Jump to content
IGNORED

U1MB, Incognito, SIDE2, SIDE3, 1088XEL/XLD Firmware 4.0 Released


Recommended Posts

40 minutes ago, Mazzspeed said:

I had one 2GB SanDisk SD card that didn't work, and it's about ten years old. The other two cards I tried work perfectly with outstanding performance.

 

The user in question ordered three cards of different makes at the one time, how unlucky does one have to be? There's either a problem with that particular SIDE3, or the cards are fakes. As I believe the SIDE3 has been tried on two different A8's and exhibited the same issues on both while FJC has verified the ISO as working.

Just to be be perfectly clear, I wasn't trying to imply that there is A) something wrong with the SIDE3, or B) that FJC's BIOS is faulty in the way it reads SD cards, and in no way am I saying you implied any of that, but it's good to clear the air.

 

It was just something I've observed on various memory card reading hardware for the A8, and then stumbled upon an alternative method that can be used for accessing SD cards which I had never seen before. Not saying that'll necessarily work any better, but in case it hasn't been tried it would be an interesting experiment.

 

Something I've wondered about for a long time now, is why do some memory cards fail to work properly, irrespective of whether they are fakes. Because I have a bunch of fakes, and a bunch of different brands, and/or speeds, having varying degrees of success at reading all of them on various bits of Atari hardware I own, whereas all of them work perfectly fine on any of my PCs and tablets. There's got to be a reason for this difference in compatibility, and it would be nice to know what that is. To be clear I am not putting down what Candle has created in the SIDE3 or FJC's BIOS, this is more curiosity, and me thinking out loud. And... I've also seen the same situation on my NUC FujiNet, as well as my XEL-CFx products - with some cards refusing to work.

 

Theoretically a memory card interface for the Atari really should be able to be as simple as the first MyIDE hardware was, and still work reliably. I know @bob1200xl who's kind of like the father of the IDE interface on the Atari, created an even simpler hardware IDE interface (no buffering) that worked with all kinds of CF cards no problem. So it would seem there's got to be a way to do this.

 

  • Like 2
Link to comment
Share on other sites

23 minutes ago, mytek said:

Just to be be perfectly clear, I wasn't trying to imply that there is A) something wrong with the SIDE3, or B) that FJC's BIOS is faulty in the way it reads SD cards, and in no way am I saying you implied any of that, but it's good to clear the air.

 

It was just something I've observed on various memory card reading hardware for the A8, and then stumbled upon an alternative method that can be used for accessing SD cards which I had never seen before. Not saying that'll necessarily work any better, but in case it hasn't been tried it would be an interesting experiment.

 

Something I've wondered about for a long time now, is why do some memory cards fail to work properly, irrespective of whether they are fakes. Because I have a bunch of fakes, and a bunch of different brands, and/or speeds, having varying degrees of success at reading all of them on various bits of Atari hardware I own, whereas all of them work perfectly fine on any of my PCs and tablets. There's got to be a reason for this difference in compatibility, and it would be nice to know what that is. To be clear I am not putting down what Candle has created in the SIDE3 or FJC's BIOS, this is more curiosity, and me thinking out loud. And... I've also seen the same situation on my NUC FujiNet, as well as my XEL-CFx products - with some cards refusing to work.

 

Theoretically a memory card interface for the Atari really should be able to be as simple as the first MyIDE hardware was, and still work reliably. I know @bob1200xl who's kind of like the father of the IDE interface on the Atari, created an even simpler hardware IDE interface (no buffering) that worked with all kinds of CF cards no problem. So it would seem there's got to be a way to do this.

 

Honestly, I've never had a problem reading SD cards on any retro platform I use - With the exception of that one card which could have been classed as retro itself.

 

I guess, you could have problems regarding certain controllers built into the SD card itself. I know that back in the days of Windows 7 it wasn't uncommon for me to plug a card into a PC and find it downloading needed drivers for the card in the background before the card appeared as a usable device - The same could be happening under the various retro platforms today, with the exception that it's not so easy to download a driver in the background.

Link to comment
Share on other sites

1 hour ago, mytek said:

Something I've wondered about for a long time now, is why do some memory cards fail to work properly, irrespective of whether they are fakes.

This is something that might be worth contacting @batari about.  The Concerto cartridge for the 7800 had similar issues early in its release life, but firmware updates were able to (AFAIK) fix the issue.  Could be a similar situation here.

Link to comment
Share on other sites

10 hours ago, DrVenkman said:

For what it's worth, I use a Samsung EVO 32GB microSD card in its adapter in my SIDE3. I haven't had any specific problems loading files from it, but none of the three files posted today at the end of the FujiConvert thread will play on my 1200XL+U1MB+SIDE3 combo, nor does a file I converted myself using the recommended settings. I get a brief message at the bottom of the Loader as it reads the file, then a black screen. 

 

And yes, the U1MB firmware and Side3 Loader are freshly updated just tonight to the latest downloaded from Jon's site. 

As I mention in the other thread, none of them will play off my 32GB SanDisk card either, but they all work on my 8GB Sony card. No clue why at the moment. I'm going to have to make a debugging version of the PDM player to see exactly why it fails.

  • Like 2
Link to comment
Share on other sites

3 hours ago, Level42 said:

Only problem is that bloody Brexit .... How can I send it over (and you back) without us having to pay import duties and costs, even worse: it staying at customs for weeks...

That is a non-issue as far as I can see, at least regarding items entering the UK for traders like myself. Although I have an EORI code, I haven't been hit with import fees on a single thing since January 1st 2021, no matter where in the world they came from. Of course most larger items are marked 'Vintage computer for repair' (since that's what they commonly are). Brexit is working out even better than hoped, therefore. :)

 

  • Like 2
Link to comment
Share on other sites

3 hours ago, mytek said:

Just to be be perfectly clear, I wasn't trying to imply that there is A) something wrong with the SIDE3, or B) that FJC's BIOS is faulty in the way it reads SD cards, and in no way am I saying you implied any of that, but it's good to clear the air.

I completely agree with this sentiment. When I conclude 'the software is OK', I mean that the algorithms which read data from the card appear to be determinate and free of bugs. Emulation is a useful way of removing 'hardware quirks' from the equation, since - if emulation is accurate - it presents an idealised version of the hardware which works determinately 100 per cent of the time. But there are many things other than 'timing issues' which can cause a piece of software to work in emulation and fail in some way on real hardware (a point tirelessly laboured in other threads where developers leave real hardware testing till the last minute after developing software exclusively in emulation).

 

It's annoying to be accused of defensiveness (not here, I hasten to add) when one is simply trying to coordinate a testing strategy which reveals problems wherever they happen to be, and when there is a known hardware issue (requiring replacement of 74LS08 on the host machine) which has not yet been satisfactorily addressed. The whole saga of media compatibility is a depressing one which has gone on for a decade or more (through MYIDE, SIDE, to a lesser extent with SIDE2, XEL-CF, etc), and it sometimes takes hardware developers several years to revisit an old project and suddently discover that compatibility could be greatly increased via a hardware of VHDL tweak ('Imagine my shock!', I will cry). I definitely don't want to spend dozens of hours trying to compensate for this in software (which is an impossible task), but at the same time, it's important not to conflate timing issues and card compatibility with possible firmware bugs.

 

Hopefully cool heads and consistent reasoning will prevail.

  • Like 6
Link to comment
Share on other sites

SIDE3 Loader 0.43:

 

s3loader.xex

 

Fixed:

  • PDM Playback error
  • 8K 'left' cart ROMs being mirrored at $8000-$9FFF

The PDM issue was a real hum-dinger, and it required several debugging builds running on real hardware before the penny dropped.

 

SD card controllers work with big-endian numbers, while the 6502 is little-endian. Obviously this is accounted for in the SD card command buffer (the 32-bit LBA numbers are flipped around before they're sent to the SPI controller), but there was one place where this caused a problem. The PDM player stores all the LBA sector numbers in big-endian format to save time when pulling them back out of SRAM during playback. Now, after the big-endian list of sector numbers is created, the player sets the ball rolling by pulling the first sector number out of the list and setting up the read before jumping into the main playback loop. Ah... BUT: it calls the common 'SetUpSectorRead' function which actually expects the LBA number in little-endian format.

 

So what was happening here is that the first sector number was never correct, but it went by so quickly it was never noticed anyway. It only became a show-stopper when the lower order bits of the sector number - when sent to the SD card controller as high-order bits - resulted in an out-of-range sector request (beyond the limits of the card capacity). And it gets worse: such a read would properly fail if only the NMIs hadn't already been disabled. This caused the timeout loop waiting for a valid response from the card to run forever, since it's governed by the OS system clock.

 

So, a real cluster-f***, no pun intended. :) The PDM player is a relatively new feature and since it wasn't extensively tested, it was probably a bad idea to draw any conclusions from it concerning card compatibility, etc. Sorry for any inconvenience.

 

Note that this won't cure Level42's messed up directory reads pictured in earlier posts, which is one reason I still think his problems extend beyond PDM playback issues. But for anyone with a working system who just hit a brick wall with PDM files, this update should fix them.

Edited by flashjazzcat
  • Like 14
  • Thanks 1
Link to comment
Share on other sites

3 hours ago, flashjazzcat said:

8K 'left' cart ROMs being mirrored at $8000-$9FFF

Looks to be fixed ;-)  Tested on the 600XL, 800XL, and 1200XL.  Each one reported 37902 bytes free in BASIC, and that was across all three revisions.  I'll add a note to the other thread to this effect.

  • Thanks 1
Link to comment
Share on other sites

6 hours ago, flashjazzcat said:

 Note that this won't cure Level42's messed up directory reads pictured in earlier posts, which is one reason I still think his problems extend beyond PDM playback issues. But for anyone with a working system who just hit a brick wall with PDM files, this update should fix them.

They sure do. However, I still am going to see if PDMs now may play back with my old SDcards.

 

its funny though, that the issue was with the faster card on your side Jon….

Link to comment
Share on other sites

Just flashed my SIDE3 to 0.43:

 

Results:

Sandisk 4GB micro SD card (Standard) which is from november 2008: PDM files now work fine on this card, which they did not before :)

Also, all other files also still load fine the first time.

IMG_4892.thumb.jpg.02ebbcf96a36e5a535662be69d064118.jpg

 

Kingston 1GB Standard SD card which is from July 2006: PDM files don't work. Two of them have the little clock running while building the file map, then yield a Stream error. The third file I tried did start but immediately crashed (humming sound, the same looking wave scrolls down the stream.
IMG_4891.thumb.jpg.3b27d5e566c1f5949d78a096baba57aa.jpg
Please note: I do NOT expect backwards compatibility for ancient SD cards, I only reported the Kingston behavior as a means of feedback for Jon.

 

So, now the funny thing is that the original SD card that I have used on SIDE3 now works 100% with all files.

 

As expected, 0.43 does not change the behavior of the CAR and BIN loading on the three modern SD cards.

 

Edited by Level42
Link to comment
Share on other sites

23 minutes ago, Level42 said:

its funny though, that the issue was with the faster card on your side Jon….

If it wasn't apparent from my explanation of the bug in post 507, the PDM issue had nothing whatsoever to do with card age or speed.

4 minutes ago, Level42 said:

So, not the funny thing is that the original SD card that I have used on SIDE3 now works 100% with all files.

If that card only had issues with PDM files, there's nothing strange about the fact PDM files now work.

Link to comment
Share on other sites

A bit more info:

Here are the results of the three new Sandisk SD cards:

 

IMG_4893.thumb.jpg.3362f65a6d3dd93fa41f89c72bc311f2.jpg

 

IMG_4894.thumb.jpg.1336638fa32f1540dbe550ad61bcc503.jpg

 

IMG_4895.thumb.jpg.21295d4b149e3a798537bcdcbdb2be5a.jpg

 

No fishy-smelling results. Manufacturer and OEM both say SanDisk. Maybe fake cards show this too. However, there is further proof of these cards being real:

 

First: packaging. Have a look at this video that points out any obvious clues between fake and real Extreme Pro packaging:
 

 

 

The one I have clearly has the SEALED BOX version of packaging. Also, the obvious other indications (the spelling errors) are not present. 

 

IMG_5271.thumb.jpg.63d3ea5fc4b5afe44023a2b4995c16b6.jpg

 

IMG_5272.thumb.jpg.45abb72260b16aee30ab315d2d9df46c.jpg

IMG_5274.thumb.jpg.c13dd9a8ffa5c29f6fe4f2cf7dca58f4.jpg

IMG_5275.thumb.jpg.579a9d33437c61729b02fafd2eaeed79.jpgIMG_5273.thumb.jpg.61f5cb1bb37a53e0a902b303675c2375.jpg

 

Also, there is a coupon with serial number included, like mentioned in the video, fake cards don't have that.

 

IMG_5277.thumb.jpg.b285094fa4e8170fa5097c8637cc42dc.jpg

Link to comment
Share on other sites

7 minutes ago, flashjazzcat said:

If it wasn't apparent from my explanation of the bug in post 507, the PDM issue had nothing whatsoever to do with card age or speed.

If that card only had issues with PDM files, there's nothing strange about the fact PDM files now work.

Remark 1: yet, still the Kingston, which I expect to be the slowest, yields (stream) errors)

Remark 2: I didn't say it was strange, I found it funny because I set out to get (a) newer card(s) only in the hope it would make the PDM files play.

Link to comment
Share on other sites

1 minute ago, Level42 said:

yet, still the Kingston, which I expect to be the slowest, yields (stream) errors

Well, if you updated the loader, we might reasonably assume the Kingston card has some other issue not related to firmware bugs. Kingston manufactured the 'Flower' CF cards which wouldn't work for shit with anything.

4 minutes ago, Level42 said:

I didn't say it was strange, I found it funny because I set out to get (a) newer card(s) only in the hope it would make the PDM files play.

Do you want a refund, or what? I sat here and helped you with your problem for some days and thanks to this, we have corrected a problem quite efficiently, I think.

 

Also: can you please not spam the thread with stuff about how genuine your cards are. It wasn't me who suggested they were fake in the first place (I merely pointed out that Amazon had been known to sell fakes in the past, in response to your assumption that purchasing media from Amazon constituted a guarantee of genuine parts).

Link to comment
Share on other sites

More about the new Sandisk cards being original or fake:

 

The cards themselves. Labels are straight, are glued well etc. There is a gray (not yellow) write protection switch:IMG_5264.thumb.jpg.4e58f4698b0b103dfd2d8c023dfdda56.jpg

 

Now let's flip them over. Fake cards apparently usually do NOT have serial numbers printed on the back:
IMG_5269.thumb.jpg.72c983e46f4a61ebd42a98d54d164956.jpgIMG_5270.thumb.jpg.19dd02d41b2369722721c2fe5885b703.jpgIMG_5268.thumb.jpg.d72388817900101dfc01553a94b4aeb4.jpg

 

I do find it interesting that these serial numbers do not seem to have any relation with the serial numbers that Jon's SDtester tool indicates.

I also tried the tool Blackmagic Disk Speed Test with all three cards.

The write and read speeds never reach anywhere close to what the cards are capable of. However, I am sure this is because I use the built-in SD card reader of my ancient late-2009 iMac.

The older Kingston and SD cards were slower (especially the Kingston). The 2008 Sandisk actually had about the same read speeds as the three modern ones.


I'd have to buy an external card reader to get reliable results for this test, however, my iMac only has USB 2.0 ports, and I expect USB 3.0 would be needed to also get the results which could be expected for the modern cards.

 

Anyway, all in all I'm convinced the cards are REAL and I hope I now have convinced anyone else suggesting they might be fake here as well. (Mazzspeed......)

 

 

Link to comment
Share on other sites

17 minutes ago, flashjazzcat said:

Well, if you updated the loader, we might reasonably assume the Kingston card has some other issue not related to firmware bugs. Kingston manufactured the 'Flower' CF cards which wouldn't work for shit with anything.

Do you want a refund, or what? I sat here and helped you with your problem for some days and thanks to this, we have corrected a problem quite efficiently, I think.

 

Also: can you please not spam the thread with stuff about how genuine your cards are. It wasn't me who suggested they were fake in the first place (I merely pointed out that Amazon had been known to sell fakes in the past, in response to your assumption that purchasing media from Amazon constituted a guarantee of genuine parts).

I agree, I don't trust the Kingston either. It's the reason why I avoid the brand.

 

I don't really understand your remark about the refund. Must be some language barrier or misunderstanding my words.

"I sat here and helped you with your problem for some days and thanks to this, we have corrected a problem quite efficiently, I think."

First: I sat here and invested also a lot of hours of my free time and ordered three cards (which I hope I can return 2 of and get my money back) to help diagnose a problem with PDM playback. I was happy to do this and I am glad my (and your) persistence helped fixing a bug.

 

5 minutes ago, flashjazzcat said:

Oh dear... it's happening again.

 

Yup. Too late. This was a cross posting. We need to rule out possible causes. Done. Now I expect not a single word about fake SD cards anymore.

Edited by Level42
Link to comment
Share on other sites

24 minutes ago, Level42 said:

Anyway, all in all I'm convinced the cards are REAL and I hope I now have convinced anyone else suggesting they might be fake here as well. (Mazzspeed......)

Hangon.

 

I asked the question if they were fake. I never implied they were fake. As stated in another thread, stop taking the comment so personally.

 

Honestly? I don't really care if they're fake or not.

Edited by Mazzspeed
  • Like 1
Link to comment
Share on other sites

36 minutes ago, flashjazzcat said:

Kingston manufactured the 'Flower' CF cards which wouldn't work for shit with anything.

Interesting. I had one of those with 32GB, which was very large and expensive back in the time, it was my main workhorse in several digital SLR cameras. It retired gracefully adapted via a CF->IDE adaptor in an MP3 player. No problems at all. I like that Kingston gives longer warranty than most other brands.

  • Like 1
Link to comment
Share on other sites

8 hours ago, flashjazzcat said:

The 'Flower' cards were quite infamous; there are lots of references to them on the forum in SIDE/SIDE2 threads.

I've been using them with my Incognito with no problems.  But then, it is an 800.  =B>

  • Like 3
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...