ACA500 tested

[dJOS]

Originally Posted by Amicol

I received mine last week, and I couldn't be happier!
Success with a CF card is dependant on the actual A500 too as I have cards that work on my A500+, but do not work on my standard A500 (rev 6a) using the same ACA500. However, I have a 16Gb card that works flawlessly on both :-)

That makes no sense at all, if a cf card works 100% on an a500+ it should also work on an a500 just as well.

Jens are you able to provide insight on this aca500 behaviour?

[Loedown]

Originally Posted by Schoenfeld

Jacques,

As stated in the manual, the ACA500 is a low-budget product where tiny flaws are expected. We indeed have a set of a few CF cards that don't work at all and others that cause trouble when re-booting (either with or without early startup menu), and we can't make those work, mainly because there's no development budget left to support those cards.

Please turn to your reseller for a full refund of the ACA500, as you're obviously not satisfied with the product. We've done our best, but there's tight limits to what we can do for 79,90 EUR.

Jens

Two points to make.

Firstly, this has been covered in another thread in regard to timing issues, no two Amigas are ever going to respond exactly the same way and in exactly the same time, as the components age they cause more and more issues. I use the example of a Roland TR808 drum machine, some units are more sought after because they have a sound which is more desirable ( sharper snare or better kick ) and this is because the components have aged in a way that is more pleasing to the ear of most people. Same with Amiga, some units will start to have tiny little flaws that may not be even discernible under most circumstances but in the case of accelerators where timing is CRITICAL then they can start to show flaws.

Second, Discussion on issues is a healthy thing but it needs to be done without any emotional content and if misinformation is being spread then it's an absolute right of the creator of the product to keep his or her name clean. I think the better way to have approached this is to say is anyone else having this sort of issue and if so what have they tried and what success have they had. I do like Jens' gear but I will also say that some products need more time assigned to them.

Originally Posted by dJOS

That makes no sense at all, if a cf card works 100% on an a500+ it should also work on an a500 just as well.

Jens are you able to provide insight on this aca500 behaviour?

CF card spec is probably as loose as everything else is, not necessarily the fault of the manufacturer using these third party cards.

[Schoenfeld]

No emotions involved in my proposal to Jacques - we've put time and money into solving this issue, but couldn't solve it for him. The best I can do now is to offer a refund. There's supposedly two other solutions for the A500 out there, and they might work better for him.

dJOS:
The CF card slots on the ACA500 are unbuffered. The "unbuffered" property explains why some CF cards keep the ACA500 from starting at all. The bus towards the A500 has data bus buffers, but no address bus buffers. If the A500 has some non-standard expansions (kickstart switcher, memory expansion?) installed, the address bus may have unusually high load and the 68EC000 can't drive it any more. This may be different if you attach an A1200 accelerator, because the address bus will then be driven by the CPU on the A1200 accelerator.

Jens

[Leffmann]

An adapter priced at €80 is hardly a low-budget product. To me it looks like the ACA500 simply wasn't tested enough and just rushed out the door, just like some of the €180 ACA12xx products that were discovered to not work on a common revision of A1200 motherboards. Is €180 also low-budget? If so, at what price would customers get premium products that are fully tested?

[Apollo]

IMHO the ACA-500 compared to other accelerator cards is a low budget product. Even more if you comprehend the background in which these devices are designed and produced.

[nikos.rizos]

Look... Jens was crystal clear

As stated in the manual, the ACA500 is a low-budget product where tiny flaws are expected. We indeed have a set of a few CF cards that don't work at all and others that cause trouble when re-booting (either with or without early startup menu), and we can't make those work, mainly because there's no development budget left to support those cards.

Next time before you buy something ask how much the development budget is going to be and how long this budget will last to support those cards.

The ACA500 was released only a couple of months ago, some time in December 2013 with a 0.xx flash version.
Right now after only a couple of months the flash version is 0.99 and there's no development budget left to support those cards.

Grab a copy of the 0.99 flash and get one of the working CF cards

[Jacques]

Whatever level the price is set at (ACA500's is fair, but not low by any mean), it shouldn't be mentioned as an excuse for lack of compatibility with otherwise fully functional Amigas and CF cards (in my case A500/500+, both with 8372/8375 ECS Agnus and genuine Sandisk, Pretec, Canon CF cards. It's the very basic functionality of Amiga that fails for me (mapped Kick 3.1's Early Menu boot problems, software reboot). I'm not the only one to experience this issue (2 other persons in Poland known, one of them being RetroAmi testing some cards, apart from mine which I sent back for tetsing), so there is some sort of popular problem. Yes, it's life, issues happen, but apparently it wasn't tested enough ahead of going out to the customers, as Leffmann said.

As for returning the card and refund, with the amount of time I spent on testing, preparing feedback e-mails and videos, I'd rather have those issues fixed (if possible), than to simply put my time into trashcan. Especially that IF WORKING CORRECTLY, ACA500 is really nice piece of hardware I'd like to have. It just happened that I really need to occasionally change the partition I'd like to boot (Early Menu) and use software reboot, so I probably was first to report the problem and maybe one of the few to use this functionality.

What I understood from Jens post is: "Issue unfixable for the price card is offered, since no more time can be spend on ACA500", doesn't necessarily sound like a proper customer support. It can't be that hard to get e.g. Sandisk Ultra II 4GB, Transcend 133x 4GB or even Canon 16/32MB cards to emulate the problem on some Amiga 500's and ACA500 cards, since RetroAmi reseller already happened to find out more than "mine" ACA500's with the same issues when tested on his Amiga's. And hey, forget trapdoor expansions causing the problems, my both A500 and A500+ without any expansion were behaving the same way in this regard.
And issue was miraculously gone, when I used hardware Kickstart 2.0 from A500+ instead of mapped 3.1 - then I could choose the partition from its Early Menu and attempt to boot without the freeze. But it's not the machine I plan to use, nor is Kickstart 2.0.

To sum it up. I really would like to wait for the actual fix and am hopeful for that. But IF it's impossible, will I get a refund of all the shipping costs I spent (that would be 3 times of sending already)?

But still, I'd really like to get ACA500 to work as expected regarding this very basic boot/reboot functionality and having a nice happy end. Amiga 500 deserves it

[TCD]

Originally Posted by Jacques

As for returning the card and refund, with the amount of time I spent on testing, preparing feedback e-mails and videos, I'd rather have those issues fixed (if possible), than to simply put my time into trashcan.

I think it's very clear now that you only have two options: Keep the card *as is* or to send it back and get a refund. Jens posted that there won't be any improvement anymore as it simply would mean a loss for him. No amount of 'persuasion' here on EAB will change that I'm afraid.

I don't mean to support either side, but it would be sad to see this thread get into an endless loop of requests and denies...

[Toni Wilen]

CF card incompatibilities are never fun..

So because I find this interesting (is this hardware problem or software or actual CF card bug that most hardware won't care)

Did any card work, 100% correctly?
Did some card work randomly?
Does some card always fail to work?

To confirm it is not contents of card, always boot from floppy. Make empty bootable floppy disk (format + install), copy software reset program to disk, boot from disk, run the reset program.

Use boot menu F1 option only:

Test with all cards.
Test without card installed. (Debugging always also includes listing cases that work, not just those that does not work!)
Power cycle after each test

Then enter expert menu and create config that has 3.1 rom, map rom, all ram options enabled. IDE driver update not enabled. Save it.

Do above tests again. Any difference?

And finally enable IDE driver (do not enable aux slot). Save and power cycle and so on..

Do above tests again. Still no difference?

And finally final one: if you have physical 3.1 ROM, select "on-board" ROM option, do not enable map rom.

And tests again. Differences?

Yes, very boring but thats how any kind of debugging works, sorry

[amiga1260]

Did you test with a beefier PSU?

[Jacques]

Hi Toni, thank you for interest and test-scenario request.

I'll describe my test:

Amigas: A500 (8372 ECS Agnus) & A500+ (8375 ECS Agnus), both tested with 0.5 MB mainboard CHIP RAM produce same behaviour, also tested with 0.5 MB slow under trapdoor and it haven't changed results.


1) Canon 16 MB & Canon 32 MB
- mapped Kick 3.1's early menu BOOT - FAIL (grey screen freeze)
- mapped Kick 3.1's early menu BOOT WITH NO S-S - SUCCESS
- sw reboot forced by HdToolBox run from floppy or ClassicWB's reboot: FAIL (grey screen freeze)

2) Sandisk Ultra II 4GB and Transcend 133x 4GB
- mapped Kick 3.1's early menu BOOT - FAIL (grey screen)
- mapped Kick 3.1's early menu BOOT WITH NO S-S - - FAIL (grey screen)
- sw reboot forced by HdToolBox run from floppy or ClassicWB's reboot: FAIL (grey screen freeze)

1) & 2) testcases are identical in both ACA500's IDE Driver settings.

I don't have hardware Kickstart 3.1 to test it with, but hardware Kickstart 2.0's early menu of A500+ allowed to successfully choose and boot selected partition even with 4GB card (no grey screen, just proper boot).

What's even more interesting, mapped 3.1's early menu BOOT, freezes Amiga even if DF0 is selected as device to be booted.
With no cards at all in any ACA500's slot, mapped Kick 3.1's early menu boot (DF0) works fine (no freeze).

best regards.

[fryguy]

I used an elcheapo Verbatim CF-card, check attached file. Had no issues.
I used an A500+ though.

[kipper2k]

Some of my CF cards which work fine with my IDE board do not work with the ACA 500 regardless of settings. I have tried 2 different ACA 500 boards (and a few Amigas. All the ones that fail in one A500 fail in them all

[Toni Wilen]

"Do not work" ("It does not work" explanation is the main reason why there is always confusion) = do you mean it does not boot at all (not even to boot menu)?

I think this is totally different problem than hanging later (after reset or similar). I have Transcend 8GB 133x that does above (no boot at all).

btw, Jacques, check your PM.

[kipper2k]

Originally Posted by Toni Wilen

"Do not work" ("It does not work" explanation is the main reason why there is always confusion) = do you mean it does not boot at all (not even to boot menu)?

I think this is totally different problem than hanging later (after reset or similar). I have Transcend 8GB 133x that does above (no boot at all).

btw, Jacques, check your PM.

do not boot, no boot menu

[Jacques]

Hi Toni and thank you.

Please check your PM.

best regards.

[dJOS]

Originally Posted by kipper2k

All the ones that fail in one A500 fail in them all

And that makes perfect sense - what Jaques is reporting is quite bizarre, slight variations in timing between various a500's (rev5-8) shouldn't effect CF card compatibility unless there is some strange flaw in the aca500 that Jens is yet to discover.

[Amicol]

Originally Posted by dJOS

And that makes perfect sense - what Jaques is reporting is quite bizarre, slight variations in timing between various a500's (rev5-8) shouldn't effect CF card compatibility unless there is some strange flaw in the aca500 that Jens is yet to discover.

I guarantee you there are differences, see my previous post....

[dJOS]

Originally Posted by Amicol

I guarantee you there are differences, see my previous post....

If timing was such an issue, both Kipper and Mika would be seeing similar issues with their IDE68k based products but they aren't.

[Schoenfeld]

dJOS:
You're drawing conclusions from no information (do you have all schematics and timing specs of the boards you're comparing?), and you're not seeing the information I have provided yesterday: An unbuffered CF card slot does affect timings by changed signal load on shared lines. If you are only half as educated as you're attemting to appear, this will give all the answers you're expecting.

Jacques,
you're claiming that there are more people that have this problem. They have not reported this to their reseller, and they haven't reported it to me. At this point, your claim is a lie. Have these people use the recommended support chain. The grapevine is not the recommended way to get product support, and claims of other people's problems are not the way to make your own problem higher-priority. Stick to the facts.

Toni has done more tests, and it appears to be a reset issue: Different A500 models have different values for the reset-pullup. Try adding a 2k2 resistor between Vcc and Reset.

Jens