#emc | Logs for 2011-01-30

[00:05:39] -!- pcw_home [pcw_home!~chatzilla@ip-66-80-167-54.sjc.megapath.net] has joined #emc
[00:42:30] <skunkworks> imagebin.org
[00:42:51] <danimal_garage> photobucket.com
[00:51:43] <andypugh> picasa
[00:53:54] -!- andypugh has quit [Quit: andypugh]
[00:54:06] <alex_joni> picasa's bad for your connection
[01:01:17] -!- servos4ever has quit [Quit: ChatZilla 0.9.85 [SeaMonkey 2.0.11/20101206162726]]
[01:04:58] -!- robh__ has quit [Ping timeout: 255 seconds]
[01:08:52] <atmega> no clue
[01:21:09] -!- skunkworks has quit [Remote host closed the connection]
[01:22:37] -!- paul11211 has quit [Quit: Visitor from www.linuxcnc.org]
[01:27:30] -!- jdj has quit [Quit: Ex-Chat]
[01:50:42] -!- Valen has quit [Quit: Leaving.]
[01:53:54] <danimal_garage> hmm, odd.... a lot of my IO pins arent working with the servo hal and ini
[01:56:01] <danimal_garage> here's what i did... i made a new config folder using the 5i20 sample servo config. Then i coppied everything from my old stepper ini and hal that wasnt stepper related and replaced the stuff in the servo files with it
[01:57:04] MattyMatt is now known as matt[NaN]
[01:58:05] matt[NaN] is now known as MattyMatt
[02:05:32] <danimal_garage> ah i fixed it
[02:24:20] -!- kb8wmc [kb8wmc!~chatzilla@] has joined #emc
[02:27:17] <MattyMatt> I've snapped the cast iron drill clamp on my drill press. is it clever to arc weld it or should I wait until I can braze it?
[02:32:12] <MattyMatt> or should I leave it broken and duct tape the drill to the press? :)
[02:35:51] <MattyMatt> does anyone know which of the top two products on http://sra-solder.com/section.php/3/1 would be strongest in tension? (in a clean break in iron)
[02:36:34] <MattyMatt> paste looks good to me, as I only have a butane pen torch
[02:44:18] <MattyMatt> 6 hours till tools up. I'll weld it then unless someone tells me I'll ruin my press in the meantime
[02:44:55] <MattyMatt> wow 15 years since I welded. I'll practice on crap first
[02:53:27] -!- tlab [tlab!~tlab@c-69-243-187-136.hsd1.in.comcast.net] has joined #emc
[02:57:39] <MattyMatt> meh, that Cu-Ph paste is for brazing brass & copper. I know nothing about brazing
[02:57:59] <MattyMatt> welding it is
[03:06:35] -!- theorbtwo [theorbtwo!~theorb@] has joined #emc
[03:19:53] -!- tlab has quit [Quit: Leaving]
[03:29:31] <Tom_itx> cast iron is tricky to weld
[03:29:43] <Tom_itx> you need to bake it first to bring it up to temp
[03:40:27] <danimal_garage> hmm i'm having a problem. The servos jump a little when i enable the drive and it throa a following error
[03:40:50] <danimal_garage> throws*
[04:22:44] -!- paul11958 [paul11958!~paul11@c-98-237-170-144.hsd1.wa.comcast.net] has joined #emc
[04:24:09] <danimal_garage> yay all fixed. I was trying to turn on all the drives with one enable from the 7i33, and apparently it was just too much. i split them up and it works fine
[04:30:49] -!- paul11958 has quit [Quit: Visitor from www.linuxcnc.org]
[04:34:48] -!- garage_seb [garage_seb!~seb@174-16-125-123.hlrn.qwest.net] has joined #emc
[05:09:10] -!- nullie [nullie!~nullie@nullie.telenet.ru] has joined #emc
[05:09:10] -!- kb8wmc has quit [Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101206122310]]
[05:21:56] -!- bootnecklad [bootnecklad!~bootneckl@] has joined #emc
[05:24:10] -!- bootnecklad_ has quit [Ping timeout: 255 seconds]
[05:24:45] -!- nullie has quit [Ping timeout: 246 seconds]
[05:39:28] -!- rooks has quit [Read error: Operation timed out]
[06:07:09] -!- ve7it has quit [Remote host closed the connection]
[06:17:11] -!- mhaberler [mhaberler!~chatzilla@extern-188.stiwoll.mah.priv.at] has joined #emc
[06:19:36] -!- Valen [Valen!~Valen@ppp226-148.static.internode.on.net] has joined #emc
[06:25:54] -!- psha [psha!~psha@] has joined #emc
[07:04:34] <danimal_garage> hmm one of my resolvers is bad
[07:10:52] -!- nullie [nullie!~nullie@nullie.telenet.ru] has joined #emc
[07:21:01] -!- garage_seb has quit [Quit: Leaving]
[07:23:09] -!- mhaberler has quit [Ping timeout: 276 seconds]
[07:23:17] <danimal_garage> ouch! those effers are expensive!
[07:33:57] -!- bootnecklad_ [bootnecklad_!~bootneckl@] has joined #emc
[07:35:51] -!- bootnecklad has quit [Ping timeout: 240 seconds]
[07:48:48] <danimal_garage> sweet, the extra resolver from my lathe's spindle works with the drives, i just need to make a mount and coupler to hook it ip to the servo
[07:53:58] -!- nullie has quit [Quit: Ex-Chat]
[08:05:56] -!- Valen has quit [Quit: Leaving.]
[08:06:24] <Emcrules_Laptop> Danimal: late night or what!!!
[08:11:39] -!- ca-kc6lbj-2 has quit [Remote host closed the connection]
[08:13:03] <danimal_garage> yea, a little past midnight now
[08:13:07] <danimal_garage> time for bed
[08:13:12] <danimal_garage> at least i have a plan though
[08:13:53] <danimal_garage> nite
[08:14:21] -!- celeron55 [celeron55!~perttu@a91-155-42-12.elisa-laajakaista.fi] has joined #emc
[08:22:21] -!- pjm has quit [Quit: TTFO]
[08:28:32] -!- pjm [pjm!~pjm@host81-146-71-253.btremoteinternet-dsl.bt.net] has joined #emc
[08:34:50] -!- Emcrules_Laptop has quit [Ping timeout: 240 seconds]
[08:49:27] -!- psha has quit [Quit: leaving]
[08:49:35] -!- celeron55 [celeron55!~perttu@a91-155-42-12.elisa-laajakaista.fi] has parted #emc
[08:51:15] -!- vezza [vezza!~andrea__@host107-41-dynamic.21-87-r.retail.telecomitalia.it] has joined #emc
[08:52:41] -!- vezza has quit [Remote host closed the connection]
[09:05:58] -!- odiug [odiug!~guido@p57BD12B6.dip0.t-ipconnect.de] has joined #emc
[09:09:58] -!- robh__ [robh__!~robert@5ace70af.bb.sky.com] has joined #emc
[09:51:58] -!- psha [psha!~psha@] has joined #emc
[09:55:23] -!- Valen [Valen!~Valen@] has joined #emc
[09:59:53] -!- Valen has quit [Client Quit]
[10:10:28] -!- Valen [Valen!~Valen@] has joined #emc
[10:13:03] -!- HDB10 [HDB10!~howard@] has joined #emc
[10:15:58] -!- HDB10 has quit [Client Quit]
[10:17:35] -!- HDB10 [HDB10!~howard@] has joined #emc
[10:19:28] -!- nullie [nullie!~nullie@nullie.telenet.ru] has joined #emc
[10:19:42] -!- HDB10 has quit [Client Quit]
[10:55:21] -!- rooks [rooks!~rooks@102-bem-18.acn.waw.pl] has joined #emc
[11:05:40] -!- odiug has quit [Ping timeout: 260 seconds]
[11:11:17] -!- Dallur has quit [Read error: Connection reset by peer]
[11:11:38] -!- Dallur [Dallur!~root@157-157-118-204.dsl.dynamic.simnet.is] has joined #emc
[11:23:53] -!- KimK has quit [Quit: Leaving]
[11:32:32] -!- Dallur has quit [Ping timeout: 259 seconds]
[11:34:10] -!- Dallur [Dallur!~root@157-157-119-88.dsl.dynamic.simnet.is] has joined #emc
[11:39:41] -!- Valen has quit [Quit: Leaving.]
[12:07:17] -!- psha has quit [Quit: Lost terminal]
[12:24:10] -!- Athlocatle has quit [Quit: υΘΟΦΥ]
[12:30:20] -!- Guest538 [Guest538!~Guest538@78-83-51-185.spectrumnet.bg] has joined #emc
[12:30:59] -!- Guest538 has quit [Client Quit]
[13:24:48] -!- acemi [acemi!~acemi@unaffiliated/acemi] has joined #emc
[13:32:22] -!- acemi has quit [Quit: WeeChat 0.3.2]
[13:49:45] -!- isssy [isssy!~isssy@78-83-51-185.spectrumnet.bg] has joined #emc
[13:49:59] -!- jdj [jdj!~jj@cpe-71-65-51-222.insight.res.rr.com] has joined #emc
[14:01:09] -!- Emcrules_Laptop [Emcrules_Laptop!~Jean-Paul@CPE001310e6c94f-CM001bd71cb794.cpe.net.cable.rogers.com] has joined #emc
[14:35:43] -!- skunkworks [skunkworks!~chatzilla@str-bb-cable-south2-static-6-412.dsl.airstreamcomm.net] has joined #emc
[14:41:12] <JT-Shop> what a lovely morning it is for roofing work :)
[14:42:17] <skunkworks> heh
[14:55:38] -!- Emcrules_Laptop has quit [Ping timeout: 240 seconds]
[15:09:24] -!- Dallur has quit [Quit: Leaving.]
[15:14:20] <Tom_itx> yeah, about 30F here
[15:41:24] <JT-Shop> 37F here
[15:42:54] <Sairon> hmm, no idea what it is here
[15:44:33] <Sairon> 31F
[16:03:37] -!- andypugh [andypugh!~andy2@cpc2-basl1-0-0-cust1037.basl.cable.virginmedia.com] has joined #emc
[16:04:38] -!- acemi [acemi!~acemi@unaffiliated/acemi] has joined #emc
[16:09:43] <andypugh> Anyone interested in a litle C-coding puzzle.
[16:09:44] <andypugh> http://pastebin.com/9asCAgca
[16:12:19] <andypugh> This is the code that converts my Resolver output to quadrature pulses. What I can't figure out is why it will, every few seconds, do a full revolution for no good reason. I can imagine a glitch in the voltage measurements which are converted to an Arctan might have that effect, but surely it ought to bounce back next measurement, before the quadrature counter has managed more than a few counts up or down?
[16:13:52] <andypugh> i suspect that there is some very clever way to handle wrap in this sort of situation.
[16:14:11] <awallin> most math packages come with an arctan2 that is slightly different
[16:14:30] <awallin> although I haven't worked with arctan enough to know the details...
[16:14:31] <andypugh> In case it is not obvious, the arctan function actually returns a number between 0 and 1024
[16:15:05] <andypugh> It is a lookup-table based Arctan2 function.
[16:15:21] <andypugh> (runs in 14ns, which I am quite pleased with)
[16:15:21] <awallin> ok
[16:20:04] -!- EDocToor has quit [Read error: Connection reset by peer]
[16:20:44] <Tom_itx> avr code?
[16:21:14] <andypugh> Pity I don't have a really long USB cable so I can reprogram the Arduino in-situ.
[16:21:27] <andypugh> Sorry, yes, I forgot to say, AVR code running on an Arduino.
[16:21:44] <Tom_itx> arduino *is* an avr
[16:22:15] <andypugh> Other bits of the code synthesise a 10kHz sine wave to excite the Resolver, and then measure the sin and cos voltages synched to the 10kHz.
[16:22:30] <Tom_itx> they just call them arduino since they have someone's convoluted code on them
[16:23:02] <andypugh> Yeah, convoluted friendly code that I have pretty much entirely bypassed.
[16:23:02] <Tom_itx> you really need 2 ints to do quadrature iirc
[16:23:59] <andypugh> The angle measurement part is 100% working, with a bit of unavoidable noise.
[16:24:38] <andypugh> What I am not getting right is converting the 0-1024 arctan values into a long int with wrap.
[16:25:20] -!- abcminiuser [abcminiuser!~Dean@2001:700:303:2:1991:db7e:4644:c237] has joined #emc
[16:25:25] <abcminiuser> Yahoyoy
[16:25:43] <Tom_itx> <andypugh> What I am not getting right is converting the 0-1024 arctan values into a long int with wrap.
[16:26:15] <andypugh> It works on the desk, all looks good, but once the motor is powered it sits there vibrating a bit (as the angle output to the drive is slightly noisy) and occasionally flips 360 degrees.
[16:26:40] <andypugh> The vibration I can live with (or fix later). The flips are a bother.
[16:26:49] <Tom_itx> with one int, you can't detect direction
[16:27:09] <andypugh> I don't get what you are saying?
[16:27:17] <Tom_itx> because you don't know the phase of the two channels
[16:27:43] <Tom_itx> if a fires before b it's one direction and if b fires first it's the other
[16:28:11] <Tom_itx> look at a quad diagram
[16:29:00] <Tom_itx> http://www.barello.net/Papers/Motion_Control/index.htm
[16:29:05] <Tom_itx> about halfway down
[16:29:30] <Tom_itx> the code probably won't work for you since yours are alot faster rpm
[16:29:58] <Tom_itx> that's not really what i was looking for anyway
[16:30:12] <andypugh> I have two voltages, V[0] and V[3] (8-bit values) which are enough to unambiguously give me Resolver angle.
[16:30:53] <Tom_itx> so this isn't a digital quadrature decoder
[16:31:17] <andypugh> No, this is a Resolver to Quadrature convertor.
[16:32:17] <andypugh> It measures rotor angle at 5kHz, then that angle is "chased" by a counter which outputs quadrature pusles every increment.
[16:33:24] -!- skunkworks has quit [Read error: Connection reset by peer]
[16:34:14] <Tom_itx> are you using this:? http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?ResolverToQuadratureConverter
[16:34:38] <andypugh> The problem is, every now and again, it gets the count wrong by 1024 counts, and the quadrature counts run round a full circle, which the motor follows.
[16:34:47] <andypugh> Yes, I wrote that Wiki page ;-)
[16:34:51] <Tom_itx> heh
[16:35:30] -!- servos4ever [servos4ever!~chatzilla@173-87-48-6.dr01.hnvr.mi.frontiernet.net] has joined #emc
[16:35:59] <andypugh> I think if I could get the absolute angle into EMC2 there would be no problem, but the only way to get it in is to convert it to simulated encoder counts, and that part is a bit flaky.
[16:37:16] <Tom_itx> what about feeding the output to a schmidt trigger to clean up the analog to digital and use that output?
[16:37:33] <Tom_itx> or comparator
[16:38:14] <andypugh> Where would the comparator go?
[16:38:19] <Tom_itx> hardware's gonna be faster than software
[16:38:29] <Tom_itx> on the resolver output
[16:38:31] <Tom_itx> what v is it?
[16:38:46] <Tom_itx> you'd need to bring it to 5v or cmos level
[16:38:58] <Tom_itx> cmos will handle more v iirc
[16:39:09] <Tom_itx> i use those on ir detectors
[16:39:15] <andypugh> You mean convert my Resolver into a 4-counts per rev quadrature encoder? That seems like a bad plan...
[16:39:35] <Tom_itx> i haven't worked with resolvers so i don't know for sure what they output
[16:41:35] -!- skunkworks [skunkworks!~chatzilla@str-bb-cable-south2-static-6-412.dsl.airstreamcomm.net] has joined #emc
[16:43:48] <andypugh> They are a transformer, so the output is two sine waves at the excitation frequency. As the resolver rotates the peak amplitudes of the two sine waves varies. The variation of the two is at 90 shaft degrees to each other, so when one has max positive amplitude the other is zero amplitude, and so on. At any time you can measure the output amplitudes, do an arctan and get the angle. However it makes most sense to measure at th
[16:43:48] <andypugh> input excitation peaks, so that you are measuring the biggest possible voltage.
[16:45:05] <Tom_itx> yeah, they sound more precise than a quad decoder but it sounds like you need a quad decoder ouput
[16:45:23] <Tom_itx> quad decoders output the same waveshape only digital
[16:45:26] <Tom_itx> 90 out of phase
[16:45:45] <Tom_itx> that was why i suggested the hardware
[16:45:49] <andypugh> They are very good devices. Tough, stable and accurate. But there is no easy way to get the angle data into eMC2
[16:46:08] -!- tris has quit [Ping timeout: 250 seconds]
[16:46:25] <Tom_itx> it sounds like the hardware is better than the software in this case
[16:46:51] <andypugh> That's why I am using the Arduino, it is being Hardware. :-)
[16:47:59] <andypugh> With only one resolver cycle per shaft rotation, converting direct to quadrature with a comparator would give me very poor angular resolution.
[16:48:08] <Tom_itx> the trigger will 'snap' into a logic state once a specific voltage is reached giving you a smooth square wave from the analog
[16:48:39] <Tom_itx> you only get 1 wave per rev with those?
[16:48:43] <andypugh> Yes.
[16:48:45] <Tom_itx> per channel
[16:48:50] <andypugh> Yes
[16:48:51] <Tom_itx> hmm
[16:50:28] <Tom_itx> thinking out loud here: take the schmidt trigger output from each channel and input that into a divider to get the other phases
[16:50:41] <Tom_itx> there may be some propogation delay
[16:51:11] <andypugh> I could just pay the $184 for the Mesa resolver convertor, but what I have is so nearly working.
[16:51:14] <Tom_itx> schmidt -> mux
[16:51:26] <Tom_itx> well how did they do it?
[16:51:36] <Tom_itx> it can be done in hardware or software
[16:51:40] <andypugh> Much as I do, measure the voltages, take the arctan.
[16:52:04] <Tom_itx> just depends where you feel comfortable living
[16:52:24] <andypugh> Then there is http://uk.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=5385520
[16:53:13] <andypugh> I don't think you can do it in hardware, really. You do actually need to solve an arctan.
[16:53:59] <andypugh> And that part of my code works perfectly. It is wrapping the angles and converting to quadrature where it all goes wrong.
[16:54:03] <Tom_itx> when the analog wave gets to ~80% 5v it will trigger the schmidt trigger
[16:54:15] <Tom_itx> ~70 to 80%
[16:54:26] <andypugh> And then?
[16:54:44] <Tom_itx> feed those digital waves to a mux
[16:54:49] <Tom_itx> and you have quad out
[16:55:18] <Tom_itx> it may not be as simple as i picture it but i haven't sat down to figure it out either
[16:55:24] <Tom_itx> just a first thought
[16:55:49] <andypugh> Yes, but that gives a 4-counts-per-rev quadrature output, which is no use.
[16:56:05] <Tom_itx> what are you looking for?
[16:56:19] <andypugh> 1024 counts per rev
[16:56:23] <Tom_itx> that's what i thought you wanted
[16:56:24] -!- tris [tris!~tristan@173-164-188-122-SFBA.hfc.comcastbusiness.net] has joined #emc
[16:56:40] <Tom_itx> heh add more mux :)
[16:56:57] <andypugh> (Or thereabouts). I need enough angle data to commutate the motor.
[16:57:04] <Tom_itx> yeah
[16:59:51] <andypugh> Anyway, this is rather besides the point. I can measure angle great, works fine. The question is what is wrong with that C-code which makes it put in spurious motor flips.
[17:00:01] <pcw_home> The AD parts do it in hardware (I used their scheme with the 7I49 firmware = detect amplitudes --> phase compare --> DPLL)
[17:01:14] <andypugh> I guess the PLL is equivalent to my idea of modelling motor velocity and plausible accelleration rates?
[17:02:02] <andypugh> I confess I am a bit disgruntled, I thought it was all working until I hooked it up to drives.
[17:02:47] <pcw_home> Yes you can tweak the P and I gains to compromise error on acceleration vs noise
[17:03:13] <andypugh> I guess I could just put a lowpass filter on encoder_position....
[17:03:37] <Tom_itx> yeah, i was getting ready to suggest a noise filter of some sort
[17:04:04] <pcw_home> did you look at how the HostMot2 encoder counter promotes the 16 bit counter to 32 bits? same problem basically
[17:04:28] <andypugh> That would kill the buzzing, but not the flips, I think. It wold just smooth-out the flips.
[17:05:39] <pcw_home> a low pass will cause a position lag at a constant velocity, thats an advantage of the DPLL
[17:05:40] <andypugh> pcw_home: Yes, but that all happens by twos-complement magic. I have the same thing in bldc.comp to promote 32 bit to 64 bit.
[17:06:39] <pcw_home> Isn't it the same problem?
[17:06:42] <andypugh> It is just 64_bit_value = 64_bit_value + (new_32_bit_counts - old_32_bit_counts) and the wrap happens by magic.
[17:07:48] <andypugh> I don't know, I don't know what the problem is.
[17:10:52] <pcw_home> what is the range of your arctan output?
[17:13:23] <andypugh> 0 to 1024
[17:13:45] <andypugh> I do wonder if it would be easier with -512 - +512
[17:13:57] <Tom_itx> 1024 would roll over
[17:14:03] <Tom_itx> 1023 wouldn't
[17:14:12] <andypugh> OK, 1023
[17:14:23] <pcw_home> probably easier -32768 to 32767
[17:15:17] <andypugh> Yes, then the rollover is handled easily. But then my peak RPM would be far too low, as the quad counter can only count so fast
[17:15:20] -!- Vq_ [Vq_!~vq@90-227-210-39-no122.tbcn.telia.com] has joined #emc
[17:16:03] <andypugh> (And I don't think you can bit-shift-divide signed numbers)
[17:16:28] <pcw_home> OK you are doing it in C where the compiler will deal with all the sign extension etc
[17:16:40] -!- Vq has quit [Ping timeout: 260 seconds]
[17:19:18] <pcw_home> sure you can divide signed numbers with arithmetic shift right
[17:20:23] <andypugh> I don't know if C differentiates between arithmetic shift and logic shift.
[17:21:59] <andypugh> Wikipedia says: ^ The >> operator in C and C++ is not necessarily an arithmetic shift. The http://en.wikipedia.org/wiki/C99 standard specifies that the resulting value is implementation-defined for a right shift in which the left operand is a signed integer that is negative. However, most implementations use sign extension, thereby making the >> operator an arithmetic shift. For instance, the
[17:21:59] <andypugh> http://gcc.gnu.org/onlinedocs/gcc-4.3.3/gcc/Integers-implementation.html#Integers-implementation is such an implementation (although if the left operand is unsigned, it will perform a logical shift).
[17:22:29] -!- ca-kc6lbj-2 [ca-kc6lbj-2!~raffi@zima.linwin.com] has joined #emc
[17:24:36] <pcw_home> I think you have to do that thing where if the difference in angles is >512 you subtract add or subtract 1024
[17:25:53] <Tom_itx> that was sorta what i was getting at with the 1024 value. the overflow flag would set
[17:26:07] <Tom_itx> and if not processed, would give you 0 as a value instead
[17:26:30] <Tom_itx> causing it to think it's off 359 deg
[17:26:38] <Tom_itx> possibly
[17:27:30] <pcw_home> Sorry, subtract the number from 1024
[17:29:58] <andypugh> Yes, that is what I am doing if you look at the code (http://pastebin.com/9asCAgca)
[17:30:19] <andypugh> But it seems like it glitches and then fails to unglitch.
[17:31:03] <Tom_itx> does it happen the same time every time or erratic?
[17:31:47] <Tom_itx> if it's repeatable, it's more likely code
[17:33:12] -!- Athlocatle [Athlocatle!~vip0@cc27x197.sels.ru] has joined #emc
[17:34:54] <andypugh> The motor just sits there ostensibly static, with random jumps of one full rev.
[17:35:59] -!- nullie has quit [Quit: Ex-Chat]
[17:37:03] <Tom_itx> can you verify the output of PORTD?
[17:37:25] <Tom_itx> put some leds or something on them
[17:38:07] <andypugh> PortD needs to go through 1024 cycles to make the motor do a full rev. I think that part is OK.
[17:38:31] <pcw_home> Does it jump at any position or only near 0 degrees?
[17:38:43] <andypugh> I am pretty sure that the problem is only in R.
[17:38:54] <andypugh> pcw_home: As far as I can see, in any position.
[17:39:54] <Tom_itx> how is R[0] defined?
[17:40:36] <andypugh> I don't think it is the sampling getting out of synch with the excitation, as it behaves itself perfectly with an unpowered motor. (outputting angle data to a terminal window)
[17:40:47] <pcw_home> If you re-arranged your wrap math (get the fixed delta first) you could check for the max delta (to bisect the problem)
[17:40:48] <andypugh> signed long.
[17:41:40] <andypugh> pcw_home: I am in the middle of that. I am going to do away with B[] and use the LS part of R.
[17:43:26] <pcw_home> thats also a possible point to insert a lp filter (on velocity, not position)
[17:44:17] <andypugh> That's the next step. Calculate a velocity, and add that to R rather than the delta.
[17:45:05] <andypugh> where the velocity is capped at (say) 100 counts and can only increment at 1 count per cycle,
[17:45:47] <andypugh> However, I am rather limited by CPU. Can't get too clever, and FP is out
[17:46:19] -!- ve7it [ve7it!~LawrenceG@S0106009027972e37.pk.shawcable.net] has joined #emc
[17:48:26] -!- rooks has quit [Read error: Connection reset by peer]
[17:50:07] <andypugh> Annoyingly, it won't misbehave on the desk with the motor unpowered.
[17:50:25] <Tom_itx> of course
[17:50:46] -!- rooks [rooks!~rooks@102-bem-18.acn.waw.pl] has joined #emc
[17:51:04] <pcw_home> Motor noise?
[17:51:05] <archivist> noise, screening, check noise immunity of ttl levels
[17:52:20] <andypugh> There is absolutely no screening on the wires. In fact they are single-strand hookup wire wrapped round the pins of the resolver connector (those plugs cost Β£23 each)
[17:52:52] <archivist> you will suffer with no screening
[17:52:58] <andypugh> But, even if the signal is noisy, the resolver should recover.
[17:53:08] <andypugh> Because it is an absolute device.
[17:53:40] <archivist> but any interface measuring it may not be immune
[17:54:26] <archivist> I would be scoping stuff
[17:54:30] <andypugh> No, indeed not, but if it measures a wrong voltage and gives a spurious angle, the next angle should be correct, or the one after that.
[17:55:00] <archivist> yes and spurious will set of an action you may not want
[17:55:37] <archivist> and it may be electrically noisier than you expect
[17:55:48] Vq_ is now known as Vq
[17:55:52] -!- abcminiuser [abcminiuser!~Dean@2001:700:303:2:1991:db7e:4644:c237] has parted #emc
[17:56:05] <andypugh> For 100uS, yes. But if I leave the debugging serial print lines in, the spurious motor rev takes a second.
[17:56:49] <archivist> if the noise is in the digital domain...
[17:57:18] <andypugh> So it appears that a glitch can add on a full rev, but then the next good value does not take that full rev off again.
[17:57:49] <archivist> so the high bit has flipped
[17:58:48] <andypugh> There isn't a "high bit" as such, except inside the software.
[17:59:37] <archivist> at some stage you convert analog to digital yes?
[17:59:54] <archivist> is that a gray code or a plain count
[18:00:04] <andypugh> Yes. But errors there should be recoverable (I think)
[18:00:25] <andypugh> Neither
[18:00:39] <archivist> what then
[18:00:45] <andypugh> Two voltages are measured, then the arctan of their ratio is taken.
[18:00:57] -!- nullie [nullie!~nullie@nullie.telenet.ru] has joined #emc
[18:01:52] <andypugh> From that point on it is all softwae.
[18:03:41] <archivist> 359.999 deg+ noise could put it a rev out
[18:06:17] <pcw_home> maybe your follower (that does the quadrature output) doesn't cope well with large deltas (from noise)
[18:06:26] -!- ries has quit [Quit: ries]
[18:06:52] <andypugh> It just blindly counts to the new target.
[18:07:46] <archivist> does it only occur when parked near 0/360
[18:08:10] <andypugh> Not as far as I can see, no.
[18:09:07] <andypugh> (It is hard to tell)
[18:25:54] <cradek> "Your clues are the these." </me, in a particularly lucid moment on emc-users>
[18:27:05] <SWPadnos> at least it's not these thoses
[18:30:12] <cradek> mozmck: I found a resonator banjo on craigslist. the fun part is it has no case, so it's sitting next to me. now whenever I cough, it's accompanied by a nice G chord.
[18:30:13] -!- motioncontrol [motioncontrol!~i@host193-28-dynamic.20-87-r.retail.telecomitalia.it] has joined #emc
[18:31:28] <skunkworks> oh look - snowing again
[18:33:05] <pcw_home> "It just blindly counts to the new target." how fast can it count? can it wrap in one A-D cycle so bad data gives it a 360 spin?
[18:35:16] <andypugh> No, it takes a few dozen A-D cycles per 1024 quadrature counts
[18:37:29] -!- nullie has quit [Quit: Ex-Chat]
[18:38:35] <pcw_home> Sure sounds like something doesn't like large deltas (from noise perhaps)
[19:04:46] -!- tom3p [tom3p!~tomp@74-93-88-241-Illinois.hfc.comcastbusiness.net] has joined #emc
[19:07:18] -!- tom3p has quit [Remote host closed the connection]
[19:08:36] -!- acemi_ [acemi_!~acemi@unaffiliated/acemi] has joined #emc
[19:08:44] -!- tom3p [tom3p!~tomp@74-93-88-241-Illinois.hfc.comcastbusiness.net] has joined #emc
[19:09:10] -!- acemi has quit [Ping timeout: 240 seconds]
[19:12:16] -!- tlab [tlab!~tlab@c-69-243-187-136.hsd1.in.comcast.net] has joined #emc
[19:22:16] -!- psha [psha!~psha@] has joined #emc
[19:35:36] -!- toastydeath has quit [Read error: Connection reset by peer]
[19:38:14] -!- toastydeath [toastydeath!~toast@c-69-140-223-139.hsd1.md.comcast.net] has joined #emc
[19:38:49] -!- odiug [odiug!~guido@p57BD12B6.dip0.t-ipconnect.de] has joined #emc
[20:04:16] -!- isssy has quit [Quit: Visitor from www.linuxcnc.org]
[20:12:39] -!- tlab has quit [Quit: Leaving]
[20:20:47] -!- mhaberler [mhaberler!~chatzilla@extern-188.stiwoll.mah.priv.at] has joined #emc
[20:28:00] -!- bootnecklad [bootnecklad!~bootneckl@] has joined #emc
[20:30:35] -!- bootnecklad_ has quit [Ping timeout: 265 seconds]
[20:44:12] -!- Roguish [Roguish!~chatzilla@c-67-174-215-143.hsd1.ca.comcast.net] has joined #emc
[20:50:11] -!- micges [micges!~micges@afoz173.neoplus.adsl.tpnet.pl] has joined #emc
[20:59:21] <danimal_garage> cradek, one of the resolvers is bad on the servos
[20:59:47] <danimal_garage> i'm making a mount so i can use the resolver i took off the spindle of the hnc
[21:00:04] <danimal_garage> i already wired the resolver up and it seems to count ok
[21:00:32] <danimal_garage> the resolvers on those servos are like $500-700
[21:01:50] <danimal_garage> i'll make this work until i can find another servo with the same resolver on ebay
[21:03:11] <danimal_garage> glad i saved that resolver
[21:10:04] <andypugh> I am wondering if it might not be simpler to swap my resolvers for encoders. But the thing is, I rather like resolvers.
[21:10:37] <danimal_garage> i liked resolvers too, till i found out how much they cost
[21:11:02] <andypugh> They shouldn't ever break.
[21:11:21] <andypugh> (No idea what has happened to yours)
[21:11:59] <danimal_garage> i'm still planning on "tricking" my drives to think there's a resolver hooled up and then going straight to the mesa card with some encoders. all the tuning would have to be done in EMC i would assume at that point.
[21:13:12] <danimal_garage> i dont know, but it was loosing position, and i did a process of elimination with cables, motors, drives, channels on the 7i33, etc, so i took the resolver off and one of the little metal plates in the winding shifted
[21:13:26] <andypugh> Tricking the drives might be troublesome if they require a resolver input.
[21:13:27] <danimal_garage> i tried pushing it back, but it didnt work
[21:13:54] <danimal_garage> i swapped in a resolver i had, and all is good
[21:14:06] <danimal_garage> why is that andypugh?
[21:14:07] <andypugh> Any way to make Halmeter text bigger?
[21:14:33] <cradek> andypugh: change your screen dpi :-/
[21:14:46] <andypugh> Do the drives use the resolver input for commutation?
[21:14:57] <cradek> danimal_garage: yuck - glad you have a spare - those sure are expensive (unless they come attached to cast iron, of course)
[21:14:59] <andypugh> cradek: I am not sure I can.
[21:15:15] <danimal_garage> andypugh, not sure what that is
[21:15:44] <danimal_garage> cradek, yea, i nearly choked when i saw what they were going for on ebay
[21:16:25] <danimal_garage> i got all 4 drives and 3 motors on ebay for $300, the resovers alone on ebay were $500 at the cheapest!
[21:16:34] <cradek> heh
[21:16:56] <cradek> have you scoped it? maybe you could figure out the problem.
[21:17:04] <danimal_garage> i saw the problem
[21:17:16] <danimal_garage> one of the metal plates in the winding shifted
[21:17:39] <danimal_garage> hard to describe
[21:18:05] <danimal_garage> tried pushing it back but i think it severed a wire in the winding
[21:18:23] <andypugh> Eeew! 640x480 was like it was 1989 again.
[21:18:45] <danimal_garage> there's no way to make the text bigger like in windows?
[21:18:50] <andypugh> danimal_garage: Brushed or brushless motors?
[21:18:56] <danimal_garage> brushless
[21:19:12] <andypugh> In that case fooling the drives is likely to be really hard.
[21:19:17] <danimal_garage> the drives will run either
[21:19:54] <danimal_garage> brushed or brushless
[21:20:21] <danimal_garage> are you talking about a tachometer feedback?
[21:20:37] <danimal_garage> it can take a tach as well
[21:20:42] <andypugh> Yes, bit if the current motors are brushless then the drive needs to know the rotor angle in order to commutate it.
[21:21:03] <andypugh> (I ought to get this page finished now that bldc is in Master) http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?BLDC
[21:21:30] <danimal_garage> the directions say that if you run a brushed motor, you need a tach, but if you run a brushless, you dont
[21:21:52] <danimal_garage> so i do not believe it gets the velocity feedback through the resolver
[21:22:08] <danimal_garage> or whatever
[21:22:08] <andypugh> Do you have Hall sensors?
[21:22:21] <danimal_garage> not that i know of
[21:22:55] <pcw_home> Which means you need to get the new resolver angle correct or you will have bad performance
[21:23:03] <andypugh> In that case the drive needs the absolute position information from the Resolver. (It also means that if you swap the resolvers they need to be mounted in exactly the same orientation)
[21:27:12] <pcw_home> The key thing here is that the resolver is an absolute encoder so can provide the drive with rotor position at startup
[21:27:14] <pcw_home> a quadrature encoder is a relative position device so cannot provide the drive with rotor position (at least not until index is passed)
[21:29:02] <pcw_home> The drive needs the absolute rotor position for commutation (knowing what angle to drive the stator magnetic field)
[21:31:34] -!- Valen [Valen!~Valen@] has joined #emc
[21:34:37] <danimal_garage> crap, how do i get it in the exact positon?
[21:35:27] <danimal_garage> i need to get the new resolver in the exact same position??
[21:35:43] <danimal_garage> by like how close?
[21:36:24] <pcw_home> good question (it does not have to be exact but should be within 5 degrees or so)
[21:36:31] <danimal_garage> ah ok
[21:36:45] <danimal_garage> how will it act if it's off?
[21:37:12] <pcw_home> unsymmetrical torque
[21:37:25] <pcw_home> (in either direction)
[21:37:44] <danimal_garage> anything i can check in hal scope to help dial that in?
[21:38:38] <cradek> can you see how the others are set up? do they have set screws on the resolver shaft or anything like that?
[21:38:59] <danimal_garage> the other resolvers are completely different
[21:39:10] <danimal_garage> hold on i'll link you to what i have
[21:39:34] <danimal_garage> http://www.dynapar.com/content.aspx?id=232
[21:39:46] -!- theorbtwo has quit [Ping timeout: 255 seconds]
[21:40:04] <cradek> funky
[21:40:10] <danimal_garage> yea
[21:41:05] <danimal_garage> these being completely digital drives, do you think they might be smart enough to detect the unsymmetrical torque and correct for it?
[21:41:22] <pcw_home> No
[21:41:23] <cradek> maybe you can scope one of the working ones, find electrical zero, and see if there's a marking or something on the motor that corresponds? there must be a way to orient them, if you only knew the secret.
[21:42:25] <danimal_garage> yea, but there's no mark on the encoder from the lathe, so i'm not sure if it would help
[21:42:32] <cradek> I'm talking out my ass - I have never dealt with ac servos - but seems like there must be a way to do it.
[21:42:45] <cradek> well you'd find its electrical zero by measuring, too
[21:42:53] <cradek> does it even have the same number of poles?
[21:43:35] <danimal_garage> yes
[21:43:37] <danimal_garage> 2 poles
[21:43:48] <pcw_home> you can align the rotors magnetically and match resolver output voltage ratios
[21:44:22] <pcw_home> ( with a 50% chance of being 180 out)
[21:44:31] <danimal_garage> the software for the drives has it's own scope, not sure if that is relivant
[21:45:03] <danimal_garage> hmm not sure how to do that
[21:47:09] <andypugh> I would expect there to be some sort of marks on the existing resolver.
[21:48:45] <danimal_garage> there is, but i dont know how that helps me as the new resolver doesnt
[21:48:52] <danimal_garage> and they're completely different
[21:49:12] <cradek> figure out what the mark means electrically
[21:49:19] <andypugh> danimal_garage: Don't suppose you have an Arduino lying about? You could use my Arduino resolver driver, make a note of the output when the motor in in a marked position, and then fit the new resolver to be the same.
[21:49:21] <cradek> then figure out how to match that
[21:49:22] -!- The_Ball has quit [Ping timeout: 265 seconds]
[21:50:04] <danimal_garage> cradek, the resolver is dead, and i already pulled it off the motor, so i don't know the relation of the mark to the motor
[21:50:20] <cradek> but you have three identical to it that work?
[21:50:29] <danimal_garage> i shoulda asked first, huh? :)
[21:50:47] <andypugh> Yes
[21:51:00] <danimal_garage> yes
[21:51:18] <danimal_garage> andypugh, i dont even know what arduino is
[21:51:32] <andypugh> Do you have a lab power supply?
[21:51:35] <pcw_home> Heres a possible way: apply DC between to stator connections on the working motor/resolver (make sure the current is no more than the motors rating)
[21:51:41] <pcw_home> measure the sine and cosine resolver voltages. Divide the sine/cosine voltages.
[21:51:41] <pcw_home> do the same with the new resolver and rotate the resolver relative to the (fixed) motor shaft so the sine/cos ratio is the same (this may be 180 degrees off)
[21:51:52] <pcw_home> s/to/two/
[21:52:16] <andypugh> if it is 180 degrees off, I think you can flip the polarity of a sine or cosine pair.
[21:52:38] <pcw_home> Yes
[21:53:33] <danimal_garage> i have a power supply, but it's not a lab one. it's a astron, with adjustable current and voltage, 50A max, 15v max
[21:53:47] <danimal_garage> it's for my anodizing
[21:54:17] <andypugh> One convention for motor electrical zero is +ve to phase A, negative to B and C. Apply that to your motor, and mark the shaft position. (I would be unsurprised to find that it turns out to be keyway-up
[21:54:41] -!- Guest63696 has quit [Ping timeout: 264 seconds]
[21:54:53] <andypugh> A very small current should do, just enough to align the rotor.
[21:55:17] <danimal_garage> there's u, v, and w
[21:55:40] <andypugh> If the motors are multipole then there will be 2 or 3 or 4 positions that the motor will sit in with that arrangement. They are all equivalent.
[21:56:01] <andypugh> abc, uvw, you are just making things hard for fun now aren't you?
[21:56:20] <danimal_garage> no, i'm not trying to, i just wanted to make sure i understood right
[21:56:20] -!- The_Ball [The_Ball!~The_Ball@123-2-12-83.static.dsl.dodo.com.au] has joined #emc
[21:56:35] <danimal_garage> i'm completely uneducated in this department
[21:56:49] <danimal_garage> i really appreciate the help
[21:56:55] <andypugh> U +ve, V and W -ve. Go for rather less than an amp.
[21:57:02] -!- davec [davec!~davec@host-174-45-229-40.bln-mt.client.bresnan.net] has joined #emc
[21:57:18] <danimal_garage> ok
[21:57:28] davec is now known as Guest81999
[21:57:31] -!- motioncontrol has quit [Quit: Sto andando via]
[21:57:33] <danimal_garage> so that'll align the rotor
[21:57:37] <danimal_garage> easy enough
[21:57:45] <danimal_garage> and then i have to align the tach
[21:58:07] <cradek> tachs don't align, they are just dc generators
[21:58:20] <danimal_garage> sorry, i ment resolver
[21:58:35] <danimal_garage> i'm a bit worn, just got back from a big bike ride
[21:59:08] <andypugh> First try it with a working motor, with the resolver powered up, and see what the (AC) voltages are on the sine and cosine outputs. (I would expect that one of the motor stable positions will probably correspond to max volts on cosine, and zero on sine.)
[21:59:30] <andypugh> Do you have an oscilloscope?
[21:59:44] <danimal_garage> no scope
[21:59:53] <pcw_home> You will need the drive to supply the resolver excitation
[21:59:54] <pcw_home> (I hope it does this regardless of drive faults which you will surely get sone the UVW are disconnected from the drive)
[21:59:59] <andypugh> Pity, they are the best way to see (and to understand) a resolver.
[22:00:06] -!- mrsunshine [mrsunshine!~mrsunshin@c-bc3de253.4551536--62697410.cust.bredbandsbolaget.se] has joined #emc
[22:00:16] <mrsunshine> http://www.danahermotion.com/website/com/eng/img/product/LeadScrew2Nut.jpg <-- how are these kind of screws in performance etc?
[22:00:26] <mrsunshine> looks like a very high pitched acme :P
[22:00:27] <pcw_home> s/sone/since/
[22:00:30] <mrsunshine> or low ?
[22:00:35] <danimal_garage> pcw, i can hook up another motor to the drive so it wont fault
[22:01:03] <danimal_garage> but i know it still measures position even if the drives are not enabled
[22:01:07] <andypugh> danimal_garage: It probably will, when it sees the resolver not moving :-)
[22:01:49] <andypugh> mrsunshine: They have their uses, but not for a CNC axis I wouldn't have thought.
[22:02:03] <mrsunshine> oki =)
[22:02:07] <danimal_garage> they dont fault if they see the resolver moving and the motor isnt
[22:02:31] <andypugh> Sounds promising then.
[22:02:35] <pcw_home> Just to be really clear here (for safety) only the resolver is connected to the drive, the motor leads go to your test DC supply
[22:02:36] <skunkworks> mrsunshine: they work great for high speed low resolution positionoing
[22:02:58] <andypugh> Hmm, what _do_ the drives show as position?
[22:03:19] <andypugh> Can you use the drive to tell you resolver alignment?
[22:03:31] <danimal_garage> andy, i am not sure
[22:03:42] <Valen> danimal_garage: you can usually get scopes pretty cheap
[22:03:43] <danimal_garage> pcw_home, i got it, thanks!
[22:04:02] <Valen> they are pretty crappy but they give you an idea
[22:04:28] <andypugh> I paid Β£100 for a 25 year old scope. I don't regret it.
[22:04:47] <danimal_garage> is there any way to check the resolver position without a scope?
[22:05:08] <pcw_home> voltmeter
[22:05:16] <danimal_garage> ok
[22:05:27] <andypugh> But it needs to be one that works in AC mode at the resolver frequency
[22:05:38] <danimal_garage> ahh
[22:05:44] <danimal_garage> likely not 60hz?
[22:06:03] <andypugh> (actually, you could power the resolver from a 5V mains transformer, then any multimeter with an AC setting would work)
[22:06:14] <Valen> going to be 40C here today
[22:06:31] <Valen> 38+ for the rest of the week
[22:06:48] <andypugh> I would expect the standard excitation frequency to be at least 10kHz, and possibly a lot higher.
[22:06:51] <Valen> 107F
[22:06:52] <danimal_garage> it's probably 70f here
[22:07:17] <pcw_home> 5V RMS at 60 Hz might be a bit much for a high frequency resolver
[22:07:26] <andypugh> I have just realised, I have completely failed to taunt Valen about the cricket!
[22:07:56] <Valen> eh, except for playing cricket with michael clarke I don't follow it lol
[22:07:58] <andypugh> pcw_home: Ah, yes.
[22:08:07] <Valen> dang boring to watch
[22:10:03] <pcw_home> danimal_garage you can also just try every 5 degree position from 0 to 180 (mechanical) and see which works best
[22:11:02] <danimal_garage> ok, so, if i am understanding this right, i line up the motor with a dc power supply. then i power up the resolver, and check the voltage on each coil. one will be zero when it's lined up. (Will the second coil be 5v if i have 5v going into it? or do i just look for peak voltage?) And that's where i tighten it down. However, it could be off 10 degrees. Will it be obvious with how it runs?
[22:11:42] <andypugh> danimal_garage: One might not be zero, you need to see what the known-good motors give you.
[22:11:47] <danimal_garage> pcw, it doesnt seem too bad to do it the right way, i'll give it a shot, assuming i'm understanding it right
[22:11:59] <danimal_garage> andypugh, ok, good call, thanks
[22:12:21] <andypugh> and it is the ratio of the voltages, not the absolute that matters.
[22:12:44] <danimal_garage> ok, just wanted to make sure
[22:12:54] <danimal_garage> hi vs low
[22:12:56] <andypugh> (though, hopefully, the voltage-efficiency of the two resolvers will be close)
[22:13:10] <pcw_home> We dont know what the voltages from the resolver would be what we are aiming at is making the new resolver voltage ratio equal the old resolver voltage ratio (at the same rotor position)
[22:13:44] <danimal_garage> ok, easy enough
[22:13:55] <andypugh> To be more sure, you could note the ratio in each of the stable positions the motor will go to with the specified field excitation.
[22:14:03] <danimal_garage> thanks a lot guys, i never would have figured that out without you
[22:14:23] <Valen> danimal_garage: whats the 25 words or less of the problem your having?
[22:14:40] <pcw_home> sine vs cosine (or better match arctan (cosine/sine)
[22:14:43] <Roguish> andypugh and pcw_home: got my bldc motors working with the latest (as of last weekend) bldc.comp and the 5i20/7i39 combo.
[22:14:50] <danimal_garage> i have to change a resolver, and i'm finding out it needs to be alighned properly
[22:14:52] <andypugh> He took off a resolver and needs to put on a new one, but didn't note the alignment.
[22:15:09] <andypugh> Excellent!
[22:15:19] <Valen> its brushless motor or some such?
[22:15:23] <andypugh> Aye
[22:15:27] <danimal_garage> yes
[22:15:31] <Valen> ahh
[22:15:42] <Valen> can you rotate it in operation?
[22:16:14] <andypugh> Might be tricky, the one he has is open-frame and bearingless.
[22:16:16] <Valen> like command the motor to spin, then rotate the resolver till it works?
[22:16:35] <Valen> thats probably a no then
[22:16:35] <andypugh> But yes, that ought to work.
[22:17:07] <danimal_garage> andy, the new one isnt
[22:17:10] <andypugh> Though as the servo is probably running at a few hundred volts, that might be a bit scary.
[22:17:19] <danimal_garage> the one i linked is the one that came off of it
[22:17:25] <andypugh> Ah, OK.
[22:17:33] <Valen> put some large resistors in line with the motor to protect everything
[22:17:51] <Valen> should only need a few volts to get it spinning
[22:18:06] <andypugh> Actually, he probably has proper connectors.
[22:18:19] <danimal_garage> yea, i can limit the current and voltage on my PS, i'll start off low
[22:18:45] <andypugh> It is probably only me running 300v motors with uninsulated terminals onto the plugs ;-)
[22:18:57] <danimal_garage> haha
[22:19:09] <danimal_garage> yea, these have really nice cables and connectors
[22:25:17] -!- bootnecklad has quit [Quit: Leaving]
[22:25:42] <andypugh> That was lucky, the cables cost hundreds.
[22:32:50] -!- odiug has quit [Ping timeout: 260 seconds]
[22:36:29] -!- vezza [vezza!~andrea__@host107-41-dynamic.21-87-r.retail.telecomitalia.it] has joined #emc
[22:39:22] -!- acemi_ has quit [Quit: WeeChat 0.3.2]
[22:40:45] <andypugh> Hmm, getting closer, but it looks like my velocity is zeroing every rollover...
[22:45:04] <danimal_garage> yea, i was stoked, all i had to make was the feedback cables from the drives to the mesa board, and they werejust regulat molex connectors
[22:45:13] <danimal_garage> regular*
[22:45:51] <danimal_garage> my buddy just brought over my new road bike frame, i'm pretty stoked, wish i had time to build it up
[22:48:46] <andypugh> I like this frame: http://www.kingcycle.co.uk/hpvs/Hachi.html
[22:53:26] -!- Fox_Muldr has quit [Ping timeout: 276 seconds]
[22:53:27] <danimal_garage> that's pretty slick
[22:54:13] <Tom_itx> seems it would be akward in turns
[22:55:01] -!- Fox_Muldr [Fox_Muldr!quakeman@frnk-4d01f5f7.pool.mediaWays.net] has joined #emc
[22:55:42] <danimal_garage> this is the frame i got: http://trekbuyukada.com/bikes/road/madone/madone49e/index.html
[22:57:15] <danimal_garage> opps it's actually not. mine's a little different
[22:57:27] <danimal_garage> still a madone though
[22:57:53] -!- tris has quit [Excess Flood]
[22:58:02] -!- tris [tris!~tristan@173-164-188-122-SFBA.hfc.comcastbusiness.net] has joined #emc
[22:59:03] <andypugh> Needs posher sprockets :-)
[22:59:14] -!- micges has quit [Ping timeout: 250 seconds]
[22:59:14] <danimal_garage> lol
[22:59:30] <danimal_garage> wish i made some for it
[23:07:35] -!- psha has quit [Quit: Lost terminal]
[23:11:52] <andypugh> I have had enough of this, tomorrow I will buy a USB extension cable so that I can re-program the Aduino without having to disconnect it from the motors/drives/EMC2 machine and move it to the main machine
[23:11:53] -!- micges [micges!~micges@adhz232.neoplus.adsl.tpnet.pl] has joined #emc
[23:11:57] -!- ries [ries!~ries@] has joined #emc
[23:12:24] <andypugh> (Though I guess I could, alternatively, install the Arduino environment on the Linux machine)
[23:13:05] -!- micges has quit [Client Quit]
[23:13:55] -!- tom3p [tom3p!~tomp@74-93-88-241-Illinois.hfc.comcastbusiness.net] has parted #emc
[23:17:35] <andypugh> night all
[23:17:43] <JT-Shop> night
[23:17:48] -!- andypugh has quit [Quit: andypugh]
[23:19:23] <danimal_garage> night
[23:19:26] <danimal_garage> opps too late
[23:20:29] <Tom_itx> meh, the night is young
[23:23:02] -!- John_f_ has quit [Quit: Ex-Chat]
[23:30:07] <danimal_garage> it's only 3:30, it's almost beer thirty
[23:50:35] -!- JT-Shop has quit [Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014]]
[23:52:20] -!- rooks has quit [Ping timeout: 272 seconds]
[23:54:31] -!- vezza has quit [Quit: Sto andando via]
[23:55:24] -!- theorbtwo [theorbtwo!~theorb@] has joined #emc
[23:58:01] -!- rooks [rooks!~rooks@102-bem-18.acn.waw.pl] has joined #emc