#hazzy Logs

Jul 05 2020

01:09 AM Joco[m]: Lcvette: well looks like I'm back to the drawing board. Just finished running a part and all outer dimensions are about 0.25mm too large. Not the end of the work as will put in big mill and bring it to spec. But still more work to do. I'm wondering if this reflects the leadscrew scale more than backlash
01:10 AM Joco[m]: * Lcvette: well looks like I'm back to the drawing board. Just finished running a part and all outer dimensions are about 0.25mm too large. Not the end of the world as will put in big mill and bring it to spec. But still more work to do. I'm wondering if this reflects the leadscrew scale more than backlash
01:22 AM Joco[m]: * Lcvette: well looks like I'm back to the drawing board. Just finished running a part and all outer dimensions are between 0.2 and 0.3 too large. Not the end of the world as will put in big mill and bring it to spec. But still more work to do. I'm wondering if this reflects the leadscrew scale more than backlash
06:02 AM JT-Cave: morning
06:02 AM JT-Cave: you can map the leadscrew
06:03 AM Joco[m]: morning JT-Cave
06:03 AM Joco[m]: JT-Cave: how would I map that lead screw short of getting a DRO and glass scales mounted on it
06:05 AM JT-Cave: 123 blocks dial indicator
06:05 AM Joco[m]: lots of them?
06:06 AM Joco[m]: I have 4
06:10 AM Joco[m]: JT-Cave: or do mean to lock em in place, do measurements then keeping the last one locked in place, "bunny hop" the others forward and lock em in place to get the next set of movement references. rinse and repeat until get to end of travel
06:12 AM Joco[m]: with reference to my message about over size. I ran the second part with all the same settings/CAM and the dimensions were all consistent with the part I was referencing. So at least the mill was being consistent.
06:26 AM JT-Cave: yea start at some fixed place and use the 123 blocks as spacers if you have a 1" dial indicator you can map quite a ways then as you say lock the last one
06:30 AM Joco[m]: JT-Cave: cool. I think I will try some simple tests. including seeing if the travel variation is consistent over a longer distance.
06:33 AM Joco[m]: I am also wondering if I should tighten gibs a little. At the moment if I put locking pliers onto the X ballscrew, set at 45 deg up from horizontal, they will drop turning the screw with their own weight. That might indicate things are just a little too loose
06:45 AM Joco[m]: right - sleeps. Nite all.
07:05 AM Lcvette: Bugger all... Man that is some issue you are fighting...
07:05 AM Lcvette: I have never seen a machine make different sized parts like that, typically if it's off it is consistently off
07:06 AM Lcvette: You have a moving target
07:07 AM Lcvette: I would start with the only thing I know is definitely different and that is your fixed bearings
07:07 AM Lcvette: Those are the foundation of the screw mounting so should be looked at
07:08 AM Lcvette: Next is the screw itself
07:08 AM Lcvette: And finally the ballnut mounting
07:08 AM Lcvette: Then look at the motors
07:09 AM Lcvette: Losing steps, although generally lost steps show as undersized parts
07:09 AM Lcvette: And once all that is sorted look at the gibs
07:09 AM Lcvette: Perhaps you are getting binding at different spots
07:11 AM Lcvette: 0.3mm is 0.011" that is absurdly off
07:12 AM Lcvette: Screw mapping will not be 0.011" off
07:12 AM Lcvette: To allow for that type is lead error the clearance in the ballnut balls would be obnoxiously large
02:23 PM Joco[m]: <Lcvette "I have never seen a machine make"> that is what has me wonder. Did the test part with you and your saw the photo of what I was cutting. sizes were pretty good. Then went on to make two "real" parts. I made two, one after each other and while the dimensions were off from the model they were off (from what I could measure) pretty much exactly the same amount on both parts. Hence why I am wondering about the steps
02:23 PM Joco[m]: per unit.
02:24 PM Joco[m]: by "same amount" I mean the differences arounf its various dimensions seemed the same.
02:25 PM * Joco[m] uploaded an image: ima_1a9a36e.jpeg (265KB) < https://matrix.org/_matrix/media/r0/download/matrix.org/GBslKJFSOyfHrTDctcpbUUDa >
02:26 PM Joco[m]: These are the parts I'm making. So not particularly complicated.
03:38 PM -!- #hazzy mode set to +oo by card.freenode.net
03:51 PM -!- #hazzy mode set to +o by card.freenode.net
03:53 PM Lcvette: Once steps per is set it should be consistent
03:53 PM Lcvette: It cam be consistently off but should repeat
03:53 PM Lcvette: Yours its not repeating results
03:56 PM -!- #hazzy mode set to +o by ChanServ
04:01 PM Joco[m]: That’s what I am going to look at tonight. Going to tray and gets some longer measurement distances in place. Plus I am seriously considering putting my 5u scales back on.
04:02 PM Joco[m]: But will see what I can assess using stacks of metric 123 blocks. I have 4 of them so can get a 300mm run to check what happens.
04:03 PM Joco[m]: I have already tightened up the tension on the thrust bearings. So no material play there.
04:04 PM Joco[m]: I will recheck the end plates to make sure they are bolt solid. Pretty sure they are.
04:08 PM Joco[m]: Will also double check motor to screw coupling is tight. After that gonna measure sections of travel from a known start. Essentially as if was mapping the screw. But will let me see if the steps/mm is correct or needs adjusting. Will be testing for repeatability on these. Cose if I am getting that consistent then I guess it comes down to something flexing like bugger under load.
04:09 PM Joco[m]: Lcvette: oh started thinking through the plugin on train trip this morning.
04:30 PM Lcvette: Not flexing 0.011" and way to far out for screw mapping to be the culprit
04:30 PM Joco[m]: Agree. But incorrect steps/mm could be
04:30 PM Lcvette: Checking step pet at different distances is a good idea
04:30 PM Joco[m]: Given. Both parts were the same post our setup process
04:31 PM Joco[m]: ie x and y dimensions on both parts were equal. So consistently out of spec. :-)
04:32 PM Joco[m]: I WILL NOT BE BEATEN! 😁
04:32 PM * Joco[m] can be very stubborn
04:35 PM roguish[m]: Joco: steps per unit distance should be simple math....
04:35 PM roguish[m]: not should be, IT IS
04:36 PM Joco[m]: <roguish[m] "Joco: steps per unit distance sh"> Agree. Just need to make sure backlash is not muddying the measurement.
04:37 PM roguish[m]: backlash is irrelavent in a single direction move.
04:37 PM Joco[m]: So will lead into the start point AND will push on the table to try and ensure all slack in nut is taken up before starting a measurement run.
04:38 PM roguish[m]: slop in the screw system is bad. a potential problem.
04:38 PM Joco[m]: Then will do this on the same distance run looking for consistency BEFORE adjusting the steps/mm
04:39 PM Joco[m]: The repeat post adjustment looking for consistency. Once I have that with confidence will look at slop/flex issues.
04:39 PM Joco[m]: Cose I will at least know the steps side is correct.
04:39 PM roguish[m]: try a constant force along the axis you are moving. tie a wight to the table and hang it over the end so it gives the table a constant force
04:40 PM Joco[m]: <roguish[m] "try a constant force along the a"> Nice idea.
04:40 PM roguish[m]: Z axes usually don't have backlash problems because gravity works.
04:40 PM Joco[m]: <roguish[m] "Z axes usually don't have backla"> Yup. Unless getting stickage/binding problems.
04:41 PM Joco[m]: But agree - Z less a problem.
04:41 PM Joco[m]: Especially if you have the vast weight Lcvette has on it. :-)
04:44 PM roguish[m]: anyway, not being critical, just helpful...... one thing that we all face is managing expectations.....and realities.
04:45 PM Joco[m]: roguish: all good. Not taking it as critical. It’s turning into a frustrating problem. So very open to all ideas.
04:45 PM roguish[m]: so, how's Kiwi land these days...????
04:48 PM Joco[m]: winter here. Wet, overcast. On the plus side, no social distancing anymore, no internal cases of COVID-19 (anything active is people returning from offshore) no one in hospital with it from memory. Professional sport play, stadiums packed with people. Pubs doing their thing. Other than no international tourists back to normal. Oh and the economy is probably fucked for the next 5 years. lol. BUT, I just paid off my
04:48 PM Joco[m]: mortgage a week ago. :-)
05:23 PM roguish[m]: and life is good.
06:02 PM Joco[m]: Yup. Pretty much not too bad at all really.
08:17 PM MrHindsight is now known as CaptHindsight
11:49 PM Joco[m]: nearly at the end of the business day