gmane.linux.distributions.emc.user

Re: Pluto-P

Peter C. Wallace — 2008-12-04T04:10:15

My choice would be whatever is most common in 100 Bt PCI add-ons. Probably a Rhine clone of some type. Peter Wallace Mesa Electronics (\__/) (='.'=) This is Bunny. Copy and paste bunny into your (")_(") signature to help him gain world domination. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Pluto-P

Chris Radek — 2008-12-04T04:04:33

I think the answer is explicitly in what Peter said above. You don't care what comes in the computer. You point them to a particular network card because it has the right chipset. It'll be about $5. Heck, you might say, as someone who sells hardware that works with it, you'll just send him one off your stack of them. They're $5, after all. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Pluto-P

Jon Elson — 2008-12-04T04:00:49

The Arm 7 chips would be fine for the stuff I want to do with them. There are a few Arm 7 chips that have 10/100 ethernet support built into the chip. On Linux, a standard network stack that was accessible from the real time environment would work in the case I'm thinking of. The target device would only speak when spoken to, it would always be idle when receiving a command so it should be able to respond quite quickly, like within 10 us or so. The scheme I am thinking of is the regular dispatch of the servo thread sends a request to the target device, it reads encoder counts and switch condition and sends back a response packet. The PID calculates new velocities and sends a command packet, there's no need to reply to that. This would allow standard protocol stacks to be used on the Arm target CPU. So, the only thing I don't know how to do is get a protocol stack that is accessible by an RT thread. rtnet is a time-sliced variant of ethernet, and requires at the very least modified protoc

Re: Pluto-P

Jon Elson — 2008-12-04T03:48:18

Having fought this problem to some extent with the par port, I can't IMAGINE the headache when a potential customer shows up and you tell him "Oh, it only works on computers with a DEC Tulip ethernet chip." Add, he says, "How do I tell, sight unseen, whether any particular computer has that specific chip?" No, it has to have a bit wider hardware support to be viable. Jon ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Ethernet, rtai, and rtnet

Ray Henry — 2008-12-03T23:46:46

http://www.ce.utwente.nl/rtweb/publications/MSc2004/pdf-files/011CE2004_Buit.pdf An interesting study of RTnet. In it they say; "RTnet communication times are mostly determined by the hardware. Not only processor speed but also architecture and type of network interface card (NIC) are of great importance. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Pluto-P

Ray Henry — 2008-12-03T23:17:53

This sounds like just the right approach. We should pick a PCI ethernet card with a chip we can work with and begin developing a HAL driver so that when you get the ethernet version of the 7I43 working we have a way to begin testing. Any thoughts on a specific card to begin testing with? Rayh ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Surface Finsishing

John Thornton — 2008-12-03T23:14:16

I think he uses belts like this 47505A444 from mcmaster carr, a simple phone call would confirm the usage with them. and here is the Aluminum Polishing 101 pdf http://www.home-machine-shop.com/Down-Load/Polishing_Aluminum.pdf John On 3 Dec 2008 at 7:51, admin-W4ay7r7vdIs< at >public.gmane.org wrote: ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: axis use of memory

John Thornton — 2008-12-03T23:14:16

I would guess you are right on the edge... If you used stepconf to generate your config try putting a larger number in for the latencey test this will allow more time for other things to happen like drawing the screen... John On 3 Dec 2008 at 7:19, ulises barrera wrote: ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Pluto-P

Peter C. Wallace — 2008-12-03T18:16:17

Actually we started a 7I43 with Luminary Micro ARM Ethernet interface CPU back in May or so but custom work kept me from finishing it. I will start again soon. As John says the real problem is getting realtime Ethernet support on the EMC host. As I said before maybe the way to ease into this is just support 1 or a few Ethernet chips, and require the user to have a add-in PCI/PCIe Ethernet card with the required chip. Peter Wallace Mesa Electronics (\__/) (='.'=) This is Bunny. Copy and paste bunny into your (")_(") signature to help him gain world domination. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Pluto-P

John Kasunich — 2008-12-03T18:03:50

As Peter pointed out, it isn't the FPGA that is the limiting factor, it is the computer interface. The reason the parallel port (in EPP mode) works for connecting to external devices is that the parallel port is simple and well defined and the SAME in just about every PC. (Even so, EPP version 1.7 vs EPP version 1.9 has caused some grief.) Ethernet on the other hand has about a thousand different chips that can be used on network cards. The vast army of Linux programmers has written a thousand network drivers, so our PCs can connect to the world using Ethernet. But those Linux drivers are not realtime capable. For EMC to support Ethernet, we need realtime drivers. There are some out there, associated with a project called RTNet, but so far nobody is motivated enough to figure out how to use them. The same thing goes for USB, only about 10x worse. Personally, I'll stick with Parport EPP for low cost, and PCI for performance. Both are extremely generic and don't need special drivers for r

Re: How can I calculate the INPUT_SCALE value?

John Kasunich — 2008-12-03T17:58:14

I glossed over quadrature 1x vs 4x, and "counts per rev" vs "cycles per rev" vs "pulses per rev", because Howard's error is not a nice simple factor of four. This is something to keep in mind in general though - encoder specs can be ambiguous. There are no "ugly" numbers with EMC2. Scaling is done in floating point, so it doesn't matter if you have 1003.267 counts per mm, EMC will move the correct distance (within one or two counts, of course). Doesn't matter at all. Regards, John Kasunich ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Pluto-P

Kenneth Lerman — 2008-12-03T17:54:09

Hey Peter, You don't happen to have an ethernet interface IFDEFed out there too, do you? Ken <..........snip..........> <..........snip..........>

Re: How can I calculate the INPUT_SCALE value?

Alan Condit — 2008-12-03T17:49:56

One thought that I had, was is the 2000 ppr before multiplying by 4 for quadrature? Another thought, while the screw is 4 tpi, is it a single start or 2 start thread? These would not get you to 1250 scale but would get you to ~1417. Alan --- Alan Condit 1085 Tierra Ct. Woodburn, OR 97071 Email -- acondit-93PhqhqcWuo< at >public.gmane.org Home-Office (503) 982-0906 ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: How can I calculate the INPUT_SCALE value?

Jon Elson — 2008-12-03T17:40:09

Are you off by a factor of 4? Your encoder may be 2000 cycles/rev, wich will produce 8000 quadrature counts/rev. I'm not coming up with anything that works out near 1250. You might turn the motor one revolution exactly and see what the change is on the display, to see if the stated encoder count is wrong. Depending on how you are reading the encoder, there may be a hal pin that produces absolute encoder counts without scaling to linear units. Also, you should crank the encoder backa and forth a few times and return to the same spot, and see if the position reading is repeatable. If the encoder signals are weak, you might be losing counts. Jon ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=10

Re: How can I calculate the INPUT_SCALE value?

tomp — 2008-12-03T17:33:26

Howard Howard Chan wrote: howard look at the numbers without the formula, just to get the relationships... i think... you have 2.25 turn of motor to get 1 turn of screw ( 36/16 = 2.25) and 1 turn of motor makes 2000PPR which is 8000 counts ( due to quadrature) so for 1 turn of screw you have 18000 counts from motor ( 8000 counts per motor turn * 2.25 motor turns) and 1 turn of screw is 0.250" (inches) which is 6.35mm (0.25 * 25.4 = 6.35 ) so, 6.35mm travel = 18000 count so 6350 microns = 18000 counts so 1um of travel = 18000/6350 = 2.8346457 counts thats a pretty ugly number but the sanyo denkii driver likely has 'electronic gearing' which is just a multiplier and a divider the incoming counts from the encoder is multiplied by the numerator and divided by the denominator and the output counts are reported to the control and there are an infinite number of screw and gear and ppr combinations that wont yield nice counts per micron so it is possible that you cannot get it with the combination you hav

Re: tool table.

Jon Elson — 2008-12-03T17:30:17

Yes, I think that is the best way to do it, without some changes to EMC to allow different files to be loaded by command. That might be a useful feature. Jon ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Pluto-P

Peter C. Wallace — 2008-12-03T17:21:10

I doubt the Pluto has enough gates for a decent serial interface along with good set of peripherals. It could be done with our 7I43 though. Its just a minor change from the USB-HostMot2 firmware to interface to a internal UART instead of the USB data. I think our parser code has a UART interface IFDEFed out at the moment. But the real problem is using it. Other than (upcoming?) Modbus support, I dont think there is any support for serial interfaces in EMC. Regular PC RS-232 serial ports are probably too slow for most real-time tasks anyway at about 90 uSec per character (10 bits 115200 baud)... Peter Wallace Mesa Electronics (\__/) (='.'=) This is Bunny. Copy and paste bunny into your (")_(") signature to help him gain world domination. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two t

Re: How can I calculate the INPUT_SCALE value?

John Kasunich — 2008-12-03T16:16:19

Your calculations appear to be correct. But obviously something is not working. The screw pitch and pulley ratios are easy to verify, but the encoder counts might not be what they should be. Here is an experiment: TURN OFF MOTOR POWER - safety first! Start EMC, then use HALMeter to examine the raw unscaled counts from the encoders. Make a mark on the motor pulley, and write down the number from halmeter. Then rotate the motor shaft exactly one turn by hand. See how much the halmeter number changes. Rotate the motor back to the original position and make sure that the count goes back to the original number. You might want to turn the motor shaft ten turns and divide by ten for more accuracy. Any non-repeatability might indicate problems in the encoder system, which could be electrical connections, noise, or even mechanical - something like a slipping coupling between motor and encoder. Hope this helps, John Kasunich ----------------------------------------------------------------------

Re: How can I calculate the INPUT_SCALE value?

Stephen Wille Padnos — 2008-12-03T16:16:03

It sounds like you have 3600 count encoders, not 2000. The rest of your math is correct as far as I can see. Try this: set the INPUT_SCALE to 1000 Leave the motor disconnected from the screw and put a mark on the pulley Jog far enough to make the pulley move 1 turn (by eye at least, more accurately if you can) This will tell you whether the motor is moving as far as you think it is. If it is, you should check the table movement by measuring how far it moves when you turn the screw 1 (or 10) turns. You probably don't need to re-count the pulley teeth :) - Steve ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: tool table.

tomp — 2008-12-03T16:15:24

Dave Dave Engvall wrote: could symlinks do it? 1 file always used by emc but that file is a symlink to a file with a useful name just a thought ( hi from hong kong!) tomp ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Surface Finsishing

Wayne Parks — 2008-12-03T16:02:52

These sites deal with the finishing and polishing of Aluminum http://www.car-accessories-magazine.com/aluminum-polishing.html http://www.irinfo.com/polish/html/polish.html A Scotch-Brite™ Cut and Polish Flap Brush in a hand held drill would work. I could also see using a small round abrasive pad with a foam backer in the spindle of your mill and run a small G code program that moves the work piece approximately 3/4 the diameter of the pad and then lowering the head to just contact the work surface and repeat this the full length of the part resulting in the traditional circle polishing affect common on aluminum. Wayne ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/