RAPIDUS ACCELERATOR

[flashjazzcat]

FJC did the resister Array fix the problem???

 

Sadly not Steve. But Lotharek has very kindly offered to post out a replacement board, and I am to post the suspect unit back to him in return. Then we should be back in business. Hopefully the replacement board will behave the same when warm as when cold, and if so, any stability problems I encounter may be blamed on the host machine. Both the 130XE and 1200XL work OK for the first five minutes with the problem board, however, so with luck I'll have the new board running in the 1200XL soon.

[Stephen J. Carden]

FJC did the resister Array fix the problem???

 

Thanks Steve

Hi FLC!

I can say this about Lotharek everything I have ever purchased from him he has backed up 100%. I had a Side 2 Cart I purchased from him and all the documentation he has provided with it he warns about slamming a Compact Flash chip into the socket. Well I did just that! I pushed one of the pins in the socket out of the socket. This was 100% my FAULT and not a design flaw. I sent him an e-mail and a picture of the pin I had pushed out of the socket and he verified that it was a pin not used by the Side 2 or the Compact Flash Chip. Well the Side 2 Was working and was also working with the Ultimate 1 Meg as well. Two weeks later Lotharek SENT ME TWO REPLACEMENT Side 2 Cart and I also got an e-mail from him saying he was sorry that happened to his hardware. There again this was something that was 100% my F**K up not his. Over the last 2 1/2 years I have purchased over $2000.00 US Dollars worth of Upgrade hardware from him. It always comes on time 10 to 14 Days from Purchase. It always comes in a good Sturdy Box and is packed the way you want all your electronic boards to be packed. The Thing I like the most about Lotharek is he stands behind his hardware. He sent me a PC version of a SIO2SD that is real neat, and it works well on the PC. I know I am laying on thick for Lotharek but he has earned it. For the Most part Atari Users like myself are very luck in that we get really nicely designed and Built Hardware. I can not think of anyone selling something to the Atari World that it could be referred to as a junk! It is like people Like Lotharek and Many others that have set that very high Standard that all of us enjoy.

 

Take Care All!

Stephen J. Carden

http://www.realdos.net

[flashjazzcat]

Yes: his stuff is pure quality, and he appears receptive to improvements (such as the 100pf cap suggested by Hias which cures U1MB reset issues on a host of machines). He's also sent me a lot of stuff in return for the firmware and tools I've developed which run on his products, and he's going out of his way to ensure my Rapidus experience is positive. One could hardly ask for more.

[Kyle22]

He sent me a new set of ribbon cables and socket adapters when I accidentally broke a pin.

 

Thanks again Lotharek.

[JoSch]

Has anyone used the Rapidus board on their XEGS?

Does anyone know how much of Antic is covered?

I'm asking because I installed a VBXE board into my XEGS and RAPIDUS will collide with it.

A possible solution I'm thinking of, would be to mount VBXE or RAPIDUS via the socket adapters Ultimate 1MB uses. (How are this adapters called?)

Which problems would this approach have?

[flashjazzcat]

There ought to be sufficient clearance in an XEGS for a stack of 40 pin sockets under one or the other upgrade.

[Mclaneinc]

Sorry to hear the problems have not been sorted as yet Jon...

 

Here's hoping the new board solves all that..

 

Paul.

[danwinslow]

Well, here's a noob question for comic relief -

What pitch is the little 3 pin connector that Lotharek did not supply?

Is that a dupont 2.54 mm pitch?

[flashjazzcat]

It's 2.54mm pitch, but not Dupont. Not sure what they call it, but I replaced it with a right angle pin header (that's how I know the pitch is standard). It's the same connector used as a power jack on the IDE Plus 2.0.

Edited May 18, 2016 by flashjazzcat

[danwinslow]

Thanks, Jon. Unfortunately, that doesn't really help me much, as I've no cabling at all. If anyone happens to know the terminology for the connector, such that I could order one, let me know. Thanks.

[flashjazzcat]

Best person to ask is Lotharek himself, I guess.

[doctor_x]

Am I correct in thinking that Rapidus isn't going to do a thing for me with legacy software and that only stuff written to run under the on board processor (cant remember the name now) will be able to utilize the speed the board provides?

 

Sounds like u1m, side2 and VBXE are still the magic three for me unless I'm wrong about the above..

[Dmitry]

DoctorX, code doesn't have to use 816 instructions in order to be accelerated at a faster clock. Whether you want your software to in at 20mhz, is another question. Certainly sometimes you do.... A compiler compiling faster is always nice. The board has a spectrum emulator, so that's a lot of software opened up. It has flag on board, and one example use of that, it was simulating a second pokey, for software written to take advantage of that. Stuff like that. I own one, but not installed yet.

[Dmitry]

Thanks autocorrect. Post is unreadable, lol..... Where to start...board has an fpga on board...

[Dmitry]

Thanks autocorrect. Post is unreadable, lol..... Where to start...board has an fpga on board...

[flashjazzcat]

For those still following this saga: replacement Rapidus arrived courtesy of Lotharek on Thursday (for which many thanks, and yes I will eventually send the other one back), but results weren't immediately more encouraging than with the first board. I can't yet tell if Impact works reliably with this board because there's no 10 pin JTAG header soldered on (and I've run out, having soldered them all to people's Ultimate 1MB boards, so I ordered more), but it does at least seem that the new board behaves consistently regardless of running temperature. By consistent, I mean that if the system is unstable when the board is warm, it was also unstable when it was cold.

 

Anyway: four hours tinkering last night (moving the board from the 1200XL to the 130XE and back again) didn't yield any positives (the old reboot on reset issue was still there, and some new symptoms like losing video completely, locking up on power-up, etc, although all this was dependent on the choice of CPU - chosen from a pool on ten). The 130XE looked good at first and then spontaneously decided to spit out endless RWCRC errors, sometimes generating NAK errors with the HDD before even writing out the test file (using U1MB PBI/SIDE 2). I tried IDE Plus 2.0 but this was arguably worse, since the system rarely booted at all.

 

I emailed Lotharek and Pasiu with my findings, although as probably the most hated member of the Rapidus user-base I'm presumably subject to some incoming email rule. In any case, I PM'd Beetle and as ever he came up with more ideas, the best of which was to look at the power supply. Now, I was using a 2A unit here for both machines which I guess is ample for a VBXE/U1MB 130XE even with Rapidus present, but is probably cutting things fine with the 1200XL. So I sat down this evening and made a new 7 pin DIN PSU from a spare 3A Lantronix device server power supply. It certainly looks better than the 2000MA rated wall-wart I had been using, and it does seem to have made a difference. The 1200XL (using its original CPU) managed to boot into the Rapidus ROM in 65C816 mode and complete three runs of RWCRC without errors, and resisted the temptation to reboot when reset is pressed. Back to legacy mode, and much the same story. I didn't spent too long testing this evening since I have developed a blinding headache, but it looks like positive steps have been made.

 

The 130XE looks like it no longer wants to play at all (although I didn't test it with the 3A PSU yet), but that's not going to be a problem if Rapidus ends up working in the 1200XL.

 

Anyway: thanks to Beetle for his invaluable suggestions.

Edited May 27, 2016 by flashjazzcat

[Dmitry]

I'm following with interest. My rapidus is not installed yet. Luckily my Atari uses one of those usb power adapters so I can swap out a heftier ps if needed. I think I normally use 2.1a

[flashjazzcat]

I think PSUs are often left until last when troubleshooting, but I guess these 20MHz have to draw current from somewhere. The 1200XL is admittedly heavy with upgrades, and it will be to Rapidus' credit if a trouble-free board ends up working well in that machine (especially since the 130XE appears beset by Phi2 issues). Regarding this virulent reboot-on-reset issue I have experienced, Pasiu said he initially cured this on an U1MB 1200XL via the pull-ups on A0-A7, but I'm wondering if in my case the current draw on system reset was simply starving the RAM of juice, resulting in bit loss and the OS deciding it was in the middle of a cold powerup?

[MEtalGuy66]

Yeah. Well, in my experience, all these modern 3.3v and 1.8v devices are about 100 times (no exaggeration) more succeptable to power-rail noise/variation than the old 5v NMOS/CMOS stuff. If you're using a little dinky "wall-wart", you should probably look at the 5v circuit on a scope, not just a voltmeter.. You need ample current capacity to avoid voltage drops during heavy load situations but you also need adequate rectification and noise filtering when coming from AC, and a lot of the "wall wart" PSUs are intended for devices that have their OWN internal secondary regulators and noise reduction circuits. If I was you, I'd try an old (known good) PC/AT PSU.

 

This problem is very prevalent in the retro-upgrade market these days. A tiny regulator and a few caps on the power input of one of these upgrade boards can only do so much. You give it input voltage that can fluctuate by half a volt or more under load going in, and it can only "smooth out" so much.. Two examples I can think of are the Indivision ECS display enhancer/flicker-fixer for the Amiga 500/2000, and the Uthernet II Ethernet card for the Apple II. Both of these devices need a ROCK SOLID, noise-free 5v supply rail with relatively low noise, or they flake all to hell and back. Both devices work fine in a stock unexpanded machine, but put them in a machine full of cards/upgrades, and they flake out.. When all the other expansions/upgrades are working fine.. The makers of these devices have gone nuts with various "fixes" involving additional decoupling/dampning caps, etc.. None of which worked for me.. The solution: fix/upgrade the 5v supply to provide more current capacity and less noise if possible.

Edited May 29, 2016 by MEtalGuy66

[flashjazzcat]

Yep. I bought the wall-wart from who knows where and while it probably had plenty of headroom for the usual machines, who knows how clean the power is or even if it's hitting the quoted Amperage. I guess the Lantronix device server PSU I'm using now should be pretty robust.

[MEtalGuy66]

Also, I read your compliment to Lotharek about being receptive to suggestions for product improvement.

 

In that spirit:

 

I just installed a Rapidus in an 800XL.. The Rapidus overhangs the POKEY and PIA. Without having the POKEY and PIA soldered directly to the PCB, its almost impossible to get Rapidus to sit low enough for the top case half/keyboard to fit back on without interference.

 

If you look at the Rapidus Board, by far the most offensively HIGH component on top is the 6502C once its plugged into the socket on top of the RAPIDUS.. The socket used on the Rapidus board is a normal profile machined socket. My suggestion: use millmax 115-43-640-41-003000 instead. This is an "ultra low profile" machined socket and it's installed height is over .040" lower than a standard machined socket.

 

If you really want to help the 800XL people, in addition to using the "ultra low profile" socket on the top of Rapidus, you could include 3 additional "ultra low profile" sockets with the Rapidus, for those who have the ability to replace the CPU, POKEY, and PIA sockets on the motherboard.

 

Mouser has these sockets for $2.67US each if you buy at least 100 of them.

 

http://www.mouser.com/ProductDetail/Mill-Max/115-43-640-41-003000

Edited May 29, 2016 by MEtalGuy66

[flashjazzcat]

Assuming that the 6502 socket on the Rapidus has direct connections to the header legs on the underside of the board (I assume it must for proper legacy operation), the VBXE adapter route could have been taken: i.e. make the through holes big enough to accept the wider (top) part of the socket legs, and the legs of the socket on the top of the board therefore protrude through the underside far enough to actually sit in another socket underneath (e.g. the CPU socket on the A8 mainboard). This lowers the CPU on top of the board and also makes the board sit lower in the socket.

 

A little too late for that now, though. As for 600/800XL installation: I didn't even attempt this, since the board's clearly too tall. This is going to be addressed by the "Adaptus" adapter board, although I'm not sure which direction this will send the Rapidus (obviously further back in order to clear the keyboard).

 

Some of these low-profile optimisations may have been overlooked owing to the fact the 130XE is somewhat ubiquitous in Poland.

 

EDIT: actually, thinking about it, I guess not all forty legs of the socket make direct connection with the CPU socket (otherwise it would be rather difficult to make the 6502 vanish under software control), so there may be very good reasons why the VBXE route wasn't taken. But the low-profile socket suggestion is still a good one.

Edited May 29, 2016 by flashjazzcat

[guus.assmann]

Hello flashjazzcat,

 

Making the 65C02 vanish from the system is very simple.

Just keep the /HALT line asserted. That's what Antic does.. So 39 pins can stay connected...

 

BR/

Guus

[flashjazzcat]

Making the 65C02 vanish from the system is very simple.

Just keep the /HALT line asserted. That's what Antic does.. So 39 pins can stay connected...

OK. So presumably that's why the processor continues to get rather warm.

[Dmitry]

I have an Atari 800XL with VBXE.

But I've also wondered, lets say someone was going to start from scratch, and their goal was to have an 816/VBXE/Hard Drive Atari.

 

Say the user has two Atari's. One is stock. The unexpanded Atari works fine for playing any game or app that requires an unexpanded Atari. So for that reason, the accelerated system can be dedicated as an advanced system. It never has to switch into a classic mode, never has to clock down to make an old game run, never needs to see a XEX loader, will use SDX only, will never switch between DOS for any reason.

 

In other words, it'll be pleasant to use, and not be mired in configuration hell - with all the limitations that imposes on which software it'll run.

 

What would someone get for that system, if starting from scratch?

 

I'm thinking then, Atari 130XE and all Lotharek solution like: 130XE/VBXE/Rapidus/Side 2

 

Lotharek has always online ordering/immediate availability/fast shipping - and never makes you wait multiple years without telling you that you are never getting what you ordered....which actually happens sometimes with other vendors. Nobody here, I'm just saying, it happens. Lotharek is awesome.

 

The only problem with that system is Side 2. It's lovely but it is cartridge based. Ideally I'd like to keep the cartridge port unused. I will eventually burn some game to cart, and then use the cartridge port for testing a game.