shansen

It's funny you posted this, I had the exact same problem last week but never got around to solving it. I would be very interested if anyone has an answer!

KEJR, I rarely need motion programs, so I haven't really looked into doing them without Delta Tau's IDE. However, if motion programs are interpreted instead of compiled, it would seem straightforward to use Eclipse as an editor (just create a custom syntax highlighting for motion programs). When you want to download it to the Power PMAC, Eclipse has built in functionality to do this over SSH. On a side note, Eclipse has a "Remote Project" functionality, which lets you store, compile, and debug a project remotely (e.g. over SSH). What makes it so cool is that you can put Eclipse on a flash-drive, and then go to any computer (Windows, Linux, etc) and develop C and C++ for Power PMAC without installing anything. The downside is that this approach requires hardware (a physical Power PMAC). Obviously the typical Power PMAC user should use DT's IDE as recommended, but if you are really looking for C++ code this is one possible way to do it.

KEJR, I have used Eclipse for a little bit of C++ development (and much more C development) and I love it. One of the big benefits is that Eclipse lets me organize my projects however I see fit. It has its quirks like any software, but I highly recommend it if you are interested in C++ (or C) development for Power PMAC. Plus, it's free. Can't get better than that!

KEJR, I can confirm that the Power PMAC IDE will not connect if you change the format of the prompt (including adding colors, etc). I had the same thing happen to me a month or so ago. The IDE expects a certain format of response string, and will not connect unless it gets an exact match.

I have one Power PMAC that seems to have a problem accessing the filesystem. If I try to make an image of the 8GB flash, it takes nearly 7-8 minutes (compared to 1.5 minutes on many other Power PMACs). Also, when I remount the file system from r/w to r/o or vis versa, it takes several seconds. The other Power PMACs seem to remount their file system near-instantly. Apart from the slowness, nothing else seems to be wrong. All software runs as expected (so far), and I'm not seeing any strange error messages. Any thoughts?

Biocat, Completely out of the blue, I had EXACTLY the same issue as you today. I was testing some new code on a PPMAC that I've used for over a year, and the processor hung. After rebooting, the kernel panicked and would not load the file system. To fix it: 1) Remove the mezzanine board from the bad Power PMAC 2) Wait 30 seconds 3) Reinstall the mezzanine board on the bad Power PMAC 4) Your Power PMAC should now be working. This seems to happen consistently when a real-time thread crashes. Not sure why this happens though....

As an update, I was able to successfully backup and restore by switching from WinImage to Acronis. Not sure what was causing the problem... I did find out that the write protect switch on the mezzanine board seems to be in the 'write OK' position by default from the factory. I think the board is in 'write OK' when the switch is lined up with the silkscreen, but I seem to be able to do basic file writes with the switch in either position. I can only do an image restore when the switch is lined up with the silkscreen.

OK, scratch that. I can create images, but cannot restore them (even with the write protect switch in the correct position). I get a "The drive cannot find the sector requested" error. I am using a PPMAC with 8GB of flash memory if that makes any difference.

I am fairly sure that USB power is not the issue. Even though I cannot browse the filesystem, I just verified that I can use WinImage to backup the disk. Also, I figured out how to manually mount the FAT32 partition, and sure enough I can see it in Windows explorer. For some reason it just isn't auto-mounting (I've tried this on several computers, both Desktop and Laptop and they all do the same thing). However, it looks like I can create disk images, which was the purpose of the exercise in the first place.

Hmmm...this seems to be a Windows issue. If I plug the mezzanine board into another Power PMAC and then mount it manually I can see all of the files.

I purchased the shortest mini-USB to USB cable I could find (12 inches), and have tried it with the mezzanine board attached to the Power PMAC and with the mezzanine board removed from the Power PMAC. I have also tried with the write protect switch in both positions. In all cases, a removable disk shows up in Windows explorer, but I cannot open it to view files. Here is a picture of my setup:

Someone from Delta Tau could probably give you much better help with this than I could... Just out of curiousity, have you made sure that the little mezzanine card mounted on the front of the Power PMAC is fully seated and screwed down? This is the card that the SD slot is mounted to, and it also contains the flash chips that the root filesystem reads from. Worth a shot! It has three small screws on the top of the card that keep it seated.

I use 115200 baud rate, 8 data bits, 1 stop bit, None for parity, and XON/XOFF for flow control. That output you are getting looks very strange, I'm not sure what is going on there. I attached an image of what my console looks like on powerup.

I just discovered a bug (well, unexpected behaviour at least!) when connecting to a Power PMAC over SSH. If I change the PS1 environment var via the .bashrc profile (for example, to simplify the prompt), the IDE can no longer connect to the Power PMAC! It is obviously looking for a particular response from the PPMAC on connection, and when it doesn't get a particular response it won't connect. Is there a way around this, or should I submit a feature request???

Yes, I use a Keyspan USB to serial converter, and it works great!

Sorry to revive an older thread, but I've been having the same problem as Studebaker above. I've unscrewed the mezzanine board from the Power PMAC and connected it to the PC via USB, but Windows only sees a generic disk drive which I cannot open or view files. FYI D18 and D14 light up when the USB cable is connected. The USB cable is 24", but I will try a shorter cable tomorrow if I can find one. However, you've piqued my curiosity. Why is a short USB cable required???

I had a similar problem where I couldn't communicate with the Power PMAC, and I finally solved it by reading debug messages printed to the serial port (it turned out the RAM wasn't fully seated in my case). The fastest way to debug those "no boot" problems seems to be via the serial port. Just plug in an RS232 cable, open up PuTTy (or a similar terminal), and power on the Power PMAC. You should get a stream of information that will help you figure out what is going on.

@symetrie: I am not from Delta Tau, but the programs under Realtime Routines cannot access shared libraries by default. As an alternative you can take your library functions and add them in to the Realtime routines directly (e.g. duplicate the library functions in your usrcode). Regarding folding/outlining, this feature is language specific in Visual Studio (e.g. specific to C#, VB, C++, etc) and I don't believe that Delta Tau has implemented this feature for C or Script languages. Perhaps Delta Tau can give you better information about whether this feature will be incorporated in future releases.

It is definitely possible from a Capp, however I don't believe it is possible directly from a PLC. If you want, your CPLC can set a bit that tells a background Capp to write to a file...check out: http://www.thegeekstuff.com/2012/07/c-file-handling/

Are you sure you are setting tga->WpKey=$AAAAAAAA before the write?

Are you sure that the printf isn't misleading you? The reason I ask is that I ran into a similar problem where I would set a register, and then immediately printf the register to read its value. The problem is that whenever you modify a register in C you are probably doing a read-modify-write operation. If your program hasn't completed the "write" operation before your printf, you'll get the wrong value printed to the terminal. Try inserting a usleep between the write and the printf, see if that changes anything. This would also explain why SetPmacVar works, but the direct write appears not to work.

Dan, Here are some Beckhoff drives, which claim they can do 16kHz on the current loop and 8kHz on the speed loop: http://www.beckhoff.com/english.asp?drive_technology/ax5000_ethercat.htm My guess is that the reason Sigma 5 drives are so much slower is because they use CoE (CAN over EtherCAT). I would bet that Yaskawa took an existing design using CAN (very slow) and just packed their CAN messages into EtherCAT frames. This could explain why it's so slow. Again, I have never used any drives over EtherCAT. Maybe it is all just big talk...theory is much easier than practice!

@Dan: This link says you can update 100 servo drives at about 10kHz using EtherCAT: http://www.rtaautomation.com/ethercat/ I'm not sure where the 2kHz limitation was on the drives you were using, but it definitely isn't coming from EtherCAT. Maybe the drives you are talking about aren't really using EtherCAT, but only Ethernet? This is possible, because you can embed Ethernet frames within an EtherCAT message. @kmonroe023: Either way you go, both MACRO and EtherCAT seem to be very good protocols in terms of speed.

kmonroe023: I think Dan was talking about Ethernet, not EtherCAT. I have never used EtherCAT personally, but I have heard of EtherCAT networks with many drives running at VERY high speeds (up to 20kHz).

If nanosleep returns EPERM that suggests that Xenomai thinks it is not in a real-time context. Are you sure that you have called pthread_setschedparam on the current thread?