Interesting Experience-5890/5972/7673 and G1701BA

[benhutcherson]

I have written up a fair bit here about my adventures with a "rescued" 5890/5971. I pulled it out of storage a while back with its old computer running "C" series Chemstation. Along the way, I added a full 7673 set-up-a tray, tower, and G1512A control box.

As a side note, for anyone running this sort of set-up, please treat yourself to a G1512A. This is the later style controller without discreet cards that also works with 6890s. It operates the tray arm significantly faster than the old style box, but that's a bit of a digression.

In any case, I'd previously just had the control box tied in by running an HPIB cable to it, setting the address in the configuration tool, and away it went.

Elsewhere, I've documented a series of upgrades to the 5971, including fitting a 5972 mainboard, PSU, and SmartCard II, and have fitted the more efficient filaments as used in the 5973 and later. I built up a computer running Win2K and G1701BA, and had it working(at home, during our WFH period) with the 5972.

I took everything back in May, and since no one would be using it just did an initial check-out to make sure everything was talking(and left it under vacuum). The G1512A needed a tweak of its DIP switches on the back to talk to G1701BA, but that was easy enough to sort out.

Actually returning to it in person(after a period of furlough) this week, though, I ran into something else. The ALS would make an injection just as instructed(and would complete the syringe wash/bottle return). Unlike previously, however, the run would not start after the injection. Pressing start on the GC front panel would have it proceed as normal, including prompting me if i wanted to override the solvent delay.

I was a bit stumped on this, but on a suggestion I looked into something else.

The G1512A has a remote port on the back of it that's identical to the one you see on a 5971/5972 with the 9-pin D-sub connector. Fortunately, I had one of the "4 way sync" adapters for the 5890(to let you connect 3 devices to the 5890s single sync port). I dug around and found another MSD sync cable, and ran that from the 5890 to the G1512A. Sure enough, it immediately started working.

I'm not sure what changed between the old software and now with quite literally the same A/S hardware to now require it. I've also serviced several running G1701BA with 18594B controllers(the old card type) that didn't require a sync cable. In fact, somewhere or another I have one or two cards for the 18594 that have that sort of port on them, but I'd thought that was for diagnostic use and I've never even used one. Most that you see "in the field" don't have anywhere to connect a sync cable.

[DR]

In addition to having the correct connections among all of those cable leads, you have to be cautious about taping or otherwise isolating the unused leads.
On a number of occasions, I've had problems caused by the wrong leads coming into contact. People tent to set these up and just spread all of the unused leads into sort of a flower looking pattern, which is great until the cable moves or is bumped by something after they've neglected the isolation of the extra leads.

[benhutcherson]

Not sure what unused leads you're talking about.

The 5890 normally only has one sync connection. There is a small available board that breaks this out to allow up to three sync signals(4 ports on board, one to the GC and the other three to cables) but it is heavily insulated with only the protected female ports accessible.

The sync cables I'm talking about have a the Molex-type connector to plug into this(or directly into the GC) on one end, and a 9 pin D-sub type connector on the other.

I've never heard of issues with the hemaphroditic HPIB cords causing problems, but I like ending chains with the ones that are male on one end. I will usually cap the other end with the purpose-made plugs for this(the HP supplied ones are usually read plastic).

This particular set-up has a rather...fun...HPIB chain since I run two 5890s plus a mass spec and the G1501 off of it. The only slightly worse one I've seen has pretty much the same set-up but A/S control boxes on both GCs.

[James_Ball]

I know what DR is talking about. I had to custom make one to connect to a Tekmar 2000 purge and trap and you can have extra wires that aren't used, if you don't make sure those are safely capped off or removed from the connector completely, they can short with another and cause all sorts of problems.

The worse ever was when we upgraded from the Tekmar 2000 to an Encon and I had to figure out the new wiring because one used a contact closure to signal ready and the other used a contact open. Luckily I had the manual with the pinouts for that connector on the GC.

[LALman]

I too, had to get the custom cable to synch my Tekmar 3000 with my 5890/5971. But I've never had to do a synch cable to us my 7673 and tray on the same system. Maybe this is a software or operating system issue where an interrupt is not being serviced by the software? But it can be taken care of with a synch cable?

I agree about the controller. I have a very old 1980's vintage and one that came with a newer 6890 GC-FID system. That newer controller hooks up the same but runs the 7673 and tray much more smoothly.

@Benhutchinson "This is the later style controller without discreet cards that also works with 6890s." You mean HPIB card? Maybe that is why mine works without an extra cable. I still use the HPIB cables.

[benhutcherson]

@Benhutchinson "This is the later style controller without discreet cards that also works with 6890s." You mean HPIB card? Maybe that is why mine works without an extra cable. I still use the HPIB cables.

There were two "7673" autosampler controllers.

The first one was the 18594B. If you look at the back, it will have power and the two tower controllers built in to the back. There's then a set of 4 vertical card slots. Every one I've seen has a big D-sub connector to control the tray. Usually a second card will be one of two things-either INET or HPIB.

The G1512A does not have the vertical cards. Instead, it has HPIB, INET, and and the tray controller all permanently built into it. There are several 9-pin d-sub connectors on it-one is marked "for future use" one is "APG sync"(this is where I connected the sync cable I mentioned above) and one is RS232. I've never actually hooked one up, but it's my understanding that normally if you were using this with a 6890 without a built in A/S controller, you use one or two of the D-sub ports for the box to talk to the 6890. On the 5890, though, you add it to the HPIB chain.

BTW, even though I have bunches of 18594B parts, when one dies in the field I usually try to convince the owner to replace it with a G1512A. It's worth it if, for nothing else, your throughput because the tray arm moves a whole lot faster.

[James_Ball]

@Benhutchinson "This is the later style controller without discreet cards that also works with 6890s." You mean HPIB card? Maybe that is why mine works without an extra cable. I still use the HPIB cables.

There were two "7673" autosampler controllers.

The first one was the 18594B. If you look at the back, it will have power and the two tower controllers built in to the back. There's then a set of 4 vertical card slots. Every one I've seen has a big D-sub connector to control the tray. Usually a second card will be one of two things-either INET or HPIB.

The G1512A does not have the vertical cards. Instead, it has HPIB, INET, and and the tray controller all permanently built into it. There are several 9-pin d-sub connectors on it-one is marked "for future use" one is "APG sync"(this is where I connected the sync cable I mentioned above) and one is RS232. I've never actually hooked one up, but it's my understanding that normally if you were using this with a 6890 without a built in A/S controller, you use one or two of the D-sub ports for the box to talk to the 6890. On the 5890, though, you add it to the HPIB chain.

BTW, even though I have bunches of 18594B parts, when one dies in the field I usually try to convince the owner to replace it with a G1512A. It's worth it if, for nothing else, your throughput because the tray arm moves a whole lot faster.

I believe the G1512A moves the arm it multiple directions at the same time, such as around and in or out instead of moving around then in or out, and it is at least twice as fast on the motor speed.

On the old 18594 there is a B and A version and the trays have a B and A version. The towers are interchangeable but the trays have to match or they don't work.