Another 130xe memory issue

[moonlight_mile]

Hey guys,

   I have this old 130xe I got when I first started getting back into the scene. After a few years I got an side2 cart and it didn’t work. There was a memory issue.   I ended up getting another 130 and 800xl. 
 

But I have gotten into the mood to try to figure out what is wrong with this 130. 
 

I have run eXended memory test and it shows only 64k and a few errors. 
 

The board is populated with all MT4264s except 1. I have some 4164s laying around so I tried to do the old piggyback method to see if I could get it to recognize the extra 64k.   Tested them on top of all the ram chips. Still same result. (I read that 4164s were somewhat compatible with the mt4264s). 
 

Now this leads me to believe the problem could lie elsewhere. Possibly the bank switching chip?

 

I know mt chips are failure prone but since nothing changed good or bad when I piggybacked them I guess there is a possibility they are ok. 
 

So what do you guys thing?  Could it be the co25953? Or something else. 
 

I have attached a bunch of pics for

your perusal. 
 

thanks in advance. 

[+Nezgar]

I would use Syscheck or shoestrings RAM tester replacement OS ROM to also ensure your base 64K is OK.

 

Next, you can swap the base and extended 64K by cris-crossing one end or R110/R111, which swaps U26-U33 to become your base 64K, and U9-U16 becomes your extended 64K:

 

 

If the system still boots afterwards, then the extended (U26-U33) 64K of ram may actually be OK, and your suspicion about an MMU issue may be true... Also test this 64K with syscheck and/or shoestrings RAM tester, as they do not test extended memory, only base memory. If the system doesn't boot, then the extended 64K is actually bad and you might as well desolder and socket U26-U33 to permit further troubleshooting by replacing some or all of the DRAM chips...

[Beeblebrox]

@moonlight_mile  Few things. If you have the patience/skills I'd remove all the MT ram and socket the all. Then ditch the MT ram chips.

 

Also look at the MMU (61618) and socket it.

 

Plus worth socketing the 74LS08 and replacing with a 74F08 while you are at it.

 

The fact you have issues and there are MT chips involved is all the more reason to replace them. Perhaps with some Samsung RAM chips, 12n or 15n

 

EDIT: Go with Nezgar's suggestions in the first instance. 

Edited April 13, 2022 by Beeblebrox

[moonlight_mile]

Thanks! I know I did run the syschek a few years ago. (This was the reason I got the syschek. ) and I know it tested ok and I was able to boot the side2 using the syschek memory but that obviously didn’t address the issue.

I will work on doing the resister swap this weekend.

I agree with scrapping all the mt chips. I was just hoping to not have to do such invasive surgery at the beginning. Lol.

When I did ponder starting this project I wanted to get an assortment of the round hole machine sockets but it seems that no one sells assortments (except Ali-express, I would rather not wait 3 months). Are the standard sockets ok to use in this case?

Thanks again. I’ll post my results from the syschek tests.

[RodLightning]

The piggyback trick doesn't always work, although always worth a try in finding bad RAM. No success here either.  A second cheap trick is to leave the XE on for a few minutes and do a touch test on each chip.  Dead ones might be very hot (shorted) or cooler (broken circuit) when compared to most of the others.  With the bad reputation of those MT4264s found in 130XEs, RAM is still highly suspect as the problem.  Best to replace all the MT branded chips.  The NEC chip has the best chance, I suppose of being good.   There has been speculation about why that first chip in bank 0 is often a different brand.  Could the flakey RAM have been a known issue from the factory?  Many including mine also have one NEC chip in that spot.

 

The one I have here is in pristine condition otherwise.  It only boots to self test and shows red blocks in RAM test.  I've been putting off desoldering all the RAM for a while now.  Reading various forum posts about poor quality circuit boards and lifting traces if the temperature isn't just right has given me pause.  You will find many threads around here with good advice on the 130XE and bad RAM.

[moonlight_mile]

Darn I forgot the touch test. I will do that tomorrow.

Also the fragility of the traces has also been a reason I have been dragging my feet. But I am looking at it if I kill it I kill it. It’s not my daily driver or back up.

What is a good place and brand to get replacement 4264s?

Thanks

[+Nezgar]

6 minutes ago, moonlight_mile said:

I agree with scrapping all the mt chips. I was just hoping to not have to do such invasive surgery at the beginning. Lol.

After doing the resistor-cross, any bad memory identified by sys-check, can potentially be narrowed down to specific chip(s). I had bad memory in the extended ram of a 130XE and got it functional with only having to desolder/socket/replace three of the chips by hand since I didn't have a desoldering gun... I really should do all of them, but it's good enough "for now".    I'm fully aware more could fail at anytime so it's definitely not a candidate for resale unless I finish the job.

[RodLightning]

I ordered some from Jameco a while back.  I notice that they are listed as refurbished.

 

https://www.jameco.com/z/4164-150-Major-Brands-IC-4164-150-DRAM-65-536-Bit-65-536x1-150ns-with-Page-Mode-DIP-16_41662.html

[RodLightning]

30 minutes ago, moonlight_mile said:

Darn I forgot the touch test. I will do that tomorrow.

Also the fragility of the traces has also been a reason I have been dragging my feet. But I am looking at it if I kill it I kill it. It’s not my daily driver or back up.
 

In that case, a method mentioned elsewhere of cutting out the chips with a pair of fine snips, leg by leg might be safest.  It sounds tedious as hell to me, but does make the pins "easier" to remove.  The chips are out of the way and each leg can be individually heated with fresh solder and a little flux, then gently pulled out with pliers.  All 256 of them!  or, 240 if we leave the NEC dram in place. 

[xrbrevin]

dont forget the Freddie could also cause this issue

[moonlight_mile]

I have a hot air and desoldering station so it shouldn’t be too too bad to get the chips out.

[gilsaluki]

41 minutes ago, moonlight_mile said:

I have a hot air and desoldering station so it shouldn’t be too too bad to get the chips out.

If you were hesitant WITH a hot air and desoldering station, I will never try it with a soldering iron alone.

[moonlight_mile]

If you were hesitant WITH a hot air and desoldering station, I will never try it with a soldering iron alone.

Lol. I was hesitant long before I got the desolder and hot air station.

Like I mentioned it just popped back into my mind recently. I figure maybe I can fix it up and pass it on to someone who could use an xe. It seems to be in fairly good cosmetic shape.

[gilsaluki]

I wanna upgrade a 65XE to 128K, but the traces are what is preventing me from doing it. 

[TGB1718]

Your going to bin all the MT chips so just cut the legs off and remove the posts remaining

in the board, will much reduce the possibility of damage.

[+bf2k+]

I would replace all the mtRam as a first step.  In October 21 when I got all of my machines out of storage and started testing, two of my "working" 130XEs went bad in the first week of testing.  That means they were working and failed almost immediately. The mtRAM WILL FAIL... get it out of there.

 

And as TGB1718 said above, just cut the legs off them... makes it easier and less likely to damage the board.

[+scorpio_ny]

I went through the same process. In my case, my computer was still working but I wanted to be proactive (I was working on another project with it and needed to open it up).  The motherboard was still in very good condition, so I did not have to clip the legs.

 

 

 

 

 

 

 

 

 

 

 

 

Edited April 14, 2022 by scorpio_ny
Remove extra images

[moonlight_mile]

Well all I can say is wow!

I ran syscheck. All passed. 
 

Then I switched the resistors and everyone of the chips are bad. Well I ordered new sockets. I had a set of known good 4164’s I got for my atr so I am going to use those and order another set of 4164s for the main bank. 
 

So I know the problem appears to be pretty obvious,  it is there a chance that something else could be wrong causing syscheck to show all chips bad?  
 

Just trying to get all my bases covered before I start unsoldering. 

[Beeblebrox]

@moonlight_mile  I've know Sys check II to report all RAM or some ram chips failing. Generally in those cases and especially if we are talking MT ram branded chips, best policy it to get rid of all MT RAM chips, socket and replace them with non MT branded ones.

 

I'd just socket and replace ram and re-check. More than likely Sys check II will then show a clean run.

 

It's also worth socketing the MMU (61618) while you are in there. May need to replace it if the new RAM is still showing issues.

[moonlight_mile]

What is the safest temps to set my hot air and desoldering station at that won’t cause too much trauma to the traces?

[+scorpio_ny]

When removing the RAM on my 130XE, I had set my TS-100 digital soldering iron to 350 degrees Fahrenheit (176.667 Celsius). This soldering iron same  a small tip. A lot of industrial circuit boards use solder that have more tin than lead, so I had applied 60/40 mix solder to each pin to make it easier to remove. After that, I then applied heat to each of the pins so I can remove the solder with my Engineer SS-02 solder sucker desoldering pump. I was mindful to not keep the soldering iron on too long on the pin.

[Peri Noid]

180 deg. Celcius is a pretty low temperature. My soldering iron, when idle, sets itself to 200 while I use it @300 deg. Celcius. I never fried anything due to temperature.

[+scorpio_ny]

On 4/18/2022 at 5:58 AM, Peri Noid said:

180 deg. Celcius is a pretty low temperature. My soldering iron, when idle, sets itself to 200 while I use it @300 deg. Celcius. I never fried anything due to temperature.

You are quite right! My apologies. I had a bit of brain fog there. I used my soldering iron at 350 Celsius on the lower side when I removed the ram chips.

 

[moonlight_mile]

I haven’t gotten to work on it yet. Been a little under the weather and want to be on the top of my game when I start working on this. Will keep ya all posted.

[moonlight_mile]

Well, I finally got the ambition to work on the chips. Everything went well not lifted pads or anything. 
 

I put in my known good chips and alas. A red screen of death. Lol. I took continuity tests on all the chips and they all connect somewhere else on the board except pin 1 on all the chips. I looked at the schematic and it doesn’t appear that pin 1 is connected anywhere but I might be mistaken. 
 

Does pin 1 get connected anywhere. 
 

I was exhausted after finishing up I didn’t have the energy to continue to troubleshoot. I have to do the hot chip test and run syschek, but as it stands it appears all the sockets and chips are in and connected.