Dead 130XE - Any advice?

[Level42]

From what I understand the Atari is not nearly as sensitive to timings as some other platforms. Don't some of the 4164 versions have time and response varieties?

Exactly. My 130XE has a mix of -015 and -020 mT's from the factory and never showed any problems.

Edited November 26, 2017 by Level42

[Guest]

 

Damn....I hope I don't get sued.....

 

I clearly explained WHY _I_ would do something else than others MAY do at their own decision....but you forgot to quote that part of the posting....

 

Oh, and I'm not paid to do this stuf.....

i quoted you word-for-word. words that clearly are at odds with each other.

as i said that's definitely a "do as i say, and not what i do" statement

 

and what has "getting paid" got to do with anything?

or are you now painting yourself as some "benevolent soul, that we should be eternally grateful for"?

news flash: we all do it for free - that's what a hobbyist community is.

 

if you can't see why your advice is contradictory - then that's another issue entirely.

Edited November 26, 2017 by Guest

[Level42]

Whatever....I will think twice giving advices here from now on as I only get shit for it.....typical forum behavior....so sad... I thought this forum was a bit better in that context than others...

 

If you want to catch me ...or pin me....on certain words ....or just don't want to understand what I mean...you know....that's really not my problem.

[_The Doctor__]

I kind of understand both sides of the quip, just as a mechanic will do certain things with his own car that he would *NEVER* do with a customers car... he might share what he does with a person whom he believes will understand the ramifications of doing both.

Edited November 27, 2017 by _The Doctor__

[Guest]

Whatever....I will think twice giving advices here from now on as I only get shit for it.....typical forum behavior....so sad... I thought this forum was a bit better in that context than others...

 

If you want to catch me ...or pin me....on certain words ....or just don't want to understand what I mean...you know....that's really not my problem.

you're over-reacting.

you challenged me on my perfectly good advice. i disagreed with you.

you then disagreed with yourself.

 

please don't imply you were misunderstood. you clearly instructed

 

"i would do A, but you should do B"

 

you talk about "this site being better" - but i'll probably get banned now for daring to challenge your inconsistency.

Edited November 27, 2017 by Guest

[TheTIGuy]

Well I don't think my Y2.23 is needed on this topic...

[Level42]

I sincerely hope we didn't chase TroyQ away.....awaiting his feedback to get this thread back on the rails...

[SoundGammon]

The first thing I would have swapped out would be the OS chip. If it's bad, the cpu goes into la-la land.

 

If there is any type of screen, red or whatever, GTIA-ANTIC have to be doing something.

[1050]

2) Are there mT RAM's => Feel for hot ones. Replace. Even better, replace them all.

My take on first read was all MT ram, but MT is

only mentioned the one time. Could be immaterial

to the rest of the sentence except for the easily

assumed assumption that all ram mentioned in that

sentence is MT except perhaps the replacement ram

I would only hope. Too much guessing required.

 

Oddly flashjazzcat states 'to be fair' about an

ideal situation which I now take to be the 'even

better' part although I didn't and wouldn't normally

go there. I didn't even understand flashjazzcat until

I made allowances for across the pond interpretation

differences possible and it's still a stretch.

 

Even though he speaks English, it isn't USA english

and we can't follow the way they do?

 

Level42 has shot up some antagonistic

pot shots at my writing which also shows a bad

attitude which I've chosen to ignore but more

to the point - he truly doesn't understand me.

So he has the same issues I do with his writing.

Not very clear on the actual intent if Level42

claims he didn't say yank them all in the above

quote.

 

Very curious.

 

Are we witness to language evolution at work and

it very well may be the root cause of the quibbles?

Due caution is advised from here on out will be

my tack.

 

Level42, thanks for the in depth explanations

of post #22.

I'd like to say yes, it's all clear now, but

I'm not sure about that yet due to above issues.

 

1) I have stated MANY times on a few threads about 130XE and RAM issues that my 130XE has mT RAM and it is still working fine and I have no desire whatsoever to fiddle with it....as long as it works.

Won't be reading all your posts in other

threads. You'll have to understand that's the

default mode for everybody wandering by too.

 

2) I have also stated several times that by my experimenting with RAMs all over the sockets on the 2nd 130XE board I have that an 130XE,WILL NOT OPERATE WITH A mT RAM IN THE TOP LEFT POSITION, no matter if that same RAM chip will work on any other position. There is something about timing or maybe voltage levels, I dont know what it is but a GOOD mT RAM on the top left position will render a non working 130XE.

Missed that tidbit because of #1 above.

But it's a juicy tidbit none the less. Can't

promise when but I'll see if I can verify

that very odd behavior here eventually.

 

3) mT RAMs do not only die because of over voltage, they just die (of age) on systems that have never been exposed to over voltages. Who says its the wires ? It can just as well be defects on the chip itself. Whatever they did, they screwed up in the factory.

bla, bla, bla. Wires were just a handy

thing it could have been, don't believe I'm

invested in it to any degree at all.

 

4) I have bought mT RAMs in the past during my arcade collecting years. I was unaware of their issues then. I stored them in an ESD safe box. Only recently, when repairing my 2nd 130xe board, I tried them. They were dead. All of them. They had done nothing for 25 years or so and they were dead. IMHO this is an induction that probably some gas or other substance got inside during production that corroded or otherwise deteriate these chips.

OMFG - why didn't you start with that MT

horror story, move over I'ma getting in your

boat now.

"IMHO this is an induction that probably"

indication, no worries I read thru it fine

and it's not a chastisement, I quote it only

to point out the real defect here. Humble?

You should be shouting this from the mountain

top - how do you do that and be humble about it?

Serious defect in getting your point across

when this horror story pops up in post #22.

If you posted this in other threads you need

to look them up and save the link to it and

send people there long before you get to post

#22. Are your logic circuits functional?

Because I don't understand how you could be

so obtuse as this. You save the punch line

for last when telling jokes only, you normally

lead with the best shot you got and this one

alone has me converted. I'm scared to death

of MT now and it's all about this 25 year old

100% dead tray of MT ram - there is something

seriously wrong with these things.

 

As to what is the issue with MT ram dying

while unused, it's likely they didn't do a

proper rinse on one of the applications of

hydrofluoric acid which is used in the

manufacture of ICs everywhere - it's pure

evil and will eat everything except wax.

And teflon. Where the fl in teflon comes

from flourine.

So there is an issue with hazardous waste

disposal and this is the big one so the

tendency would be to quit making so much

waste by using a shorter rinse cycle. Trace

amounts left on the die could eat the junction

between N type and later deposited P type

leaving a dead zone that does nothing - just

what we gots.

No function at the junction.

 

It's primarily used to eat away silicon

dioxide layer that has been covered in

photo sensitive resit which dictates where

the still intact glass like insulation will

be protected under the resist pattern generated

by the projected UV image. It eats glass

very quickly, other items very slowly but

none are long time immune outside of teflon

and wax. Flat surface frosted glass is most

likely hydrofluoric etched with a wax

resit pattern. Non flat designs can be rubber

sheet masked sandblasting.

 

 

Do you seriously expect that Atari tested the XEs in over voltage situation ? It was the Tramiels running the show then....their only goal was to get the machine to run just long enough to make it past the warranty period.

I know for a fact that they would have for a random

selection of each lot of each chip and other components

used. SOP - called Quality Control department where

several do nothing all day long but burn and catalog

their results. Supplier already did that at their

place and have sent the torture machines and free

samples to torture - again SOP. And you could read

all about it in the back of some ancient Data Books

if you had access to them where this issue is

presented in much finer detail. Insurance company

needs to know the MTBF hours so they can assess

insurance rates for both the chip maker and end

manufacture. They need duplicates and their own

techs and it's all one big burning fest. At the

component level only, and this allows 99.9 % of

assembled machines to sail right out the door to

marketing department for delivery after a power up

check passes.

 

They don't have to test the entire working

XE and burn every other one. If a larger library

near you has any IC Master books you might find

something in there about quality control

procedures as offered by the various chip makers.

 

Knowing only now what you didn't tell us

before has me convinced it would be a waste of time to

do anything other than avoid MT ram entirely.

AND replace on sight like a madman gone amok. Stark

raving crazyman mode here.

 

Any chance there is a date code on those

chips in that lot? Might be helpful if

you could post that info if it's there.

 

Would strongly suggest you posting some links to

those posts with detailed information such as

found in bombshell #4 above. Would have saved

me a lot of typing and again I'm not reading every

word you've typed - you are just too prolific and I

hope you don't burn out and just quit us.

 

Thanks for very handy link DrVenkman, it's on it's

way at 27 with postage.

 

Yep, my sentiments exactly Level42, hope TroyQ

is not put off in the slightest, just a family

squabble, since I can only speak for myself, I

will, I'm good here and looking forward to a lot

less typing.

[Level42]

I need an apirine......or maybe just a whisky....

[flashjazzcat]

Oddly flashjazzcat states 'to be fair' about an

ideal situation which I now take to be the 'even

better' part although I didn't and wouldn't normally

go there. I didn't even understand flashjazzcat until

I made allowances for across the pond interpretation

differences possible and it's still a stretch.

 

Even though he speaks English, it isn't USA english

and we can't follow the way they do?

Not sure I follow that exactly, but in any case I wasn't responding to anything you'd written. I was responding to the prior assertion that Level42 had said all MT DRAMs should be replaced. That's not how I interpret "Are there mT RAM's => Feel for hot ones. Replace. Even better, replace them all." I interpret that sentence as "Replace all the MT DRAMs if you can be bothered, as they're known to be crap. If not, just replace the hot ones." I don't find that starkly inconsistent with later admitting that "personally, I'd just replace the dead ones", hence "to be fair" (or "in fairness" or "on the other hand"). I could happily advocate completely fitting out all unsocketed A8s with precision sockets as a matter of course, but it's not something I necessarily have the time or inclination to do, however much trouble it might save later on. I would not expect to be labelled a hypocrite because of that.

 

But in any case, I'll leave you to argue other points between yourselves. I'm often faced with contradiction here, simply because this forum attracts the most experienced and knowledgeable people. Those accustomed to having their opinions left unchallenged may have a hard time.

[_The Doctor__]

You know what? I am considering Knighting Flash Jazz Cat........

in fact ....

I dub thee SIR FlashJazzCat of the United Kingdom,

 

May logic and wisdom ever be the guide that lets the spiritual journey towards common sense continue..... etc etc

[Level42]

Whatever you can say about my English, it is always lighyears better than your Dutch.....

[TheTIGuy]

And his dutch is still better than my Spanish, and that is much better than any Japanese I know.

[TroyQ]

LOL - no, I have not been chased away; the holidays (US) have taken up most of my time. The strong opinions and views don't bother me a bit; I've been in IT way too long for that to bother me.

 

I'm still waiting for RAM to arrive, so there hasn't been too much to report.

 

I did swap the banks (R111 & R112 swapped). Still had the same result - so it didn't really help. It all went smooth, no lifted trace, so I am hopeful that it portends good things for the future.

 

@1050 - thanks for the link for the pico scope. If I was going to be doing a lot more chip level work, I would seriously think about it; but with a whole board replacement of $80, I have to weigh the cost to figure it out vs the cost to just make it work. Of course, there is still the voice in my head that says "Go Ahead... its a tool. You know you need it!" I've taken pix of the RAM, so if they end up being the issue, I can post for any date code,etc documentation.

 

@DrVenkman - $20 is more in my budget , so I did get that logic probe.

 

Now, with logic probe in hand, suggestions on where to start checking?

[shoestring]

Start off with the "troubleshooting" section in the service manual & also download a TTL logic data book. This really helped me fix my 65XE.

 

http://blog.3b2.sk/igi/post/Sams-Computer-Facts-ATARI-130XE.aspx

 

I use the same logic probe for checking stuff quickly but a scope is more ideal ( it's better than nothing I guess ).

 

Also, a logic comparator or EPROM device programmer with TTL logic / RAM test will go a long way. I have an EETools TopMax which is great for checking mask roms / EPROMs , TTL logic and different types of RAM but it's quite expensive. I don't know what I'd do without it.

[_The Doctor__]

I made my own logic probe with some wire (alligator clip).. a 330 ohm resistor.. a two lead bi color led and a pin (pogo pin if you have one laying about).. total cost 2.76 including electricity and solder.... I probably could have gotten one shipped from china for less....

Edited November 28, 2017 by _The Doctor__

[CPUWIZ]

EDIT: Not worth it.

[_The Doctor__]

sure it can be worth it... but since people never believe I did a quick google search so I can be lazy and not have to draw anything....

here

http://www.seattlerobotics.org/encoder/200012/logicprobe.htm

 

if your worried about too much draw on the circuit increase the resistor, take it as high as you like so long as you can still see the led light up your golden

Edited November 28, 2017 by _The Doctor__

[Level42]

I've taken pix of the RAM, so if they end up being the issue, I can post for any date code,etc documentation.

That is great. I would like to investigate if it was a certain batch or maybe period where the production troubles happened. Clearly many mT RAMS do still work fine so maybe there is something to find out about it. Which would be cool to know if we can isolate certain date codes to be suspicious. Can you please post a picture here ?

 

Now, with logic probe in hand, suggestions on where to start checking?

Im interested in that too. Because lets assume one RAM chip is dead and drives down one of the databus or addresslines because of a short to ground or +5V...you will see no activity on one of those data- or addresslines. But how can you see which RAM IC (or any other of the ICs on those busses) is causing the short ? Edited November 28, 2017 by Level42

[+DrVenkman]

@DrVenkman - $20 is more in my budget , so I did get that logic probe.

 

Now, with logic probe in hand, suggestions on where to start checking?

Preliminarily, grab pin outs from google or a field service manual for all the main IC’s and simply verify that grounds are actually going to ground; that the Vcc for each is actually getting a steady +5V or thereabouts, and then check clock, address and data lines - comparing to the pin outs - to make sure each chip is getting or sending the signals it’s supposed to. Tedious but effective for finding a bad trace, bad socket or bad chip.

[TroyQ]

Is there any documentation or flow charts that document the actual boot process? I'm assuming it is something much simpler than a PC POST, BIOS Init and Boot. Something along the lines of power on, all chips get power and perform some kind of init, CPU begins execution at a specific memory address (in the OS ROM chip)? Is the OS ROM copied to RAM first or is it just mapped within the 64K space and then execution begins at a specific memory address?

 

For the RAM, if I probe the CAS/RAS lines, I assume I should see constant activity there? Freddie would be generating that? Would Freddie do that regardless of what is happening with the rest of the system (ANTIC, POKEY, etc)? Or to put it another way, if there is no CAS/RAS activity, does that point to Freddie or could that happen due to some other failure (OS ROM for example)?

 

The first thing I would have swapped out would be the OS chip. If it's bad, the cpu goes into la-la land.

 

If there is any type of screen, red or whatever, GTIA-ANTIC have to be doing something.

 

Is the OS ROM a common failure? Its just a ROM chip, right? I always thought ROM/PROM/EEPROM etc were fairly tank-like and mostly indestructible. The amount of abuse a cart can take is pretty substantial (cough cough digging up functional 2600 ET carts from a landfill after 30+ years). Assuming that there was not some wild voltage surge, what would cause the OS ROM to die?

 

 

Start off with the "troubleshooting" section in the service manual & also download a TTL logic data book. This really helped me fix my 65XE.

 

http://blog.3b2.sk/igi/post/Sams-Computer-Facts-ATARI-130XE.aspx

 

I use the same logic probe for checking stuff quickly but a scope is more ideal ( it's better than nothing I guess ).

 

Also, a logic comparator or EPROM device programmer with TTL logic / RAM test will go a long way. I have an EETools TopMax which is great for checking mask roms / EPROMs , TTL logic and different types of RAM but it's quite expensive. I don't know what I'd do without it.

 

Thanks for the link. There are a few steps in there that I will check, but it is pretty sparse with regards to a no boot failure. Reading it, it looks like if the CPU passes some basic RESET failure checks, the next steps are to start swapping chips until it works. And I could see that that would have made sense for a service center that probably had a stock of known good test chips; that's essentially the process for any PC/Server related repair: replace boards/drives/ram/cpu until the problem is gone. LOL.

 

 

That is great. I would like to investigate if it was a certain batch or maybe period where the production troubles happened. Clearly many mT RAMS do still work fine so maybe there is something to find out about it. Which would be cool to know if we can isolate certain date codes to be suspicious. Can you please post a picture here ?

Im interested in that too. Because lets assume one RAM chip is dead and drives down one of the databus or addresslines because of a short to ground or +5V...you will see no activity on one of those data- or addresslines. But how can you see which RAM IC (or any other of the ICs on those busses) is causing the short ?

 

 

Yes, I will definitely post pix of the RAM.

 

I recognise the board swapping issue, my education was in electronics but in my work too I used to simply replace entire units, there was no time (=money) for doing component repair, in fact in many jobs we werent even allowed to mod anything on a PCB because it would fall out of certain safety certifications.

Wonder what company you work for, you can PM me if you dont want to name it here.

 

I used to be a Network Engineer for a company called CompuCom. Probably never heard of them, they've been bought up by someone else since I left. At the time, I supported HP, Compaq, Dell & IBM servers & workstations - IMAC, setup, admin & repair. But that was long ago, now I am a software developer for a vegetable seed company.

 

Preliminarily, grab pin outs from google or a field service manual for all the main IC’s and simply verify that grounds are actually going to ground; that the Vcc for each is actually getting a steady +5V or thereabouts, and then check clock, address and data lines - comparing to the pin outs - to make sure each chip is getting or sending the signals it’s supposed to. Tedious but effective for finding a bad trace, bad socket or bad chip.

 

Thanks! I had a feeling that was the answer. I was hoping that someone would have some magic "Probe @ X & Y and result W means chip A is ok/bad" etc but I see that was probably too much wishful thinking, oh well.

 

I am still waiting for RAM and in the mean time I will start poking around with the probe and see what I find. I will try to do some A/B comparisons with probing a good 800XL and see if there are any obvious differences and post any useful results.

 

Thanks again for everyone's advice!

[shoestring]

RAS ( pin 33 ) and CAS ( pin 35 ) should both always pulse with your logic probe. Otherwise consider Freddie dead.

 

See my old thread for what these signals look like under a scope.

 

http://atariage.com/forums/topic/258601-65xe-repair-and-troubleshooting-advice/

[_The Doctor__]

if I followed your ordeal correctly you had bad ram, a bad solder joint, bad mmu, and a bad gtia..... after replacing all those it was good again

[shoestring]

Not entirely. In my case initial symptoms were a black screen.

 

DRAMs were bad. Replaced with no change.( still black screen )

CPU tested bad by substitution in the 600xl doing weird stuff. After replacing symptoms change to red screen ( signs of life )

MMU tested bad by substitution ( red screen with it in the 600xl )

 

Everything tested good at this point except I hadn't gotten around to troubleshooting the freddie chip yet.

 

Fixed the fault in the modulator which was causing intermittent video signal.

 

With some help from tf_hh, we determined that freddie was showing signs of life but was still faulty. Replaced it ( not much change )

 

Re programmed the pld with the correct MMU code ( still red screen )

 

Discovered OS rom was bad, tested good initially so I must have killed it in the process of all that was done.

 

GTIA was partially bad on one output only ( no keyboard clicks ).

Edited November 29, 2017 by shoestring