National Instruments GPIB C-source drivers HELP?!
Jarkko Lavinen
I posted a week ago a query about National Instruments AT-GPIB
drivers. Things have now moved on a little and cleared somewhat.
The NI Finland technical support said there are no new drivers
available for that card. The most recent driver I have been able
to find is version E7 for DOS from NI ftp server at 130.164.1.12. It
has also DLL's for Windows 3.0 but there is no support for 386 mode.
Their suggestion to update for a newer controller would have costed
me a bout US$ 600 which is a way too much for me - just because they
don't make new drivers...
I have then found out from NI catalog that there are a products called
ESP-488 kits, which contain generic primitive C-source drivers
for their cards, including AT-GPIB. It is meant to be incorporated
into program to access directly the Turbo488 chip and their new
ASIC which replaces the old 7210. There is no 488.2 functionality
in these drivers - only one writes it himself, so I think the code
should look like very much like the code for older NEC 7210 chip.
I think this software is next to PD because it is just about
400-lines of C code and typically presented in hardware
reference manuals. Some of the code looks like I could have
written it myself if I only had the data sheets for the chips.
So I don't shame to ask if anyone on the net has these sources,
could You kindly post it to me (or contact me for anonymous ftp transfer).
Interesting enough, I _have_ NI's C-source Unix device driver for SUN3
which is for a VME-card with a Hitachi 68450 chip as a DMA controller
and NEC 7210 as a GPIB-interface. In AT-GPIB card the DMA chip is
Turbo488. (I have included as an example from this Unix device driver as
a routine ibonl() which initialises the controller.)
As You see from the code example, if I would just know enough of
Turbo488 chip, it would be not too difficult to write the code to drive
AT-GPIB. NI's ESP-488 for AT-GPIB does exactly this if only I had
it...
I called today the NI service here in Finland and they didn't seem to be
very aware of this product. I gave service the ordering number from
their catalog and they promised to find out the availability but it
sounded like it could a week or two. I haven't seen yet the datasheet
about Turbo488 I asked them a week ago. Somehow I feel like things do
not move very quickly here. :-(
So, If You have a ESP-488 kit/C-source drivers for *ANY* NI GPIB
controller, I would ask You to mail it me, please...?
Jarkko Lavinen - jl...@cs.joensuu.fi
---
Excerpt from "ib.c", routine ibonl(). It is quite useless without header
file which explains the registers, but You can see separate the 7210
code and the DMAC code are:
static ibonl(v){
char s;
ib->cfg2 = D_LMR|D_SFL; /* local master reset */
ib->cfg2 = D_SFL; /* clear LMR */
/* 7210 setup */
s= ib->cptr; /* clear registers by reading */
s= ib->isr1;
s= ib->isr2;
ib->imr1 = 0; /* disable all interrupts */
ib->imr2 = 0;
ib->spmr = 0;
ib->adr = 0; /* set GPIB address; MTA= 100, MLA= 040 */
ib->adr = HR_ARS | HR_DT | HR_DL; /* disable secondary addressing */
ib->admr = HR_TRM1 | HR_TRM0 | HR_ADM0;
ib->eosr = 0;
ib->auxmr = ICR | 8; /* set internal counter register N= 8 */
ib->auxmr = PPR | HR_PPU; /* parallel poll unconfigure */
ib->auxmr = AUXRA | 0;
ib->auxmr = AUXRB | 0;
ib->auxmr = AUXRE | 0;
/* DMAC setup */
ib->cfg2 = D_SPAC|D_SFL|D_SC;
ib->cfg1 = (IBIRQ << 5)|D_BRG3|D_DBM;
ib->ch1.niv = IBIRQ + 24; /* 24 == autovector 0 */
ib->ch1.eiv = IBIRQ + 24;
ib->ch1.ccr = D_EINT;
ib->ch0.cpr = 3; /* highest priority */
ib->ch1.cpr = 3; /* highest priority */
ib->ch1.dcr = D_CS|D_IACK|D_IPCL;
ib->ch0.dcr = D_CS|D_IACK;
ib->ch1.scr = 0;
ib->ch1.mfc = D_SUP|D_S24;
ib->ch1.bfc = D_SUP|D_S24;
ib->ch0.scr = D_MCU;
ib->ch0.mfc = D_SUP|D_S24;
if(v) ib->auxmr = AUX_PON; /* release pon state to bring online */
}
Jarkko Lavinen
I wrote yesterday:
> their catalog and they promised to find out the availability but it
> sounded like it could a week or two. I haven't seen yet the datasheet
> about Turbo488 I asked them a week ago. Somehow I feel like things do
> not move very quickly here. :-(
I was wrong, once again... NI service called me and they will send me the
ESP kit! It was some 250k of C source code, so it was certainly not a PD
level package, but actually quite pricey - about the same as update to a
new card, but we got into an agreement.
Still, if anyone has the primitive drivers in less 1k of C source code, I
would be most interested...
Jarkko Lavinen, jl...@cs.joensuu.fi