Programming Arlan cards

Отправлено: FH-Bochum on September 18, 1997 at 13:21:29:

Dear Collegues,

please excuse my last mail, I've had some trouble to
understand your Cyrillic Mail form.

I can tell you more precise about our problem:

Our kiss-lab is part of the polytechnical institute Bochum, Germany
(Fachhochschule Bochum). We've devlopped many software and tools for our
mobile robots. We use TCP/IP under LinuX to control the robots.
Inside the robots are typical LinuX-PC's, using several Arlan A655-2400
ISA cards. To reach the outside world we've installed an AP Arlan A630-2400.

We've written a device driver for the A655 working under LinuX.
Our problem is now the following: If we transmit many little datapackets
from LAN-side to the robot the AP A630 gives the following message:
"robot4 removed, remote buffer full too long" and it deregisters the
A655 client at the same time. The client didn't realize that situation.

Our questions: How can we force an automatic reregistering in this case?
How can the client _really_ detect that is was "deregistered"?

On detecting the transmit timeout on the card we used the TXStatusVector
(00B0h). After a deregistration it results in 01h --> Transmit o.K. instead of
02h --> Timed out.
Is this the place you mean to find a transmit timeout?

In which way should the reregistration command be issued:
-by executing a reset?
-or by writing a command byte (what value?) and toggling the Channel
Attention bit ?

The second problem that we have is a question about the internal differences
between the A655 and the PCMCIA version A690-2400. Here we're interested in
the comparison of the internal organisation of both cards. For example: is the
method to issue commands the same in both cases ?
On pushing the card into the PCMCIA slot it registers very seldom (one of 30
times). Do you have any idea?

Does anybody mean that it is not possible to access the card via the shared
mem? Is the command handling the same like for the 655-version? For example:
writing CODE 83h for RxEnable into the address 0090h and toggling the CA bit?
Do all registers have the same location like those on the A655 card?

We're interested in any hint or every helpful idea.

Thanks for your help.

Yours,
Ralf

Отклики:

Написать ответ

Ваше имя:
E-Mail:

Тема:

Ваше сообщение:
> Dear Collegues, > please excuse my last mail, I've had some trouble to > understand your Cyrillic Mail form. > I can tell you more precise about our problem: > > Our kiss-lab is part of the polytechnical institute Bochum, Germany > (Fachhochschule Bochum). We've devlopped many software and tools for our > mobile robots. We use TCP/IP under LinuX to control the robots. > Inside the robots are typical LinuX-PC's, using several Arlan A655-2400 > ISA cards. To reach the outside world we've installed an AP Arlan A630-2400. > We've written a device driver for the A655 working under LinuX. > Our problem is now the following: If we transmit many little datapackets > from LAN-side to the robot the AP A630 gives the following message: > "robot4 removed, remote buffer full too long" and it deregisters the > A655 client at the same time. The client didn't realize that situation. > Our questions: How can we force an automatic reregistering in this case? > How can the client _really_ detect that is was "deregistered"? > On detecting the transmit timeout on the card we used the TXStatusVector > (00B0h). After a deregistration it results in 01h --> Transmit o.K. instead of > 02h --> Timed out. > Is this the place you mean to find a transmit timeout? > In which way should the reregistration command be issued: > -by executing a reset? > -or by writing a command byte (what value?) and toggling the Channel > Attention bit ? > The second problem that we have is a question about the internal differences > between the A655 and the PCMCIA version A690-2400. Here we're interested in > the comparison of the internal organisation of both cards. For example: is the > method to issue commands the same in both cases ? > On pushing the card into the PCMCIA slot it registers very seldom (one of 30 > times). Do you have any idea? > Does anybody mean that it is not possible to access the card via the shared > mem? Is the command handling the same like for the 655-version? For example: > writing CODE 83h for RxEnable into the address 0090h and toggling the CA bit? > Do all registers have the same location like those on the A655 card? > We're interested in any hint or every helpful idea. > Thanks for your help. > Yours, > Ralf

Optional Link URL:
Link Title:
Optional Image URL:

E-mail: sales@comptek.ru tel: (095) 785-2525