopped. This allows heat to stay in the head and not get pulled out properly. The machine needed to have a ladder change, keeping the spindle chiller on longer. This could be part of what you are seeing. In an environment like India's, and in a non-climate controlled environment, this may be especially important to look at and ask about.

Just a friendly heads up and I hope this can be solved one way or another.

Regards,
Paul
 
Considering all the money spent on the machine, and the delays due to inaccuracy and remediation attempts, I can't understand someone at least throwing a large tarp/tent over the thing and plugging in a couple room air conditioners with the heat exchangers ducted to outside the building.

Along with a dehumidifier it would cut down the humidity (a big deal in India), so lowering risk to electronics and reduce rust attack of the structure, along with improving consistency.

I'd guess under $2K USD for two A/C's and ducting, and a big tent or tarp. Maybe some lights needed if the tarp is opaque, but it could be clear plastic (just keep open flames away from it).

Under 1% (0.2%?) of the machine cost, and allows you to use it correctly. Seems a easy decision to me...

In Japan they do thermal testing / micro climate control using PVC type curtains,

I installed two 4 ton units for approx. 3000 SF for about $16K

But wonder about fire risk like you say about large tarps etc.

Not sure how expensive electricity is in India but compared to a 50 HP spindle :-)

If you look at the Doosan PDF / brochure third line in big letters they claim the machine can regulate itself , and maintain stated accuracies in an unregulated environment.

Interesting that OP's temperature changes are only 12 degree Celsius in the shop. So some of the others (Okuma, Mazak and Makino) without scales can regulate to 8 micron (i.e. smooth and correct over a day) to stay within or under 8 micron drift so a two degree difference in Op's shop ~ 10 micron shift to be mapped out over an 8 hour period - With scales . I was surprised that some of India's temperature maxima and minima (in any given day) were so stable.

I guess they (DOOSAN) understand the machine is BIG and a high % of clients won't have climate control ?

But honestly @9106872762 you could probably climate control the space around the machine for less than the cost of buying your own Renishaw ball bar testers and laser interferometers and software ?
 
thanks all for your reply
yes it takes time to figure out problem

yes machine is big with x travel 2100mm, y and z 1400 mm and spindle with 50hp and main advantage of this machine is 300 mm quill so we can machining from all side without extra long tool and can make even small job easily.
we just completed leveleling and geometry setting second time about month ago for 7 days.

we are now talking to doosan and doing laser calibration in 1 or 2 week and lets see its effect on positining accuracy.

You might invest in Renishaw ball bar and laser interferometer test kit or similar.

I think PROBE had one for sale once ;-)

they're not cheap but waiting two weeks for someone to ball bar and laser etc. may be longer than you want ?

Maybe having in house ball bar testers etc. might enable you to track and monitor the "health" and drift of the machine.


__________________________________________________________________________________________


Mathematically - some of the applications we built we have what we call "Blunder detection" and that is usually triggered when any adjusted value or residual error exceeds two times a two Sigma deviation of a predicted error. So 30 micron seems firmly into "blunder detection" versus +/- 2.5 micron to 5 micron positioning.

That was a really nice hint from @Locknut - Sounds plausible there are a few kinks to iron out like what he was saying about ladder logic and the like and sequences for thermal control and even scale reading (perhaps).

Looks like a fun machine when it gets going.

______________________

** "Peeps" at MAZAk (as you mentioned MAZAK @9106872762 ) they basically impressed upon me that on some machines (smaller machines in this case) that linear scales might not improve individual part accuracy but will improve part to part accuracy / Cpks , can aid repeatability and stability of the process you are trying to build.

I think with the Doosan machine it's supposed to operate in a fashion that's comparable to a dedicated horizontal boring machine ? Given that the Doosan machine is expected to carry out a lot of boring operations ?
 
It seems like we should be able to thermal effects in and out fairly easily. Record the temperature of the machine (non-contact thermometer) and the material before you run a unit fist thing in the morning. Measure that unit. Do the same thing at the hottest point of the day. Steel moves 10 microns per meter degree C, so for a 500mm distance, it only takes a temperature difference of 10C to change that dimension by 50 microns. See if the problem correlates to machine or material temperature. Then go from there.
 
How is the part being held down? If you indicate a hole move to next hole and indicate does the control say the position is correct when in fact the CMM shows its wrong? If not I wonder if you are distorting the part , putting in the features and when you relax the clamping the holes are no longer in the correct position.
 
Another hint while we are on the subject and these are things I would ask your dealer rep. Is the thermal comp on and is it working properly?

Paul
 
Considering all the money spent on the machine, and the delays due to inaccuracy and remediation attempts, I can't understand someone at least throwing a large tarp/tent over the thing and plugging in a couple room air conditioners with the heat exchangers ducted to outside the building.

Along with a dehumidifier it would cut down the humidity (a big deal in India), so lowering risk to electronics and reduce rust attack of the structure, along with improving consistency.

I'd guess under $2K USD for two A/C's and ducting, and a big tent or tarp. Maybe some lights needed if the tarp is opaque, but it could be clear plastic (just keep open flames away from it).

Under 1% (0.2%?) of the machine cost, and allows you to use it correctly. Seems a easy decision to me...

You need more than 2K usd AC and big tent in India. First of all if the machine has 50hp spindle you need to prepare to cool off something like 130k BTU. Maybe more if you run it flat out.
Add to that 50m3 tent in a tropical heat and humidity..maybe actually only like 10k BTU based on some tent heating calculator online. But moisture is going to start condensing on the tent. Dew point above 21 degrees celsius is not uncommon in that part of the world. So add insulation to your list.

Me thinks that the parts are correct size out of the Doosan but at wrong temperature.. perfect temperature compensation on the machine itself won't help anything if the parts machined are at wrong temperature.
 
Me thinks that the parts are correct size out of the Doosan but at wrong temperature.. perfect temperature compensation on the machine itself won't help anything if the parts machined are at wrong temperature.

No, but these are mold bases so they're pretty simple geometry. No reason one couldn't have a half dozen programs pre-deformed for the measured stock temperature. Moldbase20, Moldbase22, Moldbase24 etc. Or do the same with macros if that's your fancy.
 








 
Back
Top