Low cost GPIB-USB controller

Nice, but $125 isn't exactly low cost. It's a long time ago but we used to buy GPIB card for the old ISA bus for under $100. Think about another "de-featured" version. Most people only need them to get screen prints out of an instrument, not to control them.

Regards, Joerg

Joerg,

Thanks for your suggestion. Compared to the $500 GPIB-USB dongles from NI, Agilent and others isn't $125 low cost?

Regards,

Sure, for the hardware... but is your converter compatible (software-wise) with the NI cards? Are there LabView drivers for it? If the answer to either of those is "no," I can tell you that a lot of companies won't care about your low price -- the engineering time to modify their software wouldn't make it worthwhile. (True story -- that I've mentioned before --: Tektronix still uses PDP-11's as part of their probe manufacturing lines. Although they pay an arm and a leg to keep those PDPs alive, it's apparently cheaper than migrating to newer, cheaper hardware and porting the software!)

I was rather pleased to find that some old test software we had for an NI ISA GPIB card worked flawlessly (no changes required) with the NI USB GPIB dongles; they've maintained pretty broad compatibility between all their interfaces over time (granted, they aren't perfect -- their _original_ USBGPIB dongle was only supported up through Windows 98... grrr!). Similarly, NI provides some nice LabView drivers (no surprise there, eh?) that make it very fast to development LabView apps that "speak" GPIB.

For someone writing new software, though, I think your dongle looks like a good bargin. Although you might want to stick it in a case to make it look more like a "finished product!"

---Joel Kolstad

Hello Joel,

For USB it is, but there are (were?) ISA cards for much less.

Not that it was required but the cheap ISA cards actually were fairly compatible, AFAIR. I have to say though that their documentation was hardly intelligible but for us it worked.

Oh man. Time to make the switch before something breaks and cannot be replaced. This is like flying a critical mission with a DC3. Not that it wouldn't be safe but one busted engine can put you out of service for a long time. Sometimes a down-time of a few weeks can bring great troubles.

I have modernized quite some production gear, usually after asking questions like "What if this thing over here breaks?" and the response was an eerie silence.

Regards, Joerg

On the other hand, if his dongle crashes less frequently than the NI one and isnt' too hard to code for, I might see if I can get work to buy a few of them.

The original posters design seems to be a GPIB to SERIAL interface only.

This is not such a bad thing as you don't need LabView, MathLab, TestPoint or other big bucks instrument control suites.

This design includes a VI to interface with LabView plus all the schematics and source code to roll your own:

University of Ljubljana Faculty of Electrical Engineering Trzaska cesta 25, 1000 Ljubljana, Slovenia Laboratory of semiconductor devices

There is a article in EDN too:

No thanks - I think I'd rather pick up a National Instruments PCI card on eBay, which would be 100% software compatible with NI.

If I really needed USB, and it is hard to see why, then I'd probably still buy a NI one.

--
Dave K     MCSE.

MCSE = Minefield Consultant and Solitaire Expert.

Please note my email address changes periodically to avoid spam.
It is always of the form: month-year@domain. Hitting reply will work
for a couple of months only. Later set it manually.

One word: laptops

Why you might ask, would one want to control a GPIB bus with a laptop?

1) Salesmen need to demo GPIB instruments without dragging along a desktop computer

2) Support engineers need to be able to check out a GPIB instrument without dragging along a desktop computer or depending on the customer's equipment.

3) When GPIB equipment is installed in a rack and it is necessary to try someting out of the ordinary - use a different computer to see if the problem is with the rack mount one, or ttemporarily use GPIB to debug a system where it is not ordinarily in use, it's a lot easier to carry a laptop over to the system, crawl behind the rack with it, whaterver, than to do so with a desktop.

There is an electrical problem: Power.

The GPIB bus is a TTL-level bus with quite low impedance. There is simply not enough power available from USB to drive the GPIB according to the specifications.

--

Tauno Voipio
tauno voipio (at) iki fi

Hello Tauno,

Huh? If 2.5 Watts ain't enough then I don't know what you want to drive with it:

Ok, it won't power your espresso machine...

Regards, Joerg

What do you think the bus impedance is?

I think it's fairly high, but dependent on the number of devices. Each device should have a 3k resistor to 5v and a 6.2k resistor to ground, of course in parallel with the actual input gate. Load up a bus to its maximum device count so you have many of these in parallel and you might see problems, but typically with a USB dongle you have few devices - most often only 1 device. I don't see this exceeding the post-enumeration USB current limit.

I also think it would be perfectly legitimate to make a device which was specified in bold letters right on it, "only for direct connection to a single instrument without an intervening cable" This is after all the most common use of the USB-GPIB dongles on the market.

..or have a socket for an external PSU.

But can keep your coffee warm...

4) Space. Many workbenches are short of space - a laptop takes little room and can quickly be stashed away when not in use.