Page 3 of 5
Re: QL-SD card project
Posted: Sun Oct 20, 2013 7:41 pm
by Simon_Carr
Hear Hear! Having been away a lot recently, I've just got back to see lots of posts elsewhere in this forum regarding costs of gold cards etc...
I'm not posting to that discussion, because it seems to have become completely and unnecessarily nasty, but I think the discussion does really highlight the need, and market for an SD card based storage system. I really hope that something does come of this, because clearly this issue is going to become all the more pressing. The QL needs an SD card, RAM boost and perhaps also improved/enhanced ROM facilities.
Oh and a convertor to be able to use an Atari joystick would be nice, and much easier to deliver!
Re: QL-SD card project
Posted: Sun Oct 20, 2013 7:43 pm
by Dave
As always, I'm here ready, willing and able to manufacture anything that is a finished design, and sell it at a low price.
Re: QL-SD card project
Posted: Sun Oct 20, 2013 7:53 pm
by 1024MAK
So, there I was thinking... Does anyone know of a finished *working* design for a time machine for Dave to err, manufacture?
Seriously, thanks for your support Dave
And in-between the err, debate that is being refered to, the relevent bit about SD cards is here:
1024MAK wrote:If the QL community wants to survive, new active members are needed (see the problems with QUANTA and QL Today). Microdrives are not viable long term. Floppy disks are not viable long term. Emulation is not the answer either (explain to me how I can encourage a young person to use an emulator of an obscure {to them} old 1980's computer on an all singing, all dancing modern computer).
So an SD card system is vital. Also important is an inexpensive RAM card (with a socket for a EPROM/EEPROM/Flash chip so that people can have access to the extra commands in Toolkit).
Myself, I would also like an expansion card with all this included that slots into the expansion port.
The increasing cost of living that is affecting vast numbers of ordinary people is going to affect people around the world for many, many more years. So keeping the cost of new hardware low or at a reasonable price is important.
Mark
Re: QL-SD card project
Posted: Sun Oct 20, 2013 7:56 pm
by Dave
My gut instinct is that we should throw up a Kickstarter campaign to raise some funds to a) properly reward the first to market for their work by paying some or all of their development costs and b) using the remainder (if any) to offset production costs with the condition this be passed through to the consumer.
Re: QL-SD card project
Posted: Sun Oct 20, 2013 9:20 pm
by Simon_Carr
Good idea, although I don't know how big the QL community is, so would suggest a fairly low target to raise. I would certainly back it with my spare cash if it was to help such an item come to fruition.
Re: QL-SD card project
Posted: Mon Oct 21, 2013 7:08 pm
by Peter
This picture shows the ROMOE signal on my SuperGoldCard QL, which is used as an imput to QL-SD. The horizontal dotted line is the maximum allowed low level TTL voltage of 0.8V. The measured signal illegally goes up to 1.5V where logic level is low. The oscilloscope has a bandwidth of 60 MHz only, so the spikes visible to a >300 MHz PLD might be higher.
- SGC_romoe_large.JPG (22.89 KiB) Viewed 6532 times
The 30 year old circuitry of the QL was too slow to see this, so the fault did not have immediate effects. But the QL-SD uses a contemporary PLD, with over 300 MHz maximum operating frequency, which considers it a valid ROMOE signal and does unexpected things. Other signals than ROMOE are also far from perfect.
Quality of ROMOE is better on a standard QL.
Re: QL-SD card project
Posted: Mon Oct 21, 2013 8:03 pm
by 1024MAK
Best get some 220Ω resistors and 47pF capacitors then to low pass the filtor the signal...
Follow that with a schmitt trigger input and plug and pray
Yes, I know, it's never that simple
Mark
Re: QL-SD card project
Posted: Mon Oct 21, 2013 8:42 pm
by Peter
1024MAK wrote:
Yes, I know, it's never that simple
Mark
And here's one of the examples why: Noisy low-level address signal at 1.2V during ROMOE active-high. (Scope has persistance on, address signal on top.) To deal with this kind of noise, the adress line is sampled at ROMOE rising edge, where there is less noise. Now if ROMOE is delayed by filtering, the adress line might get evaluated at the worst moment.
- GC_addr_romoe.JPG (27.49 KiB) Viewed 6517 times
In the next step you'd filter every address line.
I was posting this because Dave asked for details, and also as a warning what needs to considered when using modern PLDs. I do not lack ideas what could be tried, but the design is finished & well tested on the standard QL, so I won't rip up the whole thing again. I have already invested far more time than reasonable. And Adrian Ives as well.
Re: QL-SD card project
Posted: Mon Oct 21, 2013 9:06 pm
by 1024MAK
Yes, with digital signals, correct timing is everything.
A question. Could some of the problems be due to a poor ground layout / tracks? Are you seeing a form of "ground bounce" where the voltage drop through the ground is helping to cause the voltage on the signal lines to be out of spec?
Mark
Re: QL-SD card project
Posted: Mon Oct 21, 2013 9:15 pm
by 1024MAK
Thinking about my comments after hitting the post button, if the 'scope is grounded at a good "solid" ground point, probably not ground bounce
Mark