ICS 8065 GPIB to Ethernet Adapter

[BloodyCactus]

Picked up this nice little box off ebay for a song. Thought I'd take some pics. I didnt expect much in there and there isnt;

the gpib interface looks like a SN75162B, SN75160B with an Nat Ins 7210 chip and an LVC16245.

Main controller is a NXP LH75411 SOC, 84mhz ARM7TDMI-S from Sharp. NXP BlueStreak.
http://www.nxp.com/documents/data_sheet/LH75401_411_N.pdf

Some spansion + cypress flash/sram and an SMSC LAN91C111 ethernet interface 10/100 MAC+PHY
http://ww1.microchip.com/downloads/en/DeviceDoc/91c111.pdf

[eas]

Hate to say it, but, how does it work once you turn it on? Hows the software support?

I'd had my eye on one of those, but didn't pull the trigger because I didn't know how much trouble it would be to get it to work with Python or similar.

[BloodyCactus]

its transparent. your talking gpib/scpi just over ethernet instead of gpib cable. no difference.

[dom0]

Hate to say it, but, how does it work once you turn it on? Hows the software support?

Probably all of these support VXI-11 so no problems there.

[Fraser]

I own an ICS USB to GPIB controller and can say that ICS make nice kit. I have heard that they are not always 100% compatible with applications that expect an NI controller however. The same is true of many non NI GPIB controllers though.

Nice find 

Aurora

[free_electron]

ICS is really good stuff.

they are the third company that makes official GPIB stuff. ( besides agilent and NI ).
they had some gpib32.dll problems in the past but are quick on the ball to fix it.

[Fraser]

This thread reminded me that I need a GPIB to Serial converter. Thanks 

As luck would have it an Irish seller has some beauties for sale at the moment at reasonable cost.

They are the ICS 4896 and provide GPIB to four serial ports with LCD display and serial buffers. 

http://www.ebay.co.uk/itm/291213457188?_trksid=p2060353.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

The unit offers far more than I truly need, but it will be very useful to have around in the lab.

Aurora

[tmbinc]

I've bought a ICS4896. It passes the self-test, but I can't get any output. I'm a GPIB beginner, so it could be something stupid on my side.

The port is set to 9600/8/N/1 (shown in the display at startup). If I send "SYST:COMM:SER:BAUD?\n" to the device, I expected to see the current baudrate, as explained in the manual.

In the channel 1 status screen, I see that the number of received "RX" bytes goes up when i send something; but the TX stays at "0B" always.

Is there anything I'm fundamentally doing wrong?

[macboy]

I've bought a ICS4896. It passes the self-test, but I can't get any output. I'm a GPIB beginner, so it could be something stupid on my side.

The port is set to 9600/8/N/1 (shown in the display at startup). If I send "SYST:COMM:SER:BAUD?\n" to the device, I expected to see the current baudrate, as explained in the manual.

In the channel 1 status screen, I see that the number of received "RX" bytes goes up when i send something; but the TX stays at "0B" always.

Is there anything I'm fundamentally doing wrong?

Try disabling hardware handshaking for the port. The port will not transmit if hardware handshaking is enabled by the connected device isn't asserting CTS (Clear To Send).

You say you are sending "SYST:COMM:SER:BAUD?\n"... you are sending this to the GPIB port, correct? After sending the command, you need to explicitly read the response. Your GPIB console software might have something like a "send and read" button. Try that. Or if you are doing this programmatically, call the read function after doing the write.

[tmbinc]

I just realize that I'm just stupid. I was thinking this device is actually a serial-to-GPIB bridge, i.e. which lets me talk to GPIB devices from (up to) 4 hosts. Instead, I now realize it's the opposite device - it lets me talk to up to 4 serial devices over GPIB.

Oops

[eas]

I picked one of these up on EBay a couple of weeks ago. After a bit of cursing about Windows, I got the firmware updated to the latest version. Since then, I've been tinkering with using it to talk to 5 Keithley DMMs using Python-vxi11 on OS X and Linux.

At this point, its working well enough that I'm running a script on my linux server to collect readings twice a second for a few days so I can see how the meters track relative to each other, and how my AD588 voltage reference tracks.

Well, almost well enough. I'm getting socket timeouts every hour or so, and partial readings (just the sign, not the value) from one meter or another a few times an hour. Not sure where the problem lies though. Could be the WiFi bridge, GPIB (had to replace a cable already).

Any one know if there is a way to query the state of the ICS 8065, maybe dump an error log?

[bson]

Make sure you don't have the NI discovery service running, or any VISA drivers.  Apart from being flaky like all h*ll they can "steal" devices from your python-vxi11 client if they discover the ICS 8065.  I've had an Intel NUC D54250WYK with Linux sit and collect data from an assortment of instruments for days using an ICS 8065 with nary a glitch.  The only glitch is a K2001 DMM sometimes locks up, but I design my tests so they checkpoint state and can resume if interrupted, and it has only happened twice so far - so no big deal.  The K2001 might need a firmware update.  (Mine's B13; I see xdevs.com wants an upload of it, maybe I can oblige.) The only problem device I have is a vintage HP8903B with HP-IB that predates SCPI, and its only means of readout is via async SRQ, which python-vxi11 doesn't support.  Fortunately I don't use that one much, but I'm definitely sticking to GPIB and SCPI for all future instruments.  Otherwise, between the ICS 8065, python-vxi11, numpy, scipy, matplotlib, it's super easy to whip up two-part tests (one to collect data, the other to analyze and plot the results).  I do keep the NUC+8065 on a VLAN dedicated to instruments.

[eas]

Thanks for that perspective.

I just checked back over the output I log when there is an incomplete reading and it seems like its the same two instruments every time, the two with the oldest firmware versions, so I suspect that has something to do with it. It looks like the 2700 can be flashed, and I just got a programmer, so all I need are some suitable EEPROMs or flash chips and I can update the 2000 as well.

Also, I installed the ICS utilities in WINE and they seem to work, and there is one to get at error logs, so that will probably help with my debugging too.

For the data I'm gathering right now its not a huge deal if I miss some readings, so I've just got the python code to log an error and continue when it gets a partial reading. For the socket timeouts, I'm running the script from an infinite while/do loop in bash. Things seem to recover if it sleeps long enough after the python script exits before restarting it.

Once I've got this a little more stable I want to get some live charting working in an iPython notebook

[bitseeker]

I picked up one of these a few months after eas. Works well. I checked with ICS regarding firmware updates and my hardware version doesn't have enough memory to use the latest.

Jerry at ICS said that I can go up to rev 17, but that version was no longer on their site for download. Does anyone happen to have rev 17, perhaps on an old ICS 8065 CD, or know of an online repository of firmware versions?

[BloodyCactus]

strange that you can only go up to a certain firmware level. How does one tell if the box has enough ram or not?

mine says 'ICS Electronics, 8065, 806006, 01.23, 130215' and I do believe I ran the update firmware on it. 1.23 is the latest.

[bitseeker]

Yeah, I wasn't expecting that. Here's the specs on mine.

Sticker on enclosure:

Model: 8065
Rev: 0.13
SN: 806005

Version info on PC board:

P.C. ASSY - 115504 REV 3
P.C. BD - 115506 REV 6

I originally made my inquiry because the ICS site mentioned a hardware mod that was needed to install the latest firmware on certain units. It initially appeared that mine fell into that group. However, when I received the engineering change note for the jumper wire that needed to be installed, the board layout in the note didn't match mine. So, I sent the above board revision info to them and got the news that mine is not compatible with the latest firmware due to insufficient onboard memory.

Jerry at ICS said that he was able to locate the last revision (rev 17) that is compatible with my unit, but that he needs to make a new installer for it. So, that's where I am thus far.

Now, what's really interesting is that it looks like your serial number is the one right after mine! Also, your PCB version numbers are the same as mine. So, what does that mean? Did the firmware update on yours really work or will it do something strange in the future because some part of the firmware didn't fit, yet there was no warning? 

[BloodyCactus]

lol consequetive serial numbers? thats too strange.

the firmware web page says 'units with revision 0 hardware cant be update by this program'. my pcb says rev 6 and rev 3, like yours. I wonder how one identifies rev 0?

i dont remember if I updated it or not. I def downloaded the zip. I'm pretty sure I did run the update, because  think it comes up in a fixed IP and that played havoc on my network. had to connect it to the wifes laptop or something so I could get it into dhcp

very weird.

I know its working fine talking to my Fluke PGIB psu's and my racal dana 1992 without problems.

[bitseeker]

It is funny that they happened to be consecutive. And half way around the world to boot.

Can you log into your unit to see what its current version is? That'd be the only way to tell if you ran the update.

I got from ICS the new installer for the old version of the firmware. However, Jerry said to follow the ECN before running the update. That ECN doesn't match my board, so I sent him the PCB rev number. I'm awaiting his reply before I proceed with either the old or new firmware versions.

[BloodyCactus]

its 1.23 as I posted;

[bitseeker]

its 1.23 as I posted;

Ah, OK. This part, below, sounded like the update was not definitive.

i dont remember if I updated it or not. I def downloaded the zip. I'm pretty sure I did run the update, because  think it comes up in a fixed IP and that played havoc on my network.