Multiply incompatibility with 48 to 128kb memory expansion

[F3nyo]

Hello,

This project is great, but I have one problem with it: the multiply SD reader does not work with a 48k to 128k expansion, I can only load ROMs, but it works with an untouched 48k machine, even with an AY interface. Is there any way to make it compatible with the above memory expansion (this allows most 128kb games to be played on a 48kb machine, at least from ROM, but of course only a limited number can fit there, and it takes a very long time to load with Maxduino)?

Best Regards; Fenyő

[mad3001]

Hello Fenyő,
You mean dandanator works correctly but if you try to load anything from SD (multiply) it does not works?
Do multiply load or sais "Multiply not detected" ?
If you can reach Multiply menu, can you load differente file format as TAP,Z80,SNA ?
Also does not works with .SCR (screen) ?
Regards
Mario

[F3nyo]

When browsing the SD card, when I enter the main directory and it says it is reading the contents, it usually freezes, sometimes resets, but now I tried it on another machine where I made a 128kb mod and it works on that, so maybe my first machine is not 100% in this setup, both are issue3, but the problematic machine also has a dc-dc mod, with which the SD reader only works in 48kb mode without any other additional interface, if I put anything else into the system, only the rom part works flawlessly. On the machine without the DC-DC mod, however, the 3 cards (128kb exetension, ay inteface, Dandanor amie with multiply in one pcb) work together with the SD reader. So this will require further investigation on my part...

[mad3001]

Hello, thank you for your comments. One more thing: make sure you do not use an additional Kempston interface (or an AY + Kempston interface that has the Kempston interface enabled on that board), as it will interfere with Multiply, because it needs to be the only interface with Kempston enabled to work properly.

[F3nyo]

Yes, thank you, I read in the description that it is incompatible with Kempston, it didn't even work with the Kay interface, even if I disabled the joystick with a jumper, only if I removed the 2 logic ICs belonging to it and left only those belonging to the AY, so that's not the problem. However, even with the unmodified issue 3 machine, it sometimes freezes during gameplay. I used version 1.1.4 of sdfat, which is the recommended one. Tzxduino, for example, was very sensitive to sdfat versions, and it also froze with newer ones. The description mentions version 1.1.4 as the maximum. Is this the optimal version?

[mad3001]

Version 1.1.4 of SDfat, by Bill Greiman, or the Adafruit fork 1.5.1 are the latest versions compatible with the multiply code. Starting with version 2.0, many internal changes were introduced in the sdfat library, and adapting it requires significant modifications to the code.

[F3nyo]

Thank you, I used sdfat 1.1.4 to compile and upload the source code, so that's not the problem. It's completely stable on a 48k base machine on its own, but the gal chips are suspicious. The memory expansion reset button only works with a 25ns gal chip (it doesn't respond to faster ones, probably because the signal is too short/sharp), just as the Dandanator Amie reset doesn't work with a 10ns gal chip, but I only had 10ns in stock, so I ordered 25ns in case the signal is just too fast for Specci, plus this way the signals of the two gal chips are different.

[mad3001]

Well, we look forward to your comments with the two 25 ns “gals”. Although nowadays it is difficult to find good “gals”.

[F3nyo]

I think the problem has been solved, but it is multi-layered and depends on the given Spectrum as well, how stable the system will be with 3 add-ons.
In the end, Gal only calculated that the reset button works without any problems at 25ns, so it can be said that we should choose the slowest Gal chips possible.
For the stability of the Issue3 DC-DC mod machine, I first had to replace the faceplate, as the old one was worn out and slightly damaged. However, this meant that it did not behave as a defined capacitive body. Then, in the case of 3 expansion cards, the order of the accessories is important: the Dandantor Amie must be first, followed by the AY and then the 128kb memory expansion card. The rev6 version is stable in the long run, but the rev3b needs to be modified with two edge damping resistors (I used 33 ohms). I placed it between gal 18 and 19 and the corresponding RAM pins (22, 30), otherwise the timing slips after a while, which manifests itself in either a reset or a graphics error.
Unfortunately, the plain, unmodified Issue 3 board does not like the rev6 128kb memory expansion if the AY is also installed, but with rev3b, even with the above modifications, it is not completely stable in the long run, only showing up later, usually after at least an hour or more. I will try the original rev3 expander to see if it works, but the lesson is that delay is critical with many expanders, and this can vary depending on the motherboard revision and specific unit. From this point of view, issue 2 seems to be less sensitive to both latency and capacitive noise coming from the faceplate.

[mad3001]

Thanks for the update. Sometimes, having many interfaces can be challenging; in fact, different delays depending on the version of the interface or the board can make it work or not.