I've seen people connecting step and dir of both x axis motor drivers to the same output on a bob. It's there some reason this isn't a good idea? If the IP-M can source/sink enough current this would work to slave, yes?
Printable View
I've seen people connecting step and dir of both x axis motor drivers to the same output on a bob. It's there some reason this isn't a good idea? If the IP-M can source/sink enough current this would work to slave, yes?
Yes your correct Gerry it doesn't and can't ever do so not even with a firmware update. I've contacted Cslabs about this and the Hardware doesn't support it so unless they change the hardware it can't ever use slaved motors.Quote:
Originally Posted by Ger21
No this isn't a good idea. While in theory and practice it does works it gives problems due to the way drives work regards handling midband resonance. Modern drives compensate for mid band resonance and apply correction, problem is the correction only happens on one motor but at same time forces the other motor into resonance so it's a bad idea and often results in lost steps and crazy weird shit happening. (If 2 motors sharing one drive)Quote:
Originally Posted by Tumblebeer
Also then there's slight timing differences and if your lucky enough to run the motors in range where you avoid resonance the timings can slowly run the machine into loosing positon and racking etc.!
Connect just like any other drive except you connect to the IP-M Output terminal connections instead of going thru a BOB. Also the outputs are differential outputs meaning each output has it's own Positive & negative signal and are not connected to the common Ground like most regular BOB's do. This gives the best protection from cross interference and noise immunity, they are also very high speed signals as they are not optoisolated which can slow down signals. This high speed and noise immunity does come at a slight cost in that you lose some protection. Like I say these drives are industry standard and designed to work with Servo drives or high quality Digital servo drives that take full advantage of the extra speed and clarity of the signals and with this it's given that the person installing is competant. They don't suffer poor workmanship and misswiring.!Quote:
Originally Posted by Ger21
Some experience from my build.
If you make your long rails perfectly parallel / using some hard stop fixture and precise square and rule or straight edge/ and perfectly level -using epoxy, then will be very easy to align perfectly the ball screws. That would lead to extremely precise movement where the only factor left would be a missed step due to who the hell knows what reason.
So doing all this correctly will lead to such ease of movement, as i discovered some days ago, that the gantry will move without racking even with one motor only and the other tripped. The tripped motor ball screw will turn by itself even with speeds like 10m per minute as i said-without racking.
What i want to say is not worry, if you like the controller, wire it like Dean says, you can use hard stops to Zero the gantry axis at some very low speed / this programmed at the soft limits tab in mach3/
Another thing is that the Leadshine stall detect works better than expected/ found with faulty LPT cable/ so if you wire the alarm to say Enable or something similar or alarm input/ the machine will stop right away. Then take the power of the motors, move screws by hand so to say to center more or less the gantry, then run again and Zero.
This is a worse case scenario which may happen very rarely if ever when the machine is tuned.
Anyways, doing the machine precise is a must and so is centering the ball screws. Otherwise there will be problems later. Like running out of square/the gantry and the job, cause the travel of rails differs between them/due to slight angle out of parallel/ or between rail and screw/for the same reason as before said/, and that means adjusting steps in Mach separately for the gantry motors, which i believe could be a nightmare, especially if you aim precision.
As to the Americans and their preferred drives and BOBs. I have Geckos on my machine and they never lost a step for 3 years, coupled with Campbel design BOB. So,for sure there is a reason why. Too bad that they are far away and we have to pay tax.
About the CSMIO 6 axis controllers. Cheap from industrial point of view. Very expensive from DIY point of view. Not worth the effort for me. On my next build which would be quite industrial i spend nights contemplating to buy or not. 549.00 EUR + 23%=667euro. Not for me. I bought 8 axis Galil for 500eur which though risque, seemed quite a better purchase with much more possibilities.
At the end it seems there is almost no choice of BOBs, though it looks at first glance that there are a lot of them.
I'm not saying to connect two motors to one driver, but rather connect two drivers to one step output. The aim of a slaved axis would be to provide two identical step pulses anyway, yes? There's no feedback from the driver to the controller anyway, so it can't possibly adjust more synchronized with separate step pulses than with two parallelled identical?Quote:
Originally Posted by JAZZCNC
I know what your saying but the answer is not simple and several factors play apart. The BOB output circuitry plays a hige part and like wise so does the drive input circuitry. Often output signals are buffered and filtered on the BOB and only have a certain amount of current available and if the drives are optoisolated which often they are there may not be enough current to drive both opto's correctly and then signals start going stray. This is just a simple version and there's more to it than this but the up shot is your playing with fire and likely prone to timing errors.Quote:
Originally Posted by Tumblebeer
IMO why bother to go to all the trouble and time building the machine only to spoil and cripple it for the sake of a few more pounds.?
Real stuff are real time linux kernel + linuxcnc + mesa fpga or any other card that supports real time.
If you want industrial level controller with low cost that is the way.
USB is not a real time interface all usb boards use buffers mach3 is not real time because windows is not real time.
I've been using "industrial" machines that cost $150,000 or more for almost 20 years. They all run from a Windows PC.
I doubt that there are any real commercially available "industrial " machines running Linux CNC.
LinuxCNC is a viable alternative. But the LinuxCNC guys need to stop knocking Mach3 every chance they get.
Mach3 users probably outnumber LinuxCNC users by at least 20:1. So Mach3 can't be all that bad.
I have seen an industrial machine 160k euros or 216k usd it is semi-realtime because it has a realtime fanuc motion controller (with buffer)
but it don't show in real time at windows program what the machine is doing, is realtime but it has no realtime feedback to windows (windows is not realtime).
The only realtime part that machine have is the fanuc controller not the windows software...
Linuxcnc is not alternative.. is the only way to get a soft real time industrial controller with low cost.
I dont want to knock mach3 but the difference with linuxcnc is huge (difference to the underlying os) that is the simple truth.
The problem is not in mach3 is to the windows which is not real time os.In the other side linux has a realtime kernel.
You can get realtime with mach3 but only with a hardware realtime motion controller and buffer.
Caveat with linuxcnc is more difficult learning curve.
Any feedback on this below,theres not that much chat about em on the web...,im not sure how long they've been around.
http://cncdrive.com/UC300.html
I think there are a lot of guys out there milling and turning parts who have a problem they don't know about, how about you don't tell them and they'll just keep making the parts ;-)Quote:
Originally Posted by vre
- Nick