Back
[02:07:08] <DJ9DJ> moin
[05:48:43] <Loetmichel> mornin'
[06:06:24] <mrsun> direct belt drive or chain drive needs quite beefy motors to drive doesnt it? :)
[06:06:30] <mrsun> on router
[06:12:17] <archivist> mass being moved
[06:26:55] <Valen> you can still have reductions
[06:29:41] <archivist> reductions often add backlash
[06:30:18] <archivist> Valen, did some googling today my ball slides are rexroth
[06:30:31] <Valen> the name sounds expensive by itself
[06:30:57] <archivist> good job they were close to free
[06:31:31] <archivist> http://www.boschrexroth.com/various/utilities/mediadirectory/index.jsp?pagesize=8&oid=15107&language=en-GB&objectlang=en-GB&ccat_id=61&remindCcat=1&publication=1
[07:09:14] <mrsun> yeah directdrive usaly is backlash free isnt it ? but then you need heck of alot more torque then when doing it with screws
[07:19:59] <jthornton> someone posted a link to my web site here
http://en.wikipedia.org/wiki/G-code
[07:21:14] <jthornton> in the External Links at the bottom
[07:21:24] <jthornton> not sure what I think of that yet
[07:27:04] <archivist> happyness/pride :)
[07:28:24] <jthornton> yea
[07:28:54] * pfred1 sticks archivist with the gom jabbar and tells him that, pride is the little killer ...
[07:29:19] <jthornton> I was just looking at all the web sites that have external links to my web site and a bunch of them are not legit
[07:29:29] <archivist> incoming links can also mean work
[07:29:42] <pfred1> jthornton seems legit
[07:30:14] <archivist> a proportion of my incoming are from rubbish sites
[07:30:35] <jthornton> really, how does that work out?
[07:30:43] <pfred1> I counter that by not having a web site I find it an effective defence
[07:30:45] <jthornton> pfred1, the wiki is yes
[07:31:17] <pfred1> I've had web sites in the past but now I figure why bother
[07:31:32] <archivist> getting work
[07:33:05] <pfred1> well I finished up my lead screw stop collars yesterday
[07:33:25] <pfred1> I'm happy with how all of those turned out
[07:33:55] * pfred1 could say happyness/pride even :)
[07:38:15] <jthornton> posted my Anilam 1100m parts on PM about 30 minutes ago and someone wants them already :)
[08:58:17] <pcw_home> jthornton: do you still have the 100 Ohm resistors in series with the analog gnd?
[09:26:47] <JT-Shop> yes
[09:32:29] <pcw_home> A extra character means there was about 2V developed across the parallel cable from one end to the other
[09:32:30] <pcw_home> so that pretty severe EMI. This could be from coupling to nearby cables (power cables near the parallel cable)
[09:32:32] <pcw_home> or a ground loop
[09:34:50] <JT-Shop> the parallel port cable crosses the cables in the back of the PC but goes nowhere near the drives or anything else.
[09:35:53] <pcw_home> Is the PC case grounded or is it on a desk? (this would make the path to ground different than via the PC power cord)
[09:36:22] <JT-Shop> the PC case is sitting on rubber feet on top of the electrical cabinet
[09:41:15] <Tom_itx> morning
[09:41:27] <pcw_home> Might need a line filter or big ferrite bead (run PC power cord through it a couple of times) to sort this out
[09:42:39] <Jymmm> SWPadnos:
http://www.ebay.com/itm/Intel-DQ77KB-Thin-Mini-ITX-MB-LGA-1155-Core-i3-i5-i7-Dual-LAN-19V-BOXDQ77KB-/130765855617?pt=Motherboards&hash=item1e72409b81
[09:44:56] <pcw_home> Also on the the 7I77 (since you have an early one) can you check that S1 (marked 000) is present
[09:44:58] <pcw_home> this is a SMD resistor looking thing just to the right of W5
[09:47:55] <JT-Shop> tiny little bugger but yes it is there
[09:51:21] <pcw_home> Doesnt fix the noise problem but an extra char error really should not be a fatal error
[09:51:22] <pcw_home> I'll look into this next week. Might be SSLBP or driver issue not sure yet
[09:52:50] <pcw_home> Or is it fatal? does the I/O still work?
[09:53:15] <JT-Shop> it still seems to work after the error
[09:53:45] <pcw_home> OK so the driver/SSLBP handle it correctly
[09:54:30] <pcw_home> Do you have a scope to look for the noise source?
[09:55:40] <JT-Shop> yes, but I'm running out of time today I can put the scope on tomorrow
[09:55:53] <JT-Shop> what do I look at the 5v?
[09:57:27] <pcw_home> plug scope into same outlet as PC
[09:57:28] <pcw_home> put scope ground lead on PC case, scope input on 7I77GND
[09:57:46] <JT-Shop> ok
[09:59:12] <pcw_home> other random things to try:
[09:59:14] <pcw_home> common mode choke on PC power cord (giant ferrite bead)
[09:59:15] <pcw_home> common mode choke on 7I77 encoder and analog signals (giant ferrite bead per cable)
[09:59:17] <pcw_home> bond PC case to equipment case _and_ bond 7I77GND to equipment case
[10:00:04] <pcw_home> bond means low inductance connection = short and wide or multiple
[10:00:18] <JT-Shop> ok
[10:02:40] <DaViruz> bond, james bond
[10:02:41] <pcw_home> random stuff like this is very often ground bumping from the PWM driving the motors (and since the PWM frequencies are not synchronized, occasionally they add together to make a bigger noise pulse)
[10:05:45] <pcw_home> but 2V or so along a cable (6 or 10 ft) with 18 GNDs and a shield is pretty severe
[10:13:08] <archivist> I have had vacuum cleaners interfere with the scsi cable between cpu and disk unit, that was cured by bonding mainly
[10:14:08] <archivist> I wondered why my backups failed at 6-7pm ...mrs mop and the cleaner
[10:14:42] <pcw_home> Yeah at high frequencies all GND is is taps on a voltage divider
[10:15:49] <archivist> it is an antenna and an inductor and a common mode resistor and... :)
[10:16:11] <pcw_home> so you make the critical segment lower drop (bonding) or the other segments higher drop (common mode choke)
[11:04:52] <evokanevo> Greetings all. I noticed that enabling the on-board network chip and having a network connection (even with minimal network usage) bumps jitter in the latency-test from about 5 microseconds to 15 microseconds. I am wondering whether the rtai patch implements network polling for network drivers, which should shrink jitter compared to a purely interrupt-driven driver.
[11:05:47] <evokanevo> If the drivers aren't patched, I will have to patch the driver myself, but I thought if someone knows, it can save me time from looking into it.
[11:16:20] <alex4nder> I'd guess the network driver for your part is just going something dumb.
[11:21:39] <pcw_home> Wouldn't you have to poll in the base thread for that to actually work?
[11:25:15] <evokanevo> I think how it works with polling is that the IRQ for the eth0 goes off, and a callback is scheduled to do the rest of the network-stuff. This "other stuff" can be context-switched out if the base-thread needs to run.
[11:25:56] <pcw_home> I dont think that actually helps (at least on PC hardware)
[11:26:34] <pcw_home> cant context switch out a bus master transfer
[11:28:26] <evokanevo> I'm looking into how bus-mastering works now...
[11:28:33] <pcw_home> if it was scheduled at the end of the base thread then the problem goes away
[11:30:36] <evokanevo> So are these bus-transfers going to be an issue for CPU initiated transfers, or also device-initiated? I see that scheduling transfers for after the base thread would work if all transfers are CPU-initiated.
[11:31:04] <evokanevo> I don't know whether a device can initiate transfers whenever it wants. I figure that's DMA, but I don't know much about the technologies.
[11:31:22] <pcw_home> I suspect (but do not know for sure) that the block transfer which is all done in hardware on PCs is the main time sink (at least for PCI interfaced Ethernet)
[11:32:06] <pcw_home> The Ethenet chip itself usually manages ring buffers in system memory
[11:33:35] <evokanevo> Can the base-thread not run until a block-transfer finishes?
[11:35:13] <pcw_home> depends on the hardware (and in a real system the DMA may allow the basethread to run but it will be blocked as soon as it attempts to access hardware)
[11:36:35] <evokanevo> Okay thanks. I'm going to give the patch a shot, but at least now I have something to look into a bit more.
[11:37:25] <pcw_home> 10 usec does rather look like a 1.5K packet DMA time (PCI)
[11:38:01] <evokanevo> Maybe I could lower the MTU.
[11:38:51] <pcw_home> you might try pinging with different sized packets and see what effect that has on latency
[11:38:53] <pcw_home> to try and separate interrupt overhead and transfer latency
[11:40:05] <rizo> i would like to make my own hal module (as described in hal handbook) - it says i should place the component in /src/hal/components. But there is no such folder - should i install the developers version of emc?
[11:40:56] <evokanevo> Good point. I'll try that too.
[11:49:16] <archivist> would it be easier just to get a network card with its own on board memory
[11:53:25] <evokanevo> Seems like pcw_home knows a lot more about this than I do, but as I understand it, the problem is transfers to/from the ethernet chip occupy the PCI bus, which prevents other transfers on the bus requird by the base-thread. So assuming the card is also PCI-based, it will still create the latency.
[11:53:32] <JT-Shop> rizo: that is for a RIP install
[11:54:17] <JT-Shop> sudo apt-get install linuxcnc-dev
[11:54:36] <JT-Shop> then
[11:54:38] <JT-Shop> comp --install rtexample.comp
[11:54:46] <JT-Shop> opps sudo
[11:54:47] <JT-Shop> comp --install rtexample.comp
[11:54:54] <JT-Shop> sudo
[11:54:54] <JT-Shop> comp --install rtexample.comp
[11:54:56] <archivist> evokanevo, if it has its own memory not really as card to pc will be faster and not at packet speed
[11:55:01] <JT-Shop> why does it do that
[11:55:30] <JT-Shop> sudo comp --install rtexample.comp
[11:55:34] <JT-Shop> there
[11:56:57] <evokanevo> archivist - I assume that pretty much every network chip buffers the packet before writing to memory, but I haven't looked into it. 1 more thing to check.
[11:57:50] <archivist> I assume nothing with cost reductions these days
[12:10:45] <IchGuckLive> hi all
[12:14:11] <pcw_home> A lot of PCI Ethernet chips have rather modest buffers (just enough to accomodate bus latencies and one pkt)
[12:14:13] <pcw_home> the DP83816 for example only has 1 packet worth with of RX and TX buffers
[12:14:14] <pcw_home> but since PCI bandwidth is about 10X 100BaseT data rate this is not a real issue
[12:22:24] <pcw_home> RTL8139 is similar: 2K RX and TX FIFOs
[12:24:45] <IchGuckLive> are you searching for a new chip to fit the rs485 needs on a 1GEnet
[12:25:50] <archivist> no, discussing latency
[12:35:12] <IchGuckLive> i moff bavarian evening in town By
[12:47:41] <rizo> i compiled and installed the sincos.comp component in example. I can also loadrt the sincos module and wore the signals. The problem is tahet the component dos not calculate the value. Any idea why not? It seem there is a missing ".update base-thread" addf function.
[12:48:39] <rizo> i compiled and installed the sincos.comp component example code. I can also loadrt the sincos module and wire the signals. The problem is that the component dos not calculate the value. Any idea why not? It seem there is a missing ".update base-thread" addf function.
[12:58:47] <rizo> i thought the addf component.update is a default function of every component.
[12:59:36] <pcw_home> wheres the source to sincos?
[13:01:07] <pcw_home> I though you always needed the addf in the HAL file except for exotic "comps" like motion
[13:01:27] <rizo> http://linuxcnc.org/docs/html/hal/comp.html#_sincos
[13:03:41] <rizo> But the .update is not the part of component, it think it should be a part of HAL.
[13:05:16] <rizo> If i try the addf .update i get an HAL:ERROR: function 'sincos.o.update' not found
[13:05:30] <rizo> If i try the addf .update i get an HAL:ERROR: function 'sincos.0.update' not found
[13:08:16] <pcw_home> how about addf sincos.0 servothread
[13:14:08] <pcw_home> comps with xxxx.update like encoder have functions by that name
[13:14:16] <pcw_home> sincos does not
[13:16:11] <rizo> I understand that, so i should write the .update function/procedure - how?
[13:17:13] <pcw_home> its already there is just has a null name so in the HAL file:
[13:17:14] <pcw_home> addf sincos.0 servothread
[13:17:27] <pcw_home> should be right
[13:19:06] <rizo> i see :) this solved my problem. Thank you very much :)
[13:20:23] <pcw_home> welcome, and it not that I really know anything other than how to look at the source for the other comps and how they are invoked in HAL
[13:22:09] <pcw_home> http://git.linuxcnc.org/gitweb?p=linuxcnc.git;a=tree;f=src/hal/components;h=c592a0c058b541b29755199bd03bb6f9e655366d;hb=HEAD
[13:22:11] <pcw_home> for example
[13:25:53] <rizo> i see, thank you
[13:33:23] <andypugh> Unusually functional reprap item:
http://things.hands.com/thing/foetoscope/
[14:45:30] <Loetmichel> hmmm... is there any movement on the CAD front? any advice on free/cheap 3d soft that is useable by a year-long CorelDraw user?
[14:47:01] <archivist> open source CAD movement seems slow
[14:48:06] <toastydeath> slow/virtually nonexistant
[14:49:00] <archivist> "virtually nonexistant" is a little harsh :)
[14:51:40] <toastydeath> I CALLS IT LIKE I SEES IT
[14:55:02] <archivist> well I see a lot of commits on the BRLCAD project...its just not for drawing parts though
[16:21:36] <DJ9DJ> gn8
[16:53:05] <r00t4rd3d> Still need to clean it up, sand, stain etc:
http://i.imgur.com/c3QxD.jpg
[18:03:44] <Tecan> http://www.youtube.com/watch?v=2JtVh_9hPh8 << slo mo kittens through the air
[18:43:20] <tjb1> Im trying to tie all homing sensors to one input but it tells me that is already in use
[18:45:07] <andypugh> tjb1: Use the signal name, not the pin name
[18:49:44] <tjb1> andypugh: So i would use net-both-home-x
[18:49:59] <tjb1> Well that was the name made for x
[18:50:00] <andypugh> Yes
[18:53:01] <tjb1> I copied the x to the z and its just hitting home and saying it is on limit, it doesnt back off and search again
[18:53:12] <tjb1> What should i look at
[18:53:25] <andypugh> Set HOME_IGNORE_LIMITS in the INI file
[18:53:37] <tjb1> It is set to yes
[18:54:28] <andypugh> Bah! I can't remember the other one, this may take seveal seconds.
[18:55:25] <tjb1> The x will search quick, back off and go slow...the z just hits and says joint on limit
[18:55:32] <andypugh> 2.10.1 HOME_IS_SHARED
http://www.linuxcnc.org/docs/html/config/ini_config.html#_axis_lt_num_gt_section_a_id_sub_axis_section_a
[18:56:39] <tjb1> No switches are set though
[18:57:40] <tjb1> Oh its tripping x switch when homing z?
[18:58:09] <andypugh> Yes.
[18:58:24] <tjb1> So that should be 0
[18:59:31] <andypugh> Z hitting the home switch looks like X hitting a limit. And despite what the manual says, I think you want "1" or "TRUE" (and in all AXES)
[19:00:36] <tjb1> Neither works
[19:04:27] <tjb1> I found this "To use only one input for all homing and limits you will have to block the limits of the axis not homing with axis that is homing in HAL."
[19:04:42] <tjb1> Doesnt say how though
[19:09:59] <tjb1> I found it, JT-Shop shows it here
http://linuxcnc.org/lucid/emc2/emc2/index.php/french/forum/38-general-linuxcnc-questions/9633-combine-all-home-and-limit-switches?start=18
[19:11:58] <tjb1> JT-Shop: Helllp
[19:14:26] <andypugh> I am sure that it is meant to be a built-in possibility
[19:14:51] <andypugh> Do your axes move off the switches as the last stage of homing?
[19:20:25] <ReadError> tjb1
[19:20:30] <ReadError> got some prints done finally
[19:20:31] <tjb1> yes?
[19:20:49] <ReadError> https://dl.dropbox.com/u/25091878/Photo%20Nov%2010%2C%207%2015%2012%20PM.jpg
[19:20:59] <tjb1> Im banging my head with these electronics…I spent 3 hours to find out my 12v wall wart wasnt enough for my fans and homing switches :(
[19:21:20] <tjb1> Thats nice ReadError
[19:21:30] <tjb1> What are you printing
[19:22:04] <r00t4rd3d> a new house
[19:22:43] <r00t4rd3d> tjb1,
http://i.imgur.com/c3QxD.jpg
[19:22:45] <ReadError> still calibrating and stuff
[19:22:57] <tjb1> You made that r00t?
[19:22:58] <mutilator> pff i print out spaceships
[19:23:15] <r00t4rd3d> yeah
[19:23:17] <tjb1> JT-Shop: jthornton are you here?
[19:24:58] <r00t4rd3d> ive since sanded and cleaned it
[19:25:56] <tjb1> Has a cool font
[19:26:11] <tjb1> Be careful doing copyrighted items...
[19:26:45] <r00t4rd3d> i have no morals.
[19:27:00] <mutilator> lol
[19:27:09] <andypugh> Neither do lawyers
[19:27:09] <tjb1> Well when they rape you in court
[19:28:05] <r00t4rd3d> you can use stuff like that for personal use
[19:28:51] <r00t4rd3d> its when you sell stuff like that panties start knotting up.
[19:28:54] <t12> the only acceptable non-rights thing you're allowed to print
[19:28:57] <t12> is richard m stallmans face
[19:29:10] <tjb1> why I gave up vinyl cutting
[19:29:15] <tjb1> everything is against the law to make
[19:29:30] <tjb1> Well everything someone wants…can you cut me the chevy bowtie? no...
[19:30:05] <tjb1> ReadError: Is the perimiter to get the nozzle cleaned and purged and to make sure its extruding good?
[19:31:11] <tjb1> What is this "13 Mai 2011 04:26" is that supposed to be May?
[19:32:05] <andypugh> C'est francaise. Ou sont les mots en question?
[19:34:21] <r00t4rd3d> french is for fries.
[19:34:26] <tjb1> andypugh: Thats on
http://linuxcnc.org/lucid/emc2/emc2/index.php/french/forum/38-general-linuxcnc-questions/9633-combine-all-home-and-limit-switches?start=18
[19:34:32] <tjb1> Nothing else is in french
[19:34:50] <r00t4rd3d> you are in the french section
[19:34:52] <tjb1> Wait...yes
[19:34:56] <tjb1> Damn google :/
[19:35:03] <r00t4rd3d> look at your link
[19:35:45] <tjb1> Im gonna go before I look any more retarded
[19:35:56] <r00t4rd3d> is that possible?
[19:36:01] <r00t4rd3d> :)
[19:36:09] <tjb1> :|
[19:36:31] <r00t4rd3d> I buy the one Android phone you cant update/root/rom :(
[19:48:01] <tjb1> To use xor2 do I need to place "loadrt xor2" at the beginning of hal file?
[19:49:57] <andypugh> Yes
[20:05:55] <tjb1> ok andypugh I have another question…probably really simple
[20:06:10] <andypugh> 42
[20:06:11] <tjb1> To use xor2 in a second instance I need "xor2.1 correct
[20:06:39] <tjb1> xor2.1.in0, in1, etc...
[20:07:02] <andypugh> Yes, and that needs count=2 in the loadrt line
[20:07:14] <tjb1> Thats what I was looking for, thanks!
[20:09:18] <andypugh> Or, if you want, you can give them all names: loadrt xor2 names=obama,romney,palin than net something palin.in0
[20:09:20] <Tecan> http://youtu.be/3FpjcOWwiI4 << tesla coil fun
[20:09:35] <tjb1> loadrt xor2 [count=2] gives me empty variable name
[20:10:06] <andypugh> don't put the [] in
[20:10:42] <tjb1> just "loard rt xor2 count=2
[20:10:44] <tjb1> "
[20:11:19] <andypugh> Tecan: The guys should have been orange and green
[20:17:22] <tjb1> Ok andypugh, last time I promise...
[20:17:39] <tjb1> This works for homing both but now limits dont work -
http://pastebin.com/pt7jpcK6 the xors are at the end
[20:21:24] <andypugh> Is xor what you want? That seems like it will set the everything pin the instant homing starts, then set it to zero when the switch trips.
[20:22:28] <andypugh> I _think_ you want the homes to be straight through, and the limits to be AND NOT homing
[20:22:42] <tjb1> I was using what JT posted in that thread
[20:23:19] <andypugh> Yes, and as I said earlier, I am not sure he is right. I thought that all-shared was trivially configurable.
[20:23:41] <andypugh> Did you try all 4 combinations
[20:23:55] <tjb1> 4 combinations?
[20:24:05] <tjb1> Of the all-shared?
[20:24:31] <tjb1> Right now im only working with X and Z but I will have Y soon also
[20:26:16] <andypugh> There are 4 combinations of ignore-limits and home-is-shared
[20:29:00] <tjb1> Like ignore limits no, home shared 1 -- ignore limits yes, home shared 1 -- ignore limits yes, home shared 0 -- ignore limits yes, home shared 1?
[20:29:17] <andypugh> Yes
[20:29:45] <andypugh> home-is-shared is per-axis by the way.
[20:30:04] <tjb1> Well if it doesnt ignore the limts its going to trip all the time when homing
[20:31:24] <andypugh> I would expect so, but at least one LinuxCNC dev is barely competent (though I haven't touched homing)
[20:44:27] <Tom_itx> one switch for both functions?
[20:45:10] <Tom_itx> i haven't been following...
[20:47:30] <andypugh> Yes, one pin handling home and limit for every axis.
[20:48:54] <andypugh> I wanted
http://www.ebay.co.uk/itm/BT30-spindle-speed-multiplier-/181012927719?nma=true&si=PQpYFLIAvsBr4Fk6DPdQbZ4pUy4%3D&orig_cvip=true&rt=nc&_trksid=p2047675.l2557
[20:49:02] <andypugh> But not badly enough.
[20:49:46] <Tom_itx> but then Lcnc wouldn't know what switch tripped
[20:49:50] <Tom_itx> just that one did
[20:50:30] <Tom_itx> andypugh, a gearhead?
[20:50:41] <andypugh> I guess so
[20:50:51] <Tom_itx> why did you want that?
[20:50:58] <Tom_itx> i've never seen one
[20:52:09] <andypugh> My BT30 mill is geared for a top speed of 1000rpm, and a lowest speed of 47rpm. I have a VFD. Low gear, minimum frequency is more like a clock than a spindle (and has enormous torque)
[20:52:21] <Tom_itx> oh
[20:52:27] <Tom_itx> yeah i bet it has torque
[20:52:33] <Tom_itx> no way to regear it?
[20:52:46] <andypugh> Yes, and I intend to.
[20:53:06] <Tom_itx> may have to change bearings
[20:53:15] <Tom_itx> depending on how high you go
[20:53:25] <andypugh> The machine had two motor options and two head-gearing options from the factory. I seem to have the special slow one.
[20:53:47] <Tom_itx> for steel work probably
[20:54:32] <andypugh> But the bearings are the same (hand-engraved and signed in France, bizarrely) for all versions, so 2800rpm is fine
[20:54:50] <Tom_itx> yeah 2800 isn't bad
[20:55:19] <andypugh> I am looking out for an _enormous_ face mill to make the lower gears worthwhile
[20:55:22] <Tom_itx> i made a high speed pulley for my sherline as well
[20:55:30] <Tom_itx> apparently it's bearings are the same
[20:55:39] <Tom_itx> like 6"?
[20:55:55] <andypugh> No, bigger than that. 47rpm?
[20:55:59] <Tom_itx> hmm
[20:56:09] <Tom_itx> with 30 taper??
[20:56:20] <andypugh> There are dogs
[20:56:26] <Tom_itx> still
[20:56:39] <Tom_itx> side loads come into play with the tapers
[20:57:23] <andypugh> You are probably right. I am actually more likely to make a fly-cutter holder for one of my lathe tools.
[20:57:31] <Tom_itx> you may be limited what you can find with 30 taper too
[20:58:15] <Tom_itx> our big machines were 50 taper and you didn't really need to worry much about it
[20:59:32] <andypugh> I hadn't seen these until last week. They look quite nice. (They do bigger for more cash)
http://www.ebay.co.uk/itm/Fly-Cutter-for-Boring-Mill-Bridgeport-CNC-Verticall-Mill-Horizontal-Mill-Fadal-/320842397830?pt=LH_DefaultDomain_0&hash=item4ab3b27886
[21:02:53] <Tom_itx> yeah
[21:03:10] <andypugh> Video:
http://youtu.be/FMNkPCtkM80
[21:04:36] <Tom_itx> the only issue would be balance and vibrations but the rpm may not be high enough to worry
[21:04:56] <Tom_itx> especially on the larger diameter settings
[21:05:56] <Tom_itx> larger diameter the slower you have to go so i'm not sure how much you really gain
[21:06:44] <andypugh> Skimming the whole part in one pass, regardless of if it is slower than 2 passes looks a lot nicer
[21:06:58] <Tom_itx> yeah you don't get the blend marks
[21:07:25] * Tom_itx goes back to his movie
[21:40:42] <roh> hm.. what are typical speeds for stepper driven ballscrews on 'small' mills?
[21:47:42] <roh> i just wonder if i should invest time in tuning it even more
[21:51:37] <br87> hey, anyone awake?
[21:51:47] <br87> why does channel EMC point me to this channel?
[21:56:22] <roh> because linuxcnc was named emc2 for about a decade
