Brooke Clarke wrote: > Hi David: > > The HP 59000 series boxes were some of the very first HP-IB > instruments made. They are what might be called glue boxes used to > build systems. As has been pointed out there is not a micro > controller and in addition there is no IEEE-488 interface chip (this > was before there was an IEEE specification).
Hi Bruce, Seems they are very old then. > > In addition they are slow. The handshake on the HP-IB slows the buss > down to the slowest instrument that's on the buss even if you are not > talking to that instrument. This has two implications, first, it's > good to put the slow instruments on a different bus from instruments > where you want speed and second, when talking to a slow instrument add > a time delay after you send it a command before the program proceeds > so that there's time for the instrument to process the command. I did think about adding a few sleep() commands, but it went out of my mind. I'll try that later tonight when I get home. The speed is a bit worrying, as it will slow things down if this device takes significantly longer to respond than the instrument making the measurement. Unfortunately the 1U server I intended using for this only has one PCI slot, so I could not add another GPIB board. Anyway, that is another issue - the first problem is to be able to get it working at all. I did think about programming the parallel port of the Sun - something that was very easy on PCs under DOS, but is far from as simple on a multi-user UNIX operating system. > > The 59306 has two modes, local and remote. You can press buttons on > the front panel in local mode and use it without a remote control. > When you want to use remote control you must first send the remote > command <$>. The front panel switches stay pressed and when the box > is put into local mode those switch positions well be restored. Where do you get the $ as being the remote command? It appears to be % in my manual, but I'm not sure if that would be sent with ibwrt or not. The fact the control lines are different in the manual for sending the % to other commands makes me think it might have to be done another way. > > The three letter line names are control lines on the HP-IB and modern > interface cards take care of most of those signals. But, since > there's no micro controller you should not send any characters after > the single command byte. No <CR><LF>, no <EOI>. You may need to > explicitly set the termination sequence so that these are avoided. You might have solved it there. I'm opening the device with dd=ibdev(0, gpib_address, 0, T3s, 1, 0); the 5th argument to ibdev is the 1, not 0. The manual for linux at least says "If |send_eoi| is nonzero, then the EOI line will be asserted with the last byte sent during writes" So it seems I might be sending EOI, which you are saying I should not do. I hope you are right!!! I'm not in a position to test now, as I am about 50 miles away from the instrument. Dave _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
