#hazzy Logs

Jun 28 2020

03:25 AM Joco[m]: Lcvette: bit of advice needed. I tuned in the scale setting for the X and Y axis. Also rechecked back lash. Seemed to be all on the money. However just checked a part and its cutting oversize. I.e a dimension which should be 42mm is coming in at 42.12mm. Possible causes? Stickage? Lost steps?
03:59 AM Joco[m]: Oh - and the material was 6061 aluminium
05:16 AM JT-Cave: morning
05:20 AM Joco[m]: Morning JT
06:12 AM JT-Cave: you have backlash or no?
08:32 AM hazzy-m: morning
08:53 AM TurBoss: good morning
12:51 PM Lcvette: oo/
12:51 PM Lcvette: * o/
12:53 PM Lcvette: Joco: too size off slightly?
12:53 PM Lcvette: * Joco: tool size off slightly?
12:53 PM Lcvette: maybe backlash setting is a tad to aggressive
12:54 PM Lcvette: cam setting?
12:54 PM Lcvette: stock to leave is on in cam software?
12:54 PM Lcvette: could be a myriad of things
12:54 PM Lcvette: if the machine is moving bang on but the programmed part is off spec i would investigate cam
12:55 PM Lcvette: that tells me the machgine is dialed in but the toolpath is not
02:12 PM Joco[m]: Lcvette: thanks. I’m going to recheck all axis responses again. Just to double check. Pretty sure cam is ok re stock to leave. Will recheck on train trip to work.
02:12 PM Joco[m]: Question: how do you test/dial-in your hub settings? Ideally without having to disconnect ball screws.
02:13 PM Joco[m]: * Lcvette: thanks. I’m going to recheck all axis responses again. Just to double check. Pretty sure cam is ok re stock to leave. Will recheck on train trip to work.
02:13 PM Joco[m]: Question: how do you test/dial-in your gib settings? Ideally without having to disconnect ball screws.
02:32 PM Lcvette: Joco: i lock the Y axis so its solid
02:33 PM Lcvette: i then loosen x until there is some wiggle play and slowly tighten just until the wiggle play is gone then lock in place with the opposite screw, do not tighten that screw too much or it will bow the gib and cause more issues
02:34 PM Lcvette: then i run the axis to see how it sounds
02:34 PM Lcvette: with steppers you can hear changes based on gib adjustment
02:34 PM Lcvette: if it sound like it is struggling, i may back the gib out a 1/8th turn until it moves freely
02:35 PM Lcvette: the unfortunately truth to these small machines is not that they are low mass, but that the dovetails are fairly poor fitting
02:35 PM Lcvette: so getting them perfect is almost an impossibility
02:36 PM Lcvette: after x is adjusted to the best compromise of free movement and rigid travel movement, i move on to Y axis
02:36 PM Lcvette: same as above, this time lock X and asjudt Y
02:37 PM Lcvette: you can even put an indicator on the edge of the table and try twisting the table by grabbing both ends (twisting in a rotational manner mimicing the spindle rotational axis
02:37 PM Lcvette: last check is to see if stiction is an issue
02:39 PM Lcvette: if you have more time than money and want to learn to scrape, i have seen someone scrape in one of these machines
02:39 PM Lcvette: it looked painful to do
02:39 PM Lcvette: well beyond my patience level
02:40 PM Lcvette: but im sure it would make a remarkable difference in how well the gibs could be adjusted and how rigid the machine was
02:40 PM Lcvette: another thing to note
02:40 PM Lcvette: is i have noticed that when overtightening the ballnut retaining set screws the Y axis Saddle with deform cause even more give fitment issues
02:40 PM Lcvette: this is one place that the lack of a rigid casting actually make a noteworthy difference
02:41 PM Lcvette: the ideal solution is a linear rail conversion using Hiwin RG series 15mm linear ways
02:41 PM Lcvette: :D
02:41 PM Lcvette: \o/
02:45 PM Joco[m]: Lcvette: hugely helpful. Looks like i will have a busy evening.
02:45 PM Joco[m]: Back to work today.
02:45 PM Joco[m]: Monday
02:47 PM Lcvette: gib adjustment is quite infuriating
02:55 PM Joco[m]: As I’m finding out
02:57 PM Joco[m]: I might also have my micro stepping too high
03:04 PM Lcvette: whats it set too?
03:04 PM Lcvette: i gind usually 1/8th microstepping is a good balance
03:04 PM Lcvette: * i find usually 1/8th microstepping is a good balance
03:04 PM Joco[m]: 16x. So 3200 steps per turn.
03:05 PM Lcvette: if you want more resolution that 1/8th servos are in order
03:05 PM Lcvette: * if you want more resolution than 1/8th servos are in order
03:05 PM Lcvette: 1/8th gives you quite a bit more torque
03:05 PM Lcvette: are you running closed loop steppers or open loop
03:06 PM Lcvette: i don't remember if you told me
03:06 PM Joco[m]: 1/8 would be 1600. I can also set to 2000 which on a 5mm pitch ball screw gives a very clean 0.0025mm per step.
03:06 PM Lcvette: keep forgetting you are metric
03:06 PM Lcvette: lol
03:06 PM Joco[m]: open loop. I have some 0.005 glass scales I have been pondering about adding for some kind of feed back
03:07 PM Joco[m]: yeah. Metric. Inch does my head in.
03:07 PM Lcvette: if open loop definitely set the microsteps lower
03:08 PM Lcvette: you may be losing steps
03:08 PM Lcvette: what size motors?
03:08 PM Joco[m]: Neman32. 4Nm
03:08 PM Lcvette: nema234nm
03:08 PM Lcvette: * nema23 4nm
03:09 PM Lcvette: thats what i used to run
03:09 PM Joco[m]: You have? I’m definitely the next size up.
03:09 PM Lcvette: 570ozin
03:09 PM Lcvette: well there is nema23 and nema34
03:09 PM Joco[m]: think 4Nm is about that in ozin
03:09 PM Lcvette: im unfamiliar with nema32
03:09 PM Lcvette: yeah it is 566
03:10 PM Joco[m]: Then it’s nema 34
03:10 PM Lcvette: nema 34 is a 86mm x 86mm
03:10 PM Joco[m]: Short length and really fat
03:10 PM Lcvette: nema23 is 57mm x 57mm
03:10 PM Lcvette: what stepper drives?
03:10 PM Joco[m]: Think something like Dm870 or 860
03:11 PM Joco[m]: Leadshine style
03:11 PM Lcvette: digital?
03:11 PM Joco[m]: Yup
03:11 PM Joco[m]: Can take DC or AC in
03:11 PM Lcvette: ok those should be ok
03:11 PM Lcvette: they won't win any races as the stubby nema34's don't like to turn very fast
03:12 PM Lcvette: when you get a chance consider the upgrade to at least closed loop steppers
03:12 PM Lcvette: they make a pretty significant difference in performance and no more worries about lost steps
03:12 PM Joco[m]: i hinted about and these ones have good Henry values. (Think that is what is). About 2.5 or 2.6
03:12 PM Joco[m]: Feeding them 48V
03:13 PM Lcvette: i run closed loop at 60v
03:13 PM Joco[m]: They can throw the mill around if I wind up the max velocity
03:13 PM Lcvette: z axis is a nema34 8.5nm
03:13 PM Lcvette: think is a monster
03:14 PM Joco[m]: I debated that. But not had an issue on Z. But it’s never moving as fast. And I have the acceleration quite sedate.
03:16 PM Lcvette: do you have the setting set for G65 P?
03:16 PM Lcvette: * do you have the setting set for G64 P?
03:21 PM Joco[m]: G64 p?
03:23 PM Joco[m]: ah - found it. Think might have. somthign like 0,01 or 0.001
03:23 PM Joco[m]: mm
03:23 PM Joco[m]: * ah - found it. Think might have. somthign like 0,01 or 0.001 mm
03:24 PM Lcvette: ok needs to be set
03:25 PM Lcvette: or else it will not give you accurate parts
03:25 PM Lcvette: lets the toolpath wander
03:25 PM Lcvette: and really bad for clipping the top edges of parts on rapid moves when it traverses from one edge to another
03:26 PM Joco[m]: pretty sure it is. But maybe I need to set itup as standard preable in post processor
03:28 PM Joco[m]: where is that set? ini file>
03:28 PM Joco[m]: ?
03:28 PM Lcvette: i would
03:28 PM Lcvette: can be in ini file yes
03:29 PM Lcvette: hazzy's brother made a post that sets the G64 Pvalue to the fusion 360 tolerance setting
03:29 PM Joco[m]: G64P0.005
03:29 PM Lcvette: but i have run into some issues with the post not wanting to post some programs
03:29 PM Joco[m]: from ini file
03:29 PM Lcvette: looks good
03:30 PM Joco[m]: I have a custom fusion PP as well. I merge in updateds from the auto desk master periodically
03:31 PM Joco[m]: i use it to suppress G43 as at the moment I use G43.1 and a G43 clobbers that setting.
03:31 PM Joco[m]: when I get TTS holders will change things back.
03:32 PM Joco[m]: Atfer this weekend I can soooo see the attraction of TTS
03:39 PM Lcvette: lol
03:39 PM Lcvette: yeah they are seriously a no brainer
03:39 PM Lcvette: its nice having all your tools setup in fusion and in your machine control and just have to cam and probe and go
03:39 PM Lcvette: all is pre set up
03:43 PM Joco[m]: right - have 3/;4" MT3 collet on order and a CBN insert for lathe to be able to hard turn it down. Will sort some TTS holders next. Have that emai you provided as well. Do you remember if they have a minimum order?
03:44 PM Joco[m]: and tonight will be out with the Dial Indicator and checking movements and travel. Going to break out the 0,001mm DI and see what get some real accuracy on the movement. "Cose I can".
03:44 PM Joco[m]: movement in travel. Not gibs - that will be chasing taiol material
04:30 PM Joco[m]: Lcvette: did you get any answer re your entry mask question?
04:54 PM Lcvette: no :(
04:55 PM Joco[m]: I can have a look tonight when get home from work. To at least see if can figure out possible path forward. Unless someone else gets to it. 9:45am here.
04:59 PM Lcvette: cool that would be great!
04:59 PM Lcvette: :D
04:59 PM Lcvette: \o/
05:02 PM TurBoss: yop
05:03 PM Joco[m]: TurBoss: if your python foo needs some flexing then Lcvette has a problem for you. :-)
05:04 PM TurBoss: whats that
05:05 PM Joco[m]: Text format/mask in vcpLineEdit
05:05 PM TurBoss: oh
05:06 PM TurBoss: he was asking
05:06 PM Joco[m]: Like the vcpsettinglineedit does
05:06 PM TurBoss: ok
05:06 PM TurBoss: i have to take alook
05:07 PM Joco[m]: I’m about 8hrs away from having free time to do it. If you have time earlier to snoop would be nice.
05:07 PM Joco[m]: All I was going to do was snoop to see what was possible.
05:07 PM TurBoss: :)
05:07 PM TurBoss: i'll try
05:08 PM Joco[m]: What might need to be done to give same behaviour. But Monday here and I’m at the start of the work day.
05:08 PM TurBoss: don't worry I leave you the fun part
05:14 PM Joco[m]: oh cheers. Lol
05:17 PM Lcvette: sorry, lcvette was left with the visiting twin 5 year old girls
05:18 PM Lcvette: /o\
05:18 PM Lcvette: im getting the 411 on paw patrol
05:18 PM Joco[m]: ROFL
05:20 PM Joco[m]: Lcvette: back on topic of cnc adjustment. What sort of dimension tolerance do you shoot for on your grizzly?
05:21 PM Joco[m]: * Lcvette: back on topic of cnc adjustment. What sort of dimension tolerance do you shoot for on your grizzly? To give me a frame of reference
05:28 PM Lcvette: part dimension?
05:29 PM Lcvette: on flat edge to flat edge surface i can usually get them about +/-0.0005"
05:29 PM Lcvette: on holes, forget it
05:29 PM Lcvette: if you need something perectly round this machine with backlash comp and cheap screws will not do it
05:30 PM Lcvette: if it is critical you can under size the hole and bore it to size if it is critical like a bearing pocket etc
05:35 PM Joco[m]: <Lcvette "if it is critical you can under "> Yup - was the plan. Lathe or boring head.
05:36 PM Lcvette: yeah
05:36 PM Joco[m]: <Lcvette "on flat edge to flat edge surfac"> So 0.0127mm. Cool gives me something to aim for.
05:36 PM Lcvette: with backlash at all, it shows in the 90deg transitions in the work
05:37 PM Lcvette: alot of times if your all setup but getting dimension inaccuracies, it can be from the cutter moving the axis back to the backlash amount
05:37 PM Joco[m]: I was hoping to get to around 0.02mm on edge to edge dimensions. But we shall see. Might be a pipe dream.
05:37 PM Lcvette: so sometimes you have to play with backlash comp and find what will be most important to be tight for tolerance
05:38 PM Lcvette: holes and stuff from drilling do fine with backlash comp but exterior and interior pockets will tend to move x and y around
05:38 PM Joco[m]: The back lash idea is a good one. Will check it all without load and then will test making a square and adjust backlash until get it on the money. If that is the issue.
05:41 PM Joco[m]: I did check the part with calipers and it was bang on across the length. So within 0.02mm across the length parallel.
05:41 PM Joco[m]: would backlash shift not give an error from one end to other?
06:18 PM TurBoss: brb
06:29 PM Lcvette: not a pipe dream
06:29 PM Lcvette: but will take some tedious time to get i there
06:30 PM Joco[m]: I can be very patient. :-)
06:30 PM Lcvette: need to do external and internal tests
06:33 PM Joco[m]: Yup. Internal and external square shape. Was kind of my play. Bit of 12mm flat bar. Then 6mm down from top and do external square and an internal pocket then run normal chip load test cuts on inner and outer profiles to a target size. And measure for error. Need to snoop and see if I can make a quick custom UI to dynamically adjust backlash settings and scale settings.
06:36 PM Lcvette: geez that top hat is cooked
06:36 PM Joco[m]: top hat!
06:36 PM Joco[m]: * top hat?
06:38 PM Lcvette: lol.. wrong box
06:38 PM Joco[m]: lol
07:22 PM Lcvette: i think once you get it dialed in it will be fine for both
07:22 PM Lcvette: you have double nuts?
07:22 PM Lcvette: or single nuts?
07:22 PM Joco[m]: single. But backlash seems to be around 0.02 to 0.03
07:22 PM Lcvette: sorry i have this conversation lately alot and cannot remember who has what
07:22 PM Lcvette: on single nuts?
07:23 PM Joco[m]: Yup
07:23 PM Lcvette: guessing that isn't right
07:23 PM Lcvette: ive never seen a single nut have such low backlash
07:23 PM Lcvette: NEVER
07:23 PM Joco[m]: that’s what I want to retest tonight.
07:23 PM Lcvette: EVER
07:24 PM Lcvette: 0.002" has been the best i have ever seen on a single nut and closer to 0.0025-0.003"
07:24 PM Lcvette: for the normal average
07:24 PM Lcvette: thi thatn may be where your oversize is
07:24 PM Lcvette: not too much backlash but not enough
07:24 PM Lcvette: that is a good test
07:25 PM Lcvette: just run your test cuts and adjsut backlash + or - the amount of half your tolerance deviation
07:26 PM Lcvette: unless you spring for C5 or better screws
07:26 PM Lcvette: * unless you sprung for C5 or better screws
07:26 PM Joco[m]: Think best plan is to recheck scale and backlash without load. Get that looking ok. Then do tests on squares. Thinking 100x100mm
07:26 PM Lcvette: that seems like exactly the amount of backlash left on the part wall
07:27 PM Lcvette: humopr me and set your backlash at 0.002"
07:27 PM Lcvette: see if your toelrance deviation begins to go away
07:27 PM Joco[m]: C7 I think. Or whatsoever the rolled version is.
07:27 PM Lcvette: yeah no C7 screw has 0.0008" backlash
07:27 PM Lcvette: the lead error won't allow for balls to be packed that tight
07:28 PM Joco[m]: Yup. Agree it’s smelling wrong. But will just be diligent and go through the process to make sure have covered the bases.
07:29 PM Joco[m]: Also gong to bring my micro stepping down as well. Get some more torque.
07:30 PM Lcvette: C7 ballscrews have a lead error of +/-50 microns over 300mm
07:30 PM Joco[m]: Like the idea of x10 for us in metric land.
07:30 PM Lcvette: the ballnut and balls are packed to accept this deviation
07:30 PM Lcvette: otherwise they would bind and eat themselves
07:30 PM Joco[m]: So 0.05 backlash expected
07:31 PM Lcvette: 0.1
07:31 PM Lcvette: +/-
07:31 PM Lcvette: +0.05 to -0.05
07:31 PM roguish[m]: Joco: just lurking........ don't forget the machine rigidity / stiffness. EVERYTHING moves.
07:31 PM Lcvette: mew
07:32 PM Joco[m]: Yeah. BF20/G0704 naturally flexes.
07:32 PM Lcvette: everyone tells me that but i find its not as big of an issue as having accurate axis movement
07:32 PM Lcvette: can cut to spec if done right
07:32 PM Joco[m]: I think more an issue when getting into steel - maybe.
07:32 PM Lcvette: not gonna deflect much on a 0.005" finish pass
07:32 PM Joco[m]: Yeah - all about how much load you drive.
07:33 PM Joco[m]: Heavier cut so more load so not flex/vibration.
07:33 PM Lcvette: rigidity in my oprion doesn't come into play until you get into harder materials or are trying to move alot of material
07:33 PM Joco[m]: * Heavier cut so more load so more flex/vibration.
07:33 PM Lcvette: if you keep the machine happy while cutting it will hit specs
07:33 PM Lcvette: upset it while cutting and it will chatter]
07:33 PM Joco[m]: correct.
07:33 PM Lcvette: but its expected on a 400lb machine
07:33 PM roguish[m]: just another thing to keep in mind when you're looking really small numbers.
07:34 PM Joco[m]: Heck you can get that on any machine no matter how big. I can get my big mill into that situation if I over egg it.
07:34 PM roguish[m]: ok, back to lurking........................
07:34 PM Joco[m]: lurk lurk
07:34 PM Lcvette: some machines are stiff enough that tool deflection is the issue and not the machine
07:35 PM Lcvette: push a 40hp horizontal with a 1/2" endmill hard in a pocket and feel the floor, its scalloped like mad
07:35 PM Lcvette: 20k lb machine
07:35 PM Lcvette: but you come back with a lighter finishing cut and all is good
07:36 PM Lcvette: same thing on the grizzly but you just won't rough at 200ipm
07:36 PM Lcvette: lol
07:36 PM Lcvette: more like 40-50ipm
07:36 PM Joco[m]: Have been careful to work in realities of the BF20. Seem to be doing ok on that front at the moment.
07:37 PM Joco[m]: I was roughing at 300mm/min at 2700rpm.
07:37 PM Joco[m]: 0.04mm per flute
07:37 PM Joco[m]: Once I get higher rpm will be able to run faster.
07:38 PM Joco[m]: 1mm width of cut and 6mm deep. I tried 12mm on an 8mm end mill and not good. 6mm and it was happy as Larry. Could probably get to 8-10mm
07:39 PM Lcvette: https://www.youtube.com/watch?v=z7JMN_EYsWQ&t=150s
07:39 PM Lcvette: https://www.youtube.com/watch?v=z7JMN_EYsWQ&t=150s
07:39 PM Lcvette: machine is complete happy cutting at this speed and taking a healthy depth and width of cut
07:40 PM Lcvette: making nice thick chips
07:40 PM Lcvette: zero chatter
07:40 PM Joco[m]: Is that a 12mm end mill?
07:41 PM Joco[m]: 10mm?
07:41 PM Joco[m]: 3/8 or 1/2”
07:41 PM Lcvette: 3/8
07:41 PM Joco[m]: Ok 10mm
07:41 PM Lcvette: bread and butter size for this machine
07:41 PM Lcvette: YG1
07:41 PM Lcvette: best tooling ive found
07:41 PM Lcvette: carbide
07:41 PM Lcvette: alumapower
07:42 PM Joco[m]: About 1 or 2 mill width?
07:42 PM Joco[m]: * About 1 or 2 mm width?
07:42 PM Joco[m]: I’ve been using HSS. Starting to move to carbide.
07:42 PM Lcvette: i always rough at 15% tool diameter
07:43 PM Joco[m]: 1.5mm
07:43 PM Joco[m]: I was doing 1mm on an 8mm 3 flute HSS cutter. Worked really well
07:43 PM Joco[m]: 6mm deep was happy.
07:44 PM Lcvette: cut as deep as you can
07:44 PM Lcvette: i cut as deep as the flutes allow
07:44 PM Joco[m]: I tried 12mm chatter city.
07:44 PM Lcvette: and try and match tool cut length to the depth i will be cutting
07:44 PM Joco[m]: End when backing feed right back to
07:44 PM Lcvette: need better tooling
07:45 PM Lcvette: my machine prefers cutting at deep depths
07:45 PM Joco[m]: Yeah getting more carbide. But have a lot of HSS I want to slowly use up.
07:46 PM Lcvette: save it for crap material to cut
07:46 PM Lcvette: iron etc
07:46 PM Joco[m]: and steel. Is ok on steel at slower rpms with most
07:46 PM Joco[m]: * and steel. Is ok on steel at slower rpms with mist
07:47 PM hazzy-m: Lcvette: what are you using the line edit for?
07:47 PM Joco[m]: Have had good results with it.
07:48 PM Lcvette: Offset entries
07:48 PM Joco[m]: I do appreciate these discussions. Learn so much. All being filed away for use. :-)
07:49 PM Joco[m]: Lcvette: that feeds calc has been invaluable as well.
07:49 PM hazzy-m: Lcvette: will it trigger an action, or just be used for holding values?
07:50 PM Lcvette: It sends the data to the tool table
07:50 PM hazzy-m: so what actions does it trigger? <DI?
07:50 PM hazzy-m: * so what actions does it trigger? MDI?
07:50 PM Lcvette: Joco: great!
07:51 PM Lcvette: actually the data is pulled into a sub routine from a subcallbutton hazzy
07:51 PM Joco[m]: 😁
07:52 PM * Lcvette uploaded an image: Screenshot_2020-06-28_20-42-49.png (10KB) < https://matrix.org/_matrix/media/r0/download/matrix.org/EFkzsRpRHKkMUcJFzYZImUlG >
07:53 PM hazzy-m: hmmmm
07:53 PM hazzy-m: need to know the data type to be able to use a format ...
08:02 PM Lcvette: Number
08:02 PM Lcvette: 4 trailing zeros
08:02 PM Lcvette: 0.0000
08:03 PM Lcvette: neg and pos
08:04 PM Lcvette: Hazzy ^^^
08:06 PM Lcvette: Can you just add the text format option in the designer properties like the vcpsettingslineedits have
08:06 PM hazzy-m: working in it
08:07 PM hazzy-m: but it's not so trivia since the data type is not known
08:12 PM Lcvette: It's just a number
08:12 PM Lcvette: It's pulled into a sub routine
08:13 PM Lcvette: I don't understand what data type you need maybe expand AMD give me an example?
08:13 PM Lcvette: * I don't understand what data type you need maybe expand and give me an example?
08:14 PM Lcvette: Ie #<lineedit_value> = #1
08:16 PM Lcvette: Hazzy ^^^
08:47 PM Joco[m]: Lcvette: I think what hazzy is getting at is that in the generic case this edit field can hold any data ? Not necessarily a number.
08:49 PM Joco[m]: And I suspect hazzy is coding to cope with all cases not just a specific number case.
08:49 PM Lcvette: ah
10:38 PM hazzy-m: Lcvette: why don't you use an action spin box?
10:38 PM hazzy-m: its designed for numeric entries
10:39 PM hazzy-m: if you don't want the errors then you can turn them off in designer with the `buttonSymbols` property
10:40 PM hazzy-m: * if you don't want the arrows then you can turn them off in designer with the `buttonSymbols` property
10:41 PM hazzy-m: ah, looks like we need a QDoubleSpinbox based actions spinbox ...
10:49 PM Lcvette: oh i thought i needed to use a vcplineedit