Breaking news: The QIMSI Interface

Nagging hardware related question? Post here!
User avatar
Peter
Font of All Knowledge
Posts: 2014
Joined: Sat Jan 22, 2011 8:47 am

Re: Breaking news: The QIMSI Interface

Post by Peter »

Pr0f wrote: Sat Oct 07, 2023 10:24 pm Some sort of Maths accelerator comes to mind...
Yeah, or stuff like Mandelbrot calculation. I could imagine stuff like that 30..50 times faster than a QL.


FrancoisLanciault
Trump Card
Posts: 169
Joined: Mon Aug 08, 2011 11:08 pm

Re: Breaking news: The QIMSI Interface

Post by FrancoisLanciault »

Peter wrote: Sat Oct 07, 2023 8:27 pm
FrancoisLanciault wrote: Sat Oct 07, 2023 5:45 pm but I MUST have this new interface!!! 68040 like coprocessor that is difficult to use! What can be more appealing!
Would you program it in C or assembler?
Assembler most likely. Would use it as computing accelerator for chess algorithm.


User avatar
Peter
Font of All Knowledge
Posts: 2014
Joined: Sat Jan 22, 2011 8:47 am

Re: Breaking news: The QIMSI Interface

Post by Peter »

FrancoisLanciault wrote: Sun Oct 08, 2023 12:16 am Assembler most likely. Would use it as computing accelerator for chess algorithm.
Oh that sounds cool - and still realistic under the given memory contraints.
I really should document the bidirectional fast data link between the QL and the MiniQ68 coprocessor.

By the way, it makes sense to have a SER driver for that data link on the QL side.
I hope Wolfgang or me will come up with something when in the mood.


User avatar
Pr0f
QL Wafer Drive
Posts: 1308
Joined: Thu Oct 12, 2017 9:54 am

Re: Breaking news: The QIMSI Interface

Post by Pr0f »

Documenting the bi directional interface would be good Peter.

I see in the manual there is also mention of I2C in the ports, are the I2C ports made available on the module itself?

I see there is a 6 pad area to the bottom left of the interface (with edge connector pointing towards you), and wondered what these pins are for?

I'm also wondering if the work Jan did on the Q68 drivers for Minerva could be used to provide a working PS/2 interface for the legacy QL OS...


User avatar
janbredenbeek
Super Gold Card
Posts: 633
Joined: Wed Jan 21, 2015 4:54 pm
Location: Hilversum, The Netherlands

Re: Breaking news: The QIMSI Interface

Post by janbredenbeek »

Pr0f wrote: Sun Oct 08, 2023 8:09 am I'm also wondering if the work Jan did on the Q68 drivers for Minerva could be used to provide a working PS/2 interface for the legacy QL OS...
Just a little patience, work is in progress... it was held up by the lack of a PS/2 to USB keyboard converter on this side, but I've been able to create a makeshift one from an old USB cable and an old-style DIN jack (sorry Peter, still haven't received the one you sent me).

I have taken the source code from the Q68 driver and adapted it for QIMSI. On the first test it loaded OK without crashing, but when I tried typing something I got characters which didn't even remotely resemble the keys pressed. So still a little debugging to do, but it's under way.

Jan


Miguel angel rojo
ROM Dongle
Posts: 9
Joined: Sat Apr 27, 2019 9:10 am

Re: Breaking news: The QIMSI Interface

Post by Miguel angel rojo »

Interesting interface, I was told about it in a group of QL friends in Spain, I don't know if with the few messages I have I can opt for one.


User avatar
Peter
Font of All Knowledge
Posts: 2014
Joined: Sat Jan 22, 2011 8:47 am

Re: Breaking news: The QIMSI Interface

Post by Peter »

Pr0f wrote: Sun Oct 08, 2023 8:09 am Documenting the bi directional interface would be good Peter.
Done. Updated my initial post so it is easier to find. Also uploaded mouse driver, sound player with examples, sourcecode.
Pr0f wrote: Sun Oct 08, 2023 8:09 am I see in the manual there is also mention of I2C in the ports, are the I2C ports made available on the module itself?
As far as I am aware, I2C is not mentioned in the QIMSI manual. But you are right inasmuch I2C is implemented on the MiniQ68 side. The signals go to the bottom side of the PCB, where a DS1374C-33 and SuperCap could be soldered. Currently there are no plans to offer such a hardware variant. At your own risk, you could access I2C_CLK at pin 1 and I2C_DATA at pin 16 of the DS1374C-33 footprint (near the large SMD pad of the SuperCap).
Pr0f wrote: Sun Oct 08, 2023 8:09 am I see there is a 6 pad area to the bottom left of the interface (with edge connector pointing towards you), and wondered what these pins are for?
They are unsupported. (I used them for experiments as video output.)
Last edited by Peter on Sun Oct 08, 2023 11:11 am, edited 1 time in total.


User avatar
Peter
Font of All Knowledge
Posts: 2014
Joined: Sat Jan 22, 2011 8:47 am

Re: Breaking news: The QIMSI Interface

Post by Peter »

janbredenbeek wrote: Sun Oct 08, 2023 8:26 am I have taken the source code from the Q68 driver and adapted it for QIMSI. On the first test it loaded OK without crashing, but when I tried typing something I got characters which didn't even remotely resemble the keys pressed. So still a little debugging to do, but it's under way.
You are aware that the UNLOCK register has to be read on the QL side of QIMSI, not written as on the Q68?

Do you think you can squeeze this into 48 KB, maybe sacrificing the QL matrix keyboard driver?


User avatar
janbredenbeek
Super Gold Card
Posts: 633
Joined: Wed Jan 21, 2015 4:54 pm
Location: Hilversum, The Netherlands

Re: Breaking news: The QIMSI Interface

Post by janbredenbeek »

Peter wrote: Sun Oct 08, 2023 11:09 am You are aware that the UNLOCK register has to be read on the QL side of QIMSI, not written as on the Q68?
Yes, I have replaced it with a 'tst.b' instruction ;)
Do you think you can squeeze this into 48 KB, maybe sacrificing the QL matrix keyboard driver?
Will be difficult as the local keyboard translation tables need to be included. The current binary is almost 4K and includes US, UK and German tables. Maybe it would fit if I discarded the MDV driver and all low level routines associated with it, but that requires replacement of the Minerva ROM and one of the charms of QIMSI is that it can be just plugged in without any hardware modification. Also it's not very difficult to include the necessary commands in your boot file to load and activate this driver (the only thing is that you have to wait for the F1/F2 message to time out if your original QL keyboard is no longer working).


User avatar
Peter
Font of All Knowledge
Posts: 2014
Joined: Sat Jan 22, 2011 8:47 am

Re: Breaking news: The QIMSI Interface

Post by Peter »

janbredenbeek wrote: Sun Oct 08, 2023 12:27 pm Also it's not very difficult to include the necessary commands in your boot file to load and activate this driver (the only thing is that you have to wait for the F1/F2 message to time out if your original QL keyboard is no longer working).
Yes the idea to use the external keyboard already for F1/F2 was why I was asking. But I agree that this is very unimportant. I mostly use my QLs as they are, within their wonderful original case. External keyboard was never a key feature of QIMSI, it is not a big deal to wait for the F1/F2 timeout.

By the way, we tried (with Wolfgang's help) to optionally "feed" an F1/F2 into the keyboard buffer from the driver ROM, but it never worked reliably with all hardware setups.


Post Reply