well you are at the right place...i'm a commercial user/machine designer
router or mill?
It's a Jet JVM 830 mill
let me hunt up a picture
www.matronics.com/cnc2/ is a jvm 830
and I see you are on the wiki...so am I
I put myself in there a year ago and then fizzled out...
ok got a picture.
what is the travel?
So much for the live cd. Funny though EMC would run on the hosed CD but on this one it fails to connect, might be that I don't have a driver attached.
Play with it more later.
That mill is not mine. I emailed the owner and he didn't upgrade to ballscrews. I think the travel is approx 27x10 or so.
ok so 20-25 mm screw
I was thinking of get screw stock and machineing my own ends. Not sure about dealing with the heat treat and I'm have machined anything since high school (1988)
Well, I machine ballscrew ends fully hardenend
but it's slow
one trick is to grind off most of the case with a hand angle grinder
I have a 13x40 lathe and a little bit of tooling. Not sure about feeds-n-speeds.
the middle is softer
That's a good tip...Thanks.
cemented carbide, lots of coolant slow with light feeds
a tool post grinder helps.
I do have some carbide insert tools...cemented is better for this application?
I don't have a tool post grinder...yet.
yes because they rapidly dull due to interupted cuts in rc60 hardend steel. You can keep a diamond file handy to renew the edge
once the cut is continuous things get better
See what a little experiance will do!....I have much to learn.
and the core is usually rc30-40....so you can single point cut threads easily
one caution: on many screws the lands are not a good mechanical datum
so collets might not work well
sometimes have to use a 4 jaw and indicate the raceway while in threading mode...matching the pitch of the screw
I have a 4-jaw. Another good tip...I need to index all my work off of the grooves...which are the whole point of a ball screw in the first place.
many rolled screws have an uncontrolled burr on the lands
Some of the ballscrew vendors have different end machine designs. I was going to go with two angular bearings.
ground too unless it's removed
that is what I use
db configured 40 degree angulars
to 5% dynamic
it's tight lathe work...tenths matter.
Probably beyond my skill level at this point. Perhaps I can buy them machined.
you can hone stuff to size with sandpaper if you don't have a grinder
What do you do for the other end of the screw?
make stuff about .001 oversize
My last ones are fixed-fixed...double angulars on both ends
but the are very long
for you a floating deep groove will be fine on one end.
Do you recommend any different brands/sources for screw stock?
ball or acme?
I use HIWIN.
very good quality, low price
I think I looked at those a while back. Rockford & THK were the only names I could remember right now though.
I need to go do some research again.
HIWIN is often half the price
look at ABBA too
although I have not tried them
I was planning to go with ground screws and pre-loaded nuts.
ok. good choice but of course expensive.
Hiwin sells an oversize ball preloaded precision rolled that is very good
last quote I got was 2.5 meterx25mm with 2 nuts for $400
they are shipped in that way, then choped for mill retrofit
they were oem for bridgeport cnc!
They should be long enough for mine then. Mine is a little smaller.
but mine mapped out much better
5 tpi is what my current acme screws are.
they are 5% of max dynamic load preloaded
sounds good ...they are a little big at 25mm...but the price is hard to beat
4140 chrome moly steel
I had planned to make my ball screws "the same" as my acme ones so I could swap them in and put the handles back on.
I need the mill to make the parts to hole my servos. Kind of a catch-22.
well that's just the ends right?
so no problem...turn em down to whatever you want
as long as there is room for the nut
That's my current unknown.
I have no idea what the yoke looks like in this machine.
just have to have a look I guess
Are there any tricks to removing a mill table?
an engine hoist to hold it up
jmk_away is now known as jmkasunich
heh...I'll tell you soon...mine is about ready for a rescrape
some kind of hoist
use T-nuts and clamps to allow you to attach the hoist above the center of gravity without straps or anything underneath the table to get in the way
I have an eninge lift that folds up. It might fit through the door ;-)
although the jet might not be that heavy
at least 150-200 lbs I bet
two people might easily take it off
well kinda easily
the advantage of a hoist is control - you don't have that "get the damned thing off (or on) before I drop it" factor
You have to "slide" it all the way off one end don't you?
I've done some machinery moving (1600 lb lathe, 1800 lb mill, comletely disassmbled and reassembled the mill) and I perfer to do it by myself
that forces me to use hoists and such instead of brute force, forces me to plan things out ahead
heh I moved that whole shop from chicago
yes, slides off
I don't ever ever want to do that again
I did mine one at a time, with months between each one ;-)
for the mill table, you probably need to remove the handle and dial and such from one or both ends, then remove at least one end bracket
mine took 2 overloaded 24 foot trucks...the biggest I could rent
probably better to do both end brackets
thn the table slides off, leaving the screw behind
don't drop the table on the screw (that's why you need the hoist)
The plan seems simple enough. The execution...we'll see.
take notes when disassembling so you remember later
I have a lot of machines coming in so I have to deal with a bit of rigging
It's time for me to take the nike approach and "just do it"
some previous owner of my mill put the X leadscrew bearings together with both angular contact bearings facing the same way
That doesn't do much good.
got bucket forks for the light stuff...forlift from the apple packing house down the street for the rest
result: next heavy cut popped the bearings out of their ways (force was applied to the "weak side" of the bearings
this mill is a horiz/vert combo, in horiz mode it sometimes needs heavy feeds
so the feed mechanism is strong
I have an escalera stair climbing hand truck for moving pinball machines. It has a forklift attachment that would be perfect for sliding the table on to.
they interlock the feed with the spindle, cause otherwise if the spindle stops the feed will break tools, or bend the horizontal arbor (1") or worse
better to lift from above if you can
do you have a clamp and t-nut set?
I'm ready for a second mill. I was toying with a used vmc, but I fear if it didn't have a high speed spindle it would be useless for production
lower forks onto table, clamp them down ;-)
and not needed for prototypes
That's what I was just thinking.
I'll have to come at it from the side though. Should still work. The hand truck is rated for 1500 lbs. I just have to be mindful of my CG.
I can lower the knee all the way down. That might help a bit.
For engineering prototype work I think an emc converted bridgeport boss would be fine.
yeah I lower the table and rotate the head upside down to move...
This is great. You have given me all kinds of ideas that I just wasn't coming up with before.
so, when designing a microstepper
how important is equal torque between steps?
I'm thinking about driving with current amounts that are like.. (0,1), (.4,1), (1,1), (1,.4), (1,0)
the torque at (0,1) is obviously (1/root(2)), or 71%, and at the next step it's 76% or so
the ideal currents are points on a sine/cosine wave
yah, I know that
that would give you constant torq
I see, you're asking how bad will it be if you approximate
Rather than using a DAC per channel, I have in my current design two comparators per channel, and 2 global dacs
Some of the data sheets, like the L6228, for quarter stepping just do very naive things-- full current or half current
I'm a servo type so not sure...but I am using stepper on a paint robot....inherently explosion resistant
trying to understand what you mean by 2 comps/chan and 2 global dacs
jmkasunich: well, rather than having a DAC that i modulate for the stepping
you have a micro or something? (you can build a simple microstepper drive without a micro or dacs)
I have 2 DACs on the board that produce preset voltages that correspond to certian current levels
I'm using a CPLD
anyways, I compare the current on each phase to both the DACs, and feed the result to the CPLD
a mux in there picks which current level I want per channel depending on the stepping sequence
so trying to do as much as possible in the digital realm
* jmkasunich is analog oriented
Trying to avoid a lot of DACs for microstepping. Those two DAC outputs are shared for all 4 phases on my board
why use DACs at all then, just use a couple of fixed resistor dividers
That's what I probably will do initially
But i'm putting down footprint for a SPI DAC and a SPI connector in
I like the idea of software programmability of current levels, but I'll need another board to actually speak the SPI and do smart things with it
(like lower current levels on inactivity, etc)
can I suggest sticking with one comparator per winding, and using a crude resistive "DAC"?
there's no microstepping then
ie, 2 or 3 digital outputs, with resistors connected to them.
the resistors all join at one node, and that is the reference for the comparator
oh, i see.
if all outs are hi, maxV, if all low, minV
the resistors can be 1K, 2K, and 4K, gives you a crude 3 bit DAC
Yah, I had thought about that or a parallel DAC
I mean, most dacs are exactly what you describe, with input buffers and an output amplifier
you could actually use custom resistor values, to get the exact voltages you want for sine/cosine
for instance, 45deg = 0.707, you can choose resistors so that a particular bit pattern gives you that
* icee nods
it would increase my component count though
and wouldn't be software programmable for current levels
the component count issue depends on your goals
I'd rather have a bunch of $0.005 resistors than one $2.00 DAC
how many microsteps per full step are you aiming for?
I guess what I'm asking is-- what is better-- microstepping with imprecise angles and precise torques.. microstepping with varying torques and precise step sizes, or not microstepping at all
i'm pretty sure the imprecise angles is a no-no
half stepping sure is easy
its a vector thing - the two are related
cradek: yah, but with half stepping the naive way you have nonuniform torques too
you're alternating between 71% and 100% torque
torque is proportional to the magnitude of the current vector times the angular displacement between that vector and the rotor position
so changing either the magnitude or the spacing will result in torque ripple
jmkasunich: you also need to factor in rotor inertia
the reaction of the rotor to the torque ripple wil be affected by rotor intertia (actually by total inertia), but the actual ripple won't be
i guess what i'm trying to ask is.. what real world performance problems can I expect from varying torque? I figure it must not be that bad, because the stupid half stepping drivers are doing this even worse than I am
well the main reason to go to microstepping is for smoothness
halfstep is smoother than full, and micro is smoother than half
as speed goes up, it's the roughness of full or half step drives that make them start to lose steps
the real-world impact of roughness depends very much on your application
* icee nods
my initial application is to run a some 200 oz-in steppers on a mini mill
but I hope to eventually use the same board to drive a full size mill, and my brother in law has interest in using it for a plasma cutter
feed rates and accelerations in all of these applications are not likely to be too extreme
you're building the drives because you want to? (because if you just want drives, it's very hard to beat the commercial stuff with a homebrew drive)
mostly because I want to, yah.
seems like you would enjoy servos...
It'll be a lot cheaper before opportunity cost is figured in.. but since it's something i enjoy i don't think opportunity cost really matters
(I do power electronics for a living, I have dozens or hundreds of power mosfets laying around, and for me it still makes more sense to use Xylotex (for small motors) or Gecko (for big ones)
yah, I know about economies of scale, and the benefits of using a proven design.
it's like scraping machines...can't justify it for business...so when I do it i'm "off"
its not just scale - especially if you are a digital only guy, you will fry some FETs before you're done, and/or have nasty problems with electrical noise
so you better double or triple the material cost for your first couple drives
jmkasunich: well, I've designed DC-DC converters before-- similar space.
OK, didn't know that
not so much back EMF, but.. i'm using IRFI1310N's with nice body diodes etc
* jmkasunich googles that part
reminds me...am I going to get all manner of rfi from that vfd?
depends on how you define "all manner"
should I clean it up with some inductors?
enough to totally fsck up AM radio reception, yes
before it is laying next to the encoder wires etc in the conduit
enough to fsck up a computer in a decent case, probably not
don't run encoder and motor wires in the same conduit
oh the servo amps pretty much trash any radio reception
jmk I pretty much have to as set up...but all encoders are diff shielded twisted pair
I've used off-the-shelf RFI filters (corcom or similar) to clean up VFDs as far as radio goes
much of the radio stuff is radiated from the power lines, so filtering at power input helps a lot
can you get shielded 3-conductor cable for the motor leads?
I know they don't like capacitive reactance on the outputs...how about some small inductors?
yeah I will shield the motor if I can find the wire
they have to not saturate at peak motor current
means they're not neccessarily as small as you might like
and anytime you have L plus (stray) C you tend to have ringing too
motor has plenty of L already
exactly - lots of L, such that it doesn't resonate with a few nF of stray C
Here's another minor issue I'm having, btw
put a few uH between the drive and the motor leads and things will be different
How do I know what voltage I can drive a stepper with?
I know stepper voltage is specified based on winding resistance, and that if i'm controlling the current that doesn't matter
that is a function of the motor, and the PWM frequency
but things cant still arc internally, etc
need to keep the PWM ripple down
yah, I know about the ripple currents and all. it would just make me feel better to know 'the coil in the motor won't arc < 30V'
ripple current is proportional to DC supply volts, and inverse to PWM freq and motor inductance
I seriously doubt arcing is a concern until you get well over 100V
you get what I'm saying, right? Even if the motor inductance is a billion henries, there's some voltage that if you give it to the motor instantaneously where it'll arc
yeah, I understand, I just don't think arcing is a concern at such low voltages
OK. It'd make me feel better if it was specified tho
in this case, it is more like di/dT = V/L, V is determined by the DC supply (clamped by the freewheel diodes)
if you are interrupting the current thru an inductor with a switch or something then you have to worry about nasty V spikes
but with a FET bridge, they are clamped
yup. power flows from ground to vcc or ground to ground when the high side is off
make sure the loop from low fet to high fet to power supply cap and back to ground at the low fet is low inductance
yah. I'm considering having planes.
I put transorbs on my servo amps, because estop disconnects the motors and shorts them. Copley said the frewheel diodes couls handle it though.
icee: how much current/voltage are you using anyway?
jmkasunich: 2.8A/phase initially. But I've designed the board to tolerate up to 8A/phase
those look like some pretty hefty FETs, more gecko sized than xylotex sized
(xylotex is 2.5A/phase I think)
because of DAC, comparator, etc errors and the magnitude of the ripple, I figure I have about 100-200mA of absolute current error
you are sensing current with a resistor in the source lead?
with a resistor on the ground side
big TO-220.. .3 ohm I think
.3 ohm is pretty high
10x the on resistance of the FETs
yah. it's a tradeoff between the offset voltages/DAC errors and the resistance value
at 8A, you'll have 2.4V and 19 watts in the resistor
actually, it's .1 ohms
.1 ohms 10W is how I have it specified
still kinda hi
partial duty cycle on those 6W, too
(i leave the low side turned on and only modulate the high side.. so things should flow directly from fet to fet without appreciable current across the resistor when 'off'
what are you using for high-side drivers?
I have 2mV of offset error in the comparator and 10mV of error from the DAC (conservatively)
don't forget tens of mV (or more) of ground bounce
I'm using the 1310N for all 4 legs, and driving it with an ir2304
hundreds of mV if you aren't carefull about layout
jmksaunich: I have a monostable so I don't sample the comparator for a microsecond or two after I turn thigns on
that's good - cause you'll have diode recovery current spikes
I'd still recommend resistor "DACs" instead of dual comparators
fast, sensitive comparators are gonna have a higher cost and higher component count than the resistive dacs
figure at least a bypass cap and a couple resistors per comparator
plus the comparator itself
is the monostable a timer inside the CPLD, or a physical circuit?
it's a timer, yah.
ok, so more-or-less free
I'm a little scared of what happens if the clock should go away
yah, pretty much.
you are gonna have the CPLD drive the IR2304 directly
(four CPLD outputs driving 4 FETS)
most of the time the clock going away isn't so bad
but going away during that 'ignore the comparator' timespan is very bad
if you want to see, btw
[01:36:22] <icee> http://lyle.org/~mlyle/stepperboard.pdf
make sure the power up and power down states of the CPLD outputs are "FET OFF" ;-)
I don't see a cap on VGATE
this is slightly out of date btw, it's before i moved the stepper state machine into the cpld
lower right of the power page
I'm accustomed to the output caps of a supply being drawn at the supply, even if smaller bypass caps are all gathered at the bottom
needed to maximize the window
LM2675 is one of those "simple switchers", right?
Yah. It's my first time using one of the national parts
shouldn't the FB pin be connected to the output?
But it seems to be pretty nice
I've used parts in that family, dunno if it was this one or not
Reminder to self: changing net names around and not doing it everywhere is bad.
built a switcher for a VFD (day job)... 300-1200VDC in, +24, -24 and +15 out, 200W total
used the simple switchers to derive additional voltages, + and - 12 from the + and - 24
jmka: I've used mostly the linear parts. But uh, the Linear data sheets lie heavily and the parts are expensive
Vgate is for the gate drivers, about 10-12V, right?
Vcc is logic supply, 5V?
12V, minus whatever decay in the charge pump and a diode drop
and Vstep is the main motor power
I hate it when people send signals flying thru the ether (personal gripe)
looking at page 4, "where is the reference voltage coming from, I just see a cap"
I prefer to draw the lines whenever possible ;-)
hehe.. yah.. net naming can be bad.
but so can pages of maze
but many circuits can be layed out so the signals flow, rather than just being a maze
I would scale the DAC outputs... at 8A, you'll have 0.8V at the comparator input
I suspect the DAC output goes 0-5V, so you'll only be using 16% of the DAC range
yah.. i could give it a lower Vref
the DAC, that is
though because of the structure of the output amplifier, etc, that won't give me 5x the resolution
no, put a 4.7K/1.0K voltage divider on the DAC output
why not just change the dac's reference?
let the DAC continue to swing 5V, minimizes the effect of DAC errors
that way it's nice and low impedance still
yah.. I'll consider doing something about it
it's a tradeoff
the DAC probably has low DC output Z, but for high frequencies not so good
rely on the caps at the comparator inputs for low Z
I need to simulate that, btw
the caps out at a distance from the DAC
to make sure it doesn't oscillate
the resistor divider would do a nice job of decoupling the DAC from the cap
yah, good point.
I'd make the caps much bigger too, unless you specifically need fast changes in threshold
* icee nods
they're just small to lower the amount of loading on the DAC. if I divide it or otherwise terminate it they can become bigger
0.1uF with a 4.7K/1K voltage divider gives 100uS time constant, that should be fast enough
another thing - I seem to recall LMx39 comparators as being pretty slow (several microseconds)
but I'm a little rusty there
I recall using LM311s or similar when we needed speed
1.3 us for a 100mV input step, 5mV overdrive typical
is that fast enough for your application?
how fast is your PWM gonna be?
I think so; the offtime is a fixed 100us
and the ontime is a minimum 1-2 us
so less than 10KHz PWM?
gonna have a lot of ripple current, and audible noise
most drives I'm aware of use 20KHz or so
well,i can turn down the offtime just by resynthesizing the cpld
why fixed off time? easier to do in the CPLD as opposed to conventional PWM?
unless I misunderstand, you mean that the period = 100uS plus whatever time it takes the current to rise to the comparator trip point
so the frequency is current dependent
and each bridge will be running at a different (and changing) frequency
yah, easy to do, and it's what the ST micro chips I used as a reference do
hmmmm, I mean
I could also count the time on and adjust the offtime
I have lots of extra gates
turn all windings on at the same time (say every 50uS, or every 100uS)
actually, i can just run the off time counter
at that time start a blanking timer for a uS or two
afterwards, start monitoring the comparators and turn off when you hit threshold
that means one global timer for overall period, one global blanking timer
It'd save lots of gates to do that, but it'd make the input ripple currents a lot worse
and it ensures that you aren't turning one bridge on while another is sampling the comparator
(which increases the risk of noise tripping the comp)
hm. yah, makes sense.
caps are your friend as far as input ripple goes
you want caps local to each bridge
from the bottom of the sense resistor to the top of each high side FET
yah.. i was planning on a couple 470uF's plus a 0.1uF ceramic local to each
ideally the bottoms of both sense resistors are physically close together, and that is the "star ground" point, where the power and control ckts meet
comparators should be as close as possible to that point
is it better to do that than to just have a plane and keep all the ground very very low impedance?
VREF bypass caps also close to that point
you want a plane too ;-)
I'm wary of dividing up power and signal grounds despite it being a common practice
g night all
you don't really have to divide it up
just make sure that power currents don't flow thru the control part of the plane
yah, that i'm careful about.
damn, this is where I really wish we were both looking at the same piece of paper and I had a pencil
I'll feed in power inbetween the logic and steppers, and put the capacitors inbetween
assume that the two sense resistors enter the ground plane near the center of the board
put the comparators, and the vref bypass caps to the right of center, with the caps as close to the sense resistors as possible
put the power bypass caps to the left of center, again as close to the sense resistors as possible
feed power in from the left
you using 2 or 4 layers for your PCB?
i'm thinking 4
split power plane, ground plane, and ground pours on top and bottom where possible
power plane is VCC on the control side of the board, VSTEP on the power side
yah. maybe a little island of vgate, too
though the currents there are not very significant.
I wouldn't bother with ground pours on the power side, I'd just make the traces from high source to low drain, and from low drain to Rsense as wide as possible
minimze the inductance in that loop
well, pours are easy to do and improve EMI
I'll guard ring the comparators, too
the uninterrupted plane is far better than a pour
I would use pours for MT1, MT2 and SENSE
I'm visualizing a line down the middle of the board
on the left, the FETs, power terminals, and motor terms
straddling the center line are the IR2304 drivers
on the right is all the control
for sense I was thinking, instead of a pour, to use coaxial routing
left of the line, the outside layers are pours of MT1, MT2 and SENSE, the plane is VSTEP
i can run a fat trace, and put a ground box all around the trace
traces on the sides are much less effective than traces under or above
well, there'll be plane under. nothing above, though
the part from Q15 source to R5 to Q16 source should be wide, short, and above an un-interrupted ground plane
the part going to the comparators can be a skinny trace going off into the control section, it should be attached near the top of R5, not near Q16 as on the schematic
floods around the skinny part are usefull (floods in the entire control section are usefull, just not so much in the power section)
doing something like this in 2 layers would be a challange
yah, i got that idea when i did some calculation on the ripple currents etc
and figured out that i needed inch wide traces
question - is the LM239 good for input signals that approach it's negative supply?
don't have a data sheet handy, and most of my work with comparators is wiith dual supplies, so it's not an issue for me
the bounce etc i can expect is well within the absolute maximum ratings
whether there's any weird behavior with stuff saturating when it goes below vss, i don't know
the saturation issue is what I was talking about
some comparators actually reverse their output state when you go below the rail
pg 15 of the data sheet shows a simplified schematic of the LM239
going more than one diode drop below ground would be a bad thing
dunno about simply going down to ground
hm. maybe i should just give up and buy a xylotex board :P
it's ok... the input common mode range goes down to zero
on the high side you aren't allowed to get higher than 1.5V below the supply, but that won't be an issue for you
hmmm... just thought of another way to address the VREF issue that avoids needing two comparators
I'm making a sign and it skipped steps (0.90")... How can i "reset" the position w/o mucking with it's "current position" ?
jmka: how's that?
dual 4 to 1 analog mux
yah, but why bother? using a quad comparator for 2 phases is cheap
yeah, the more I think about it the more I tend to agree
the only prob is that you only have two levels
hard to get precise microsteps that way
I'm trying to figure out a way to get more precise levels (and more of them) cheaply
well, I could just get a proper DAC per phase
gets pricey (and for serial dacs, slow, or complicated, or both)
the idea of using only 1 for all windings is good...
the idea of only changing it when you want to power down or reduce current is also good
just wish you could have more thresholds scaled from that one DAC output
so there's a big need for more than quarterstepping, eh?
thats what I was thinking with the mux - DAC drives a divider string, MUX selects which tap of the string feeds the comparators
depends on your applications
some work fine with full steps
Xylotex uses 1/8 steps
gecko uses 1/10
if you're gonna go to the trouble of building a drive, might as well build a high class one
Maybe I should just buy the xylotex board for now, and then build one for my full sized machines
if your current needs only go up to a couple amps, that makes a lot of sense
is the xylotex stuff robust?
Xylotex does everything in one chip (practical at that power level) and you can't possibly beat that in cost or performance with a discrete design
pretty much so - none of these drives (including Geckos or your design) will survive direct shorts on the outputs and other such abuse
I have a xylotex 3-axis board, I've used it for testing and such for a couple years
* icee nods
running how much current?
1 to 1.5A
It puzzles me a bit that they're selling 2.5A drives and bundling them with 2.8A steppers
they are adjustable
and I don't know what chips they're using, but if it's the ST micro parts, etc, I'd be wary of really running them at 2.5A
2.5A into a 2.8A stepper you don't lose much torque
they need a fan at full current
I don't use a fan on mine because they're only part load
I did manage to kill mine shortly after I got it, but it was my fault
one of the output traces runs close to a mounting hole, I used a screw with an oversize head, and it shorted the trace to grouhd
jmka: thank you for all your help with this.. it's been very useful.
I think I'll still build something.. but it might make sense to get a xylotex board to have something to play around with. it's certainly cheap enough
yeah, then you can concentrate on a high performance high power drive for the one you design
I realize I've been pretty nit-picky, sorry about that. In my day job I work on things that run hundreds of volts and hundreds of amps, so I tend to be obsessive about noise and layout issues and such
much of what I was talking about might be overkill for something at this power level
jmka: oh, it's understandable
and you found a big bug with the dc-dc converter, too, which I owe you big for :)
fresh eyes always helps... I dunno how many times I've overlooked the same mistake, then somebody else looks and it jumps out at them
that one would have been tough to fix dead-bug style.. and it would have smoked pretty much everything as vgate climbed up close to vstep
jmk: you really ought to give up this day job thing and start a business. Starving, the rolling in money, then starving again is fun!
rolling in money is nice, but starving isn't
lately the thought has crossed my mind more than once (especially when dealing with big company beaurocracy), but I know it would require intense commitment - I prefer to be able to spend my nights doing other things (like EMC)
Well business has to be something that's fun...then you just play all day.
the "business" part of it would not be fun
the "looking for customers" part especially
Although I must admit I am getting stressed out some...too much "play"?
But it sure beats corporate beaurocracy
at least I am getting stressed out over making things rather than posturing, politics, and other bs
First.... Y axis skips .90". Ok, I get that "mostly" aligned back up. Then I'm cleaning up aroudn the table and the power supply unplugs from the computer. SIGH
who is Aaaarrrg ?? O_O
just waking up .. *_*
this morning i will go to the postal office to get a card (hope is good for international payment...)
i need to buy some gecko ..
solved the electronic issue ?
I havent tried it yet
maybe after work today
les: you around ?
i'm losting in the abba catalog :\
I'm around now
g mornig :)
hi jacky. An office day for me.
i need some help.. when you can ..
Writing boring engineering reports
mmhh .. yeah
whan youve a little of time
i want to order the slides from abba
but i dont know wich type
ok. explain a little again about the machine capabilities you want
1x1 mt wood cnc machine
we talked ..
slides 15 or 20 mm
but i dont know which model
also wich lenght
let me bring the abba site up
if youve not time now
we can see later , dont worry
this morning i got a visa card from postal office .. i think its ok for online payment
so, i'm starting buy
im looking at some self lubrificant model, but im not sure ..
ok let's see...
15mm BR series
lowest precision grade will be fine
i think so
length need to be axis travel + bearing spacing
im looking here: http://test.machinesources.com/e-catalog/e/abba/
understand the these are not structural....they must be bolted to a stiff straight surface
les: do you know how to fight stiction in huydralic systems?
now...what bearing spacing are you thinking of?
les i dont what they mean with 'preload' ..
anon: florocarbon o-rings...
preload bearing-> |<--SPRING-->| <- bearing
jacky...ok i'll explain preload
les: i was thinking 1x1mt of working area..
les: how much of friction does that take care of?
preload is an interference fit...i.e. the bearing is smaller than the rail slightly and is forced on...so there is no play even with forces on the rail
anon: I just figure teflon like materials typically have a lot less stiction
VITON comes to mind
ah, wont help too much in that case..
the valve stem seals in one of my cars was viton...it was very slick
why isnt there seals pre-made in it?
even though the o-ring tends to wipe off lubricant film
well I think lots of viton orings are made
that's probably what's there..
I dont have the material specs..
jacky: moving gantry or moving table?
but it sticks too much
viton is usually green...
les: no idea about gantry :/
Well, I would suggest moving table anyway
we are thinking of replacing the cylinders with slides off a huydralic pack.. instead
les .. you know .. i need a machine that can work all day ..
I would imagine reducing seal preload will help too..until it leaks!
ah, cant have any leaks..
I understand jacky
they are precision ground..
and leakage is minimal
but so if friction also..
yes the hydraulics on my new tractor does not leak a drop
les: moving gantry should be better.. i suppose
cant really compare this to a tractor :/
Moving gantry has some advantages....but it is much harder to build and much more expensive
Moving table is easier...does not require twin screws
but it's bigger
i prefer moving gantry
les: how close can you get the fit with a toolpost grinder?
for you, one meter travel moving table will require about 2 meter rails
no problem about build it.. my cousyn have 3 lathe ..
smallest medium big
anon: I can do .002 mm with care
so.. moving gantry should be better
I wonder how close I'd end up at..
anon: depends on how good your lathe is
les: trying to decide if it would be possible to manufacture the parts instead of taking them off a huydralic pack..
you can always hone the last bit
les: what twin screen ?
gantries usually require two syncronized ballscrews...one on each side
its not a prob
just let me now what i need
ok lets do moving gantry twin screw...like my machine
les: .. i'm annoyng you
bearing spacing for x will need to be about....300 mm
but, dou you have some photos ?
to see summary assemblement
getting a link
be patience ..
[15:08:29] <les> http://www.lmwatts.com/cnc.html
anon, I am having to hone little ultrasonic resonators
assuming a machine like that then
of course with my sizes ..
a 2 mm rod slides and turns in a 1.902 hole...
no problem with honing
ok jacky 300 mm spacing...so your travel plus spacing plus ends....
equal about 1.5 meter rail lenghth
so i can get 1 mt work area
if i understand well
ballscrews can be shorter...just travel plus ends
ok les ..
so 1.5 will have a little left over...1.25 could squeak by
now i try to plain all
with the help of my cousin
it's good to make travel a little bigger than 1 meter anyway if you want to machine all of a 1 meter workpiece
i will show all to you before to buy ..
at the end i will send you a big spaghetti package :)
[15:16:31] <les> http://news.thomasnet.com/fullstory/20965/1303
les: gecko 340 should be ok for 250 servos motor
250 oz in?
les: will a oil film make that tight?
yes it would be fine
hm, maybe I should get some high precision bearings and machine a toolpost grinder.
yes I have two tp grinders...they both use class 9 bearings
themac j-40 and j45...
I think I've seen them before
[15:19:46] <les> http://www.aegis-ceramics.co.uk/themac.htm
unfortunately they cost several thousand dollars
some pop up on ebay from time to time
dumore is another good brand
too bad ebay & shipping means insane $$
and yeah you can make one....doing a class 9 spindle would be funn
everything must be .002 mm accurate
and perfectly balanced
well, I guesss that's out of the question ;)
mine runs at 40,000 rpm
if you breathe at any part it's misaligned
you really have to watch temperatures
if you do that kind of work :D
[15:23:45] <an0n> http://www.blocket.se/view/6129678.htm?caller=nbl_s&l=0&c=1&city=11
anyway...for occaisional use...just hone bores and shafts with sandpaper
I might be better having somone edm the hole and grind the shaft..
* an0n is a toolfreak
so am I
I cant help wanting to have a cylindrical grinder/surface grinder
and a edm..
it's not like I have any serious production with machine tools..
but, I still want a edm ;)
but when you need one you need one...my surface grinder is one of the least used tools...but I must have it there
I hate to be unable to machine stuff..
tool steel ;)
I could run it 24/7 to make parts..
milling some stuff isnt possible
not with my resources..
well abasic shop with screw cutting lathe, tool post grinder, surface grinder, mill, bandsaw, and welder...
and you can make most anything
les: how much the difference of cost between 15-20 mm rails ?
dou you remember ?
is it convenient 20 mm maybe ?
jacky they are probably almost the same price
20 is fine
les: still, a edm is a nice tool ;)
you could machine your cutters from solid carbide ;)
edm is handy...esp wire
I sharpen a lot of carbide with diamond
I were just kidding ;)
but, with a edm you could
Too bad I have an office day...would rather be working on mounting the new spindle
jacky, read my post here...it's about ganties
[15:39:33] <les> http://groups.yahoo.com/group/CAD_CAM_EDM_DRO/
les: sure, thanks
i'm translating .. :/
i will disturb you after .. if i can
I need to pull the gearbox apart so I can check on getting new bearings..
or order them rather
yeah, chineese crap..
replacing them with SKF counterparts
the bad skf bearings are better then the good chineese ones
I have seen plenty of good chinese bearings...but I have alo seen used bearings or used grease used in new chinese tools
well, skf is good :)
and it's the easiest thing to get..
since we sell skf at work ;)
I wonder if I should buy bearings for 10krpm..
for the mini mill?
not a minimill..
I remember now
most greased bearings around that size will be good to 10k
hm, trying to acheive 10k/rpm silently ;)
I want want my machine walking around
6200 is the most used bearing in the world I think
might be some in the gearbox
good to 12k
the gearbox is going away also..
spur gears at 10k might not be happy
I am mostly aiming to machine alu..
5kw and 10k/rpm
should make some chips
My new spindle is rated for high speed machining of al
specs say "with caution"
avoiding end loads
even though the 5kw is at 2800rpm..
I reach 4 kw at 18000
but it should be plenty anyway..
18000 to 24000 constant power
below 18000 constant torque
but I dont think I need to gear that much
I can reach 6000rpm with the VFD I am going to buy..
I don't think I will do much aluminum with this...it costs a lot and I don't want to damage it
the motor I am going to buy is cheap..
new bearings/balance is $800
so I can afford to change it..
does your mill use tapered roller spindle bearings?
those generally run at under 5k
beyond that 40 degree angulars are used
the colombo has db 15 degree angulars at the nose
I'll give SKF the specs when I order..
hence the low end load spec
so they can match my bearings for me..
that's probably the easiest way to do it
at 2800rpm I'll have 5kw..
well...I have to order a vfd today
I will call them now
maybe dropping one when I overdrive it to that speed..
should be plenty of both torque and speed for al..
where have you been all this time ??
was picking my nose
not realy. was busy forming a mold for a speaker box i am building
tried focal ?
always on rcf are u playng ?
it's a Bill Fitzmaurice designed box
it's quite the bugger to build
the thing is that i need like 4 of these boxes
and i didnt feel like shaping them all out of wood
so i made a fibre glass mold, poured it full of cement, and now i have my final mold :)
abit over kill... but the mold is going to last for ever :P
hm, how is the sound of fibreglass?
should be better than wood considering it's dencer
hm, are you sure?
yes... it's solid... far more solid that the same thickness wood
how thick are you making them?
and the more solid the box the better the sound
it's going to be about 6mm thick
trying to make it the same strength as 20mm ply
and acording to the fibreglass place... 6 mm will be good :)
considering they stop bullets with 10mm :)
maybe I should consider it for my car stereo install..
we made a awsome box out of fibre glass
for a car
it has 2x10" speakers in it
will post the pic's as soon as my server comes back online
lol.. it's online.. let me find the pic's and post them for you
do u want me to upload the building process? or just the final product?
seems i am talking to myself
rayh: question for you
rayh: you you have the source doc of Hal for Integrator ? could you send it to my mailbox please ?
not pdf ..
I have the Lyx for that. It is in sourceforge
sorry phone dragging my attention away.
rayh: good .. np
i just wnat to try to translate it in 'IT' .. im note sure but
could be funny
can you give me the url ?
I can dcc the file to you, I think.
i'm using nat
not sure i can receive it ..
you can send it in mailbox
maybe better ..
17:08 DCC SEND from rayh [220.127.116.11 port 1025]: Hal_Introduction.lyx [176kB
infact .. im tryng but wont pass..
There is the dcc.
rayh: can you send it in email please ? email@example.com
On the way to your addy.
ok, thank you
It's out of here. Don't know how long it will take to get there.
There is a very active LyX user list if you need help with that system.
I could also make a plain text file if that would be better.
which format it is ?
latex, or openoffice wuold be nice :)
I believe that there is a lyx to latex converter but don't know anything about it.
let me play here a bit.
latex conversion on the way.
I thought there was an rtf converter also but I don't see it right now.
oh .. dont worry .. its ok
found a latex to rtf converter. will try it.
rayh: file received :) thanks
conversion done. will look at with oo.
rayh: its ok as is .. np ;)
I've got an open office file here if you want it.
ccjoe_ is now known as ccjoe
It preserves most of the lyx formatting. Chapter headings don't look all that good.
rayh: thanks, i'm annoyng you
Not at all. It is a challenge
rayh: ok, 100% received
i will try to translate it in italian in the next days
then i will send you the return file
The office doc diagrams do not look very good at all.
oh .. I know ..
its not a prob
I suppose these could be replaced with the real docs.
I'll look forward to the document.
but its enough to start
the .swx is on the way.
rayh: receivedsxw, thanks a lot
Wedcome. Glad that you can make use of the info.
catch you all later.
blah. I got bored from writing reports and tried running a web telerobot somewhere. It paints on an easel.
I made a big mess. Dipped my brush too much and had paint running down the paper.
Found a pic of les http://pumpkingutter.com/pics/margepump.jpg
before morning coffee of course
hey there les.. u mind if i pm u?
Was making a sign last night... 90 minutes into it the laptop power supply came unplugged =(
if i have to write any more reports I will look like that
[18:49:14] <Jymmm> http://pumpkingutter.com/
i was just bored and painting with a web puma telrobot.
Yuga whispers sweet machining into les ear.
les meant to ask you... are you pinning your workpiece on the table on all four sides? I ask becasue my machine is 24" MAX on top/bottom so if I use 24" material I can only do left/right edges.
look what a mess I have made:
[18:53:46] <les> http://pumapaint.rwu.edu/
les. any ideas on the pm?
[10:14:15] <les> yuga: ok
Jymmm... saw that, just getting no reply :)
yeah you weren't recieving
seal heater vents
use car interior for cabinet volume
roll down windows for port
drivers in door
ok... so there must be a pannel built behing the drivers seats with the drivers in?
you know that flutter you get when you open a window at speed
must be 5-10 hz!!!
what about it?
one side of the speaker faces outside...one in
just think of the car interior as a box!
comms are a little clow here I think
problem is i dont want the speakers to fire out the boot
les check it out now - the paint robot
I can't get on...because you are
there's my artistic taelent for the day
jymmm what a nasty mess!
many try to knock over the easel and throw paint around
not like I started with a blank canvas or anything =)
or hit a PUMA singularity
well, that sucks if they abuse it like that.
oh they love it.
now...tear the paper off.
I don't know..but people have figured it out. see bloopers
hey it's movin around. that you?
but it's suppose to be yellow =)
I mean paint
ok, let me grab another color
there a happy face
new paper please!
uh...they have to change it...
or start a stroke near the edge and the paper comes off
ow you almost tore it!!!
gnam gnam time ...
les 2HP 20K rpm 120IPM 1/4" downcut into 3/4" MDF - how much stepdown can I really do w/o pushing my limit?
greetings from sweden
ok..going to bed
that bum is always sleeping! lol
oops I missed 'em.
which one? They're all running back under their ricks!
trying out the new wiki stuff, thanks to alex, paul, steve and matt.
I hate wikis.... just a bunch of electronic post-it notes.
"oh running away are ye, come back ere and I'll bite your knee-caps."
rayh you said you were goin fishing yesterday... you live near a lake?
They do tend to lack a certain bookishness.
Yes I'm right on a little lake 170 acres.
170 acre lake?!
that's a BIG lake! lol
I don't know what the conversion factors are for your measurements.
an0n a lil warning next time if you please.
that was the final warning
an0n is now known as anonimasu
rayh where is this lake of yours?
alex: you bastard! come to lule�!
anonimasu oh go make some swiss chocolate or something!
Jymmm: he's in the other end of the country
90 minutes into a sign... the power connector on the laptop falls out last night =(
I feel like making parts..
rayh ah, ok. never been there.
anytime you want to visit email. We've got spare beds.
where is home for Jymmm?
22" x 18" Harley Davidson sign... was looking great... till the power supply got disconnected
rayh 30 minutes south of San Francisco... San Jose, California
Jymmm: your first scrap piece?
Okay. I went to hs in Watsonville.
anonimasu : My first table top now.
it still feels shit even if you've scrapped 100 ;)
Spent a few growing years in Turlock.
Jymmm: nice :)
rayh oh wow... Watsonville to Turlock... what no Fresno or Stockton in there?
College in Napa valley
Girl friends in each.
rayh musta been some commute =)
Those were the early sixties.
Hey gas at 0.18 made for fairly easy travel.
anonimasu yeah, 90 minutes into it, 90 minutes togo. real disapointment
rayh try $3.10/gallon now
*chews a part of of the table with he's bare teeth*
and I drive a Expedition too =(
anonimasu it's MDF, so you REALLY dont want to do that =)
rayh BIG ouch
anonimasu chew the tabletop
Steppers arrived this morning, a lot bigger than I thought thwy would be :)!
nah it's real wood..
I may have to start traveling by thumb again.
anonimasu ...and cat urine.
rayh better watch it... some real freaks out there... even the females
I didnt really know they made mdf out of that
I resemble that remark :)
I know a bit about those folk from my college years.
anonimasu If you look at the MDF glue contents, you'll see what I mean.
rayh no, not those freaks (we have hippies in Santa Cruz), these ones are pissed, escaped from a psycho ward, and out for blood and penises for the mantel!
"It places the lotion in the basket"
Heh, califonrnicate, I have heard they gather fruits and nuts down there LOL
K`zan: No Senior... no hable engles
* K`zan wants a controller NOW :).
met a few of those along the way also.
rayh icee live in Morgan Hill
Heh, heard that also. I grew up with a military family and it was considered MOST *RUDE* to be in another country and expect them to speak your language...
What you mean? Spanish is the #1 language in California
Now, if they could vote, it be official.
Because of the average rudeness of those people I refuse to learn spanish...
has anyone ever hit 1/16" aluminim with a router bit?
anonimasu any damage to the bit?
I do however enjoy watching the results of yelling "EMIGRATIONE" in places where they gather :).
It does seem to me that they were further between back in the 60s.
someone pointed out that we are only 6 meals away from total anarchy.
Jymmm: didnt I tell yoy you were going to be machining alu?
Jymmm: no idea, it'll probably work..
anonimasu: That's not it... my machien has a 24.5" travel in the Y axis, BUT (big butt)... I can only place a 24" piece of material with no room for dowels, so I was thinking of getting a oiece of 1/16" x 3/4" x 3/4" angle and mounting it under the table as a guide but I didn't want it to muck up the bit if it hits it.
I'd grab some plastic angle if needed, but harder to find.
oh, well you will be hitting it sooner or later..
well shit... now to find some plastic piece that's angled.
I'd use drywall corner strips, but way too weak.
20x20mm square stock :)
anonimasu nah, I want something easily replacable
hm, I was thinking make like 10
then I have to store them.... not alot of room in an apartment for storage.
oh well, buy them then
maybe trim molding
if I can find it thin enough
Jymmm: Right on the storage problems in an apartment, gotta find someplace to store the copy stand between uses, I need a warehouse :-).
* K`zan in driver thrash :-/.
New steppers are burning a hole in my pocket :).
K`zan what size? nems23?
K`zan These are what I have... http://www.xylotex.com/StepperMotor.htm
These are the KP4M4-029s I got to play with, I see more capable ones in my future though :-).
Thrashing between the xylotex, HobbyCNC and building my own at this point.
well, Jacky^ has been building/testing his own for a couple months now.
hobbycnc = unipolar, xylotex = bipolar
[21:24:49] <K`zan> http://www.pminmo.com/l297-8/l297-8.htm
Look easy enough and with good prices on the chips from Futurelec, attractive.
I just see unipolar as the better choice at this point ?!?!?
Still dispelling a lot of ignorance :-/.
Heh, reading the included paper that came with the steppers, "WE'RE ALL GONNA DIE" they contain LEAD, oh heaven I don't have a permanent guantee of infinite life and health :).
K`zan: bipolar is always best, and the l298 is a bipolar driver...
YEAH BABY!!!!!!!!! http://www.professionalplastics.com/cgi-bin/main/co_disp/displ/carfnbr/25/prrfnbr/85429/Micarta-Brass-Bullet-Resistant-Laminate
cradek: SO much to learn :-(.
insmod: error inserting '/home/rayh/emcdevelop/test2/rtlib/rtapi.ko': -1 Invalid module format
K`zan: you'll learn a lot building your first driver.
Ahyone know about such an error while trying to start emc2 with bdi 4.29?
K`zan: I don't even know if you can run your 5 wire motor bipolar since the centers of the windings are tied together.
cradek: Yes, thanks much for your site and helpful comments.
cradek: Can you use a unipolar as a bipolar if you don't use the commom connection(s) ?
Uh, never mind, maybe with the 6 wire, but I doubt with the 5 wire ones...
I think :)
yes you can with 6 wire
I'm not sure what 5 wire will do. I think it might be bad.
I guess that means that my 5 wire ones are shelf decorations then :-(.
maybe you can open them up and make them 6 or 4 wire
Heh, one pays for an education one way or the other :-).
remember not to pull the armature out of the magnets.
just pull off the cap nearest where the wires enter.
* K`zan hauls one out of bag and grabs phillips...
one web page suggests hooking up the center tap in a surprising way, but I think it just gives you a needlessly complicated unipolar drive... "To correctly connect a 5-wire to a bipolar drive, the center tap must be connected to the motor supply." -- http://www.motiongroup.com/steppermotors_basics.shtml
Wires disappear into the coils...
Might me a good thing to meter out the leads on one so I have a clue what is what.
or google for a diagram :-).
Heh, for as "popular" as these steppers are supposed to be there appears to be NO data out thereon them other than people asking for it...
dosent the manufacturer have the specs?
Apparently not, odd for a stepper that has been in goddess only knows how many million floppy drives...
K`zan: xylotex sells these to go with his allegro-based bipolar boards (XS3525/8S-x), do you have documentation explaining how to hook those motors to those boards?
oh, wait, I got confused. Jymmm is the one with the xylotex motors, not you? nevermind.
jepler: No inforation at all, abs nothing on the japan servo company site either...
K`zan: if your meter can read the low coil resistances, it is easy to find which is the center tap.
then for unipolar it's just a matter of figuring out which order makes it rotate clockwise
Got the docs on how to do that, working on it now.
I don't know any way to find the correct operating current, though
I have japan servo steppers..
they work great :9
if they're from floppy drives, then a good guess is whatever current +12V gives you.
jepler: Fortunately they were cheap (<$3) so they were acquired to be educational and maybe even useful :).
this one? http://www.doc.ic.ac.uk/~ih/doc/stepper/kp4m4/
+12v dc, four-phase, unipolar, permanent magnet, 3.6° per step
is it salvaged off a floppy?
Probably, that is the only reference I can find.
I didn't got them from here: http://www.goldmine-elec-products.com/prodinfo.asp?number=G14781&variation=&aitem=3&mitem=6
tou know, I think I bought some of those
cant they give you the datasheets?
I was going to use it to give my etch-a-sketch finer motion, but I got bored and never did
(it's got 7.5 degree motors now)
there seems to be plenty of info about it
Where are you finding this ?!?
[22:29:59] <K`zan> http://www.goldmine-elec-products.com/prodinfo.asp?number=G14781&variation=&aitem=3&mitem=6
Got it metered out...
anonimasu: have you heard the news? mr Joni has come to town :)
the other end of town ;)
yeah, the scary skåneland
very scary place :/
[22:37:47] <ValarQ> http://www.spelcenter.com/gbs/index.php
hm brb, going out to remove the gearbox off the mill..
need to get it togther I want to make chips as soon as I can
no reply ...
Jacky^: sorry, evening
anyone here speak engllish well ?
ValarQ: yes .. :\
damn bad english ..
let me know ..
Jacky^: no, we are no native speakers either :(
this morning i send an email:
(sorry for my bad english.. i'm from Italy)
I need to buy soon 3 Gecko G340, but i don't have credit card (just bancomat with Cirrus,FastPay,PagoBancomat circuits).
Do you accept some alternative payment, like postal office payments ?
Is there some Geckodrive affiliate in Europe ?
Waiting for your reply I thanks you a lot.
is understanble ?
well, its pretty ok
what i meant ?
well .. no reply ..
i guess you mean some instead of soon
soon = fast ?
yeah, but then it's in the wrong order
but it should be clean ..
"I soon need to buy..." would be more correct
thats ok ..
that why i wrote: (sorry for my bad english.. i'm from Italy)
will wait tomorrow ..
well, i can still make out what you wan't
there shouldn't be too much trouble with that :/
i need to know
if i can pay with visa card
probably yes ..
its postal office card
i'm off to bed now, goodnight folks
it seem they accept Visa card
I didnt get the mill into molecules today
I failed on getting a shaft out..
ended up making special toold..
and I coudlnt find the die.. for the end of the tool ;)