Printable View
That's a lovely machine I'm interested in why the z ballscrew/lead screw is so big?Quote:
Originally Posted by AndrewMawson
This is what I know as SY cable Attachment 21878 metal braided armoured cable it really doesn't bend much.
This is not what's in your energy chain is it?
No idea !
I bought this machine as a barely working home made table, wiring was atrocious, much of the mechanics were ropey, but it worked. (I needed to cut some bits for a 6 foot flail mower I was repairing), but I bought it in the knowledge that I'd eventually have to re-build it.
So as not to hi-jack Chaz's thread there is a re-build thread here:
http://madmodder.net/index.php/topic,11797.0.html
Looking good, try tuning manually - adjust the I & D values, there are documents online about doing this but its also in the CS-LABS booklet. The delay between reverse is a homing error - its not settling at at the zero point fast enough you should see the numbers very slowly retuning to zero and then the green LED comes on in the tuning page. Tweak the values and you should see following error drop and the cycle time rise - in the end it will be going bang-bang-bang one way then instantly the other. You need fast homing (return to zero) and not a slow creep.Quote:
Originally Posted by Chaz
The auto-tune is pretty pants really, always tune with the final load, rough tuning is fine for tests but when finished and assembled tune again with all load attached.
Thanks. Ye, its dwelling too long before it finds 0 / Green LED. Playing with the values will improved that. Ill sort that out once its mounted properly and I have limit switches to stop a run away / something going wrong.Quote:
Originally Posted by Davek0974
The Panasonic drives also offer a lot of tuning. I am not sure how the CS Labs and Panasonic tuning will go exist. Ill likely set the drive to be on a mild auto tune and then use the CS Labs for the overall tuning.
Very curious too to understand how the setup can deal with inertia / whiplash. I refer to demos where they move something that has a ball on a stick and tune the drive that basically shows no 'pendulum' effect when the device stops and starts in different directions.
Cable ordered.
Any ideas for bellows / metal guards for protecting the main table itself?
Was thinking of making some custom stainless plate ones. I know someone that has sheet bending stuff that was kind enough to offer me access if need (many thanks), is there value in DIY for this or look at something else?
An example of something that can be bought - https://www.aliexpress.com/store/pro...920640239.html - not cheap but it may take me a lot longer to make something like that than is needed.
Thoughts?
I was told you should tune the drives and motors first then tune the controller and drives.Quote:
Originally Posted by Chaz
Always tune with average working load on the table and a tool in the head.
My drives were from China and the manual is terrible so i have not been able to tune them yet - i just relied on some serious controller tuning and it seems ok so far
Ye, also heard this. The CS Labs manual states the same. My Kinco drives are like that, dont know if they are tuned or not. The Panasonics offer an ongoing autotune feature which Ill try but also concerned that it might mess things around.Quote:
Originally Posted by Davek0974
Chaz,
For the kinco drivers you have you should tune the velocity loop with the autotuning with the load applied. (connected to the ballscrew and table mass) and once done, you can then do the tuning with the cslabs controller. the tuning fails, reset the driver to defaults, set the driver up again and tune again.
Quote:
Originally Posted by Chaz
Gary, thanks as always. I've not been able to get the serial comms working and even though I've tuned the drives manually, I am uncertain if the value it pops out after auto tuning is actually used.
I had also tuned them with no load which is incorrect but I'll realpolitik at that.
Quote:
Originally Posted by Gary
you just need to have an RS232 connection to use the servo software, without Rs232, you will need to get a USB to RS232 adaptor.
I have never used the interface on the driver to do any tuning, because it is so much easier using the software.
Tuning with no load is not right and may cause the controller tuning to fail, or make the axis very spongy, so make sure you have the mass of the table there while tuning.
Also make sure you save it after, or you will delete the tuning, once you cycle the power.
Ye, ive been unlucky. Ive got 2 machines with native RS 232 ports as well as a USB one and PCI Express one. I cant get any of the 4 to work.
I need to do it again, but busy finishing Thor. If all goes to plan it might do some basic movement this weekend on all 3 axis.
I've got a cheap USB adapter, and have never had any problems. I did have to destroy the case to get it to directly fit onto the drives though.
Are you using the correct type of cable?
You'll need to check if it should be a straight through cable, or crossover cable (I would guess it should be a crossover for serial comms). Probably best to test the one you've already got, and then order the opposite.
I had tried a cable and had cut / moved the wires manually. I should just order one of each and test again.Quote:
Originally Posted by m_c
Chaz do you have the cable diagram for it. RS232 receive & transmit are on pins 2 & 3 sometimes they have to be reversed. Ground (if I remember) is on pin 7 on a DB25 plug and pin 5 on a DB9 plugQuote:
Originally Posted by Chaz
So what I've done is order a new F / M straight cable. The Kinco drive manual shows as default 2-2 3-3 and 7-7 wiring. Will try that first. Struggling to find (on Amazon) and null modem M / F cable in case.Quote:
Originally Posted by Clive S
Problem Ive had is that on my Panasonic drives that use Serial / proprietary, Ive also struggled to get it working.
I have a stack of about half a dozen HP DC7600 small form factor PC's that will shortly be going to the tip - you are welcome to all or any if you wish FOC. They have RS232 ports and multiple USB's - all have had their hard disks removed for security reasons - they are ex-corporate machines.
(the clear out continues !)
Please. I wouldnt find 2 for using as basic PCs / Mach 3 / Serial device machines. Ill make a plan to collect from you or similar.Quote:
Originally Posted by AndrewMawson
sweet jebus Chaz, don't do things by halves do you!
great work, look forward to seeing this thing in action
Thanks, nothing like your workmanship however. Perhaps Thor can build some pretty parts.Quote:
Originally Posted by Zeeflyboy
Hi Gary, would a USB to RS232 interface not suffer the same issues as standard USB when it comes to timing's? Or is there something I am missing here.Quote:
Originally Posted by Gary
I also wonder if Linux is more reliable than windows (Especially when using USB) especially a Linuxcnc Live cd? I have no idea if you can get Mach 3 for Linux.
Chaz I'll be interested to see if it was a crossed cable causing your issues, fingers crossed for your weekend movements ;)
USB serial works, its low baud rates.Quote:
Originally Posted by Desertboy
Mach 3 for Linux, no. There is Linux CNC which is meant to be good, but won't support my cslabs stuff as far as I'm aware.
Thanks, will update over the weekend.
RS232 is only used to set up the kinco ac servo driver using the kinco servo software, it has nothing to do with cnc.
Quote:
Originally Posted by Desertboy
Is the CNC ran through Ethernet? I think I must have missed a link somewhere along the wayQuote:
Originally Posted by Gary
yes, he uses the csmio-ip-a that links to the PC with an ethernet cable.
Its alive !!!!
Still got to mount the spindle, wire some limits, tidy, PID tune, test. But getting there.
https://youtu.be/iJt-4_rUux4
So I tried to do the 'double up' with the limit switch for X. Make it home and limit. It goes home, but then when you try and drive the machine manually it says 'limit hit'. I know there is a manual override on the settings page, will try that.
Holly Jaws!!!:excitement: you are almostthere!
So, managed to make a small cut, nothing exciting. I need to sort the table. X and Y limit switches mounted and wired in. Having an issue with my Y motor holder, its not 100% level and causing issues, will sort.
Any thoughts on what I could use as a table? Ive seen alu T plates and similar for sale on ebay.
https://youtu.be/B7rVrqa3GqI
https://youtu.be/1cgwDLBofAo
Argh, good progress, getting tuning right and then with the limits / home switches I was doing some test. I ended up somehow going past the X ++ endstop which has fortunately only cost me lost time as I now need to repack the ballnut and some of the cart bearings. Could have been a lot worse. Not sure why it ignored the limit switch.
I was setting the X and Y axis to be reversed as they are the wrong way around ( + and - direction ). Do people normally do it on Mach 3 or reverse either the -10 to 10V or swap motor leads (and encoder feedback)?
Bad luck Chaz
However I do seem to remember posting this about a week ago (post #283)
"However my personal preference on big powerful machines is to have the home switches inside the boundries set by the limit switches, and have the limit switches trigger powering off of the drives. Also to have the limit switches within the boundries of mechanical stops that limit travel before ball screws over travel."
Hi Andrew, would a latched relay be suitable to perform the functions you describe?Quote:
Originally Posted by AndrewMawson
Andrew, yes, point made. I thought of the same.Quote:
Originally Posted by Desertboy
Desertboy, fairly easily done with the CS Labs. It has sub 100ms error checking and can instantly disable the servos in a fault / trip condition.
I think the main point is to have a physical limit to save the ball screws if there is a runaway - preferably with a hard neoprene or other suitable plastic buffer.
As for the latched relay then yes that's the easy way to do it, but it needs to be a fast acting relay.
Suppose this means it works.
https://www.youtube.com/watch?v=QAv-q9P4FjA
Speeds and Feeds
24000 rpm - 6mm endmill
Cutting Rate 4000mm/min
Lead / Ramp at 1440 mm/min
This was run at 80% speed. Overall happy with the cut considering there is no finish pass.
DOC is 5mm, WOC is 1.2mm
Still got more to do, but happy with the initial results.
https://c1.staticflickr.com/5/4239/3...34ca42b0_b.jpg
Excellent - congratulations !
BTW do you still want those Compac PC's ?
Please, ill make a plan to come and collect. Dont chuck them please.Quote:
Originally Posted by AndrewMawson
I haven't felt any disturbance in the force yet, I think you need to up the ante.
Milling cutters start to perform around 3/4" diameter and up, anything less than that is squidgy :very_drunk:
Agreed.Quote:
Originally Posted by Robin Hewitt
Just did some test cuts, 10m/min with 10mm endmill. 0.5mm Stepover (wimpy, im still testing) and around 20mm of depth. The surface finish is superb.
Will do some more cutting tomorrow with 10mm endmill.
https://youtu.be/dqnT4AKWgvw