SIDE 3 preorder

[flashjazzcat]

11 hours ago, dostrow said:

Edit: To clarify, after typing CAR and hitting Return I get the black screen, SDX starts up just fine.

Are you sure you have the SIDE3 switch in the upper position (nearest the rear edge of the shell)? Make sure you're also using a stock XL/XE OS.

 

Other than that, I'm not sure what to suggest at the moment. Does your SIDE3 loader appear when you turn on the machine with U1MB's SDX disabled (and with the switch in the loader position)?

[dostrow]

6 hours ago, flashjazzcat said:

Are you sure you have the SIDE3 switch in the upper position (nearest the rear edge of the shell)? Make sure you're also using a stock XL/XE OS.

 

Other than that, I'm not sure what to suggest at the moment. Does your SIDE3 loader appear when you turn on the machine with U1MB's SDX disabled (and with the switch in the loader position)?

I am running the default XL OS and with SDX disabled I get the same behavior, it boots to a black screen. Interestingly I tried one other thing which was inserting the cartridge while in the U1MB settings screen. When doing this the screen gets all scrambled and the system locks up, no further keyboard input works.

[flashjazzcat]

1 hour ago, dostrow said:

I am running the default XL OS and with SDX disabled I get the same behavior, it boots to a black screen.

So that cartridge doesn't work, basically? BTW: I would not insert the cart while the machine is switched on; even if you don't damage anything, you won't accomplish anything good.

 

You may simply have a corrupt loader on the ROM; a re-flash will in that case fix it, but that's a tad difficult when you can't start said loader in the first place. I am assuming it is not a timing issue such as the one I experienced on a stock customer 65XE I tested the other night (when hurriedly preparing the loader flash tool for deployment). That machine variously either didn't register the presence of SIDE3 at all, or plunged the system into blackness until I replaced the Motorola 74LS08 with a different one from the spares box. Every other cartridge (including SIDE2) worked fine with the original Motorola chip.

 

Anyway: if you're able to boot the cart's on-board SDX (after disabling everything but extended RAM on the U1MB), let me know. You could follow an immensely convoluted series of operations in order to run the loader update tool from the FAT partition of the SD card via the FATFS.SYS and SIDE3.SYS drivers.

 

Apologies for the hassle...

[dostrow]

2 hours ago, flashjazzcat said:

So that cartridge doesn't work, basically? BTW: I would not insert the cart while the machine is switched on; even if you don't damage anything, you won't accomplish anything good.

 

You may simply have a corrupt loader on the ROM; a re-flash will in that case fix it, but that's a tad difficult when you can't start said loader in the first place. I am assuming it is not a timing issue such as the one I experienced on a stock customer 65XE I tested the other night (when hurriedly preparing the loader flash tool for deployment). That machine variously either didn't register the presence of SIDE3 at all, or plunged the system into blackness until I replaced the Motorola 74LS08 with a different one from the spares box. Every other cartridge (including SIDE2) worked fine with the original Motorola chip.

 

Anyway: if you're able to boot the cart's on-board SDX (after disabling everything but extended RAM on the U1MB), let me know. You could follow an immensely convoluted series of operations in order to run the loader update tool from the FAT partition of the SD card via the FATFS.SYS and SIDE3.SYS drivers.

 

Apologies for the hassle...

No worries at all, I understand completely that it is a new product that you are clearly still ironing things out with, nothing is perfect at launch, its the price early adopters pay, I totally understand. I'm also more than happy to return this for a replacement to lotharek once he is back from vacation if we can determine that it is faulty.

 

To remove the U1MB from the picture I went ahead and put the original ROM and MMU back in place, made sure the system was booting back to the original BASIC ROM just fine and then shut the system off, put the SIDE3 in the cartridge slot and attempted to boot it, again just a black screen, no luck. The two things I cannot isolate (yet) are:

 

1). Does this 130XE properly load any cartridge? I purchased it off eBay at roughly the same time I bought the U1MB and SIDE3 from lotharek and it didn't come with any carts so I really can't say for certain, all I can say is that it passes all the built in tests, both with and without the U1MB installed.

2). The only additional upgrade this system has is a SOPHIA DVI rev. C. I have not tried removing it to see what happens there...

[candle]

to disable SIDE3 on power up simply turn Atari power on while pressing SIDE3 button

if case of some misshaps inside flash memory, you'll be able to recover from SIO

 

[+remowilliams]

Hmm. I did the U1MB / SIDE3 updates in #290 and using 'L' from the U1MB menu is loading the SIDE3 Loader, but it won't boot any ATRs.  They are listed but throw a 144 error when trying to boot them.  xex is loading fine.

[+DrVenkman]

24 minutes ago, remowilliams said:

I did the U1MB / SIDE3 updates in #290 and using 'L' from the U1MB menu is loading the SIDE3 Loader, but it won't boot any ATRs.  They are listed but throw a 144 error when trying to boot them.  xex is loading fine.

Make sure you have the PBI BIOS Hard Disk enabled. 

 

EDIT: Nevermind - they wouldn't even be listed if you didn't already do that. Hmm ...

[dostrow]

1 hour ago, candle said:

to disable SIDE3 on power up simply turn Atari power on while pressing SIDE3 button

if case of some misshaps inside flash memory, you'll be able to recover from SIO

 

For what it is worth when holding down the button on power up (with SDX disabled) it drops me right to BASIC (which I suppose is the correct behavior for that). Unfortunately at the moment I don't even have any SIO devices (though there is a Fujinet on it's way). I'm also going to try and get my hands on a known working cartridge of some other nature to verify that loading any cartridge works as expected.

[candle]

yes, this is expected behaviour

 

[+remowilliams]

Issue solved.  Original test card was 1GB Sandisk, formatted as either FAT16/32 it exhibited the same symptoms.

 

Switching to a 2GB Kingston or 2GB Transcend, and ATRs are loading as expected.

[candle]

I just wonder, why anyone would want to compromise preformance of their mass storage device using old SD cards, when new 16GB card costs $3

use at least SDHC cards

2GB cards are probably the worst choice since special handling might be needed for those, and only few were tested (I have only one)

[Marius]

1 hour ago, candle said:

I just wonder, why anyone would want to compromise preformance of their mass storage device using old SD cards, when new 16GB card costs $3

use at least SDHC cards

2GB cards are probably the worst choice since special handling might be needed for those, and only few were tested (I have only one)

You are right. This has probably something to do with stuff people have at their home yet, which they do no longer use in 'regular' equipment. 

 

I have a bunch of that stuff too. But it is good to know that modern SD cards are fully supported. If I might get into Side3 one day, I'll use a new SD card.

 

Interesting what you say about 2GB cards. I have a bunch of those... why are especially 2GB cards asking for special handling?

 

 

[flashjazzcat]

12 hours ago, dostrow said:

I'm also going to try and get my hands on a known working cartridge of some other nature to verify that loading any cartridge works as expected.

This is a very advisable course of action, in light of the fact the machine itself is a bit of an unknown quantity.

[candle]

2GB cards are on the edge of SD card specifications - this is largest allowable size for their specs, and special measures have to be taken for multi-sector transfer support as done by SIDE3 on both hardware and software layers

 

[Jimbaloid]

10 hours ago, candle said:

I just wonder, why anyone would want to compromise preformance of their mass storage device using old SD cards, when new 16GB card costs $3

use at least SDHC cards

2GB cards are probably the worst choice since special handling might be needed for those, and only few were tested (I have only one)

I'd suggest that the logic goes something like this;

 

"I don't need massive storage space and 80MB/s for an 8 bit machine; Since all my previous experience of SD cards has been 'they just work', whatever I put them into, I anticipate no issues and so don't need to buy anything new; I have this quality, branded SD card already, it would be great to press it back into use rather than send it to land fill; I've done a full format on it with no errors, it's working great in my PC." - They are not aware of any technical limitation that might exist for this specific use case because they've never seen an SD card limitation before. And despite that they are likely some-what technically literate to have ordered a SIDE 3 in the first place, they are not electrical engineers up to date with SD card specifications and special handling needs.

 

I suggest this because it's pretty much an account of my own internal monologue on receipt of my SIDE 3...

 

And immediately I ran into trouble, and no it wasn't a 2GB card initially, it was a 4GB Transcend SDHC card, only when it didn't work did I go to my drawer and pull out a variety of older, and smaller cards. Surely one will work, but no, five more cards that didn't work - all sorts of strange behaviour - did I get a duff cart? So I wonder, maybe SD class 4 is the issue, so I obtain a class 6 Panasonic card from a friend only to get NAK errors back from that. So now I'm waiting on two new 16GB cards in the post, with no idea if they'll actually work.

 

And if we do spend just $3, I've already seen comments on here that suggest that we'd be fools to be "not thinking they are almost certainly clones from China", so what are we to do!

 

Anyway, believe it or not, this is the short version and I'll be reporting back in a few days when the postie has brought my new cards, hopefully with some good news. But I do hope I've just been incredibly unlucky or maybe I do have a dodgy cart, because it does all seem a little surprising. 

 

Please be kind, my first post here! :-)

 

[Marius]

1 hour ago, Jimbaloid said:

And immediately I ran into trouble, and no it wasn't a 2GB card initially, it was a 4GB Transcend SDHC card, only when it didn't work did I go to my drawer and pull out a variety of older, and smaller cards. Surely one will work, but no, five more cards that didn't work - all sorts of strange behaviour - did I get a duff cart? So I wonder, maybe SD class 4 is the issue, so I obtain a class 6 Panasonic card from a friend only to get NAK errors back from that. So now I'm waiting on two new 16GB cards in the post, with no idea if they'll actually work.

 

Please be kind, my first post here! ?

 

Usually this kind of trouble is 'tackled' in a phase of beta-testing. I have all kind of SD cards here, from completely obsolete sizes like 32mb and probably even smaller, and brandnew ones. I also have SD cards from totally obscure brands, and high quality stuff. Next to that I have many Atari 8bit computers in all flavors and setups. And if that is not enough, I have this completely insane and 'troubled' way of testing stuff. Probably because of some mental disorder on my side, but for some reason I always do the unexpected with my stuff, so that brings up all kind of unexpected behavior (in the atari, or in the add-on... or in me haha).

 

But unfortunately I am not in the beta-test team, so it seems someway that the users who purchased the SIDE3 are now inevitable and automatic buyer AND beta-tester at the same time. I wished I was in that beta-test team though. I could have prevented this problem. 

I think no one needs to worry though. This problem will be addressed. I know the hardware developers will, it might take some more time. I hope the hardware developers take the issue serious though.

 

It might be another issue by the way. If that 'less' cards work, isn't there a chance that you have some stability issues in the atari itself?

 

 

[+Stephen]

Same.  When we had the multitude of CF cards before this, the answer was "why the hell would you try to use an 8GB card when a few hundred meg works fine".  2GB CF cards were the "sweet spot".  Now we get a new $100+ untested device and we're told condescendingly (which seems to be the only way candler talks to anyone - down to them) that only an idiot would use a 2GB SD card when much larger ones are available.

 

Live and learn I guess.  Maybe one day we can all be at his level of god-like awesomeness and we won't have to post these problems.  You know - the kind that usually get caught when beta-testing a product.

[Jimbaloid]

9 hours ago, Marius said:

It might be another issue by the way. If that 'less' cards work, isn't there a chance that you have some stability issues in the atari itself?

I have three Atari machines... So; multiple Atari, multiple SD card, the only single common factor right now is the SIDE 3. Let me see what luck the new cards bring.

[Roydea6]

I have two Side One cartridges and since the Bios flash for Side3 use I can not get my APT partitions until I run the side.sys driver.

 

Also the Fat area of the CF card does not show any ATRs'

[candle]

Jim,

starting this project i went for a shopping and bought every card type/brand i could in two towns - I didn't care for anything but just availability and to have a test sample i could use and check for

i've ended up with 20 cards ranging from 16GB to 64GB - i didn't want to spend more than $20 on a card, so bigger ones wasn't my target

then i gathered every card i had laying since stone age and asked friends to help me with my quest

result were 16MB cards, 128MB cards (MMC, MMC+) and all were tested for read/write performace and surely they were some mishaps in the way, but always it was a software isue and card was simply requireing some kind of special handling during the init or during the usage phase to work reliable

some had to have chip select signal brought high (deselected) then low (active) before each command, otherwise they would not work at all (Kingston)

others had to be clocked eight times before they release the  bus after being deselected (ie SanDisk)

others did require  full protocol-dependand initialisation phase with no shortcuts made - especially if they were of older type (MMC)

Jon brought to the test table another ones he could find locally and this was our test sample

surely we can't have them all, yet it always been, and i belive it will be a case of some special initialization or handling between commands to make them working

 

Roy,

Use some kind of SSD migration tool to migrate your data between CF and SD card - you could use DD for *nix systems, or some other software for Windows

 

[TGB1718]

33 minutes ago, candle said:

you could use DD for *nix systems

Cheap Raspberry Pi does these sort of things really well, added advantage the Pi runs Respqt too if you need it.

 

Just power it up, log on using Putty in Windows, piece of cake   no keyboard/mouse/monitor needed on the Pi.

 

And for the Pi GUI, VNC when using running Respqt

 

Edited October 22, 2020 by TGB1718
Update

[elmer]

2 hours ago, candle said:

i've ended up with 20 cards ranging from 16GB to 64GB - i didn't want to spend more than $20 on a card, so bigger ones wasn't my target

then i gathered every card i had laying since stone age and asked friends to help me with my quest

result were 16MB cards, 128MB cards (MMC, MMC+) and all were tested for read/write performace and surely they were some mishaps in the way, but always it was a software isue and card was simply requireing some kind of special handling during the init or during the usage phase to work reliable

 

some had to have chip select signal brought high (deselected) then low (active) before each command, otherwise they would not work at all (Kingston)

others had to be clocked eight times before they release the  bus after being deselected (ie SanDisk)

others did require  full protocol-dependand initialisation phase with no shortcuts made - especially if they were of older type (MMC)

 

FYI, none of what you mention there sounds like any kind of unusual investigation or special handling was really required at all for those issues ... the things that you've described, the chip select, the 8 clocks, and the initialization, are all part of the well-documented MMC/SD card access standards. See here for SPI mode, or you can download some of the official standards docs online (although they can be hard to find without paying money).

 

That doesn't take away from the fact that "yes", it can be a total pain to actually get reading and writing and error handling to actually work reliably!

Edited October 22, 2020 by elmer

[flashjazzcat]

20 hours ago, Jimbaloid said:

So I wonder, maybe SD class 4 is the issue, so I obtain a class 6 Panasonic card from a friend only to get NAK errors back from that.

Thanks again for sending the Transcend and Panasonic cards in the mail, Jim. Although the Transcend worked out of the box for me (which it did not for you), the Panasonic 4GB card produced NAK errors here as well. Problems occurred during high-speed (fast clock) operation; the card model (which is obtained at the slow ~400KHz clock, although actually ~440KHz on the SIDE3) was reported OK, but subsequent sector transfers would fail. Inspecting the R1 response code didn't reveal anything especially illuminating, but two minutes of Googling the issue did. I'm delighted to report that writing an extra $FF to the SPI data register immediately before asserting chip select on the SD card (prior to issuing a read command) allows the card to work perfectly:

	lda	#$ff
	sta	SIDE3_SD_SPI_DATA	; fixes Panasonic 4GB
	ldy	#SIDE3_SD_SPI_CS|SIDE3_SD_SPI_CLKSEL|SIDE3_SD_SPI_POWERON
	sty	SIDE3_SD_SPI_CTRL
	lda	#$ff	; fix for certain cards: eight extra clock cycles to prepare for response
	sta	SIDE3_SD_SPI_DATA
	sty SIDE3_SD_SPI_CTRL
	ldy	#0
	sty	SIDE3_SD_CRC7_GEN

This is in addition to the commented fix for SanDisk cards (write $FF after chip select), for which we have Candle to thank.

 

The problem I have (well, one of many) is that although I have over a dozen assorted SD cards (including microSD cards in adapters), the last time I tested them, every single one worked. However, I'll now have to test every one of them to ensure they still work with this fix (although I have no reason to suppose they won't). After that, I'll pull another half a dozen Ataris out of the cupboard and repeat the same tests on each machine with a sub-set of the SD card collection.

 

After that, I'll upload the update so that folks can try it. Fingers crossed it will be a significant improvement.

[+Stephen]

Oh man - just checked tracking info.  My SIDE3 was in New York as of the 18th.  Should be seeing it any day now.

[fenrock]

1 hour ago, flashjazzcat said:

After that, I'll pull another half a dozen Ataris out of the cupboard and repeat the same tests on each machine with a sub-set of the SD card collection

Them bags under the eyes are not disappearing anytime soon then.

Still, what else can you do during lockdown?