Side3 issues....

[w1k]

wow, version 073

[flashjazzcat]

23 minutes ago, w1k said:

wow, version 073

Yes. Because I'm writing it, I get to try the latest developmental versions before anyone else.

 

[Beeblebrox]

1 hour ago, flashjazzcat said:

Yes. Because I'm writing it, I get to try the latest developmental versions before anyone else.

 

You must be best mates with that Jon fella then?!  😉😁

[Booterror1973]

On 6/18/2024 at 2:27 PM, flashjazzcat said:

Can you tell me what the JED version and hardware revision are, as per this screen?

Here the versions of SIDE3 and U1MB

[flashjazzcat]

12 minutes ago, Booterror1973 said:

Here the versions of SIDE3 and U1MB

Headings are useful, but thanks.  Imagine my shock that your SIDE 3.1 (technically 3.2 if it has a purple PCB) is equipped with the special JED version 2.3 designed specifically for the GoWin CPLDs. I've had a lot of trouble with that JED here, and am (still) waiting for a GoWin platform cable ordered on my behalf by Lotharek in order to downgrade the JED (to version 1.3, which I have found to be reliable) on a customer SIDE 3.2 which I cannot send back to said customer with his machine since it doesn't work even when the XE has 74F08 fitted. The cartridge doesn't work in my rock-solid U1MB 600XL either, and - oh joy - the GoWin lacks the 'wishbone' interface and is therefore not flashable in software.

 

So please follow Beeblebrox's advice and try 74F08, and if that doesn't fix the issues, please get back to me so we can try and figure out a solution.

Edited June 19 by flashjazzcat

[Booterror1973]

31 minutes ago, flashjazzcat said:

Headings are useful, but thanks.  Imagine my shock that your SIDE 3.1 (technically 3.2 if it has a purple PCB) is equipped with the special JED version 2.3 designed specifically for the GoWin CPLDs. I've had a lot of trouble with that JED here, and am (still) waiting for a GoWin platform cable ordered on my behalf by Lotharek in order to downgrade the JED (to version 1.3, which I have found to be reliable) on a customer SIDE 3.2 which I cannot send back to said customer with his machine since it doesn't work even when the XE has 74F08 fitted. The cartridge doesn't work in my rock-solid U1MB 600XL either, and - oh joy - the GoWin lacks the 'wishbone' interface and is therefore not flashable in software.

 

So please follow Beeblebrox's advice and try 74F08, and if that doesn't fix the issues, please get back to me so we can try and figure out a solution.

I just received the SIDE3 from Lotharek around last week, read something on his site about Cpl , but thought I got straight a good one. So the version I got is not good?

i am still busy with the VBXE so exchanged quick the original TexasInstruments 74LS08 for the Malaysian 600XL version, looks bit more stable in the few minutes I tried. 

[flashjazzcat]

41 minutes ago, Booterror1973 said:

I just received the SIDE3 from Lotharek around last week, read something on his site about Cpl , but thought I got straight a good one. So the version I got is not good?

As far as I can tell, pretty much all SIDE3 stock from all official vendors is now equipped with GoWin CPLDs and JED 2.3. As to whether it's good or not: I've had a SIDE 3 with a GoWin CPLD and JED 1.3 for a year or so and it's absolutely fine; it's just the customer unit with JED 2.3 which has been a dead loss, hence the need to downgrade the JED (and I already tested my cart in the customer's machine, an it works perfectly, but - perhaps understandably - I'm not keen to give my cart to the customer and be left with something that doesn't work in any machine I have to hand).

 

Feedback from other users with the same CPLD/JED combo has been mixed, ranging from 'works fine' to 'it's a mixed bag', all the way to 'doesn't seem to work'. Since SIDE3 seems to require 74F08 in the host machine at least half the time, that's not necessarily unusual, but I do find it concerning that this particular SIDE3/JED combination doesn't work in the machines with which I would normally have tested any pre-release JED, had I been provided with the opportunity to do so (instead, it seems the JED was coded up earlier this year and flashed to all production carts in order to 'tweak timing', without my even being invited to try it out first).

 

I must admit to being explosively annoyed about this situation when I first encountered it over a month ago now, and continually irritated by the fact the customer's equipment has been sitting here ever since, awaiting the GoWin platform cable which is presumably lost somewhere in the postal system.

Edited June 19 by flashjazzcat

[snicklin]

Was there ever any documentation produced about utilising the extra onboard RAM?

[flashjazzcat]

3 hours ago, snicklin said:

Was there ever any documentation produced about utilising the extra onboard RAM?

No, but I'll be glad to produce something if it's of interest.

[snicklin]

On 6/22/2024 at 9:26 AM, flashjazzcat said:

No, but I'll be glad to produce something if it's of interest.

That would be great. I got very excited at the thought of 2MB of RAM and bought the Side3 because of that. It blows my mind at what the Atari could do with that as RAM. I am sure that many others would love to be able to use that memory for their coding.

[flashjazzcat]

23 minutes ago, snicklin said:

That would be great. I got very excited at the thought of 2MB of RAM and bought the Side3 because of that. It blows my mind at what the Atari could do with that as RAM. I am sure that many others would love to be able to use that memory for their coding.

The newer revisions have 8MB of SRAM, although the loader doesn't yet exploit more than the base 2MB. The most exciting thing about the SIDE3's SRAM is that the DMA engine can fill it with data from the SD card at 1.7MB/s, which is exactly how cartridge images are loaded. Unfortunately DMA writes (which would be rather useful now that the loader is capable of writing to the FAT file system) still don't work despite my requesting this functionality from Candle more than two years ago. The feature was implemented but does not work for reasons we could not figure out, and fixing it appears to have fallen off Candle's list of priorities (I asked a while ago if the problematic DMA write logic is even present in the latest production JEDs, and Candle was unable to recall and apparently unable to refer to the source code in order to have a look).

 

One problem with using the cart windows ($8000-9FFF/$A000-BFFF) as an access window for RAM is that unless the cartridge ROM is switched on in the same address space, SRAM access tends to cause display glitches unless the screen RAM is carefully relocated out of the way first. This limitation would go away with the External U1MB, etc, if such hardware ever sees the light of day (which seems an increasingly remote possibility given that two years have now elapsed since I received some prototype boards and the only person who periodically has any sense of urgency about progressing these products is the vendor who wants to sell them), since SRAM would then supplant the entire base memory of the host machine and provide all extended PORTB RAM. One could then read the SD card via DMA using any buffer address one pleases (it would be nice to write via DMA as well, but - as with SIDE3 - the likelihood of that ever being possible looks similarly remote).

 

What with firmware license payments being forthcoming later and later every month, the apparently regressive JED 2.3 update for the GoWin SIDE3 carts (which has obliged a customer machine to sit here for a total of five weeks now since I completed the work on it while I wait for a GoWin platform cable to rescue the client's SIDE3 by reverting it to the working JED 1.3, at which point I can hopefully charge money for a working system), the lack of progress on derivative hardware which would rely on ported versions of the firmware I have already written, DMA writes apparently being beyond reach, and persistent stability issues which require tweaking of the host machine and which would presumably still persist on the derivative 'plug and play' hardware (External U1MB, etc), wading through 47,000 lines of SIDE3 Loader source code (as I have been doing over the past few days) in order to undertake significant code refactoring, implement video playback capability, and add lots more esoteric functionality certainly seems like a thankless and potentially pointless undertaking.

Edited June 23 by flashjazzcat

[flashjazzcat]

On 6/21/2024 at 9:07 PM, snicklin said:

Was there ever any documentation produced about utilising the extra onboard RAM?

I just remembered Phaeron's documentation of the SIDE3 hardware, and am pleased to see this has found its way into the Altirra Hardware Reference Manual. The SIDE3 section starting on page 182 includes all the information you could need.

[Booterror1973]

On 6/19/2024 at 9:17 PM, flashjazzcat said:

As far as I can tell, pretty much all SIDE3 stock from all official vendors is now equipped with GoWin CPLDs and JED 2.3. As to whether it's good or not: I've had a SIDE 3 with a GoWin CPLD and JED 1.3 for a year or so and it's absolutely fine; it's just the customer unit with JED 2.3 which has been a dead loss, hence the need to downgrade the JED (and I already tested my cart in the customer's machine, an it works perfectly, but - perhaps understandably - I'm not keen to give my cart to the customer and be left with something that doesn't work in any machine I have to hand).

 

Feedback from other users with the same CPLD/JED combo has been mixed, ranging from 'works fine' to 'it's a mixed bag', all the way to 'doesn't seem to work'. Since SIDE3 seems to require 74F08 in the host machine at least half the time, that's not necessarily unusual, but I do find it concerning that this particular SIDE3/JED combination doesn't work in the machines with which I would normally have tested any pre-release JED, had I been provided with the opportunity to do so (instead, it seems the JED was coded up earlier this year and flashed to all production carts in order to 'tweak timing', without my even being invited to try it out first).

 

I must admit to being explosively annoyed about this situation when I first encountered it over a month ago now, and continually irritated by the fact the customer's equipment has been sitting here ever since, awaiting the GoWin platform cable which is presumably lost somewhere in the postal system.

Trying to find a 74F08 in The Netherlands....to no avail yet.

[flashjazzcat]

5 minutes ago, Booterror1973 said:

Trying to find a 74F08 in The Netherlands....to no avail yet.

Really? If you get stuck, send me your address and I'll post you one (although I'm starting to think the vendor should put a 74F08 in the box with SIDE3 just in case, although they'd probably prefer you to purchase a proprietary 'fixer' instead, despite the fact it's complete overkill for the problem at hand and does absolutely nothing whatsoever that the $1 chip can't accomplish in this context).

 

[PE1PQX]

39 minutes ago, Booterror1973 said:

Trying to find a 74F08 in The Netherlands....to no avail yet.

Contact https://sinuss.nl/ you might succeed.

[w1k]

i have problem with demo rewind 2..

side 3.1, jed 1.3.. PBI OFF/SDX OFF.. when i turn on it, nothing change.. tested with 1mb, latest altirra OS, or original XE/XL OS

 

 

[tep392]

@flashjazzcat  I just finished watching your new video and I am very curious what was the exact purpose of the 2.3 firmware and removal of C30.  Did Lotharek ever elaborate on what the specific issue was that these changes were supposed to fix?

[Peri Noid]

I think, the designer is the one to be asked. 

[flashjazzcat]

6 hours ago, tep392 said:

@flashjazzcat  I just finished watching your new video and I am very curious what was the exact purpose of the 2.3 firmware and removal of C30.  Did Lotharek ever elaborate on what the specific issue was that these changes were supposed to fix?

He did. Lotharek told me on Wednesday evening during a two hour conversation and immediately before publishing his video that he commissioned JED 2.3 after testing a SIDE 3.2 cartridge with JED 1.3 after receiving a customer complaint and finding that 'it didn't work for shit' on his test machine(s), which was/were very likely equipped with O2 Fixer and/or 64K SRAM. This happened some months after he sent me my pre-production SIDE 3.2 which he presumably 'assumed to work' and which has given me no issues. He also claims that the designer - Candle O'Sin - mandated that C30 be 'no stuff'. So one can only conclude that JED 2.3's purpose was to correct the perceived issue of the cost-reduced GoWin CPLD-equipped SIDE 3.2 not working for shit on the vendor's testbed, regardless of how effectively it might function on anyone else's Atari with the updated JED, given that the revised JED was not submitted to anyone else for pre-release testing.

Edited July 20 by flashjazzcat

[flashjazzcat]

I will add that Lotharek (in addition to withdrawing sale of SIDE 3.2 pending a fix, which actually seems like responsible action) - without notifying me via any means other than the nebulous threats outlined in his video - has withdrawn sale of ALL FJC firmware products, including those not blighted by hardware revision-related issues (U1MB, incognito, and SIDE3). This act can be motivated by nothing but the desire to enact punitive material punishment.

 

I sincerely hope this decision does not adversely affect sales of the aforementioned devices, which I have championed, promoted, advertised and recommended for over a decade. Not to mention the added value and revenue generated by the firmware I spent years developing.

 

Needless to say this development ostensibly signals - in the absence of any indication to the contrary - the termination of my 'official' involvement in the development of firmware for the aforementioned products.

 

Sincere thanks to all users of my firmware over the past fourteen or so years for your support and encouragement. I enjoyed every minute spent adding value to Candle's wonderful hardware.

Edited July 20 by flashjazzcat

[flashjazzcat]

Correction: he forgot to withdraw FJC SIDE2 (which I mistyped as SIDE3 in the prior post), presumably because no-one buys it now and it makes no money.

Edited July 20 by flashjazzcat

[Mazzspeed]

1 hour ago, flashjazzcat said:

I will add that Lotharek (in addition to withdrawing sale of SIDE 3.2 pending a fix, which actually seems like responsible action) - without notifying me via any means other than the nebulous threats outlined in his video - has withdrawn sale of ALL FJC firmware products, including those not blighted by hardware revision-related issues (U1MB, incognito, and SIDE3). This act can be motivated by nothing but the desire to enact punitive material punishment.

 

I sincerely hope this decision does not adversely affect sales of the aforementioned devices, which I have championed, promoted, advertised and recommended for over a decade. Not to mention the added value and revenue generated by the firmware I spent years developing.

 

Needless to say this development ostensibly signals - in the absence of any indication to the contrary - the termination of my 'official' involvement in the development of firmware for the aforementioned products.

 

Sincere thanks to all users of my firmware over the past fourteen or so years for your support and encouragement. I enjoyed every minute spent adding value to Candle's wonderful hardware.

May I ask what Lotharek video you're referring to FJC?

[+Stephen]

1 hour ago, flashjazzcat said:

I will add that Lotharek (in addition to withdrawing sale of SIDE 3.2 pending a fix, which actually seems like responsible action) - without notifying me via any means other than the nebulous threats outlined in his video - has withdrawn sale of ALL FJC firmware products, including those not blighted by hardware revision-related issues (U1MB, incognito, and SIDE3). This act can be motivated by nothing but the desire to enact punitive material punishment.

 

I sincerely hope this decision does not adversely affect sales of the aforementioned devices, which I have championed, promoted, advertised and recommended for over a decade. Not to mention the added value and revenue generated by the firmware I spent years developing.

 

Needless to say this development ostensibly signals - in the absence of any indication to the contrary - the termination of my 'official' involvement in the development of firmware for the aforementioned products.

 

Sincere thanks to all users of my firmware over the past fourteen or so years for your support and encouragement. I enjoyed every minute spent adding value to Candle's wonderful hardware.

Well then - I will make it 100% clear, that Lotharek can go fuck himself.  I'll never spend another penny at his store, and if updates he sells are only available there, I will go without them.

[Mazzspeed]

4 minutes ago, Stephen said:

Well then - I will make it 100% clear, that Lotharek can go fuck himself.  I'll never spend another penny at his store, and if updates he sells are only available there, I will go without them.

This 100%.

[flashjazzcat]

For clarity, I should point out that Incognito - both stock and FJC versions - has been replaced by 'Incognito Gold' which does have my firmware pre-installed, but which is limited to the final run of fifty units. So when they've sold, and since no-one is now buying SIDE2, that will be the end of my firmware revenue from Lotharek's shop (unless SIDE3 ever sees the light of day again). It's unclear to me whether FJC U1MB will be replaced by the 'revised' purple boards. 

 

I see RetroLemon hasn't withdrawn anything yet, although Jakub is away the the moment.

 

I infer all of this purely from inspection of Lotharek's website, having (as usual) been informed of none of the changes. On reflection I think it would be wise to await clarification of the long-term situation before drawing any conclusions based solely on the 'further actions' hinted at in Lotharek's 'SIDE3 situation' video, and I apologise for doing so.

Edited July 20 by flashjazzcat