Atari 130XE Glitchs and Hang

[enito]

Hi,

 

Im trying to try to find the trouble with this 130XE. The adventure started with install of U1MB, i get glitch and hang with some demos. For ex (i show photos) in fornever.xex demo, i get glitchs and after some time the atari hang (all the time same glitchs and same time hang). I tryed too with boogienights demo, when demo start like the head and start uncompress process, after this atari hang. Now i tryed to restock the chips to original place, removing the u1mb and same stuff. The u1mb is not the problem. I tested too mem check, and looks fine. What can be the next step for check? I checked too traces of removed chips, and all looks fine. I attach some photos for see the problem. Thx for your time for read.

 

Greets

enito

 

 

[santosp]

Hi Enito.

Finally you make the ram test.

I see cause your monitor is pc type 31Khz horizontal scan rate. How do you manage and sees on that? Have you try to use standard composite monitor?

Edited May 9, 2014 by santosp

[enito]

Hi Enito.

Finally you make the ram test.

I see cause your monitor is pc type 31Khz horizontal scan rate. How do you manage and sees on that? Have you try to use standard composite monitor?

 

Hi Santos, yeah ram test for many hours and look right . I touched rams and looks cool, not hot. Im using a convertor from composite out of atari, to svideo and a little hardware, i show in a photo. I used too a normal composite monitor, but this one is green screen (talent old one) and for show mem test is better show with vga convertor. Thx

 

[enito]

Hi Again, i run memory test radon. With this tool apprently check extened ram, and i found problems.Now this tool test 2 banks , this ones are memory and extended one? now how i detect the bad bank? i attach a photo..

Greets

 

[Marius]

This is base ram. If you have an atari 130xe with 16 ram chips (two rows of 8 chips next to eachother). Problem is in left row.

Obviously your selftest does not check the extended ram. Only last batch of atari XE computers test extended memory in memtest.

 

As far as I know you can not say which ram chip is fault with this test, but I might be wrong.

[TXG/MNX]

I have never seen a ram test that pointed out which chip is broken.

[enito]

This is base ram. If you have an atari 130xe with 16 ram chips (two rows of 8 chips next to eachother). Problem is in left row.

Obviously your selftest does not check the extended ram. Only last batch of atari XE computers test extended memory in memtest.

 

As far as I know you can not say which ram chip is fault with this test, but I might be wrong.

Hi, thx for reply,

 

Yep exactly is 130XE with 16 ram chip.. Now, i see too BASIC rom in RED, i have some problems now to boot, dont get basic only selft test. This is a sympthom about rams and too the basic chip? thx!

[enito]

I have never seen a ram test that pointed out which chip is broken.

you mean about red points?

[Marius]

Hi, thx for reply,

 

Yep exactly is 130XE with 16 ram chip.. Now, i see too BASIC rom in RED, i have some problems now to boot, dont get basic only selft test. This is a sympthom about rams and too the basic chip? thx!

 

As far as I know this can be:

 

1) a basic rom with invalid data

 

2) a bad MMU chip

 

3) a problem with the PIA chip (I don't expect this)

 

Both BASIC and MMU can lose their contents/functionality.

[enito]

 

As far as I know this can be:

 

1) a basic rom with invalid data

 

2) a bad MMU chip

 

3) a problem with the PIA chip (I don't expect this)

 

Both BASIC and MMU can lose their contents/functionality.

 

2 ) I changed with other MMU and same result

3) Dont hope so.

 

Basic chip can be reprogrammed? eraseable via UV?

 

thx!

[Rybags]

Don't forget, 130XE also has the EMMU - it decides which of the 2 rows of chips an access has to go to.

 

Basic can't be reprogrammed, it's ROM, not Eprom.

 

Does the system work if you just start up with no disk device and go to Basic?

And not forgetting the old technique - let it sit a while and see if 1 or 2 Ram chips get hotter than the rest.

 

In theory a Ram test could point out which chip/s could be bad but there's quite a few configuration possibilities.

16-chip 130XE, an entire column of 8 contributes 1 bit for each Ram address.

4-chip 130XE, column of 2 chips contributes 4 bits for each Ram address.

Then you have the various others like 800XL with the different available expansions.

Edited May 14, 2014 by Rybags

[enito]

1)Don't forget, 130XE also has the EMMU - it decides which of the 2 rows of chips an access has to go to.

 

2)Basic can't be reprogrammed, it's ROM, not Eprom.

 

3)Does the system work if you just start up with no disk device and go to Basic?

And not forgetting the old technique - let it sit a while and see if 1 or 2 Ram chips get hotter than the rest.

 

In theory a Ram test could point out which chip/s could be bad but there's quite a few configuration possibilities.

16-chip 130XE, an entire column of 8 contributes 1 bit for each Ram address.

4-chip 130XE, column of 2 chips contributes 4 bits for each Ram address.

Then you have the various others like 800XL with the different available expansions.

Hi, thx again to all the replys,

 

1) I checked EMMU, with other atari working and same trouble, i can think emmu is not the trouble

2) Right is a rom

3) System is working, but cant enter to BASIC and mems have errors. If i power on i get only SELFT TEST. Yep i applyed the old technique of hot ram and nothing, still cold.

 

thx!

[Rybags]

Long shot - maybe try taking the Basic Rom out. Theory being it might be bad and outputting random bits which mess the system up.

 

A Basic Rom which reads bad but otherwise doesn't mess with the system shouldn't cause self-test to start though - Basic doesn't get checksummed on system startup.

[russg]

Hi, thx for reply,

 

Yep exactly is 130XE with 16 ram chip.. Now, i see too BASIC rom in RED, i have some problems now to boot, dont get basic only selft test. This is a sympthom about rams and too the basic chip? thx!

It is known that 130XE RAM chips 'MT' are bad drams.

[atari8warez]

I had a 130XE that showed very similar symptoms (or was it an XEGS?). Bad chips don't always get hot, one way to find out which chips are bad is to piggyback them with good ram modules. Its a long process but you can generally identify and isolate bad ram this way. For the bad BASIC ROM you can either replace the ROM with a good one, or if you can't find a BASIC ROM you can burn a compatible EPROM or EEPROM if you have a programmer handy

Edited May 19, 2014 by atari8warez

[enito]

Hi, thx for replys, today arrived mem repleacement, this ones:

http://s8.postimg.org/jbamsmto5/2014_06_02_23_23_42.jpg (Nec one piggyback in bank)

 

I tested one using piggyback technique in all roms (put in one and do test, but in the next do test..etc) and i get same error of bank

 

http://postimg.org/image/xofshe8m5/

 

Now , BANK 0 Is extended area? (there i get error) (and after i get basic rom error) , extended area is left side?

 

Thx!

[enito]

I think too can be a broken trace, cause i changed BASIC ROM and still trouble of no detect basic rom. And the same point of red cubes in memory test, can be some broken trace from bank0 to basic rom? or something like this? thx!

[1050]

Where did you get that Glenn M. Smith memory tester as I would like to take a look at the code it's running. Bank 0 should be the main memory and Bank 1 should be the extended memory. Main memory is the left column of ram chips and the extended memory is the right column.

 

In this picture:

http://s8.postimg.org/jbamsmto5/2014_06_02_23_23_42.jpg (Nec one piggyback in bank)

you have the two CAS\ memory bank select lines shown as the two resistors side by side on the right side down low. You can unsolder them closest to the ram chip, pull the leads out of the hole and swap them by bending one to fit the other hole and then with a jumper wire for the other to effect a swapping of the columns of ram chips - now the left becomes the extended memory and the right becomes the main memory. It's a tool you can use to isolate the problem if you need to go there or feel like doing that.

 

It may be that Smith's memory tester is showing a problem with BASIC rom select signal when it shows the bad ram locations, depends on where he is using ram to show the green or red blocks at and just how he is testing the ram as well. The orginal memory tester is well known to be lacking in a full test, Smith may have tried to change that and in the process is now showing a problem with the ram under BASIC which may come back to the problem with BASIC itself. Fix BASIC and your ram problems may disappear? If this much is true then swapping the banks as oulined above will do nothing to solve the problem, but would indicate that BASIC is where the real problem is if you get the exact same screens as you have now with Smith's memory tester. And the real problem then would probably be the BASIC rom enable signal isn't quite right for some reason yet to be figured out.

[enito]

Hi, thx for your time for explain me the trouble.Now i think for start i need to try to fix first BASIC. I tryed with 2 chips of basic and same troubles. Now i know the basic chip is fine. Whats gonna the next step for check basic? i get only self test when i power on the atari. I tryed to check traces of basic and near chip (using schemas) and not see apprently a broken trace.

 

Greets!

[1050]

BASIC chip enable (pin 20) on the BASIC rom U4 comes from U3 MMU (pin 13), a 20 pin PAL 16L8, so you need a good working MMU chip to replace yours with and if this fails to effect a change with BASIC then replace your ANTIC chip U7 which sources the REF\ signal (pin that the MMU is using for it's clock signal. I would think it would be the MMU, since troubles with ANTIC would be massive and system wide, but it could be either the MMU and/or ANTIC. And you may not have any luck even then. Even a dodgy solder joint at pin 13 (or clock input pin 11) of the MMU might be the cause. Remelt them first to see if that alone might fix it. And pin 8 of ANTIC while you are at it. Double check all traces between the pins outlined above too before swapping out chips - this is a delicate board, very easily burnt and ruined by too hot of a soldering iron. You might seriously consider putting in sockets to prevent burn damage opportunities from happening in the future.

 

A better test for your BASIC roms would be to put them in an 800XL and see how good they fly there as well. Common sense tells us that you have eliminated the BASIC rom itself from being the problem by swapping it out, but you could easily have two bad BASIC chips, apparently Europe was the destination of many of them somehow, because I've seen this issue before, but mainly over there. Just how did your donor machine get to be relegated to donor status in the first place? Perhaps because it had a bad BASIC rom chip and did not work? Nice thing about the 800XL was that most of them had full sockets already so if nothing else they can serve as handy chip testers for XE fans. I know they are hard to come by over there too, another distribution mystery. Since you've already swapped BASIC rom, it seems you have some skills with your iron, but best of luck anyway. I am never at ease working on the 130XE, just way too easy to delaminate a trace or a pin via and then you've got a further repair job to do. Until your current BASIC rom is proven to be good in another machine, you can't eliminate the BASIC rom itself from being the main problem, I would start there first since there is precedence for a lot of bad BASIC roms on that side of the pond, somehow someway. I know it's like starting all over again, but it's really the best move to make at this point - you need PROOF that it's good, before doing what amounts to a heart transplant on your machine. Imagine your grief if it all were done for nothing and you did have two bad BASIC roms. There are special sockets designed to take a standard 28 pin 2764 eprom that could be burnt with the BASIC code and fit right in the BASIC rom 24 pin layout - this would allow you to have a BASIC rom that would not be so likely to fail over time as the originals seem to want to do over there. I would use a 150 nanosecond 27C64 eprom if you can find one for this as that speed is more in line with the speeds used way back then, faster is not necessarily better for these OS roms. These 28 to 24 pin sockets can be hand made as well, notice that if you cut off pins 1, 2, 27 and 28 of a 2764 eprom that most of the pins on both layouts will now register exactly with the same pin use if you line them up bottom for bottom. You would then need to jumper cut off pin one of the eprom to cut off pin 28 and cut off pin 27 and then on to pin 26 of the eprom which is now pin 24 of the 24 pin layout - this takes care of the 5+ supply and you have only two more pins to play with. Pin 20 (full count) of the eprom gets bent over, trimmed to fit and soldered to neighboring pin 22 which is now renumbered to be pin 20 of the 24 pin layout and then you need to solder a jumper from cut off pin 2 of the eprom to the hole left by bending the two eprom pins together, hole number 21 on the 24 pin layout. Use two 24 pin machine socket for this, the first to eliminate any further soldering done to the motherboard and the second one to put the 28 pin eprom into which will overhang the 24 pin socket by two cut off pins places. BASIC roms seem to be in short supply over there too, and not a coincidence.

 

And again the source for the Glenn M. Smith memory tester? I may have it already, but just to be sure post a link or source for it, TIA.

[enito]

BASIC chip enable (pin 20) on the BASIC rom U4 comes from U3 MMU (pin 13), a 20 pin PAL 16L8, so you need a good working MMU chip to replace yours with and if this fails to effect a change with BASIC then replace your ANTIC chip U7 which sources the REF\ signal (pin that the MMU is using for it's clock signal. I would think it would be the MMU, since troubles with ANTIC would be massive and system wide, but it could be either the MMU and/or ANTIC. And you may not have any luck even then. Even a dodgy solder joint at pin 13 (or clock input pin 11) of the MMU might be the cause. Remelt them first to see if that alone might fix it. And pin 8 of ANTIC while you are at it. Double check all traces between the pins outlined above too before swapping out chips - this is a delicate board, very easily burnt and ruined by too hot of a soldering iron. You might seriously consider putting in sockets to prevent burn damage opportunities from happening in the future.e.

 

And again the source for the Glenn M. Smith memory tester? I may have it already, but just to be sure post a link or source for it, TIA.

 

Hi, first thx for this big reply and very explained one, thx for share your time with me Now i check the traces you tell me, BASIC Pin 20 - MMU pin 13 and fine, i checked too with antic the REF with MMU and fine. Now this MMU im sure is fine, its was swapped from a working machine, and with original 130XE MMU same result. I checked solder of pin 13 and 11 of MMU and looks fine, i check traces and not find any anormal. Yep when i unsoldered the chip for put socket (cause this 130XE have all chipset without socket only CPU have sock), i damaged some trace but fixed with cables, i show you in photos. Now i can re check all pins from MMU To BASIC and ANTIC for see some anormal thing. Basic chip i swaped from working machine and too a 0 miles i purchased time ago from B&C. This 130XE is NTSC in case.

 

The source of GLEM memory test, SURE, here

 

http://atariage.com/forums/topic/203450-how-depressing-my-xe-fragged-who-can-fix-this/page-3 (Here in this post)

 

I show now photos of front and backside of board, hope one day i get working this nice machine. Thx!

 

 

 

Here you see my little repair with red cables of this damaged trace when i unsoldered the chips.

 

Sorry my solder is not perfect

[1050]

OK, your soldering is a bit rough around the edges but that's pretty good for working on a 130XE, just one boo-boo. A little flux might help you make really good connections as I see a couple in need of a closer look. Pin 16 of the MMU appears to be a cold solder joint in that half of the pin is not wetted by solder, flux will really help these kinds of problems. And there appears to be a solder bridge between pins 17 and 18 of the OS rom chip, THIS would make a difference I would think, if it's really there. See if you can feel it with a knife point and perhaps just cut it out of there with the knife if it's really there at all.

 

Does the old BASIC rom work in the other machine? Would be nice to know if it's really bad to begin with. Maybe you said before and I've forgotten. What is really needed to test the connection between the BASIC rom enable pin 20 and the MMU output pin 13 is about a 20 ma load like a small LED light driven with a 1.5 volt battery in series with a 35 ohm resistor, if it lights like it does when connected directly then I would have real faith in the connection being good enough to work. As it is there could be partial trace contact under the MMU socket and we just can't see the bad connection for the socket being in the way. Even an analog VOM meter wouldn't be able to tell if the connection was good enough to carry operational currents, it would show a good connection even when it really was a bit iffy.

 

Now I'm wondering if having the U1MB installed might have overloaded the PIA somehow? And it should be changed next? Like ProWizard said, I normally wouldn't expect it to go bad and double that doubt about ANTIC, but... Unless removing the whisker between pins 17 and 18 fixes it, I'm getting stumped too.

[enito]

Only for clarify pin see distribution i added numbers looking from backside, for to be sure the pins to check (late tonight i do it )

 

 

thx!

[enito]

Okey, guiding with the pins you tell me , i checked 16 with flux for fix, and other too in case. Pin 17 and 18 not bridge, checked twice and nothing, is fine. The next step is check with diode, but with multimeter i get beep of continuity of pin 20 and 13. I put all chips again for check, but same result, the system goes direct to self test Basic chip is fine, i checked in a 800XL and works thats is a hard mission, but dont wanna loose

 

greets

[1050]

Could you post for me a new picture, bottom view of the OS rom area? Different angle if possible. I can't believe it's not a whisker.