Rambo XL memory problems

[Gunstar]

OK, I've got a Rambo XL 256K right now in my 1200XL, and when I use eXtended RAM Test 0.21.0 from Satantronic it shows 15 banks=240K which I'm sure is right, I'm assuming it doesn't check the 16K under the OS. But, after verifying that all my ram chips are good, and swapping positions of all of them numerous times, I still get a memory error at bank 10000011 or $83 as shown in the picture(s).

 

So, my questions are; in which specific chip on the 1200 mobo is this memory location? Is there any other reason I'd see a memory error at $83 if the chip is good? I need the actual chip so I can see about cold solder points, but could it be a problem with one of the on-board rambo support chips?

 

I can load most any program and they seem to work fine, including some demos that take upto 128K, I haven't found a demo or program yet that uses 256K that will load on my NTSC 1200XL, so I haven't tested anything above 128K. But since what I have tried works fine, I went ahead and upgraded the Rambo XL to 512K, and tried NUMEN which will work on my 1200 since I have a PAL ANTIC, but it wouldn't load past the title screen saying I need 320K. So, back to the RAM test program, which still shows only 15 banks of ram and still shows an error at only $83. I'm stumped on this one, again, all ram chips seem to be OK since I've swapped them all into every possible socket and the error is still at $83 regardless of the chip.

 

Any ideas? besides cold solder point which I can check once I know which chip position actually physically has the $83 location on it.

 

Thanks for any help in advance!

Edited October 20, 2011 by Gunstar

[+bf2k+]

Normal behavior for the Rambo. Both of mine do it. It has to do with the fact that 64k of the 256k Rambo is not extended memory but rather the normal 64k RAM in the machine and Satantronic is an extended ram checker. One of the smart guys can explain it better than I.

[sloopy]

XRAM isnt detecting it correctly... a RAMBO has 192k of ext ram, 12 banks... which should be on bits 2,3,6... which means the row of ext mem your getting errors on, doesnt exist, because XRAM is incorrectly thinking there is another set of banks on bit 7, and there isnt...

 

sloopy.

[Gunstar]

XRAM isnt detecting it correctly... a RAMBO has 192k of ext ram, 12 banks... which should be on bits 2,3,6... which means the row of ext mem your getting errors on, doesnt exist, because XRAM is incorrectly thinking there is another set of banks on bit 7, and there isnt...

 

sloopy.

 

That sounds...sorry, what does that mean? I guess I'm asking WHY XRAM would be seeing the memory this way, is it a bad program or a bad rambo?

[+Roydea6]

You have a good Rambo and you also have a good xram program. So relax and quit testing XTENDED ram.

 

That E block is normal for all Rambo's 256.

The first tested block is not Xtended Ram.

[Gunstar]

Ok, thanks guys, at least I know my rambo is OK now, and the fact that Numen won't load lies in the extra 256K ram upgrade to the Rambo, probably bad soldering some where, there's SO much of it piggybacking those ram chips! But I can live with 256K and just get rid of the extra 256K for now. After all, I can only use the 512K with demos like Numen...

[Rybags]

It's just the way the bank-selection works with some expansions.

 

With "Normal" RAM expansions, the only RAM that can appear at $4000-$7FFF is either the normal RAM corresponding to that area or a region of extended RAM. Some expansions will map in other parts of the normal 64K there, ie $0000-$3FFF, $8000-$BFFF, $C000-$FFFF and many programs will assume it as extended RAM and tests can fail or the machine crash because changing it also changes stuff in the base 64K.

 

I suppose the one big advantage though is that it gives you access to that normal 2K from $D000-$D7FF that's not normally accessable, it would be remapped to $5000-$57FF.