Rapidus Accelerator (rev.1c) for Atari XL/XE

[drac030]

5 hours ago, CommodoreDecker said:

have had fun problems when Rapidus starts in 65C816 mode in my 800XL --  after a delay, it goes into memory self-test and randomly fails ROM and RAM tests; each reboot is guaranteed to show different self-test failures

I assume you must have switched the ROMs because the default OS for 65C816 mode (Rapidus OS) does not contain self-test at all. But in any case, this indicates problems which communication between the CPU and the Atari motherboard. I would check the following:

 

1) contact between the Rapidus and the motherboard: you need precision sockets, the normal sockets which are in 800XL are too loose for the connectors Rapidus has.

 

2) the exact type of 6502 plays a role: some seem to change characteristics with temperature, e.g. all works stable, when the computer is cold, or the reverse, it is unstable when cold, then stabilizes (I am not sure what 6502 is used for, but I guess that it supplies Phi2 signal or something crucial like that). On mine when I insert the CPU from the (otherwise seemingly working) 800XL I have here for experiments, it stops working. This CPU is the "Mexico" type, 1987 production year.

 

3) check the connectivity of the 3-wire cable going from the motherboard to the Rapidus. When this one is not connected properly or pulled off, all sorts of weird things will happen.

 

In any case, on my website (in my sig) I have a PDF document which I update from time to time, about such issues and possible solutions (it is not complete yet, I am still working on this stuff).

 

If that above does not help, contact the vendor.

Edited March 30, 2021 by drac030

[CommodoreDecker]

16 hours ago, drac030 said:

I assume you must have switched the ROMs because the default OS for 65C816 mode (Rapidus OS) does not contain self-test at all. But in any case, this indicates problems which communication between the CPU and the Atari motherboard. I would check the following:

 

1) contact between the Rapidus and the motherboard: you need precision sockets, the normal sockets which are in 800XL are too loose for the connectors Rapidus has.

 

2) the exact type of 6502 plays a role: some seem to change characteristics with temperature, e.g. all works stable, when the computer is cold, or the reverse, it is unstable when cold, then stabilizes (I am not sure what 6502 is used for, but I guess that it supplies Phi2 signal or something crucial like that). On mine when I insert the CPU from the (otherwise seemingly working) 800XL I have here for experiments, it stops working. This CPU is the "Mexico" type, 1987 production year.

 

3) check the connectivity of the 3-wire cable going from the motherboard to the Rapidus. When this one is not connected properly or pulled off, all sorts of weird things will happen.

 

In any case, on my website (in my sig) I have a PDF document which I update from time to time, about such issues and possible solutions (it is not complete yet, I am still working on this stuff).

 

If that above does not help, contact the vendor.

Thanks!

 

I had set the system to accelerator mode, but it still flipped into self-test mode. (The ROM was removed from the mobo and put into the Rapidus expansion socket, of course.)  Your 3 points probably explain everything-- and I did have to power off the XL, sit there whistling the theme to "The Brady Bunch" to ensure nothing was residual in RAM after 1 minute 9 seconds, then powered up again and got different and inconsistent results. 

 

1) Yeah, something seemed a bit wonky and loose. I suspect that alone may be sufficient as the cause (if not for not being able to exist with RamboXL).  The ROM chip was the only one to fit solidly in its new socket. The Rapidus board's pins would not stay put and easily moved about.  Strange that the same type of socket was't being used.

 

2) Both had the same ID number, with -3 at the end (made in Taiwan). Which, from a search, is the better 6502 to have.  One was sub-labeled 8313 and the other 8323, year/mo date codes I'd presume. 

 

3) Just saw this one.  From one jiggle at one point (the ground wire came loose, yay)  I could tell that was a problem with the three-wire setup. Noting all the "fun" with crimping and then soldering the little things on, next time I'll replace the socket with wires directly. Continuity was correct since inverse/Reset brought up the menu accordingly, and was the one time I squealed with joy... I recall the middle wire got nicked loose and I saw the screen flicker and sieze up too. So that is a biggie. Will just use wires and solder directly when the time comes.

 

The vendor website states not to contact them since there are a lot of variables involved, including (paraphrased) fun with the soldering iron and possible mobo damage. Given their website also recommends installing Rapidus as a standalone and the unit without the Ultimate 1MB expansion or other goodies and I dived in to try it on has RamboXL on it anyway, I'll try again using a stock XL - by then I'll have a proper desoldering gun since my spare unit lacks sockets... which might be a good thing in retrospect as I'll solder in the Rapidus directly or find an appropriate socket. 

 

 

[+Stephen]

1 hour ago, CommodoreDecker said:

The vendor website states not to contact them since there are a lot of variables involved, including (paraphrased) fun with the soldering iron and possible mobo damage. Given their website also recommends installing Rapidus as a standalone and the unit without the Ultimate 1MB expansion or other goodies

This is such an uncool way to do business.  Why keep selling hardware that is known to be busted?  The fact they are issuing this disclaimer adds insult to injury which is then further worsened by the fact that no support is offered.  Yet they find it just fine to charge $200 USD for broken hardware.  Ponderous.  Pretty good hardware match for the known to be  un-shielded video cables too.  Someone needs to start maintaining a list of known shit that is being sold so people know what to avoid.

[drac030]

9 hours ago, CommodoreDecker said:

Your 3 points probably explain everything-- and I did have to power off the XL, sit there whistling the theme to "The Brady Bunch" to ensure nothing was residual in RAM after 1 minute 9 seconds, then powered up again and got different and inconsistent results.

Just a remark, these phenomena have hardly anything to do with RAM, the weird things happening during powerup and/or reset have their origin in conflicting ROMs. I was told that this conflict is sure to occur when the three wire cable is loose or not connected where it should at the other end.

 

9 hours ago, CommodoreDecker said:

(if not for not being able to exist with RamboXL).

Mine works with a Rambo-type RAM extension (to 320k). But in fact there may be as many Rambo RAM extensions as there are many hardware guys constructing them, and people sometimes seem to have weird ideas on how do design such thing. So I already have seen a 130XE which, after it was expanded to 1 MB, stopped to cooperate with PBI HDD (IDE+). The reason being the phi2 signal being screwed/flaky by the RAM extension having been badly designed and without paying attention to possible ill effect on the PBI bus. So, IMHO it is best to start from a stock computer, then first eliminate all possible installation issues (like loose contacts etc.)

 

9 hours ago, CommodoreDecker said:

Both had the same ID number, with -3 at the end (made in Taiwan). Which, from a search, is the better 6502 to have.  One was sub-labeled 8313 and the other 8323, year/mo date codes I'd presume. 

Yup, besides, as I said, a "bad" 6502 is rather responsible for various effects depending on (its) temperature. I once had a stock 130XE which ran Acid800 fine or crashed while the program was testing 6502 illegal instructions, depending on the temperature, so Also a Mexico CPU, IIRC.