Fehlmann Picomax 51 CNC 2, reference info, pictures
Close
Login to Your Account
Results 1 to 14 of 14
  1. #1
    Join Date
    Mar 2017
    Country
    LATVIA
    Posts
    269
    Post Thanks / Like
    Likes (Given)
    70
    Likes (Received)
    143

    Default Fehlmann Picomax 51 CNC 2, reference info, pictures

    Since the other topic was about the manual version - the P50, I thought it would be more appropriate to make a separate topic for this version of the machine, since there are quite a few differences, idea of the thread is to provide some reference material since there seems to be not much about these machines on the web

    History of the particular machine, I bought it from an auction in Switzerland, just on a whim, out of curiosity more than a need (and because Milacron praised the quill of the P50 so high ), auction pics:
    10203002.jpg10203007.jpg10203006.jpg10203005.jpg

    Fehlmann support were gracious enough to send me users manuals for the P51 and the TNC155P, and they also were able to provide me with machine parameters for the TNC, all of which are available here:
    Fehlmann P51-CNC2 operators manual
    Heidenhain TNC155 operators manual
    machine parameters file

    The machine has 2 CNC controlled axis (X and Y, feedback via glass scales - Heidenhain LS 903), quill has a digital readout only (via a rotary encoder Heidenhain ROD 456 5000ppr), but has the normal automatic feed modes (coarse and fine, 3 feeds each, min 0,02mm/rev, max 0,30mm/rev), "milling mode" with a fine feed hand wheel on the left side of the spindle housing, and screw thread tapping mode with spindle reversal

    After some troubleshooting, mainly caused by me not wiring the 3 phase rotation correctly, and making sure everything on the machine works, it was time to take it apart for cleanup, in the next picture you can see why... X axis slide removed:
    img_20181226_215323.jpg

  2. Likes ballen, Colt45 liked this post
  3. #2
    Join Date
    Mar 2017
    Country
    LATVIA
    Posts
    269
    Post Thanks / Like
    Likes (Given)
    70
    Likes (Received)
    143

    Default

    so the X/Y table got torn down for cleaning (note the Y gib adjustment screws, 2 of them):
    b1.jpgb2.jpg

    bottom view of the Y slide, micro switches are aprox. 20 x 6.5 x 9,5mm, add-on lever actuator seems to be the JE-5 type (and they are pricey it seems, 13EUR/piece for a piece of sheet metal stamping and a roller...), and the pinout of the wiring connector for the end position and reference switches, there were some terminals on the connector connected with a jump wire, forgot to write down which ones or take a picture, but I'd guess they were 6.-7.-5.-8. - making it a common terminal for all switches and then monitoring which one brakes open (or closes and opens in case of reference switches)
    b3.jpgb4.jpg

  4. #3
    Join Date
    Mar 2017
    Country
    LATVIA
    Posts
    269
    Post Thanks / Like
    Likes (Given)
    70
    Likes (Received)
    143

    Default

    Y slide fitted back on the machine, black blocks are anodized aluminum lubrication distribution manifolds, they are sealed with O rings (9 of them, OD ~16,3mm, d 2,4mm), there are one way metering valve inserts screwed into the Y slide under the distribution blocks, "00" are designated as "slow" and "3/0" extra slow, picture shows their designated positions in the Y slide, info on the valves - http://www.idilsa.com/files/DOSIFICADORES.pdf
    c1.jpgc2.jpgc3.jpgc4.jpg

    this is it for now, more to follow when I get some free time to work on it

  5. Likes ballen, Mirek Mysicka liked this post
  6. #4
    Join Date
    Mar 2017
    Country
    LATVIA
    Posts
    269
    Post Thanks / Like
    Likes (Given)
    70
    Likes (Received)
    143

    Default

    So I put the machine back together and was writing the first programs to see how she cuts and got quite good results, there are some issues with ball screws, they are loud, and when I took them apart, the balls inside looked like this:
    img_20190421_212354.jpg
    screws themselves seemed fine, didn't see any wear, the nuts might need replacing because there was some odd looking areas near the return sections that looked quite rough (no idea where to start to look for them though, I doubt Fehlmann will have those at the price I would be willing to pay, even if they would sell them separately from the screws and sorry, forgot to take pictures), there are 3 circuits with internal returns in each of the 2 nuts for each screw, nuts are preloaded with a tightening nut via belleville washer, and if I counted correctly, 25 balls (3.000mm spacers and 3.175mm load bearing ones) in each circuit, total of 75 per nut, or 75 3mm and 75 3.175mm per axis, and the local bearing shop only had 120 of the 3.175 size in stock... so the X axis got the new balls first, after putting it back together it got very quiet, but when I try to preload the nuts I can feel some binding/crunching, so either there is damage to the nuts where balls return, or I made a mistake packing in new balls, will need to investigate...

    in the meantime I tried making some light test cuts and wrote first couple programs, here are the results (10mm 3flute R0.5, 6000rpm 500mm/m feed and 2mm step down), difficult to see in the pictures, but the finish is quite nice, side cutting is very smooth in the X direction, but has some ripples on the Y (I suspect because of the non-preloaded nuts), will see if I can improve that after repacking those lead screw nuts again to be able to put reasonable preload on them
    first_cuts.jpg

    And when I got bored of punching in the numbers by hand, decided to check the RS232 hoping to test some adaptive tool paths which I definitely don't want to punch in by hand, so I put in all the parameters as per the Heidenhain manual and couldn't get it to connect it to the PC (TNCserver), I could get the TNC to complain about erroneous lines when I put it into receive mode and sent some gibberish from the PC, so I put a scope on the Tx/Rx pins and the Tx signal voltage from the TNC to the PC only goes only from around -12V to -3V, from the PC to TNC signal level is -12V to +12V, used the same cable (just rewiring Rx/Tx) to connect 2 PCs together, and signal in both directions is -12V to +12V as expected and could send data in both directions.
    The definition of bits in the TNC->PC signal looks good, it is stable and everything, just that the peak level doesn't reach the minimum +3V for the receiver to trigger, almost like the output buffer (probably an opamp) only has -15V and 0V of power supply, instead of -15V and +15V, hence not being able to modulate the output to the required level. Haven't got around to opening it up to see what the circuitry looks like, and can't find any schematics for it either, otherwise the TNC itself seems to behave perfectly fine, just this RS232 signal level issue...

  7. #5
    Join Date
    Sep 2011
    Country
    GERMANY
    Posts
    2,552
    Post Thanks / Like
    Likes (Given)
    1546
    Likes (Received)
    867

    Default

    ... the local bearing shop only had 120 of the 3.175 size in stock...
    I've got a bag of 500 matched 3.175mm hardened steel G10 balls which I don't need, can you use them? These are round and accurate to 10 microinches = 0.25 microns. Drop me a PM if you want them.

  8. #6
    Join Date
    Mar 2017
    Country
    LATVIA
    Posts
    269
    Post Thanks / Like
    Likes (Given)
    70
    Likes (Received)
    143

    Default

    thanks, but I already had ordered a 100 online, and I think they've already arrived at the post office, just hadn't have the time to go and pick them up

  9. Likes ballen liked this post
  10. #7
    Join Date
    Dec 2002
    Location
    Benicia California USA
    Posts
    7,931
    Post Thanks / Like
    Likes (Given)
    1764
    Likes (Received)
    2288

    Default

    So i will relate what i have been told bu professional ball screw re-builders ("Express Ball Screw")
    First off, the first part of the screws that wears and fails are the balls (as you indicate)

    Second, the balls are "fitted" to the screw and nut.....That means that often the size is not an exact number but rather the balls are fitted and tested for resistance on
    instillation and the ball size is adjusted to give the required results of resistance and backlash.

    In the repair world , final ball size is usually established empirically. Re-builders have balls is steps 10 or 20 millionths to try ...the fit is everything.
    Too tight and mechanical damage will follow. Too loose and poor performance and early wear will result.

    Have rebuilt (re-balled) a number of ball screws and my technique is to purchase a range of balls. based on the original ball as removed and test assemble is steps until i get the fit i need.
    The size steps here are small enough that most shops can't accurately and reliably measure the differences of either the new balls or the originals........

    YRMV.....

    Earlier posting here:
    FP-NC Ball Screw Repair(photos/large)

    Note that the photos have disappeared, but think you can get the idea from the text.
    Cheers Ross

  11. Likes jz79 liked this post
  12. #8
    Join Date
    Mar 2017
    Country
    LATVIA
    Posts
    269
    Post Thanks / Like
    Likes (Given)
    70
    Likes (Received)
    143

    Default

    I've red all I could find here on forum about ball screws before went to work on mine, including your thread, thanks for all the info! I was a bit puzzled when I took these apart, I tried measuring the larger balls, but they are in such a poor condition that they can't really be called round any more, they are that bad... single ball may measure anywhere between 3.155-3.165mm, and since they were basically no help finding out what was the original size, I went with what was readily available - 3.175mm

    and the screws sounded awful in general, that was one of the main reasons (second only to chips being packed in every crevice) why I decided to take the machine apart, after installing new balls the X screw is almost silent up to like 3000mm/min, doesn't mean much since for now I left gibs and ball screw nuts almost loose, but still, huge improvement over what it was like before

    I have too little experience to judge if the 3.175 is the correct size, but everything went together ok, felt much tighter than the old setup, but zero signs of binding, the nut screwed on with no effort, almost loose (which I figured was by design, because there are 2 nuts under spring tension there to get rid of backlash), the only trouble was when I tried to preload the set, then it started to feel like the balls have hard time skipping over the crest of the thread on the screw, there was no sign of that with no preload, and I think I might have screwed up counting the balls when I fitted them, I might have put in too many, resulting in 3 large balls being next to each other
    anyway, next up is the Y screw, will pay more attention to the count of the balls and see how that one behaves with preload

    worst case scenario - look for new screws, they are 2505, so pretty common, the only good thing is that the feedback loop is closed via glass scales, so the precision of the replacement screws doesn't need to be ultra high, and the machine is not for daily production anyway, so basically any new screw will last me for years, what I do want to avoid in this case is having to machine original parts to fit new screws, that is out of the question, I don't want to be burned at a stake for such sacrilege

    there is another bearing related problem that I didn't notice before since I didn't run the spindle above 2000rpm, there are odd sounds coming from the top side of the spindle assembly at speeds from around 3000 to 7000, they are intermittent in nature, and there is no correlation between the sounds and finish on the part, no waves of streaks, nothing, but the load indicator of the spindle motor is in sync with the odd sound, there are tiny load spikes when the sound happens, it is present under load and no load conditions, changes somewhat with the quill extending out more, so I suspect they might be bearings that hold the "fixed" pulley that the belt is on, and not the spindle cartridge bearings themselves, unfortunately the manual Fehlmann sent me doesn't contain parts diagram for the spindle, so exloratory surgery will have to be performed at some time
    I thought maybe the drive splines might be dry, there is an oil line dripping oil at the very top to lubricate it, I flooded it will prescribed overall lubricating oil and that had zero effect on the sound, when the spindle is off and I turn it by hand, there is absolutely no sign of any irregularities in the way it rotates

    and Ross, you were right about that 8000rpm tacho generator on a 9000rpm spindle, the motor has a bit larger pulley than the one on the spindle

  13. Likes AlfaGTA liked this post
  14. #9
    Join Date
    Mar 2017
    Country
    LATVIA
    Posts
    269
    Post Thanks / Like
    Likes (Given)
    70
    Likes (Received)
    143

    Default

    so I had a little spare time to tinker with the machine, 3,175mm balls arrived, removed the Y axis nuts and to my surprise all balls in it were equal size, a bit worn, but all the same size (~3.165mm), 23 balls per circuit, 3 circuits per nut, 2 nuts per axis = 138 new balls were put in, screw installed in the machine - and no crunching noises I had with the reballed X nuts!
    this meant that someone before me had monkeyed with with the X axis nuts and put 2 different sizes in... took X apart, put new same size (3.175) balls in, preloaded slightly and no crunching, smooth operation, no jitter (previously TNC was showing some jitter when X was stationary, like 0,005<>0,000, now with light preload it is rock steady 0,000

    after this I decided to try some faster feed rates and got "gross positioning error" error messages on TNC when going above 2500mm/min, wasn't a milling operation, just table movement, so will have to investigate that a bit, but <2000mm/min - no errors

    and regarding the COM port signal level, I hacked together a quick and dirty level shifter with an opamp and an old ATX power supply to get the needed +/-12V power, and immediately got connection with the PC, could download/upload programs with no problems, so will need to open up the TNC up to see why it was outputting -12V;-9V impulses instead of -12V;+12V, so when started working it meant I could upload some CAM generated toolpaths!

    at first the control was stuttering, basically pausing after each block in that CAM generated tool path, some tinkering with the parameters got it almost smooth, found the solution in one of the cnc related forums where someone said these settings solved the stuttering on their TNC155 controlled Bridgeport:
    mp53 0.5 (default 3)
    mp60 1 (default 0)
    mp91 10 (default 2)

    pics of the lead screw arrangement, you can see the preload nut in the flange, the locking block for it, spring washer behind the flange that is pushing the nuts together, couple closeups of the nut ball races and a vid of the adaptive cut, the video is stuttering for some odd reason...
    img_20190512_205600.jpgimg_20190512_205614.jpgimg_20190512_203815.jpgimg_20190512_205116.jpg

    10mm Z3 R0.5 S6000 F750 ADOC10 RDOC2

  15. #10
    Join Date
    Dec 2002
    Location
    Benicia California USA
    Posts
    7,931
    Post Thanks / Like
    Likes (Given)
    1764
    Likes (Received)
    2288

    Default

    Internal ball return tracks similar to the ones used in the Deckel FP-NC ball screws.
    Wonder if you could take a bit of time and explain how the pre-load setup works.....I can see the toothed ring and retaining stop in the photos, can you explain how
    it works.
    Cheers Ross

  16. #11
    Join Date
    Mar 2017
    Country
    LATVIA
    Posts
    269
    Post Thanks / Like
    Likes (Given)
    70
    Likes (Received)
    143

    Default

    black flange is a retaining piece that the light part you called toothed ring is screwed into, and that toothed ring is what is putting the pressure on the belleville washer that is pressing the two nuts together, the two nuts have a key that keeps them aligned - that forms what basically is the classic split nut on a leadscrew, but instead of a wedge forcing the two halves open thus taking out the slop, the "two halves" in this case are forced together turning it into a zero backlash setup, and the spring washer helps with the uneven wear of the screw when it develops over time

    I guess the proper way to adjust it would be to screw the preload nut (toothed ring) all the way in compressing the washer completely and then backing out somewhat, but for now I adjusted it to have a light preload until I sort other things out

    the thread on that toothed ring is quite fine, maybe 1mm pitch, and there are like 36 teeth on it, and the retaining piece can be rotated and flipped, which adds 4 divisions for each tooth, so the Swiss thought that something like 0,007mm adjustment steps was what was necessary here

  17. #12
    Join Date
    Dec 2002
    Location
    Benicia California USA
    Posts
    7,931
    Post Thanks / Like
    Likes (Given)
    1764
    Likes (Received)
    2288

    Default

    Thanks for the explanation. I like the idea of the spring washer....allows some compliance.
    Assuming that the nut is fitted into a bore that has a shoulder at the end away from the tension adjustment.

    Cheers Ross

  18. #13
    Join Date
    Sep 2002
    Location
    Lawrenceville GA USA
    Posts
    6,175
    Post Thanks / Like
    Likes (Given)
    697
    Likes (Received)
    1315

    Default

    Great thread and interesting to read, thank you for taking so much time to document and share this.

    Charles

  19. #14
    Join Date
    Mar 2017
    Country
    LATVIA
    Posts
    269
    Post Thanks / Like
    Likes (Given)
    70
    Likes (Received)
    143

    Default

    small update:
    I was using Fusion 360 so far to generate code (Heidenhain conversational, not ISO!) for the control and there was only a generic TNC 155 post available, it was a bit pain to use, had to manually edit out Z moves and other stuff, but it worked, and I couldn't find a way to eliminate those moves while creating the tool paths in the CAM, so today I finally sat down and made some modifications to the post processor:
    1) all Z moves are filtered out, so even if the tool path contains them in the CAM, in the post they are removed, so be very mindful about using this post to avoid crashes and what not,
    2) 9998 mm/min rapids have been replaced with 2000, my machine doesn't seem to like moving faster than 2500mm/min, so by default the post processor will use the reduced value for rapid movements, this can also be done via parameters in the TNC, but I decided to edit that in the post as well,
    3) generic post processor used to use TNC canned cycles for drilling ops, I edited the PP so that it simply does puts out a move to the X;Y and then STOP command which allows for the manual Z op after which you press "start" to continue the program, a quick edit of the program on the TNC (M03 to M23 and appropriate RPM) allows to switch to tapping and use the same program to tap the holes if needed,
    4) tool changes - generic PP used to send the machine to home/tool change position, which is now removed, and there is now an M05; STOP sequence in the program before the TOOL CALL, which allows for manual tool change (be mindful where the spindle stop happens to avoid rubbing end mill bottom on a freshly cut material for instance), and after that only need to press START to continue the program

    warning, I haven't had much time to test the post processor for any glitches, so proceed with caution, if I find any bugs, I'll post an update

    link to post processor on google drive - Fusion360_p51tnc155_noZ.cps


Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •