Haimer Measurement Probes

Michael,

Thanks for posting those Renishaw links

Our own equipment may be in a lower league, but it's always useful to see what is being done "at the high end".

The White Paper on angle encoders is extremely well written, and packs a lot of useful information into nine pages.

MichaelG.

Edited By Michael Gilligan on 01/08/2012 07:08:46

Hi Pete,

Further good information, thanks! Yes, I've read the manual, mostly makes sense but I'm still none the wiser on how to adjust the runout. The translation certainly leaves something to be desired; like understanding.

Both my CNC and vertical mill are R8, not ideal but that's what I've got. My Tormach CNC mill utilises what they call TTS 'quick change' tooling which uses a modified 3/4" collet. All end mills, drill chucks, ER collet chucks etc are mounted in holders that have a 3/4" parallel shank. They also have a shoulder that is pulled up hard to the spindle nose by closing the collet, giving repeatability of the tool height, unlike a normal R8 collet. I use the same collet system on my Bridgeport. The Haimer is fitted in a special holder that converts its spindle (10mm) to 3/4". Again not ideal, but that's life.

I've done some measurements on my Bridgeport mill with the Haimer. Truely awful! With the same procedure of measuring at the 4 quadrants I get a total runout of 0.12mm. If I turn the Haimer through 180° then the runout follows it. Runout of the spindle measured in the R8 taper is a bit under 0.01mm. However, measuring on the flat underside of the spindle nose I get a runout of ±0.01mm. By contrast I can't see the needle on the DTI move when measuring the underside of spindle nose on the CNC mill, so probably less than 0.001mm?

Simple geometry shows that an error of 0.01mm at the spindle nose multiplied by the length of the Haimer and holder (120mm) gives an error at the measuring ball of ....... 0.066mm. That seems pretty close to what I'm seeing, not sure it's a coincidence.

The next task is to check the tram of the mill, and if that is ok, contemplate skimming the bottom of the spindle nose to remove the runout.

The info on the Renishaw probes is interesting, but I'd like to make clear the Haimer is not a Renishaw style probe. If nothing else Renishaw probes are waaaaay outside my price range. The Renishaw website doesn't mention prices, so I can't afford them!

Regards,

Andrew

Hi Andrew,

A few thoughts that may be worth far less than you paid.

I bought mine a couple of years ago from a U.S. dealer where imperial measurements, tooling, and what I'm most comfortable working with are much more comman than most other areas of the world. The shank on mine is a true 3/4" and I'm using a Bison built end mill holder for mine.

After going back thru my Haimer manual and with a bit of thinking. Even though Haimer say's they should only be used with the same spindle you zeroed it to, just maybe zeroing your unit to the machine you use the most would be the best. Then with that preset zero, you could retest it in the second mill and record your deviation from a true zero in both X and Y? No it wouldn't be ideal since you'd have to always remember to compensate for that deviation on that machine. But these units are accurate enough that I think you could still depend on very good and accurate readings. Since I haven't yet used or tested mine on a second machine, I could easily have missed something with my logic.

Due to the way their designed, a dead true tram on both your milling heads probably isn't required. But I did do that while adjusting mine just as a precaution. I'll also 100% agree about the manual and Haimers description for zeroing with the built in adjustments. Frankly they could have done better. My manual shows the use of a standard dial indicator for setting that zero, since my .0001 reading indicators are the D.T.I. type, that's what I used. And I hope you've got a real good .0001 reading, or it's metric equivellent DI or DTI that your positive will give 100% accurate and repeatable results.

Your unit is newer and non digital, so the following may not match up exactly with what you have. The zeroing procedure in my manual do look to be the same for each type though. Setting for zero on these isn't something you want to rush. Maybe this will help you understand it a bit better? Think of the procedure as the same as what you'd do while setting an ultra precision ground rod in a lathes 4 jaw chuck for .0000 runout. It's basicly the same idea, except your now working vertical instead of horizontal, and with a lot finer thread pitch for those adjustments. From what I can tell, Haimer use an internal and fairly heavy spring loaded gimbal that the system uses to measure the probe movement. So, set up your DI or DTI with it's ball end centered as close as possible by eye to the center of the Haimers ball on the probe. Loosen all 4 of the Haimer adjustment screws, their located in the 4 holes just above the shank on your unit. (mine were pretty tight from the factory) Zero the dial on your DI or DTI. Ok, what you want to do is zero one axis at a time for either X or Y. We'll use X as an example. rotate your spindle by hand 180 degrees and check your DI or DTI indicator readings. Adjust those two Haimer adjustment screws till your indicator readings are the same each time you rotate the milling heads spindle by hand 180 degrees. Lock those screws down. Now do exactly the same for the Y axis screws. Your final check along with probably some very minor corrections on each of the adjustment screws is to rotate your spindle 360 degrees without getting ANY deflection on the DI or DTI indicators needle. It will take a bit of time to get this done. Good enough ISN'T for this job. You really do want .0000 if your patient enough. The really nice thing about this tool is that it doesn't depend on any real imperial or metric measurement system. Once the unit is properly adjusted for zero to your spindle centerline, then all your needing is to move your part edge for either axis till the needle or display reads zero. There's nothing to remember to do or compensate for like the standard offsets on the wobble type edge finders. Once your unit is properly adjusted for a dead on zero, then when the needle or display reads zero, that's exactly where you are. Your also not depending on these to "visually kick off" like the standard types do. Leaving aside a normal mills table flatness and overall machine rigidity. With one of these and a very good DRO, you can in a home shop replicate locational accuracys quite compareable to what some of the very accurate and more than very expensive jig borers like Moore Tools built could do.

Given that your also using CNC, overtravel well beyond the units safety limits could I guess easily happen. I hope you ordered or will now order a spare probe. You can't damage the measurement unit itself unless you had a sudden and unexpected Z axis crash, but you could break the probe due to that overtravel on either X or Y.

I really know less than nothing about CNC. But it's my understanding these can be used as a 3 axis probe system to manualy reverse engineer almost anything with a few different computer programs that are avalible. Yes it's much, much slower than something like a proper setup such as the Renshaw probe system uses. But it can be done with a lot more time involved.

You have my apologys for how long this turned out to be. But I hope this helps a bit.

Pete

Michael,

No doubt your correct. Since I'm not using CNC, I had to post how I set up mine the old school way. I envy those of you who have the knowledge and abilities to fully utilise what CNC can do. I was more than happy the day I installed a X axis power feed.

Pete

Hi Pete,

Thanks for the description of the zeroing method. I think it's beginning to fall into place. I'll have to re-read the Haimer leaflet in the light of your notes, which I think make things clearer.

The next task is to test the Haimer again in the CNC mill, but this time taking care to clean everything beforehand. If I get a reasonable result (say less than 0.02mm) I'm going to assume that in due course I can zero that out with the adjustment screws on the unit. Once I'm happy with the CNC measurements I can concentrate on the Bridgeport. I'm pretty sure the appalling results on the Bridgeport are due to the spindle nose, but there are some tests I can do that may prove that. After that it's a question of cleaning up the spindle nose on the Bridgeport so that it runs true.

Regards,

Andrew

Anyone got one of these from Arceurotrade? **LINK**

Martin.

No, but I have recently bought the equivalent from Haimer:

In due course, once I've tried it out, I'll report in.

Regards,

Andrew

I've made a few more measurements using the Haimer Zero Master; same technique as before.

On the CNC mill the maximium runout was 0.02mm. It is difficult to tell exactly, as I'm relying on the Mach3 screen readout for comparison. At each step one can hear the stepper motor clicking, and the display increments, but the Haimer needle doesn't move. After a few clicks the table might suddenly move 0.005-0.0075mm. I guess this is a static friction versus dynamic friction effect, plus a little bit of free play in the drive system.

I also did a measurement on the Bridgeport, but with the holder not pulled up tight on spindle nose. Total error in this case was 0.03mm. Again not easy to judge as the resolution of my DRO is 0.005mm.

From these tests it seems clear that the spindle nose on the Bridgeport is a bit out. As for the rest of the measurements I feel that I'm operating at the limits of my equipment. Time to cogitate; I feel that if I jump in a try to adjust the errors I'll end up chasing my tail as some of the 'errors' will be in the measurement system itself.

I have adjusted the Z axis zero on the Haimer and I tested that against a 10mm gauge block on the CNC mill, pretty much spot on, better than 0.005mm indicated.

Regards,

Andrew

Finally got to test the Haimer Centro this evening, now that I have a 16mm R8 collet to hand. I followed the instructions and centred on a hole in an aluminium plate. The hole had been previously bored in the lathe as part of a jig, so I am assuming that it is round to a better degree than I can measure. Once the movement of needle on the Centro was minimised I zero'd the DRO in absolute mode. I then replaced the Centro with a 3/4" collet and a short length of 3/4" silver steel. Having put the DRO in incremental mode I then zero'd on the hole centre by using my normal method. In other words, touch the silver steel on one side, by trapping a fag paper, and zero the DRO. Then touch the opposite side and use the '1/2' function on the DRO to go to half way, ie, the centre. Repeat for the other axis.

I then turned the handles until the DRO showed zero according to the Centro and switched from absolute to incremental mode to see what the difference was. The answer was 0.025mm in X and 0.05mm in Y. I would have hoped for better than that. The question is, which one, if either, is accurate?

Some possible errors are: the runout on the silver steel at the tip is about 0,01mm. Judging the degree of 'trap' on the fag paper is open to error. The fag paper measures 0.02mm thick. If you overshot the touch point with the silver steel slightly, one might change the alignment of said silver steel very slightly.

I've had a think about how I check the results. This is what I've come up with. I bore a recess in a flat plate on the lathe. I then use one or other of the centring methods to set the DRO at zero on the nominal centre on the mill. I then drill/bore a smaller hole right through with the DRO set on zero. Then replace the plate in the lathe and set it to run true using the original bored recess. The DTI can then be used to measure the runout of the smaller bored hole. This should tell me what the concentricity between the holes is, and hence the accuracy of the initial centring.

Regards,

Andrew

After a foul up at work today, on the part of the project manager, I found myself with a couple of hours to kill, so I decided to sort out the spindle nose on the Bridgeport mill.

First, I trammed the mill in X and Y. I don't seem to be able to achieve the miniscule figures that other people report, but I was happy with ±0.02mm over 200mm diameter in both X and Y. Next a test on the bottom of the R8 spindle confirmed that the nose was repeatably running out ±0.01mm. Using a carbide tool in the machine vice I skimmed the bottom of the spindle, like this:

After skimming the spindle runout barely moved the DTI, so around ±0.002mm? Using the Haimer Zero Master now gave an error of around ±0.015mm, as measured on the DRO, rotating the spindle round the four cardinal points relative to the first point. That is getting close to the limit of what I can measure, given that the resolution of my DRO is 0.005mm. The quoted accuracy of the Haimer is ±0.01mm so the results are probably as good as it is going to get, and more than adequate for my needs.

Regards,

Andrew

I've now also done the experiment using a plate with a hole bored in it to test the accuracy of the Haimer Centro. I did the test as described in a previous post. The result is that the second hole partially bored on the centre of the first, as determined by the Centro, had a TIR wrt the first hole of 0.15mm. Not good.

I then repeated the experiment, but using my method of a piece of silver steel, a fag paper and the '1/2' function on the DRO to determine the centre of the first hole. The TIR of the second hole in this case was 0.13mm. Not good either.

From these results my working hypothesis is that the Centro and fag paper method of centring are consistent, but there's a systemic error somewhere in the system to do with boring the second hole.

Regards,

Andrew

Hi Andrew,
You have probably done these tests but I will suggest them anyway. After boring the second hole did you check that the first hole showed as still being concentric while still on the mill. My line of thinking is that the cutting forces moved the table after setting things up.

Les.

Les,

Errr, actually no I didn't. I rather assumed that since the DRO still showed zero in X and Y, and both axes were locked during boring, the table wouldn't have moved. Nevertheless, it's a good idea, and will be the next thing to try.

Regards,

Andrew

Hi Andrew,
I didn't think about the fact that you would have noticed the change in the DRO reading. The other two possible causes are the work moving on the table which we can probably discount and the knee moving with respect to the mill body. The fact that the knee would be in a lower position when the boring head was fitted than its position when you lined things up with the short length of silver steel made me think of another possible cause. If the spindle axis was not parallel with the Z axis it would change the centre alignment at different knee positions. I suggest the following. Verify the concetricity of the first hole straight after centering it using a dial gauge mounted on the spindle with its probe the same distance from the spindle nose as the tip of the piece of silver steel. (I don't see that this can possibly be wrong.) Next fit the boring head but mount a dial gauge on it with the probe at the same distance down as the tool tip would be. Check to see if the first hole still seems to be concentric. If not there could be the axis alignment problem or the knee has moved the table in the X and/or Y directions when it was lowered. I hope you read this before trying to bore the second hole as it will save wasting any more bits of metal.

Regards Les.

A further update on the Hamier Centro: I have recently cut 63 and 69 teeth 6DP gears as part of a differential centre for a friend of mine, who is making the same traction engine as me. The jig I made to cut my gears uses the differential centre to hold the gear rings. My friend had machined all the parts, but not drilled and tapped the nine holes on a 8.375" PCD that hold the whole assembly together. So it seemed like an ideal chance to test out the Haimer Centro to set the DRO to zero over the centre of the assembly before drilling and tapping.

A noteworthy thing is that I got the needle on the Centro to move rather less than one division when centred. On my previous experiments with an aluminium test piece the best I could get was 4-5 divisions. A check with a length of silver steel in a collet in the spindle, and a fag paper, gave a maximium variation on the DRO at the four cardinal points of half a thou. That's worse than the Centro claims, but the 'trap a fag paper' measurement is a bit subjective. If I'm centred within half a thou I'd be pretty happy. The finish on the cast iron differential centre didn't seem much better than on my aluminium test piece, but clearly something was amiss with my test piece.

When I drilled and tapped the assembly I made a boo-boo and got the differential centre back to front, so on final assembly I turned it over. All the screws fitted, so I must have been pretty well centred. The screws are M6, and the clearance holes are drilled 6mm.

I've sort of been ignoring the Centro recently, but this exercise has given me some confidence in it, and a feeling that may be I hadn't wasted my money after all.

Regards,

Andrew

Edited By Andrew Johnston on 20/05/2013 20:26:05

Update: I have recently countersunk/drilled/tapped a set of 6 holes in a circular pattern in three parts that screw together at final assembly. The parts were drilled independently, using the Centro to pick up the central bore of each part. The holes are drilled 6.0mm for M6 screws, which fit in any orientation chosen at random. Overall I'm now pretty happy with the Centro and the accuracy I'm able to achieve with it.

Regards,

Andrew

Hmmmm, I'll bet he's never even seen a Haimer probe let alone used one.

Andrew