New products for C64 and Amiga from Individual Computers

[anachronicus]

I've been on the fence about this product, but I have to admit....my old A500 would be pretty happy to sport one of these....and since I have 3 of Jens's products already, might as well make it an even 4...

[elpiloto]

Originally Posted by Schoenfeld

Hey, I want to sell an ACA500! Just contact me through the contact form on the website, and we'll sort out the rest there.

Jens

Do you want to sell another one jens?

[Schoenfeld]

Did the final step of expansion port verification today: The Delfina soundcard is probably the most piccy thing I have ever built for the clockport, and it works after removing the two small capacitors on the solderside of the Delfina (they fix an A1200 timing bug that's obviously not there on the ACA500).

The second picture shows where the file comes from: An auto-mounted, FAT-formatted CF card in the second card slot (left slot). When was the last time you've heard the Blues Brothers soundtrack?

Jens

[Photon]

Originally Posted by Schoenfeld

This transparent material is shock resistant, and although it's not as stiff as the aluminium that I used last week, it stands my highly scientific test of "put down the coffee mug accidently on the inserted card".

What do you think?

Jens

The transparent plexi should guide the cards vertically, but you'd still have to hit the CF card holders pretty straight horizontally since they're tight. Not quite as tight as PCMCIA but pretty tight.

I would remove the plastic (since it sticks out beyond the PCB edge) and put it in a case. I need to be able to bring it with me in a backpack and leave it unattended at demoparties. The latter should go for families with children as well

As a compromise you could have just nuts under the PCB and fasten (with the top screws) "half" a case instead of standoffs (open underneath) which reaches down to the table. Slots in this larger piece of plastic would be the guides for card insertion instead of the plexi.

I totally agree with having "indirect lighting" instead of flashing LEDs in your face, a very good thought. The wide Amiga LEDs are quite nice but if the LEDs are very bright they can be distracting, not fun at all.

[Calabazam]

Speaking of Delfina, you should take the opportunity to release final stable drivers and documentation. Everyone knows that board is a real nightmare.

[Schoenfeld]

Originally Posted by Calabazam

Speaking of Delfina, you should take the opportunity to release final stable drivers and documentation. Everyone knows that board is a real nightmare.

Delfina drivers are stable, and they're even open-source. Last bug removed was quite a few years ago - everything else was/is hardware-related. If you feel your card is not stable, add some good ground connection. It's a day-and-night difference in stability. The connection that you see on the photo is the worst you can make.

Baby-step success on hot-plugging the aux CF card ("secondary" or "left" card): A small addition made to the CPLD logic allows us to detect if the card supports hot-plug in IDE mode. Some cards do (even reliably), some cards don't. If you find a card that does, there's no need to restart the computer if you just want to exchange data with a PC.

Jens

[desiv]

Originally Posted by Schoenfeld

A small addition made to the CPLD logic allows us to detect if the card supports hot-plug in IDE mode. Some cards do (even reliably), some cards don't. If you find a card that does, there's no need to restart the computer if you just want to exchange data with a PC.

Wow!!
Nice! I didn't know that was an option with IDE/CF!!!!

desiv

[Schoenfeld]

Originally Posted by desiv

Wow!!
Nice! I didn't know that was an option with IDE/CF!!!!

Like I wrote earlier, it is not within the specification of the CF standard for a simple mechanical reason:

The pin that tells the card to go into IDE mode is a short one. However, it must be grounded on power-up, and the power pins are longer than the other pins in order to ensure the "first make, last break" contact. As a result, the specification does not cover the case "hot-plugging in IDE mode", because the card is already considered to be powered up when the "IDE mode"-pin makes contact. While it does not have contact, it's pulled up by a resistor within the CF card, so according to specification, any CF card must come up in memory-mapped mode if it is hot-plugged.

What we are now looking for is a type of card that violates the standard (in a positive way). Our software (mostly) does what it's supposed to, only small things to be polished. I will most probably have the high-brightness LEDs to experiment with later this week, and we'll probably be ready for a broader test next week.

If you want to be a tester, let me know what types of A500s you have, and how many different types of CF cards you have at home. Testers should be able to put 2-3 hours per week at least into testing the ACA500, and there should be a basic understanding of how to write bug reports that let us reproduce a possible problem. Please understand that I cannot give away boards for free. The best I can do is to offer free shipping and an extended return period of two months.

If you feel like you're up to the task and you're the kind of guy who can do st**pid-enough things to make our software crash, please contact me. Please refrain from becoming a betatester just for "having it first", like some X-Surf-100 customers did. There will be enough for everyone later on, and I really don't want to have to say no to a real tester, just because a collector pretended to be a tester.

Jens

[dJOS]

I just bagged a couple more a500's, I think I might put an aca500 on one just for Hdd installable adventure games.

[elpiloto]

Originally Posted by Schoenfeld

Like I wrote earlier, it is not within the specification of the CF standard for a simple mechanical reason:

The pin that tells the card to go into IDE mode is a short one. However, it must be grounded on power-up, and the power pins are longer than the other pins in order to ensure the "first make, last break" contact. As a result, the specification does not cover the case "hot-plugging in IDE mode", because the card is already considered to be powered up when the "IDE mode"-pin makes contact. While it does not have contact, it's pulled up by a resistor within the CF card, so according to specification, any CF card must come up in memory-mapped mode if it is hot-plugged.

What we are now looking for is a type of card that violates the standard (in a positive way). Our software (mostly) does what it's supposed to, only small things to be polished. I will most probably have the high-brightness LEDs to experiment with later this week, and we'll probably be ready for a broader test next week.

If you want to be a tester, let me know what types of A500s you have, and how many different types of CF cards you have at home. Testers should be able to put 2-3 hours per week at least into testing the ACA500, and there should be a basic understanding of how to write bug reports that let us reproduce a possible problem. Please understand that I cannot give away boards for free. The best I can do is to offer free shipping and an extended return period of two months.

If you feel like you're up to the task and you're the kind of guy who can do st**pid-enough things to make our software crash, please contact me. Please refrain from becoming a betatester just for "having it first", like some X-Surf-100 customers did. There will be enough for everyone later on, and I really don't want to have to say no to a real tester, just because a collector pretended to be a tester.

Jens

jens, i have some a500 and a500+ systems with differents configurations, and add ons, like indi ecs, hxc2001, 4 cf cards, and i can do spme tests on an a1000 too, i would expend more than 2 or 3 hours per week, ill expend 3 or 4 per day, please feel free to contact me if you are intrested

cheers

[Retrofan]

@elpiloto I doubt he can contact you, as long as he doesn't have the option for PM's here.

It has to be you to contact them using their page http://www.jschoenfeld.com/home/impressum.htm

Edit:

Oh, well he can see your email if you have, sorry, but anyway I think it has to be you.

Originally Posted by Schoenfeld

If you feel like you're up to the task and you're the kind of guy who can do st**pid-enough things to make our software crash, please contact me.
Jens

Although that doesn't seem the best way to recruit testers

[elpiloto]

Theres allways a way! Thank you retrofan.

[Schoenfeld]

Modern LEDs rule. I decided to use this colour combination, which is nicely visible even in bright daylight.

The red LED is cut down to about 65% of it's highest brightness, and the yellow LED is at 100%. This way, they both appear with about the same brightness. Different CF cards control the LEDs in a different fashion: Some pulse the LEDs shortly, so they appear more dim, and others keep it at full brightness until there's no more action for a certain period of time.

Hotplug works a lot better now. The CF standard gives CF card makers the freedom to switch to IDE mode if they detect a low-level on the mode pin for an extended peiod of time. A longer timeout has turned surprisingly many cards to hotplug-compatible!

There are drawbacks, though: Hotplug ONLY works in the left slot ("aux slot"), and it ONLY works with FAT-formatted cards. Since Amiga-formatted cards have an RDB and potentially more than one partition, it would make the mounter extremely complex. For data exchange with an Amiga, you might as well use a FAT-formatted card.

I've found one CF card that doesn't even support IDE mode. It's a "Hama"-branded 4G card (30MB/s speed rating) and the whole computer behaves funny if you insert it (no matter if hot-inserted or already in the slot when powered up). Seems to be shorting something - that's plain incompatible, and with the large number of low-cost cards that even work with hot-plug, I won't worry about this one.
All CF-cards that I have tried will be documented in the Wiki including a photo. I'm asking all testers to do the same: If you have a report about a card, please include a photo, so I can put that in the Wiki for others to make an easier buying decision.

In addition to that, all testers should have a nullmodem connection between the native A500-serial port and another machine. The current aux.scsi.device outputs debug messages on the serial port which will help us decide if a card is really incompatible, of if there's something we can do in software to make a certain card compatible. It's actually really interesting to watch the device driver negotiate with the card :-)

Jens

[FOL]

Originally Posted by Schoenfeld

Modern LEDs rule. I decided to use this colour combination, which is nicely visible even in bright daylight.

The red LED is cut down to about 65% of it's highest brightness, and the yellow LED is at 100%. This way, they both appear with about the same brightness. Different CF cards control the LEDs in a different fashion: Some pulse the LEDs shortly, so they appear more dim, and others keep it at full brightness until there's no more action for a certain period of time.

Hotplug works a lot better now. The CF standard gives CF card makers the freedom to switch to IDE mode if they detect a low-level on the mode pin for an extended peiod of time. A longer timeout has turned surprisingly many cards to hotplug-compatible!

There are drawbacks, though: Hotplug ONLY works in the left slot ("aux slot"), and it ONLY works with FAT-formatted cards. Since Amiga-formatted cards have an RDB and potentially more than one partition, it would make the mounter extremely complex. For data exchange with an Amiga, you might as well use a FAT-formatted card.

I've found one CF card that doesn't even support IDE mode. It's a "Hama"-branded 4G card (30MB/s speed rating) and the whole computer behaves funny if you insert it (no matter if hot-inserted or already in the slot when powered up). Seems to be shorting something - that's plain incompatible, and with the large number of low-cost cards that even work with hot-plug, I won't worry about this one.
All CF-cards that I have tried will be documented in the Wiki including a photo. I'm asking all testers to do the same: If you have a report about a card, please include a photo, so I can put that in the Wiki for others to make an easier buying decision.

In addition to that, all testers should have a nullmodem connection between the native A500-serial port and another machine. The current aux.scsi.device outputs debug messages on the serial port which will help us decide if a card is really incompatible, of if there's something we can do in software to make a certain card compatible. It's actually really interesting to watch the device driver negotiate with the card :-)

Jens

All I can say is "WOW". It looks like a nice bit of kit, .

[lukassid]

Hey Jens. Can we hope one day for Indivision ECS with RTG chip so it can display Picasso screen modes?

[elpiloto]

i saw in a german forum that its possible
[ Show youtube player ]

[lukassid]

Yes but you need better chip in Indy, this was just experiment. Great coding skills but too slow and unstable to use.

[elpiloto]

true

[Schoenfeld]

Originally Posted by lukassid

Yes but you need better chip in Indy, this was just experiment. Great coding skills but too slow and unstable to use.

Indivision ECS is a flickerfixer, not a graphics card. It is in the wrong place (on the wrong bus) for a faster framebuffer mode; this chip is great already: With the hardware we have on the unit, we can move more than 160MBytes per second, which is plenty for a 12-bit framebuffer. It's just that the ECS chip bus with it's low bandwidth of 3,5MBytes/s isn't good enough for a gfx card.

In other words: It is technically impossible to add a gfx card on the Denise socket. You need a bus that has a much higher performance, and that's not available in the A500 at all. You therefore need to design an accelerator with a fast local bus, which brings us back to the topic of the last few pages: ACA500.

The local expansion port of the ACA500 is still not fast enough for a gfx card (in theory it gets 7MB/second). It'll be barely enough for USB and/or networking.

Jens

[lukassid]

Thank you for the answer. Now I understand.
Hoping for official support for SuperPlus mode some day.