Setting up a RTD Card

[Controls Guy]

Hey guys, how bout an RTDIN box?

[dieseltwitch]

Thanks for you help by the way. i know how much of pain it can be pushing newbies up the steep learning curve!
however i will have to comeback to this later, I have an 11 page paper to write and I'm on page 3! and its due tomorrow!

[MikeS]

we've stayed away from making IBoxes for specific I/O modules, opting instead for making general purpose ones that can be used by a class of modules (like 4 ch ana input).

You could use ANLGIN to setup the first two parameters (count/format & pointer) and use LD/OUT pairs to do the remaining setup, but that's not much better IMHO.

[Controls Guy]

we've stayed away from making IBoxes for specific I/O modules, opting instead for making general purpose ones that can be used by a class of modules (like 4 ch ana input).

You could use ANLGIN to setup the first two parameters (count/format & pointer) and use LD/OUT pairs to do the remaining setup, but that's not much better IMHO.

It's odd the way you word that because there's a "class of modules" that can use the I-box method and another "class of modules" that cannot.  I think rather than introduce the anomaly, I just wouldn't have done any of them (or maybe only do analog output modules).

A kind of generic approach occurred to me where you would fill in all the stuff in the present ANLGIN box, then have another couple boxes for "Parameter 1", "Parameter 2", and so on, which would get loaded into consecutive registers corresponding to the slot specified.  This could cover a variety of types of modules, where the additional parameters might mean different things to different modules, and would have to be filled in by which register they went to.  This could be in addition to or a revision of the present ANLGIN box.

Besides, since I'boxes compile to the old instructions anyway and don't require a processor change, what's the cost of adding another one?

[Controls Guy]

we've stayed away from making IBoxes for specific I/O modules, opting instead for making general purpose ones that can be used by a class of modules (like 4 ch ana input).

You could use ANLGIN to setup the first two parameters (count/format & pointer) and use LD/OUT pairs to do the remaining setup, but that's not much better IMHO.

Actually that gives me the idea of using MOVEW rather than LD/OUT, which is at least a little cleaner.

[b_carlton]

As far as my understanding of IBoxes a set of mnemonic instructions were introduced, which to the processor are essentially NOP's, which serve as delimiters for the set of standard instructions. These actually wouldn't be necessary except that an IBox has to be reconstructed just by reading the processor contents. Without the delimiter it would just come back as a set of standard instructions, not as an 'IBox group'. So a unique identifier has to exist for each IBox in the processor and it has to know to essentially ignore them. Adding another which had not yet been defined in firmware means a firmware changes, unless the previous change left a lot of 'spares'.

[Controls Guy]

That's essentially my understanding as well, except I thought the NOP's were generic, not specific to particular I-Boxes.  IOW, "If you see 'LD-OUT-NOP', render it as a MOVEW".  I never really understood why even that was necessary, because DirectSoft could simply default to I-Box mode and render all compatible sequences as I-Boxes without any delimiters at all.  Then if someone uploaded with an old version of DirectSoft, they'd get pure code, with no other delimiter boxes mixed in.  Course that takes the preference for I-Box vs. old-way out of the hands of the programmer unless you introduce a configuration setting into DirectSoft, and even then, you couldn't mix the two modes, so maybe that's the reason.

[MikeS]

you're right, 'class of modules' was bad choice, it's more like 'family of modules' because the F2-04RTD works fine with the existing ANLGIN IBox because it does not have the additional configuration parameters in v-memory like the F0 module does, (the F2 version uses on-board jumpers), that's what I was trying to convey.

another good example is the f2-8ad4da combo analog module, although it only has 4 real output channels, it has 3 additional DWORD registers that consume I/O space that are the runtime update registers. What you'd need here is a setup ibox that had 96 bit positions (3 dwords of bits) to initially setup the module, then an additional ibox to manipulate those same bits at runtime.

admittedly this is the extreme example, but our goal was to make as many generic setup iboxes as possible and let the ungeneric ones slide ... for the time being.

[Controls Guy]

Yeah, the extreme cases can get pretty weird if you try to address absolutely everything.