is it possible that navel produce cotton ?
I'll have to remember that radian/second is nearly equal to rpm
uh -- wait a second. my units gives a very different answer.
$ units '1200 radians/sec' 'rpm'
11459 is right by my book
I wonder how chris got his number
I transcribed the number manually, but it's hard to believe I messed it up that bad
$ units "1193.6621 rpm" "radians/second"
I'll have to remember that radian/second is nearly equal to 10 rpm.
jepler: Eeeesh, everybody knows that.... Even les_w chicken knows that! ;)
jepler ps.... whats a rad ? =)
Jymmm: It's a unit of radiation, I think.
03cradek 07branch1 * r1.7 10/: trigger
argh it requires <files><file>file1</file><file>file2</file2></files>
somebody should have taken away the XML authors <> keys
03cradek 07branch1 * r1.7 10/file5: trigger
looks like that's the ticket
somebody should have taken away the XML authors
the big strong men in the long white coats are coming to take me away ha-ha, hee-hee, ho-ho
the nice young men in the clean white coats
how many brain cells are dedicated to remembering stuff like that?
I dunno whats worse, that I thought of it or that you know it well enough to correct me
haha both are worse
03cradek 07HEAD * r1.34 10/scratch.txt: another test, should generate a commit email and a cia message
well look at that
did that used to say module/file? the / looks odd
14:28:27 <CIA-10> 03jepler * 10emc2/src/depend.sh.in: rework script, fix TEST vs DEST typo which led to an error removing the directory .tmp on farm machines
so it didn't have the revision
hmm, I don't immediately see how to get the module
we only have one, fake it and worry about it later ;-)
not true, we have four
well five if you count "test"
aha it's in last_commit
yep, already checked that
Update of /cvs/test
oh, you mean the module name is in there?
test is the module
its in the email too
Modified file documents/scratch.txt
03cradek 07HEAD * 10documents/scratch.txt: another cia test, sorry
that's the ticket
I think we're back up to the functionality of sf on its best days, and then some (better cvsweb)
anyone know what a oxy-propane torch setup would cost - something small
I'm seeing ~$250,
am i looking at the wrong thing?
search.ebay.com/4453015813 that's what i'm imagining
go to a mom and pop hardware store
lerman_ is now known as lerman
is the oxygen tanks refillable or disposable?
like little coleman tanks
I've seen the propane tanks, but never disposable oxygen ones
what part of "mom and pop" didnt you understand?
you arent making houses you're doing industrial stuff
fenn: I meant for the replacement o2, not the torch itself.
home depot only has crap for styrofoam/vinyl houses
its amazing how much stuff they _dont_ have considering the size of the store
Yeah, I know... It's just a bitch logistics wise
make your own o2
Getting the O2 isn't the problem, it's getting it back into the tank
you dont need no stinking tank
use an innertube
or a bucket upside down in water
there's a depressingly small volume of o2 in those tanks anyway
I have no idea
I'll hit Ace Hardware tomorrow though
pressure * volume = number of atoms * 0.812 * temperature in kelvin
oops .0821 not .812
It's funny that there is only ONE listing on ebay for that type/size torch. I'm thinking there's some BS law came out to stop seeling those.
After seeing that pic, I remember seeing those years ago.
on a regular torch, I've seen a long lever, what's that for?
turning up the o2 for cutting steel
[06:10:32] <Jymmm> http://www.bernzomatic.com/bernzomatic/consumer/jhtml/category.jhtml?catId=BernzoCat100039
[06:12:24] <Jymmm> http://184.108.40.206/search?q=cache:kGXeUlwy5YUJ:www.midwayautosupply.com/detailedproductdescription.asp%3F11631+070042196146+&hl=en&gl=us&ct=clnk&cd=2
can you use propane in place of maap gas?
[06:17:11] <Jymmm> http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=3687
prolly not, the nozzle orifice is usually sized for a particular gas
i think mapp is a blend
i guess you could fiddle with the valves
thats a cute little torch kit
have you seen the real one?
[06:19:17] <Jymmm> http://www.littletorch.com/
[06:19:43] <Jymmm> http://www.littletorch.com/kits.html
hehm it's availabel on ebay or the local welding supply store.
Now, this is interesting --> http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=45989
yeah and they made a special thread on those tanks just so people couldn't refill them
lol, so this lil adapter is a good thing?
I suspect that was for safety/profit
Or someone refilling 20K times and the tank being rusted out
well dont be a nitwit
lol "Do not refill when air temperature is above 85 F
you have to CHILL the small tank to fill it
yep otherwise the liquid would never get in the tank
have to do the same thing with paintball tanks
Eh, I guess a ice chest would work
TorbaX is now known as _TorbaX_away
_TorbaX_away is now known as TorbaX_away
are you at the computer cradek?
i was going to discuss this "pminmo" board. i'm considering givin' up my construction and go for that one
i guess you cnc'd the board
Bo-Dick: yes, he cnc'd the boards
Bo-Dick: Other options are etching them yourself, or sending them to online services like custompcb.com (which I used twice several years ago; the results were good, but the first time I was taken by surprise by the depanelizing fee because I didn't realize I was going to get over a dozen boards in my order)
i'm struggling through the datasheet for the L297 and 298 right now
03flo-h 07HEAD * 10emc2/tkemc.tcl: added more i18n support, and it works..
whats new ?
i think all the parts are pretty easy to obtain except the L297 and 298 themselves
wich parts ?
well what's new is that i've finally put my design to rest
the fact that the PNP transistor in the current limiter blew up when i applied 50 volts to it just was too much for me
are you using L298 as H bridge ?
i'm planning to build the "pminmo" board
arent they limited to 40 V ?
and if i'm not wrong the L297 supports compensated half-stepping which is almost as strong as full step. that is one very important detail that makes me love the L297
i hope i'm right about that
the TIP32C that i used in a linear current limiter is rated 100V over the collector and emitter. when i applied 50 volts it gave 0.5 amps for a couple of seconds then it exploded.
i've never seen a component blow like that ever before.
that was the last thing i needed
it worked when i applied 20 volts. but 50 volts was too much apparently.
btw, the "pminmo" design goes up to 45 volts and 2 amps. so i could use the driver for powerful and expensive motors if i wanna upgrade the motors. but for a start i guess i could buy motor that are maybe 25$ each or somethin'
the bottom line is that a cnc stepper driver doesn't have to cost a hundred bucks per axis.
Bo-Dick: I realize you didn't have time to measure it, but what's your guess about the voltage drop (Vce) when the transistor blew? Were you operating it in the linear region, or should it have been saturated?
well supply only fed the current regulator with 50 volts so how could the Vce be more than that?
Bo-Dick: For example, if it was dropping 10V then it was converting 10V * .5A = 5W of power into heat. That translates into a temperature increase of about 300 degrees at the junction, which is well outside the operating temperature range.
gives thermal resistance junction-to-ambient as 62.5 degrees per watt)
(assuming you didn't have a heat-sink)
i had a heat sink with thermal grease between and stuff
it operated for a couple of seconds and everything seemed to work fine. then a sudden POOOF and a lots of smoke came from it.
when operating transistors in the linear region, you can need big heatsinks. My favorite example (from a linear stepper driver design): http://www.piclist.com/images/member/RB-ezy-Q33/jelly01.jpg
well i wasn't drivin' anything. it died during the test phase
What was your test load?
the tranny is still attached to the aluminium heat sink: http://www.carmi.se/misterstarshine/img/sucker.jpg
the test-load was the multimeter measuring the current and it was 0.5 amps
look at the image and you'll see the hole in the component!
all the circumstances where within the operating conditions for the tranny. not too high voltage nor too high current.
btw, the heatsink comes from a scrapped atx pc power supply
my theory is that the PNP tranny were in a working but not healthy condition due to earlier experiments so it couldn't handle this higher voltage maybe.
so you touched one probe of the multimeter to +50V and the other probe to the transistor collector, and it measured .5A for a few seconds before the transistor went boom?
that was what happened
A multimeter measures current by measuring the voltage drop across a low-value resistor. For instance, if the resistor is 1 ohm then the voltage drop for .5A is .5V.
If the emmitter of the transistor is hooked to GND, that means the remaining voltage -- 50V - .5V = 49.5V -- has to drop across the transistor.
well the tranny is rated 100 volts for Vce
If there's another low-value resistor "below" the transistor (typically there's some other current measuring transistor in a current supply), then the drop will be a little less in the transistor. 49V.
We know the current, too (.5A)
so the power dissipated (turned into heat) inside the transistor is .5A * 49V = 24.5W
you said you had a heatsink. What is its thermal resistance in C/W or K/W?
well it is an aluminium heat sink that came from a pc power supply. i had thermal grease between too
the transistor's operating range is up to 150C, and I'll assume ambient was 25C, so you can afford a 125C temperature rise at the junction, or 125/24.5 = 5.1 C/W
well the heatsink weren't even extremely hot after the blow up
the rise junction-case is 3.125 C/W, so you only have about 2 C/W left
looking in a catalog at heatsinks intended for TO-220 parts, I don't see any that are as low a thermal resistance as 2V.
so I really think you violated the maximum junction temperature of the transistor
but hey, the tranny is rated 40W and i put only 25 watts in there.
...and the heatsink was large aluminium and i had thermal grease in between. what could i have done better?
that 40W is a theoretical figure -- 125C / (3.125 C/W)
what heatsink could be more effective than aluminium?
i.e., if you attach it to a perfect heatsink and operate it at 25C ambient
giacus is now known as giacus_afk
so the aluminium sink i used was supposed to be inefficient? remember i had a lots of heat-conducting grease in between!
the sink was a lot larger than a standard TO-220 sink
giacus_afk is now known as giacus
Bo-Dick: how old are you ?
follow my suggest
never , never be married
you can't be an hacker and be maeeried at the same time !
i do believe you. well i don't feel like a hacker right now. everything is just not going according to plans
that's a good thing
my stepper driver is crap.
who dont feel an hacker, he probably is
btw, i'm not even sure the L297 actually _has_ compensated half-stepping
Bo-Dick: the pminmo design doesn't have compensated half-stepping.
well it couldn't since that is not implemented in the 297
that means weak half-stepping
A URL I gave yesterday shows the additional circuitry to give compensation
"Bo-Dick: page 13 of this document shows how to add compensation to the half-step position: http://www.st.com/stonline/products/literature/an/1679.pdf"
it's not the positioning that is of major intrest it's the torque i'm concerned about
i'm gonna print that document anyway
it's funny that one can add-on this feature when it's not involved in the 297 sequencer
if i'm not wrong signal shaping is required in some way
By modifying the reference voltage you change the coil current
but the upper voltage limit will still be barrier
wait a minute, i think i understand
thats pretty smart!
it's a shame it's not already involved in the "pminmo" design
i would wait patiently to see an update to the "pminmo" board
btw I've only seen the torque compensation in this pdf, never in real life
is it considered front end technology or it there a catch with it?
giacus is now known as giacus_afk
cradek uses pminmo's design and I use his mill
so it works perfectly fine without this compensation. the bottom line is then that non-compensated are still strong enough?
does anyone here know anything about how interupts are being handled?
terr: for microcontrollers?
Bo-Dick: It's sure possible that shaping would increase the top speeds, but the mill's XY movement (32IPM he said yesterday) is 2x as fast as the feed rates usually used while cutting
no - for parallel port driven "cheap" controllers like Sherline's
terr: I don't understand the low-level details, but there's basically a timer interrupt that fires every BASE_PERIOD and executes some realtime code that is loaded as a kernel module.
i'm kind of rusty about interrupts but i remember the c command "int86" or "intvec" or somethin'
terr: the low-level details are handled by the particular realtime kernel (rtai or rtlinux)
Its not "how" the interupts are handled - the question is "howmany". I am under the impression we need an interupt per step (interupt per byte)
At least one interrupt per parallel port value change is required.
jepler: I know that - but any linux kernel is going to bog down over them
terr: it works surprisingly well
that is what I thought. 24,000 interupts per sec is crazy.
pardon - minute
that is 400 bytes per sec from a port that runs at ISA bus speeds
terr: all modern computers handle that with ease using realtime linux
i've got a lots of datasheets to read through now
recently alex_joni ran with BASE_PERIOD as small as 7 microsecond
compared to that, 24000/minute is nothing
btw, where can i get the L297 and 298 parts?
i guess i have to order them somehow
i don't know who has the best deal. www.mouser.com has them
I think I got them from mouser.com last time
I recommend you get at least one more L298 than you need if you can afford to
I blew one up before I got everything right
it's always a good idea to have spare part i know, believe me
back - mkin some coffee
cradek: I don't think so....
jeplar - do your math - 24000 i/m is 400i/s
off topic - Is anyone selling portables anymore with serial ports?
are there other similar devices just like the L297 and 298 that could substitute them?
cradek: no modren computer can handle the interupt load
terr: yes, you can usually do 100 times that many with ease
400 interupts per sec on a 200 step motor is 200 steps per sec = 2 revolutions per second = 120 RPM. THis is a linear travel of about 1 foot per min.
our base configuration runs with an interrupt period of .000050 seconds, and can run on a pentium I
cradek: prperly programmed I can drive 10000 times that on a 286
20,000 steps per second?
is "Mouser" a online shop?
that would be 10 feet per minute
Bo-Dick: Yes. www.mouser.com
cradek: what kind of controller?
terr: that depends on the number of steps per foot, of course
terr: just a parallel port
craket - I'm assuming 200 steps per revolution and 10 revolutions per inch - ball parks only
I have run 25000 steps per second with ease
cradek: when I talked to Sherline they were talking about 1/10 of what you are using.... IE 120 RPM
my machine only requires about 4000 steps/sec
You are running 1200 which is in line with the mechanicals
terr: that's a limitation of their motors and drivers, not emc
what is its resolution?
ya - I know.
.0003125" per step, top speed of 32 inch/minute (top speed limited by the motors, not emc)
to be truly honest. i've never shopped from an online shop ever before. how could i just click a "buy" button and get the parts shipped here?
I'm looking into what kind of controllers I need and the EMC website does not discuss the issues of interupts.
Bo-Dick: I've bought from mouser many times. In the US, I can buy on Monday and have the parts on Wednesday or Thursday.
what kind of machine do you want to run?
Bo-Dick: they've always been exactly what I bought
well i'm not in the US thats the problem. there a little atlantic ocean in between
Bo-Dick: yeah. Your choices are "expensive" or "slow", it looks like.
cradek: that is about 1667 interupts per second
the problem is that not even my expensive local dealer has these parts.
$37 via fedex (1 to 3 days) or $14 via mail (3 to 5 weeks)
cradek: how much jitter
terr: I've never measured it, but it works perfectly
cradek - I'm planning on a lathe and milling machine (probably sherline unless ppl suggest others) and then I want to build a gantry router table.
terr: many, many people use emc to run machinery like that using only parallel port running step/dir stepper drivers
for my router table I want to get high speeds since it will be at least 4x4
we run 200ipm with .0003937 step resolution. step and direction
you will probably want servos then
Bo-Dick: www.digikey.com seems to have a branch in belgium (be.digikey.com) which reports having the "E-L298N" in stock. http://be.digikey.com/scripts/DkSearch/dksus.dll?Detail?Ref=116528&Row=173664&Site=BE
I don't know much about big machines
cradek well I can see why but I can program a $1 CPU with a buffer to do it in a day... and then I don't even need a real time kernel.
Bo-Dick: surely buying and shipping within europe would be OK
terr: I think you're underestimating the task at hand; just keep emc in mind if you want to try something else first
cradek: issue is not servo verses stepper - but an encoder will be nice. (closed loop)
terr: the parallel port solution is nice for many people because of its simplicity (all complexity is in software, no special hardware is required)
cradek I _think_ I should be able to use emc to drive a closed loop system and even a smart controller
cradek: on the surface it appears nice.
terr: emc can close the loop itself, it just needs encoder counts as input and has internal PID
cradek I thought so too until sherline told me they can only get 120 RPM out of their motors.
terr: emc was first made to run closed loop servos. Steppers where an after thought
cradek: I read that - but it will starve for interupts I think.
terr: like I said, that's a limitation of their low voltage unipolar drivers.
03flo-h 07HEAD * 10emc2/tcl/bin/genedit.tcl: added i18n support
terr: you're mistaken about that, that's what realtime linux is about: predictable interrupts
what is a good driver? Do I have to build my own or pay with my 1st born?
cradek: I tried to find some good docs on RTlinux. Didn't find what I need.
terr: I agree the RTAI documentation is terrible
cradek: I ahve NO idea how they can get predictable interutps since the clock fires 18.2 times per sec.
terr: how fast do you expect a sherline to move with travel of only 4-6 inches? That's the question to answer
terr: The clock does not fire 18.2 times per second
The modern pentium PC is not an IBM XT
cradek: well - 17" on hte lathe bed with a 14" y (wide table) and the mill mounted on the lathe
jepler: let me look that up again.
Sherline makes a 17 x 14" lathe?
terr: we all use this every day, it's silly to try to convince us it can't work :-)
[14:52:02] <jepler> http://download.gna.org/rtai/documentation/kilauea/html/group__timer.html#rt_set_periodic_mode
"rt_set_periodic_mode sets the periodic mode for the timer. It consists of a fixed frequency timing of the tasks in multiple of the period set with a call to start_rt_timer(). The resolution is that of the 8254 (1193180 Hz) on a UP machine, or if the 8254 based SMP scheduler is being used. For the SMP scheduler timed by the local APIC timer and for the MUP scheduler the timer resolution is that of the local APIC timer frequency, generally the bus
frequency divided 16. Any timing request not being an integer multiple of the set timer period is satisfied at the closest period tick. It is the default mode when no call is made to set the oneshot mode.
I checked the type 08x interupt - 18.2 times/sec there is also a 01cX
jepler: where did you find it?
your interupt will be the 01Cx I think...
terr: I searched for 'apic timer' and noticed that one of the links was rtai documentation.
I have no idea what the number of the interrupt is
jepler: I searched Leo Scanlon assembler programming
I have the book open on p229
Timer issues a 1C interupt every "timer tick" which typically points to a IRET
terr: I'm happy RTAI hides these details from us
I cannot find how often a "timer tick" is issued by the 8254 or equivalent (that is an old 286 style chip)
jeff pasted above that it is 1193180 Hz
cradek: this means we get 1193180 interupts per second then.
that is full parallel port speeds.
I agree that sounds more likely than 18.2
what is your CPU system load when running 2000 interupts per sec (on a PI - 200 mHz?)
it would be very low, since emc users regularly run 10x that rate
I haven't worked with anything that slow lately
but most of us have newer machines
cradek on an old Pi we'll get 200 clocks per interupt then which is about 100 instructions - lots for the interupt handler.
I know one guy has tested emc lately on an original pentium 200
with a Pentium "M" 1.5GHz, I can run BASE_PERIOD 20 microseconds and my X desktop is still fully usable
cradek: ok - if we can run 20,000 interupts per sec then I don't need to worry!
I think I use a base period 25 us on my pentiumIII 666 and I notice no change
terr: that's what we've been trying to tell you all along!
emc reports that the number of cycles typically used in the base thread is around 1100, iirc (this is using the pentium timestamp)
cradek: ya - I had bad info. Good thing I asked.
terr: emc2 + Axis - I love it http://www.electronicsam.com/images/KandT/axisubuntu.png
terr: yes it sounds like it will work MUCH better than you expected
cradek - sounds like Milan over at Sherline is pointed in the wrong direction since he told me 120 RPM is max
I don't know the details of you conversation, but he is probably talking about a limitation of their motors or drivers
cradek 100,000 interupts per second is lots and we've got over a million accodring to the timer
remember you have to run the rest of the operating system (X, gui) while that is happening, so you do want lots leftover
vradek - he told me it was a limit in the number od steps and then I checked over a ##C and of course - they are not using RT kernels.
Um. The sherline motors run out of power to drive the leadscrews.
This has nothing to do with how fast an emc parport can handle the drives.
cradek true - but the I/O is pretty light weight - just needs a context switch and a dual processor board solves that nicely
rayh: terr knows that now :-) he had some bad information about how emc works
It will probably take a while for the "new guy" at Sherline to get hold of the system.
we are talking steppers - if you go servos you don't need as much proccessing power. (i think)
SkunkWorks: right, not nearly as much, you don't need that fast interrupt at all
skunkworks: I don't know of a servo motor source - what are the costs?
some day I'll ahve to look at that RT interupt code - there will be problems when interupts are disabled.
Terr - there is a lot of supported hardware out there. http://www.pico-systems.com/motion.html
skunkworks: how much is a decent entry level system - not junk!
I could sure use some good advise on what to buy - and I am interested in more than 4 axis.
I am probably not the one to give out info as I am looking also - http://www.mesanet.com/
servos that is
cradek: I have been leaning to sherline but maybe I should be looking at something more powerful. One project is to make an english wheel for my brother and the large wheel will be 10" in diameter.
I figured I could _probably_ figure out a way to do it with many light cuts.
that's going to be pretty big for a sherline. they are good for smaller things.
cradek: I know! exactly. But I only need to make one. it will be about 2" in width. I can probably do it on the 2000 mill
cradek however I did think it is a HUGE part for such a small machine.
[15:15:59] <SkunkWorks> http://www.vitalsystem.com/web/motion/motionLite.php
if you're a creative machinist, you can often work beyond the limitations of your equipment, but it's frustrating if you have to do that all the time. You should consider what you will *usually* use it for, and buy equipment based on that.
cradek well - I want to make a gantry router table - that is going to need some small parts. Most of the parts will be under a foot in diameter and if that large maybe 1/8 to 1/4 in thickness
cradek: I see sherline has excellent resale on ebay so if I find it is too small I can always trade up and at least I'll know why I'm doing it.
cradek: I have very little experiance at this point
that's good to know, I have never looked for sherline on ebay
cradek: if I could rent time on a lathe/mill I would.
sherline is very good at keeping their machines compatible over time, so you can buy an old one and put new accessories on it
that is probably the resale value is high
cradek over $1000 for a 5400 mill with small cheap steppers and a switching power supply and no accessories.
terr: do you have a local community college? often you can take inexpensive classes and get your hands on machinery that way
also they are very flexible
cradek ya - and I know some of the ppl there.
* Jymmm ponders... If rental mill/lathe is treated anything like rental cars with the extra insurance.... gawd help ya!
I'm going to take a short at building a plastic injection mold - I figure I'll have lots of set backs...
$3000 for the equipment seems cheap.
If I try to hire a mold maker it'll cost me over $10,000
terr what would be the MAX mold dimensions?
Jymmm: about 8 "" in diameter overall.
terr you going to mold anything specifically?
Jymmm: I figure the biggest part will be about 6-8" and made from aluminum and about 1/4 thick.
Jymmm: yes - I know what I want to make.
you going to mill out the molds?
have you seen ShapeLock? seems like it'd be good for 1-off sample parts.
the cavity will be a washer about 1/8 in thick and about 7" diameter quite forgiving in the dimensions actually
U am planning on it. Might be a failure since the project is quite ambitious
I've never heard of shapelock - what is it?
a moldable plastic, goes soft in hot water, hardens at room temp.
[15:26:14] <bill2or3> http://www.shapelock.com/
oh - that stuff - won't work - I need HDPE
ta - have heard of it.
I think so, but Im not %100
I think it'\s new, they're still offering sample packs.
for $4.95 for shipping
bill2or3: not that new - its a low melting point product - about 150F
Machine-able - Easy to saw, drill, tap and mill.
you gonna charge me $5 for one ounce of powder?! KMA
[15:28:57] <Jymmm> http://www.shapelock.com/page3.html
put it in an envelope!
I get pissed when companies do that... "FREE! Just pay $27 to cover shipping" or "Lifetime Warranty - include $8 to cover return postage" The damn flashlight only cost me $10
yeah, lots of that on ebay too.
That was Pelican flashlight by the way
buy it now: $.99, shipping, $25.
Maglight, Zippo, Cross Pens all have awesome lifetime warranties
I put my cross pen in my back pocket and sat on it, send it on and they replaced it with a brand new one - no questions.
jymmm: did you see this? http://www.electronicsam.com/images/KandT/DSC_0242.JPG
Those are what I am looking at drives for.
Hey, there is one company that makes (saddle, tank) bags for motorcycles... This is their warranty... If your bag gets damaged, even if it is your fault, we will replace it free, you can send in just the handle if that's all that's left.
SkunkWorks geckos wont work for that?
the peak amp is 80 Geckos are 20 amps - looking at rutex but would really rather had a real servo system with a motenc card.
oh thats a servo motor?
What is a really good motor/controller combo to look at? I looked at Sherline and I am not impressed. There seems to be a lot of junk out there.
steppers or servos?
I'd say pminmo l297/l298 plus any 1.2 degree steppers with <2A current. http://pminmo.com/l297-8/l297-8.htm
skunkworks: dunno actually. I'm interested in both and would prefer closed loop unless it is very expensive.
there are also a few companies that are making servo drives that take step and direction input.
like geckos and rutex. (although I consider it cheating ;))
well - I know so little that I really need to be wound up and pointed in the right direction.
I don't yet have any equipment and I know so little that I can easily make many mistakes.
(that is the problem - there is no right direction)
my experiance is with things like Calcomp (stepper - then servo) adn HP plotters (servo) - I've never run an NC or CNC mill or lathe
I think the biggest issue is what to avoid.
there are some systems at www.stepperworld.com - specifically an FET4 with 190 in-oz motors.
but I do not know if they are any good.
THen I see on ebay a chap who says he has some chopper drivers and motors up to 250 in-oz
I don't know anything about FET4. If it's not a "chopper" design then it probably will give disappointing performance.
Jymmm: what control are you using? xyltec or somehting like that?
jepler - well that is good advice - stay away from a power-transistor design
you want pwm and if you can - microstepping (for steppers)
the FET$ says it microsteps.
how much should I expect to pay for a motor / driver?
could not remeber how to spell it
terr, costs varies widely. I got mine for about $100 per set, but that's pretty low-end.
If we are dealing with a chopper design - then why can't they just be plugged into the mains?
terr: excluding the circuit board itself, I think the parts for each axis of a pminmo l297/298 board are around $20-$30
jepler - that is reasonable.
terr: the transistors that do the switching have a maximum voltage
terr: because the transistors typically have lower maximum voltages than mains voltage.
e.g., l298 is 46V IIRC
terr: not to mention you need DC
jepler: I read that 30-40volts is fine - what of 170 or 340?
and if you're not crazy, isolation
in fact the dc power supply is often an expensive part of the total.
cradek: put a transformer and box rectifier and big capactitor in
if your looking for that high voltage - you are going to have to look at industrial drives.
terr: An L298 will smoke if you give it 170V.
we have some compumotor drives that run 170v to the steppers.
jepler - what about using a dimmer switch for a light bulb?
but they have there own internal powersupply also.
skunkworks: those have to be very powerful = expensive
depends on where you find them ;)
what is the web site for pminmo?
terr: A 24V transformer + rectifier + big capacitor gives you around 35V .. that's going to be fine for running L298
[16:11:16] <jepler> http://pminmo.com/l297-8/l297-8.htm
OH! I was readign that site last nite
I can build that board - is it good?
cradek uses it on his mill. http://timeguy.com/cradek/cnc http://timeguy.com/cradek/cnc/stepper-drivers
If I use a 24V transformer - then how many amps does it need to produce? How many amps can the L297-8 switch?
L298 can do 2 amps per motor winding IIRC
cradek: check your server logs - I was on your site last nite
I believe you
cradek - couldn't find an email addy!!!
it's on every page
cradek how much per axis for the controller if I build it myself? the L298
depends how well your junkbox is stocked
cradek - I don't have one at this time
I think I paid around $100 for parts for 4 axes considering a well-stocked junkbox
that is reasonable.
you'd have to check current prices on the chips
terr you haven't seen cradek's junkbox
Your transformer and capacitor must be sized so that the ripple is not "too big". Compute the motor power, multiply by 2 for good luck, and divide by the supply voltage. That'll give you a ballpark figure for average current drawn from the supply
if I buy a 24V transformer - how many amps out should it have for 4 axis?
motor power = nominal voltage * nominal current
Jymmm: I have a friend who helped me rebuild my car engine one weekend (complete rebuild - on the WEEKEND) and he used grain bins for his spare parts... more than one
sample calc then. 1.3a per coil * 3v = 4 watts *2 = 8 * 4 = 32 / 32v = 1 amp?
oxygen torch pellets?
Jymmm: bolts, crank shafts, connecting rods - you name it. Sorted by granery bin
I presume then that a 24v 1 amp = 24watt out transformer is sized big enough for 4 motors?
that's not that much, actually.
bill2or3: I figurerd it was too small too - was on the phone with Digikey this AM - no help
I got a 24v 8A power supply for 3 steppers.
bill2or3: will it run 4 sreppers?
bill2or3: I presume you don't use a driver with a power resistor...
terr: I think you need to figure 32 W / 24V = 1.mumble amp
bill2or3: ok - suppose I get a 24x8. THen I get 4 box rectifiers from radioshack and 4 big capacitors and I've got power for 4 axis...
Can use use propane in place of acetelyne (sp) ?
honestly, I dont really know that much about it, but I think 1A is pretty low.
but 24V 2A (48VA) should do for the motors in question
also, everyone told me 24v was borderline adequate, and I should get something higher.
jepler - I wasn't sure if I need to figure the 24V is going to be used at 3V or 24V since it will be chopped down
but there were $$$ consideratiosn, so I got what I could.
Jymmm: yes - but you need a different little thingy in the head
bill2or3: I hear you
bill2or3: if you need high RPM then you need higher volts - this gets chopped down to about 3 volts and it only is used to overcome the capacitance of the motor
IE - charge the coil real fast so it starts working before its shut off.
terr: inductance, I think you mean
ya - right.
jeplar coil capacitance is the inductance I think...
coil and everything else - like it might be negatively charged and you have to reverse that too.
I've never taken any EE but this makes sense to me.
I am left with the idea that for high RPM that a 3 volt motor can be driven at MUCH MUCH higher volts - like maybe over 50 for a millisecond at say 1000 RPM because it is only going to be on for a short pulse.
I think that's the basic idea.
3 volts hold - very high volts when at high rpms.
bill2or3: no - the choppers I read about still current limit to 3volts - they just get the coil charged in milliseconds
As for sizing the capacitor, you can use a formula like the one shown here: http://www.kpsec.freeuk.com/powersup.htm#smoothing
heh! I thought of using a couple batteries and tying in a couple float chargers.
A chopper driver shouldn't care about the input voltage because it is going to chop it anyways.
it's true that it's less sensitive to ripple than, say, the CPU in your computer
jepler I joined the local robotics group and I know I'll meet some keen folks there who can help me with the electronics.
Using the formulat on that page, 5 * 1A / 24V 60 Hz == 3472 micro F
terr, my advice is plan on spending a lot more than you plan on. :-|
terr: are you also considering to build the "pminmo" drive?
terr: Don't underestimate the value of learning on your own or from the web
terr: there's a lot of info out there
03flo-h 07HEAD * 10emc2/src/po/de.msg: updates (created from xgettext, poedit, msgfmt) description follows
Bo-Dick: yes - the guys here speak highly of it so I'll follow in their footsteps
terr, I've been making some notes (mostly just a list of urls) as I collect info, I could send the list to you if you'd like.
not just stuff about drivers, but general DIY-Mill stuff.
i think it could be something. it's rather simple. the only problem i have right now is that i don't know where to get the L297 and 298
bill2or3: this will be GREATLY apprecited
msg me an email address.
I don't mind posting my addy because if a spammer finds it welcome to the flock and I do not consern myself with viruses
another must is to implement the half-step compensation too
I didnt get a message.
I just posted it - scroll back
heh, got it
on it's way.
Bo-Dick when I read the poop on that driver I was left with the impression he knows what he is doing and it is a good design.
are you building from scratch, or converting an existing mill?
keep in touch. :-)
bill will do!
step 1. double your budget.
step 2. add some more.
step 3. square that number to get hours needed
I bet it's actually cheaper to buy a used one.
but that takes the fun out of it
local supplier has L297 @ $10 bux and figured L298 is $7 bux. He suggested a www.solarbotic.com L293 kit for $17 bux
how big of motors are you going to be using?
I can call Dave over at solarbotics - he's in town - I was planning on buying the equipment through him actually and he never even suggested I buy any of his own controllers!!!
is the L293 just a smaller version of the L298?
bill2or3: that is what the website says.
I've kinda sorta been thinking about doing a second mill, tiny one for small engraving, so 293's would be good.
if the first one works, I'll use it for that. :-)
l298 is a 20 year old chip - gecko has far better solutions.
136 in oz motors are a good start but more power is better.
I'll phone gecko
I think gecko's are pretty much the best, but they're $$$
I'm on the phone with them now...
on hold of coursr=e
03flo-h 07HEAD * 10emc2/tcl/mini.tcl: fond some mor words for i18n
wow 150 bux per axis
for the stepper drivers?
for gecko's? the 302?
I got some older centent drivers used for $50 per, keep an eye on ebay.
One thing to watch on old centent is that they sometimes fail to output one microstep per step.
Mariss at Gecko designed those.
But centent wanted to protect the design so the potted 'em without cleaning the flux.
and thousands of them died like that after a while.
can we clean em up?
lousy. I'll keep an eye on that.
The potting means you can't get into the components where the damage was done.
these were used-pulls, supposedly working fine, so hopefully they're ok.
rayh how did he "pot" them?
I've got a stack of freebees from a company that had to replace them under warranty.
like, covered with epoxy, I think.
why can't we remove the expoy?
Some kind of compound that hardens. Black stuff.
plastic of some sort.
maybe we can pull the component and install a nekid one
you can, but it destroys takes the parts under it when you do.
I wonder what he used.
Wasn't mariss that did that. He did the design if I've got the story right.
[17:32:03] <Jymmm> http://www.google.com/search?q=epoxy+potting&sourceid=mozilla-search&start=0&start=0&ie=utf-8&oe=utf-8
bill2or3: I have a referal to a p/s guy - $135-190 for a 48v x 4a - 12a p/s
how big of a mill are you going to be making?
gecko drivers are up to $150 per card.
card = driver
card = driver = axis
'card' confused me, aren't they external devices?
like in thier own case and all?
need to mount them in something I think. possible a keenex box
If you want a lower cost per board, you might consider pmdx or xylotex.
rayh depends on what the board does.
Not rated for 80 volt
but okay for small steppers.
only need high volts for high speed
* Jymmm would have bought geckos instead of xylotex, if I knew then what I know now.
[17:38:04] <bill2or3> http://www.geckodrive.com/photos/g202_large.jpg
If we can get 600 ppkl together we can get them for about $92 each
I had a similar experience with xylotex but that was a long time ago. Jeff has improved the product a lot since I got them.
If you can live with headaches, not worried about having to do things over becasue of problems, go with xylotex.
I've never had any trouble with pmdx.
rayh It's primarily mid-band issues.
Jymmm: I prefer to worry about other things.
Great product, great guy. And besides he's logging here.
That's good to know Jymmm.
what do you mean "logging here"
here in the #emc channel, recording what transpires.
steves_logging is the owner of pmdx
possibly for later review.
rayh: I had a LOOOOOOOOOONG discussion with mariss about mid-band compensation in his drives
Oh - so he can sue us? or just so we say nice things about him?
one of those.
jymmm what is "mid band compensation?"
terr have tou read mariss whitepaper?
Because in his "heart" he knows EMC is right.
[17:43:10] <Jymmm> http://www.geckodrive.com/photos/Step_motor_basics.pdf
That's NOT a public link - one of these days he'll get a real website.
read that like 5 times.
steves_logging: Man, what's the deal with the ancient centronics connector?! =)
steves_logging DB25 Baby!!!!
I've listened to steve's logic. I think it has something to do with mil-spec.
rayh: Maybe, but availability of cabling is another issue.
Being a collector, I've got a pile of those cables so haven't thought about purchase in a long time.
rayh: Same here, but the damn collection is starting to bite back now adn crawlign out from the boxes they're in.
I was thinking of mentioning that it looked like a box full of snakes.
rayh: Yeah, I just keep the damn good cables now =)
rayh: I do have a obscene collection of wall xmfrs though... too many new types of connectors these days
I saw a universal adaptor for these. Sorta like an octapapus.
but you have to worry about ac v dc out and voltage and power and ...
rayh: It's far easier tu cut off the existing tip and solder shrink wrap to a new PS.
They are like $15 each these days - and nothing fancy eihter.
There you go.
I dread digging into my box of wall-warts.
mine are all spread out.
mixed into piles of scsi cables I'll never use again, but cant bring myself to throw out.
I can toss the SCSI cables easily, it's the SCSI cards I can't =)
I have an arbour press and about 1000 bux worth of components
Hey A-L-P-H-A, LTNS
rayh: Have you played with any electronic tool setters?
Not much. I service some machines with em
rayh: can your (basically) describe it?
what's a "shottky" diode?
[19:00:45] <SkunkWorks> http://en.wikipedia.org/wiki/Schottky%5Fdiode
i'm reading through the L297 and 298 datasheets
by the way, has anyone of you modified the "pminmo" design for torque compensated half-stepping?
Jymm What I'm thinking of is what most makers call a tool measuring stand.
and the tool makes contact with it?
[19:08:44] <Jymmm> http://www.penntoolco.com/catalog/products/products.cfm?categoryID=4512
Bo-Dick_: I think I'm the only one seriously using those boards to run a mill, and I used them pretty much as-is.
is it considered inadequate for the job?
rayh: I'm in need of a tool setter, want electroinc so I cna automate the process. Was thinking of making something like that url. requires contacting top and bottom aluminum plates (called and asked them)
rayh They said it's teh same plastic as cutting boards are made from, but I can't see HDPE being that thermally stable to give a repeatability of 0.0002"
03cradek 07HEAD * 10emc2/src/ (Makefile Makefile.inc.in): cut and paste errors, thanks Paul for the patch
CIA-8: how do you change color on the text?
CIA-8 is not a person... It tells us of changes to the source
* rayh looking
* Jymmm waiting
html tags then?
show me a simple example
This may be the system your irc client uses to show colors---I don't know for sure: http://www.mirc.co.uk/help/color.txt
Jymmm, You would certainly have to find a plastic that didn't change size with temp.
That and I'd be more than a little afraid to cnc a move to it. There is no give.
rayh Yeah, I'm concerned with overrun for sure.
I'd also be willing to bet that their "repeatablility" means location at the same temp.
14,15 oh yeah
Take successive readings
rayh: Yeah - bastards. I'm not sure what polymers are thermally stable either. Is delrin?
fwiw I think color should be used not at all^W^W^Wsparingly .. don't get too carried away
Not a clue.
well anyway, so the 3pminmo1 design wasn't so widely used after all?
rayh any idea on PTFE (teflon)?
cradek Go ahead, I'll be here when you get back in say 14 hours.
cradek: did you implement the torque compensation for half-step?
Bo-Dick_: you asked that at least once already, and I answered it at least once
"Thermally stable [insert plastic here]" isn't something you search for, it's known knowledge as it's all relative anyway. It's what you can get away with is the real question =)
cradek: do you think i'd see an advantage doing that?
Bo-Dick_: I'm not sure how much it matters, I haven't studied it at all
cradek: you said you used it "as-is". have you experienced any real disadvantages with the drive that wouldn't have existed for a more expensive driver?
cradek: If the graphs in the ST application notes are to be believed, it gets you around 40% more torque at low speeds
holy lord! 40% more torque
Bo-Dick_: it is very reliable. I haven't had any problem.
Thanks rayh =)
there is LOTS of torque at low speed already...
i'm still tempted to add the feature if it isn't too extremely difficult
(page 6 of http://www.st.com/stonline/products/literature/an/1679.pdf)
Bo-Dick_: maybe you could even modify the board with the free version of Eagle
what are the limitations of the free version of eagle?
yeah, the board is within the size limitation of the free version
(eagle is commercial software)
SkunkWorks: 80x100mm board size, I think
and 1 schematic sheet
rayh: One idea I had was using ceramic substrates as the insulator.
and only two layers
and no commercial use
rayh should be thermally stable enough for avg work.
does it do autorouting?
(but I have no idea how good its autorouting is compared to any other product)
but you CAN give away your modified design
as you should since the design is "open source"
Ceramic should work.
strange that the pminmo hasn't been updated yet on the site
Bo-Dick_: hasn't been updated?
rayh: You know, I keep trying different methods, the best so far has been using a bright flashlight and lowering Z till the shadow meets the tip. I've gotten it down to about 0.001"
rayh (could be better than that, but I have nothing that'll mesasure any smaller =)
I wouldn't blindly apply the schematic shown on page 13 of that pdf; I think that if the "REF" input is produced by a resistive divider (say a 10k trimpot) it isn't going to work
I'd build the pminmo design exactly as it is
03rayh 07HEAD * 10emc2/tcl/bin/halconfig.tcl: added tuning from ini variables
it's nice to see these commit messages again. I'm glad we're back to development...
the capacitor C7, intended to keep the reference voltage from rippling, will also interfere with the half-step-compensation circuit (the time constant of R16/C7 is around 1ms, big compared to the step time)
Well I cut crude and use a slip of paper. Measure it with micrometer.
If you go to deep the paper will tear if just right you'll feel the cutting point drag.
rayh: Yeah, I've been mucking up using that method. either too deep or not deep enough.
good feeler gauges are not much more expensive than paper...
rayh I'd love to attach to Z limit switch and zero out that way.
cradek it's not the tool as much as the method itself.
Oh you're looking for a solid repeatable home postition.
prone to user errors.
rayh: kinda, yeah. I'm workign with some thin materials.
I spec'd automationdirect prox switches on my last project.
Without the index pulse, I could get about 0.0008 repeatability
With large .75 cube steel blocks for pickup.
I have some switches coming that have a 0.0003" repeatability (mechanically), so well see how that goes.
With the index pulse, I could get a lot closer.
indec pulse? Are these roatary?
And in HAL setup of such a thing is almost trivial.
Yes 500 ppr encoders.
rotary or linear?
So the index pulse is 1/500 of a revolution.
The usdigital rotary is what I used.
is this about avoiding plunging into the table, or about finding the top-of-work? It seems like the latter is where you need the greater accuracy (to cut .003" into .032" nominal plexiglass that may vary .002 or more from piece to piece)
when cutting so shallow you always have a flatness problem too
giacus_afk is now known as giacus
yeah I'm amazed that didn't affect the stuff I did with "stippler"
hello everybody :)
jepler: Ok, here's an example... Lets say I'm engraving cast acrylic. I leave a backer on to let me "cutout" the pieces without having to put bridges in the workpiece. I need to get thru the material, but not thru the backer paper.
rayh: any idea about a new project to start ?
let me know :)
a simple one
Ah. You need some sacraficial hardboard underneath.
rayh: Actually no (already have that), but if I cut thru the paper, the cutout piece will move and get damaged.
would someone be tempted to perform the torque compensation modification to the pminmo?
I've seen hold down systems that were vacuum and used MDF.
you know i dont have the eagle software
Bo-Dick_: there's a free version, runs on linux
how is going emc2 GUI work ?
rayh: Right, but you loose vacuum when cutting irregularlly shaped pieces.
There would be some loss that depends upon the width of the cutter.
I've wished for a vacuum system too but was worried it wouldn't work when it came to cutting out the circuit board from the blank.
But I've seen stuff cut that way pretty easily.
rayh When I'm cutting out 50 pieces from a 12" sq workpiece, theere's a LOT of vacuum loss.
rayh: wich hardboard ?
You'd have to find a surplus milking pump
Medium Density Fiberboard.
rayh not alot of dairies up here =)
rayh: any websites with advice for diy vacuum hold-down?
jepler LOTS of them
Don't have any listing.
I thought maybe somebody could recommend a good one. There's a lot of crap sites out there, I dunno if any of you have noticed taht.
Roberts Hydraulics always lists a couple of GAST vacuum pumps for not much.
I havent really seen much diy vacuum stuff.
ariable-sized dots �
its a ice name for a project to start
ci sei ?
jepler: What I NEED is a laser engraver (which doens't need any clamping), what I could AFFORD was a Gantry router =)
Jymmm: you bough a laser ?
what does it cost to get a diy laser etch setup, assuming you shop for bargains/used/etc ?
say, something that'll mark aluminum.
now you're thinking I missed 't' in bough
instead I did it, because so K4ts translator cant get the term
he doesnt read english?
yeah she read using babel
but she do not want learn it
I sayd, a poor english is better then nothing
as the mine
but she wont understand :((
so, K4ts if you'll use the translator forever you'll never learn nothing :(
K4ts: you tired ?
ho no.. she's singing
sorry for the break
rayh: acn you talk me about projects to start ?
emc related, of course
K4ts: Jymmm is busy !
anche senza traduttore
rayh: you married ?
K4ts: ci sei ?
secondo me o ci hanno messi in ignore ..
ma chi ha fatto nulla?
K4ts: you muted the chat :(
Sorry giacus. I was off working on software.'
K4ts: again here ??!
rayh: ti parlava
K4ts: find 'pasqua' in english
Hi giacus. You wanted to talk about emc projects?
rayh: well, I found a name :P
next step should be about the project
What do you want to do with it?
Ah. We say happy easter.
rayh: first question I've, is better an interpreted language(python, Ruby..) or C ?
in the case of emc
Jeppler has done a lot with python. Most of my work is with Tcl/Tk.
and what could be useful ?
I know jepler is a Python developer
I understand ruby is very compact
giacus: It's perfectly possible to use interpreted languages like Tcl and Python to improve emc
tkemc, axis, pickconfig, etc., are all written in interpreted languages.
rayh: as compact Ruby could have limits ?
meant this ?
Yes and it might be difficult to apply to the way emc is setup.
I played a bit with Clipper
I've a bit of knowledge about C
could be better ?
Say you want to write a ruby program that can turn estop off. You will have to create a way for ruby programs to send NML messages (for instance, by creating an interface to the C++ library). This might be a difficult project if you aren't already very familiar with ruby
Probably Python is the most flexible for us.
if you pick tcl or python, you can use 'iosh' (tcl) or 'import emc' (python) to do those things.
latest language I used was borland delphi (windows)
The borland stuf for linux (Kylix) was interesting.
I tried kylix for some time
as you robably know, it faild in open source development
03cradek 07HEAD * 10emc2/src/emc/task/iotaskintf.cc:
Add a missing free (although it does no harm to not have it - we're in
the process of exiting anyway.) Thanks paul for the patch.
rayh: yeah ..
but seems obsolete yet
was a rad
I've some book of Python here
I seen the samples, code seems very neat
but I like Ruby
I don't know python but it seems that if there were a library of standard emc kinds of routines.
the problem is, Ruby has not so mab libs as Python
It would make coding easier for all.
rayh: as part of axis, I've written interfaces to some parts of emc
Right I saw a bit of that.
like this? http://www.madlibs.org/cgi-bin/madlib?linux.ml2
NIST did some work on modules that exported their own graphical info.
is "Intestines" a singular noun?
it might be, but that's bizarre
look photo my dog a rayh
show the photos of your dog to rayh ?
giacus: "Former French Interior Minister Charles Pasqua is put under investigation over the Iraq oil-for-food scandal."
what were you saying about Pasqua earlier?
just the wishes
its a bad moment here ..
some day yet, and I hope goverment will change
sunday wi'll vote
so, in or out
rayh: a question about documents
I suppose you use lyx to write documentation
is right ?
Yes a lot.
any xhtml converter for it ?
There are lyx docs in the cvs repository
I found an html-converter
yeah, I know
rayh: you familure with the term "Coefficient of Linear Thermal Expansion"?
there are several html converters and a docbook shtml also.
I have a book that lists those numbers for different numbers, sorta cool.
rayh: I found lyx2html
rayh Do you think 3.8 would be low enough to provide .0001" stability across 40 - 120F ?
that's is ok for html, not for xhtml
I don't know how to do those computations.
and I was thinking to try LaTex
rayh: Ah, ok. Me neither! lol.... LES!!!!!
that already have xhtml converter
You got it Jymmm
Well latex is the intermediadiate in lyx to most anything else.
Jymmm: to be honest .0001 is a lot
A lyx file is latex with a wrapper.
rayh btw... I think I might have found a polymer to use =) We'll see
giacus thermally stable
Jymmm: kidding ..
rayh: "Poor man's electronic tool setter"
jepler: what you approach with C instead ?
As long as it's less than $60 like that other one.
and what you think about it
rayh that other one is a POS
rayh like you said repeatability of 0.0002" is not compensated across temp range.
I'd like to see it made with springs to absorb overtravel.
rayh Can't do it with springs for that price range
Rather than squish out the sides like a hot icecream sandwitch.
I know it's there somewhere.
but still want "something" for overtravel though
I think that approach would be awesome with emc.
slap it on the table, hit a button, adn your done sorta thing?
The probe routine does an excellent job of recording position when the circuit closes.
Yep and I'd be quick to build that ability into the tool table setting routines.
it's finding a good cheap switch that's the problem.
jepler: sorry.. I remake a question, I do not want to start a 'discussion war about languages', just to know your opinion, what you think about C and interpreted languages ?
* giacus need to have a clean idea to start with something
before spend lot of time :)
giacus my dog saturday
as well, i'd prefer interpreted languages for my own stuff
return my home
I'm happy for you and the dog
the puppy *
giacus: I think that I can develop software faster in Python than in C, but C is sometimes useful for code which must be 'high performance'.
giacus: (it would be a frustrating exercise to try to write realtime components in python, for instance)
jepler: so, the 'highest level' side of Axis is written in Python
but it interface and calls low level routines in C, is right ?
giacus: Right. And some lower-level parts are written in C and C++ -- the parts that interface to emc are written in "C" by necessity
giacus: of course, it uses many libraries which are written in compiled languages (OpenGL, for instance; parts of that are probably even written in assembler)
giacus: One part of axis is written in "C" not out of necessity but for better speed. This is the part that gets an updated position from emc once every 10ms.
03rayh 07HEAD * 10emc2/tcl/bin/halconfig.tcl: small fix to display errors from halcmd in show.
jepler: is axis using Gtk libs too ?
jepler: how is the mod coming that does faster updates?
SkunkWorks: it's in the CVS and it works well.
giacus: no. It uses Tk for the graphical user interface.
testing or head?
jepler: Tk should take some advantage too in old HD machine I suppose
so, should be better for that
as requirements I meant
I chose Tk because I know it much better than gtk, and because the python-to-tk interface has been a standard part of Python for many years.
while I don't know pygtk well at all, and it's an add-on package for many linux distributions
(including the one where I started developing axis)
so, interfacing procedures and functions (and RT components) by Emc should not be hard by Python at the end
except for something, as you sayd
I still don't understand what you just said.
Might be worried about interfacing with NML
I meant interfacing low level functions in emc should not be hard By Python
you sayd in some case you used C code to implement high speed routines
its right ?
As an example, someone familiar with writing new Python modules in C would be able to make all the things 'halcmd' does possible from Python too
and then you could use Python+Gtk, Python+Tk, or whatever, to write a GUI for configuring HAL
is that what you mean?
last question :)
what's the biggest difficult you found developing Axis ?
or fixing it
There are two problems that come up again and again. The first one is units---so metric users have often run into problems that I never saw while testing the software myself.
the second one is staying compatible with emc1, bdi4emc, and emc2, because they're all a little different, and I only test regularly on emc2 now.
so far I've found nobody willing to help me by sending patches; they just throw rotten tomatoes at me
sometimes including me
oh yeah and there's always the problem of my good-for-nothing codeveloper
and me too ;)
jepler: nice, thank you
at least the tomatoes part
giacus: if you ever have specific questions about python I'd be happy to try to answer them.
jepler: , to be honest I know just a bit about Python
I tried a bit of Ruby code, and I found it great
* alex_joni heads to bed
I haven't learned ruby
good night all
I'm headed out too
go to sleep all you europeans .. it's late!
maybe tomorrow it will be a better day..
Ruby/RoR is like that latest pop star, Real hot till the next one comes along.
tried and failed on most I did today :(
Jymmm: you mean they never make it to v2? :D
jepler: I think Ruby for you should be a joke
ok, I'm out.. bye all
alex_joni never make it to Alpha
the only problem , is'nt have too much libs at the moment
but yes, I also like Python, because it can save lot of time developing
its very cool
for what I learned, a same thing should not be written two times
I like this concept
and the true problem its the time :(
K4ts: you there ??
I always have doubts starting something
because I know, if it don't like me I will change it later
that's wahy I ask a lot
but to change its losted time too
so, wonder if to change K4ts or not ..
K4ts: wanna change ?
losted K4ts ..
jepler: what I've here is, Programming with Python updated release 2.3
should be a good manual to start
togheter with Axis code
giacus is now known as giacus_afk
giacus_afk is now known as giacus
perchè scrivi qui
che mi sforzo
e poi mi bannano
che parlo italiano
you're not identified to the server ..
I have to ask a pleasure
sorry.. I deleted it :/
you should be able to find this book in naples: Revolution OS II
[22:30:36] <giacus> http://books.google.it/books?ie=UTF-8&vid=ISBN8850323271&id=vLbSCFblSbgC&pg=PA9&lpg=PA9&dq=revolution+os&sig=YZK_LHIF9XSTvzklg8iYZhy3-WQ
can you buy it for me please ? :)
Revolution OS II. Software libero, propriet� intellettuale, cultura e politica. Con DVD
2 not 1
II not I
mi compri quel libro/dvd per piacere ? :P
ma scusa fai prima a parlare italaino
la vers. 2
mo vado a letto
quando vieni a pasqua?
lo compriamo a pasqua o forse vieni?
ok, thank you
se non vieni niente regalo
sorry for my italian
K4ts: your italian is bad !
00:35 < K4ts> sorry for my italian
maybe she was meant my english ..
k4ts is niceee :P
she's an hacker, in the kitchen
dunno in the bed ..
PDF ruler generator: http://bald.unpy.net:8080/
I had a good chat with one of the techs in gecodrive.
giacus: is she bare foot?
He tells me that a stepper can be poistioned to 1/2000th of a circle using micro-stepping
terr: let me guess; you should buy a geckodrive
I just cannot see how this is physically possible
jepler: of course!
hacked geckos ? :P
MY MIND IS FULL OF NUMBERS !
WHo here usues linux and more specifically has tried to install wine?
I like wine
terr: I have my doubts about whether it's possible, but also whether it's useful. Take a 16TPI thread and a 400 halfsteps-per-revolution motor. That's .000015 inches per step.
er, I got an extra 0 in there
jepler: it is not useful. THere is no machine that has less backlash
it's already small compared to the tolerance of the parts you're likely to mill
backlash is typically 2 thou
on the phone -
off for dinner
terr: many people have much less backlash (I suspect I have <.0002)
gawd - my kid wants me to convert mmcf nat gas to cu meters
then he wants to go to tcf
Canada produced 5.18 trillion cubic feet of nat gas last year
we are down from 6.2 to 5.18
% units '5.8 trillion feet3' m3
gawd - he wants me to work out all the tar sands profitability numbers - in my head - over the phone.
ya - I'm enjoyiung a glass of wine.
that after figuring out the motors I need... or dont' need
I found some that run Nema #23 282 in-oz and cost $39 bux new
Then I was told if you drop a stepper from a meter it demagnitizes. I think there are a few laws of physics that prohibit that idea.
that sounds like a nice motor at a very good price
Yesterday I was told I could not get an interupt per byte sent out the parallel port.
cradek thankyou for restoring my faith in humanity and intelligence
terr: did you get the bad information about emc directly from sherline?
I got misleading - I won't say bad - yet
gecko is -ve on emc
I'm sad to hear that
I know why
I presently don't know enuf to judge anything.
well hang out here and you're bound to pick up a lot. I still do.
I'm going to buy some motors and start to do some stuff. I'm thrilled under Linux that I can get isa bus speeds on the parallel port
cradek: What's your opinion on this motor: (need to find it - hang on)
yes the parallel port is a surprisingly flexible interface on the modern PC, even after all these years
<span class="pagetitle2"><span style="">L23H276-30-8B (Dual Shaft) <a href="http://www.kelinginc.net/KL23H276-30-8B.pdf">Specification</a><span
style="">�� </span>Price: <span style="color: red;">$39</span>
(<span style="">� </span>in stock)</span></span>
wow, not a good paste
but I see a url in there
[23:26:10] <terr> http://www.kelinginc.net/SMotorstock.html
- #9 KL23H276-30-8B (Dual Shaft) $39
yup - sorry!
Off it went and I didn't get a chance to edit the crap.
2 Nm is 283 oz inch, which I think is pretty strong for nema23
I'll have to run the conversions one day.
low voltage, high current is good
8 wire is good so you can run it bipolar parallel
terr: what did the person at gecko suggest to you, instead of emc?
ok - I didn't knwo that. and I don't know why
jepler: didn't suggest anything.
terr: oh -- that's not very helpful
he's having fun writing asm for the G100 series controllers
has 6000 lines of ASM working... I have writtin more than 25000 lines of asm and I hate it!
I wrote VGA drivers in asm - uggh - not somting I want to support
I'm no expert but this looks like a good motor to use.
At one point I heard that gecko's "g-rex" actually contained some emc code (the g-code interpreter)
if it's in ASM then maybe I got bad information
oh - maybe - who knows
gcode is probably in ftn (at least partly)
apt surely must be
I'll be happy if I know a simple controller to buy and gecko's is fine as far as I'm conserned - and some good motors
then I need to buy a lathe and a mill.
I bet those motors on 50-60V on geckos with emc would be very nice.
cradek that is my conclusion - if they actually work.
oh wow, they're dual shaft too
Chap at gecko suggested 24-36V dc is fine...
* cradek drools
I can use 2-3 car batteries with chargers. lol
cradek: that's more current than your L298 system can deliver
terr: I remember hearing talk about 36v car batteries - do you know if those ever happened?
last eveing I did a simple wiring diagram to hook up float charger to 3 12V batteries and run the batteries in series and sat ther for an hour wondering why this simply cannot work - and found the short! on paper! gawd.
nope I don't see any evidence of it on google - maybe I imagined that
no 48 is common
oh is it? ok I bet that was it
I did figure out how to make a DC step up transformer... not too hard
yeah open up any old tube car radio
I don't know about that.
but why not use a normal power supply of a big line transformer and big capacitor?
cradek: Do you know the L/R of your motors?
I understand the dynomotion interpreter (probably written in C, for the TI dsp) is based on the EMC (www.linuxcnc.org) interpreter. If the manufacturer took a recent EMC version and started with that then it should be under the GPL and the dsp code thus should also be. But if the manufacturer started with an old emc interpreter (public domain) then they might keep their dsp code closed source."
I wonder if this ^^^ is what I heard of
its easily done by charging caps in parallel and discharging in series
terr: yes, but that's not in any way a transformer
cradek: I don't care what its called - it'll work - maybe even well
cradek: what would you call it?
cradek: seems like your feed rates are limited not by the holding torque but by the high-speed torque. That depends more on the supply voltage and L/R of the motor than anything else.
terr: A C-W voltage doubler? http://home.earthlink.net/~jimlux/hv/cw1.htm
damn - I just got another email on how to increase size. damn - women had this figured out more than a million years ago
jepler: I dunno
"The classic multistage diode/capacitor voltage multipler, popularized by Cockroft and Walton, is probably the most popular means of generating high voltages at *low currents* at low cost. "
jepler: I am not sure I understand that. I know next to nothing about electonics
I have this feeling great joys are in store for me
stupid dsl is so unreliable when it rains
what did I miss?
jepler: I am not sure I understand that. I know next to nothing about electonics
jepler: I am not sure I understand that. I know next to nothing about electonics
I have this feeling great joys are in store for me
jepler: Fred stopped by and said he's going to fully comment and fix all the original code.
Jymmm: oh? who's he, then?
(and anyway didn't cradek take away his cvs access?)
jepler_: Who's Fred?
terr: I don't know much about electronics either.
terr: but I can search for web pages and paste parts of them
jepler_: Fred Proctor is the guy that WROTE EMC originally
jymm I think that is a good idea. Lets keep the original of course
jepler_: I want to get some stuff going. I get lied to a lot.
jepler_ is now known as jepler
jepler: I really need good advice what to buy
already I learned that 18.2 interupts per sec is low and that I don't want power resistors
oh channel split or everyone took off for the beach
terr: I stand by my advice---I've seen the pminmo l297/l298 in action and it is quite reliable.
jepler: hey - I'm not challenging you.
terr: but I don't have any specific mill or motor recommendations.
I'll bne happy to build it and find out why I need better
terr: if I disappear it may be because the rain has knocked me offline.
jepler: you need a paradyme MVL.
The gecko sure is better in two very quantifiable respects: 8A (vs 2A), 80V (vs 46V).
jepler: that is not important.
The drivers only go to the trough if they need feed. Its how they feed the pig that counts.
microstepping and/or half-step torque compensation (bo-dick's most recent concern) may give higher step rates in practice, but I haven't ever used a system with those characteristics.