Re: [Ql-Users] QL SD/MMC-Card interface survey
Malcolm Lear wrote: I was looking at the QL schematics a few weeks back and noticed the Microdrives use a serial bus. I guess this is bit banged by the software to transfer data. Could this be connected to the SD SPI serial interface using level shifters (5V - 3.3V)? As always, things are not that easy :-) I'd have to leave at least 3 of 6 lines in their original use (drive selection) so other microdrives don't get totally confused. The other three lines are just one output and two data lines (both seem to change their direction at the same time, according to that output). And I don't know wether the data lines can be bit-banged at all. Even if I construct something useful from those 3 lines - which would of course involve more that just level-shifting - there is still risk to confuse the microdrive portion of the OS. Peter ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm
Re: [Ql-Users] QL SD/MMC-Card interface survey
On 12/01/2011 12:30, Peter wrote: Malcolm Lear wrote: I was looking at the QL schematics a few weeks back and noticed the Microdrives use a serial bus. I guess this is bit banged by the software to transfer data. Could this be connected to the SD SPI serial interface using level shifters (5V - 3.3V)? As always, things are not that easy :-) I'd have to leave at least 3 of 6 lines in their original use (drive selection) so other microdrives don't get totally confused. The other three lines are just one output and two data lines (both seem to change their direction at the same time, according to that output). And I don't know wether the data lines can be bit-banged at all. Even if I construct something useful from those 3 lines - which would of course involve more that just level-shifting - there is still risk to confuse the microdrive portion of the OS. Speaking from pretty little low level knowledge, how about abandoning the microdrives completely (I thought you were) and making your interface emulate microdrives? I wonder what the O/S would make of a giant sector count (8-)# I suppose speed would be an issue as well. I wonder whether the 8302 could input/output faster. I know Laurence wound up the data lines to the 8749 to a remarkable degree. We never found an upper limit as it could clearly go faster than the PIC. It was so fast that it took a long time for us to realise we *were* getting a response. I still have the polaroids somewhere taken on my oscilloscope. Tony ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm
Re: [Ql-Users] QL SD/MMC-Card interface survey
- Original Message - From: Ian Pine ilp...@tesco.net To: ql-us...@q-v-d.com Sent: Tuesday, January 11, 2011 2:33 PM Subject: Re: [Ql-Users] QL SD/MMC-Card interface survey - Original Message - snipped Anything which works on the Q40 would be good, but something which could be connected as a slave IDE device in place of the CD-ROM would be best. If it could be mounted behind the bay cover panel with a slot cut in it, would be very useful. It would be nice if it would work with the existing SMSQ/E WINx_ driver. Ian. ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm I've changed my mind on this. Peter's option A sounds better to me; it is portable, and if the interface details are published, could also be put to use in other home-grown projects. Ian. ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm
Re: [Ql-Users] QL SD/MMC-Card interface survey
On 12/01/2011 12:46, Tony Firshman wrote: On 12/01/2011 12:30, Peter wrote: Malcolm Lear wrote: I was looking at the QL schematics a few weeks back and noticed the Microdrives use a serial bus. I guess this is bit banged by the software to transfer data. Could this be connected to the SD SPI serial interface using level shifters (5V - 3.3V)? As always, things are not that easy :-) I'd have to leave at least 3 of 6 lines in their original use (drive selection) so other microdrives don't get totally confused. The other three lines are just one output and two data lines (both seem to change their direction at the same time, according to that output). And I don't know wether the data lines can be bit-banged at all. Even if I construct something useful from those 3 lines - which would of course involve more that just level-shifting - there is still risk to confuse the microdrive portion of the OS. Speaking from pretty little low level knowledge, how about abandoning the microdrives completely (I thought you were) and making your interface emulate microdrives? I wonder what the O/S would make of a giant sector count (8-)# I suppose speed would be an issue as well. I wonder whether the 8302 could input/output faster. I know Laurence wound up the data lines to the 8749 to a remarkable degree. We never found an upper limit as it could clearly go faster than the PIC. It was so fast that it took a long time for us to realise we *were* getting a response. I still have the polaroids somewhere taken on my oscilloscope. Tony ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm Yes, that may be possible using an AVR or PIC to perform the emulation. The big up side is no modification to the QL hardware. There is also enough FLASH in most microcontrollers to store a Microdrive image, so on reset the OS could a boot file on MDV1. This could patch the OS to cope with large sector counts or load new drivers for MMC1-8. Malcolm ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm
Re: [Ql-Users] QL SD/MMC-Card interface survey
Tony Firshman wrote: Speaking from pretty little low level knowledge, how about abandoning the microdrives completely (I thought you were) So far, I could have used just one SD card interface, keeping one microdrive. Abandoning the microdrives completely is an option I could think about. But the task of creating SD card signals is still not trivial. I don't know exactly how the ZX8302 behaves internally, and which timings are acceptable. For example, the OS even inserts delays when just bit-banging the drive select daisy chain, and I have no idea why this is required. Line lengths? Noise? ZX8302 internal requirements? ... Using the ZX8302 for a completely different purpose might require a lot of time for investigation and experimentation. and making your interface emulate microdrives? Hehe, nice idea. Unfortunately it would be a pain in terms of speed, also size would be limited by the MDV drivers I guess :) Peter ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm
Re: [Ql-Users] QL SD/MMC-Card interface survey
On 12/01/2011 13:48, Peter wrote: Tony Firshman wrote: Speaking from pretty little low level knowledge, how about abandoning the microdrives completely (I thought you were) So far, I could have used just one SD card interface, keeping one microdrive. Abandoning the microdrives completely is an option I could think about. But the task of creating SD card signals is still not trivial. I don't know exactly how the ZX8302 behaves internally, and which timings are acceptable. For example, the OS even inserts delays when just bit-banging the drive select daisy chain, and I have no idea why this is required. Line lengths? Noise? ZX8302 internal requirements? ... Using the ZX8302 for a completely different purpose might require a lot of time for investigation and experimentation. and making your interface emulate microdrives? Hehe, nice idea. Unfortunately it would be a pain in terms of speed, also size would be limited by the MDV drivers I guess :) I bet the drivers could be patched. You could also keep one mdv if the interface was fully compatible. It would be quite hard though switching between the two speeds. I suppose if simultaneous access was barred, the mdv driver could be patched on the fly. As Malcolm suggested, a PIC could be used, and may have enough on board storage to buffer one boot image. Failing that it could be a bootstrap to load from the card. Laurence Reeves knows a *lot* about the logic. Tony ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm
Re: [Ql-Users] QL SD/MMC-Card interface survey
If I remember right, emultating a microdrive completely is probably not a very good idea - The electrical intarface to the drives is typical minimalist Sinclair technology and consists of 3 adress lines to adress the drive (0-7), a motor start line and a read and write line. There's no way to adress a specific sector on a drive, and the technology relies on all sectors of the tape passing the r/w head in a reasonable amount of time - The logic just waits until it sees the correct sector header passing by and then starts reading or writing the specific sector. This worked quite well with small capacities of several hundred sectors. I guess nobody wants to wait until all sectors of a 2GB SD have passed the emulated head.. Cheers, Tobias -Original-Nachricht- Subject: Re: [Ql-Users] QL SD/MMC-Card interface survey Date: Wed, 12 Jan 2011 21:46:33 +0100 From: Tony Firshman t...@firshman.co.uk To: ql-us...@q-v-d.com On 12/01/2011 13:48, Peter wrote: Tony Firshman wrote: Speaking from pretty little low level knowledge, how about abandoning the microdrives completely (I thought you were) So far, I could have used just one SD card interface, keeping one microdrive. Abandoning the microdrives completely is an option I could think about. But the task of creating SD card signals is still not trivial. I don't know exactly how the ZX8302 behaves internally, and which timings are acceptable. For example, the OS even inserts delays when just bit-banging the drive select daisy chain, and I have no idea why this is required. Line lengths? Noise? ZX8302 internal requirements? ... Using the ZX8302 for a completely different purpose might require a lot of time for investigation and experimentation. and making your interface emulate microdrives? Hehe, nice idea. Unfortunately it would be a pain in terms of speed, also size would be limited by the MDV drivers I guess :) I bet the drivers could be patched. You could also keep one mdv if the interface was fully compatible. It would be quite hard though switching between the two speeds. I suppose if simultaneous access was barred, the mdv driver could be patched on the fly. As Malcolm suggested, a PIC could be used, and may have enough on board storage to buffer one boot image. Failing that it could be a bootstrap to load from the card. Laurence Reeves knows a *lot* about the logic. Tony ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm
