• Please review our updated Terms and Rules here

IBM 5150 and Everex RAM extender

stangman517

Experienced Member
Joined
Dec 15, 2014
Messages
243
Location
Georgia, USA
Greetings all.

New to this site, and I do not know my way around. Took me 20 mins to figure out how to post, and I hope I'm not posting in the wrong place or opening a forum already open with this subject. If so I humbly apologize. Within time I will get with the program.

Here's my question. Just obtained a 5150 with the Everex board with six banks populated. I understand each bank is 64K totally 384K. When I first received it the HDD ST-225 didn't boot. I used an old DOS pgm. from Microprose that actually let me low-level format it; then booted again with MS-DOS 3.3, fdisk'd it, and then formatted it. Booted fine; I was so proud. :p But on next reboot I kept getting 4055 201. You 5150 vets know right away what this is. I knew 201 was memory, and it took me a while to understand the 4055 and how to decipher it. Originally I thought it was a problem with the memory on the motherboard, but before the last boot MicoProse said all RAM was good on the motherboard, and I believe the Everex board(not certain on the board though). I found the 5150 Technical on how to set the MB up for 256KB and that's a-ok, but I couldn't quite decipher the guide on what the SW1 and SW2 should be when using this Everex board, nor was I sure of the settings on the Everex board; I couldn't find any docs on the Everex at this time.

Does anyone know why the board first worked when I first received the board then after formatting the ST-225 it I then started getting the 4055 201?? I have SW1 and SW2 set for 256KB on the MB and the system boots fine with the card in it but I cannot figure out what the switches should be set to for the 5150 to see the extenders RAM. Any help would be very much appreciated!

Thanks!

P.S. Always wanted a 5150 when I was younger but was wayyy out of my range then!
 
Welcome to these forums.

There is lots of IBM 5150 information at http://www.minuszerodegrees.net

We know from your post that you have the '64KB-256KB' type of 5150 motherboard.

Just obtained a 5150 with the Everex board with six banks populated. I understand each bank is 64K totally 384K.
I found the 5150 Technical on how to set the MB up for 256KB and that's a-ok, but I couldn't quite decipher the guide on what the SW1 and SW2 should be when using this Everex board
Refer to [here].

For RAM, SW1 concerns the motherboard only. On SW1, you would set switches 3 and 4 to the "Banks 0/1/2/3 populated" setting to indicate that all 4 motherboard RAM banks are populated.

Your Everex board needs to be configured to start its 384 KB of RAM at the 256K address (and it sounds like it already is). In that way, you will have 640 KB total conventional RAM. You then set the SW2 switches (total conventional RAM) to the 640 KB setting.

Does anyone know why the board first worked when I first received the board then after formatting the ST-225 it I then started getting the 4055 201??
There is no link between those two.

The '40' in '4055 201' indicates a problem at address 256 KB. The '55' in '4055 201' indicates many bits in error, many bits typically seen when there is no RAM at all at the address.

So, with SW2 set for 640 KB, we expect '4055 201' to be be produced on power-on whenever the Everex card is absent.

I have SW1 and SW2 set for 256KB on the MB and the system boots fine ...
Per my earlier statement, SW2 is set for total RAM. By setting SW2 to 256 KB, the only RAM that the 5150 will recognise (of what is fitted) is the first 256 KB.
 
Ok I think the board has problems. After trying and trying to get this board to work using your advice modem7 I think it's time to believe the troubles are with this Everex board. I took a closer look with a magnifying glass and bright light I saw some blue like material on some of the ICs and resisters close to the battery. It doesn't look like the battery has leaked. I very carefully cleaned some of the blue like stuff off but on the RP2 block of 22ohm resisters close to the battery I saw one where of the leads that goes into the resister was partially open. It looks like the cap on the resister has opened. Its strips are red red black gold. I am not an engineer or electronics guru, so if these are not 22 ohms please let me know. When I test resistance of all the other resisters of this type they measure 22 ohms, but the one with the lead slightly separated from the resister my meter registers above 100 ohms. I do not have a clue why. Does this mean it's bad? If you think the board is basically gone passed its days I may just try to replace the broken resister with a new one for practice. If you know what that blue like material is please let me know too. I've seen blue corrosion on car batteries before.

Thanks
 
Ok I think the board has problems. After trying and trying to get this board to work using your advice modem7 I think it's time to believe the troubles are with this Everex board.
Everex produced various cards. Are you sure that the switch settings information that you are using, matches your Everex card?

If you know what that blue like material is please let me know too.
Assuming that the "blue" that you are seeing is the result of a leaking NiCad battery, then the answer may be in the discussion at [here].

A related photo is [here].

I very carefully cleaned some of the blue like stuff off but on the RP2 block of 22ohm resisters close to the battery I saw one where of the leads that goes into the resister was partially open. It looks like the cap on the resister has opened. Its strips are red red black gold. I am not an engineer or electronics guru, so if these are not 22 ohms please let me know. When I test resistance of all the other resisters of this type they measure 22 ohms, but the one with the lead slightly separated from the resister my meter registers above 100 ohms.
Photos will help us.
 
I have included a few pics; they're not the best, but I believe they'll show - hopefully - some helpful information.
I used the settings based on the pic from total hardware 1999 and your pointing out the settings for 4, 5, 6 from the doc was wrong. So comparing my card with the pics from total hardware 1999 I believe I have them set correctly. I know they're correct on the motherboard.
The link on how to clean the board was very helpful; I cleaned it with hydrogen peroxide and used it sparingly, dried it well. NO MORE blue corrosion.
I also downloaded the mf-100 utils like setclock.com and getclock.com and they pull the date/time and I use them to set the system date/time for my 5150 fine. I also tested the battery and it holds power - don't know if this means anything, but it holds a charge.

I still get the 4092 201. I hope the pictures show you something.

I appreciate your time modem7!

1228142245.jpg
1228142245a.jpg
1228142246.jpg
1228142246b.jpg
 
I also downloaded the mf-100 utils ...
So this is the Diamond Flower MF-100 card that you are discussing in a different thread, not an "Everex" card. Up until now, I thought that you had two separate cards that you were trying to get working.

We know that your MF-100 switch settings (on/off/off/on/on/off) are correct for your situation, and your photo shows that they have been set correctly on the MF-100.

We know that your 5150 motherboard switches (the RAM related ones) are set correctly. The two ones in SW1 are at least set correctly, because the 5150's POST is finding and testing the 256K on the motherboard. SW2 is set for something above 256K, because the 5150's POST is attempting to test RAM above 256K.

Everything is indicating a RAM related problem (RAM or supporting circuitry) on the MF-100.

You will have pulled out and plugged in the MF-100 at least a few times, and so I am confident that a poor edge connector connection does not exist. It would not hurt to try a different slot.

I still get the 4092 201.

The "40" component indicates a problem at the 256K address, so, with other possibilities ruled out, we know that the problem is in the MF-100.
The "92" indicates bits 1, 4, and 7 are in error (92 hex = 10010010 binary)

In an earlier post, that "92" was "55", but I suspect the 55 appeared when the card was removed (55 hex is one of the test patterns used in the RAM test).

I just now revisited the 'other' thread on this, in which some posts have appeared. There, you indicate "4098", not "4092". Is the code changing periodically, or is there a typo somewhere?

Some thoughts of mine:

* Faulty switch block
* Many bad RAM chips
* Wrong RAM chips installed
* RAM chips installed in wrong orientation
* A failure on the MF-100 that results in the RAM (all or part) not being seen.

The switch block can be tested using a multimeter in resistance mode. For each switch in the block, put a probe on each of the switches two contacts. There show be no (or low, eg. less than 1 ohm) resistance when the switch is in the ON position, and high resistance when the switch is in the OFF position. Because the switch is being measured in-circuit, other components may affect the readings for the OFF position, but you should definitely measure less than 1 ohm for the ON position.

The other thread indicates that you have been swapping out the RAM chips that are indicated as failed, to no avail (same error code seen). The act of swapping should have also cleared any poor electrical contact that may have developed between chip and socket.

If some RAM chips are in the wrong orientation, I am sure you would have spotted that. If all are in the wrong orientation, then I would expect many more bits in error.

One thing that is significant is the battery leakage.

The link on how to clean the board was very helpful; I cleaned it with hydrogen peroxide and used it sparingly, dried it well. NO MORE blue corrosion.
Of course, the question is now, was any damage done by the leakage?

but on the RP2 block of 22ohm resisters close to the battery I saw one where of the leads that goes into the resister was partially open. It looks like the cap on the resister has opened. Its strips are red red black gold. I am not an engineer or electronics guru, so if these are not 22 ohms please let me know. When I test resistance of all the other resisters of this type they measure 22 ohms, but the one with the lead slightly separated from the resister my meter registers above 100 ohms.
I hope the pictures show you something.
I would need to see in-focus photos.

I also downloaded the mf-100 utils like setclock.com and getclock.com and they pull the date/time and I use them to set the system date/time for my 5150 fine. I also tested the battery and it holds power - don't know if this means anything, but it holds a charge.
That functionality will be implemented by different circuitry to the RAM related circuitry. There could be one or more common chips involved, but if so, I do not think that it would be many.
 
Sorry meant 4098 201.
Tested the switch block on the card; results: On = 0.2 ohms, Off = about 5.08K ohms.
RAM Chips: variations of different manufacturers; concerned me from the beginning - maybe means nothing?
Manufacturers: MCM6665BP15, NEC 8507RK158, TMS4164-20NL, MT4264-15, KM4164A-15.
Orientation: All correct.
Concern about one 22 ohm resister - lead was separated from base - pushed it together. May be a problem or the problem.
When I first received the 5150 the card worked because I was able to run software to low-level format a ST-225 drive.
Damaged from leakage: appears very likely, but I wait for your expert ideas.

Thanks modem7!

Was unable to take clear close ups due to cell phone quality of the resisters.
I found plenty of NOS MCM665BP15 chips; didn't check on the others. Do you recommend I buy all NOS of one manufacturer of a chip?
 
RAM Chips: variations of different manufacturers; concerned me from the beginning - maybe means nothing?
Manufacturers: MCM6665BP15, NEC 8507RK158, TMS4164-20NL, MT4264-15, KM4164A-15.
That "NEC 8507RK158" will be an NEC upD4164, per the list at [here].

On modern computers, there is a reason why same-make/model RAM is required, but for these 'ancient beasts', I am unaware of a reason for same-make/model.
The speed rating of the chips you have quoted is adequate for the 5150.

Sorry meant 4098 201.
On the IBM 5150, 5155 and 5160, you would logically think that if non-existent RAM was tested, then all bits would be reported as being in error, i.e. bit indicator of FF hex. But in my experience, you do not see that; you see multiple bits in error; multiple but not all.

And, different motherboards can produce different bit indicators. According to my notes, one of my 5150 motherboards produces 94. But interestingly, I just now pulled out one of my 5150s, one that has a fully RAM populated 64K-256K motherboard, and a RAM expansion card filling 256K to 512K. I removed the expansion card, and guess what? The 5150 shows on screen, "4098 201", a bit indicator of 98, just like yours.

That is not proof that the RAM functionality on your MF-100 card is bad (in a serious way), but it certainly raises the probability. 98 corresponds to bits 3/4/7. In bank 0 of your MF-100, if you swap the bit 0 and bit 3 chip and then see the bit indicator change from 98 to 91, then you have chips at fault. If it stays at 98, then I think we conclude: the RAM functionality on your MF-100 is bad (in a serious way).

When I first received the 5150 the card worked because I was able to run software to low-level format a ST-225 drive.
Maybe the battery leakage has done damage, subsequent to the low-level format.

Concern about one 22 ohm resister - lead was separated from base - pushed it together. May be a problem or the problem.
Could very well be. It would certainly be high on my investigation list.

Was unable to take clear close ups due to cell phone quality ...
So, no camera that has a macro facility?
 
Replace that resistor before continuing; pushing it back together will not make a connection.
 
Hello modem7.

Ok took some more shots (macro) of the bank with the resisters. I can actually see them better from the pics than with my magnifier/bright stick, AND they don't look too good to me. Please tell me what you think.

Re: the RAM chips, I found a guy on eBay and I bought 60 MT4264-15 chips. Was planning on replacing all the chips not just a single bank. What 3pcedev said makes sense. I guess I was hoping pushing the lead back toward the center of the body of the resister MIGHT make some contact. I believe I will give it a try and replace that resister. I ordered some off eBay, and will attempt it as soon as I get the. Eh my PCB experience is neophyte, but I'm learning - esp. from this site. Gotta say again, what an amazing and fun site I found!!!!

DSC05315.jpg
DSC05316.jpg
DSC05317.jpg
DSC05320.jpg
DSC05327.jpg


Please all let me know your thoughts.
 
Ok took some more shots (macro) of the bank with the resisters.
What a difference.

AND they don't look too good to me. Please tell me what you think.
The ones that measure 22 ohms must be okay. They might not look 100%, but they are still achieving their electrical purpose/specification.

However, down the track, that could change. Battery leakage may have already started a damage process which is still progressing. Yes, you've cleaned the card, but the battery leakage may have already worked itself into various components. Time will tell.

I will give it a try and replace that resister.
Fingers crossed.
And you need to remove the leaking battery.

What is interesting (to me) about your photos, is that the shown FCC sticker indicates an Everex EV-100, but the card is elsewhere labelled as "MF-100" and matches the TH99 entry for the Diamond Flower MF-100. Possibly, Diamond Flower acquired the card rights from Everex, or perhaps Everex made the card for Diamond Flower.
 
modem7 said:
What is interesting (to me) about your photos, is that the shown FCC sticker indicates an Everex EV-100, but the card is elsewhere labelled as "MF-100" and matches the TH99 entry for the Diamond Flower MF-100. Possibly, Diamond Flower acquired the card rights from Everex, or perhaps Everex made the card for Diamond Flower.


Thats normal.. Back in the days most stuff where produced by DFI and sold by Everex or an other resellers.. Same for the `brand` Modular circuit technology that was the home brand of JDR Microdevices. But actually most of that stuff was made by Diamond Flower INC. I can tell, because i have two same 286 boards here, one have the ID of MCT (modular circuit technology) and the other doesnt, but comes directly from the DFI factory i guess..
You can see it on that oval grey sticker(quality check) thats sticked on the innerside of the cards bracket..

attachment.php


$_57.JPG


That paper sticker above says MCT236631.. So this one was sold by JDR microdevices as `modular circuit technology`. And you also see its have the same quality check sticker on the right.

$_57.JPG


This board doesnt have the MCT branding.. But you see its the same board as above with the same quality check sticker.. This one is made by DFI.. So back in the days resellers, made their own fantasy brands, but in fact they came from the same factory.

So the conclusion is, that his EV-100 Multifunction board was sold by Everex original, but it was produced by DFI as the MF-100 board.

Diamond Flower acquired the card rights from Everex.
Yes this could be possible, but i think that Diamond flower made (or relabled some card for everex) (why should everex produce these cards if DFI already have the facility / factory.?)
 
Last edited:
I'll remove the battery and replace the 22 ohm resister as soon as they arrive, so yes cross your fingers for me. ;-)

Thanks modem7.
 
Hello 3pcedev.

I'll take off the battery but will miss its feature; I use the real time clock on this board to set the 5150 system time using the setclock.com in the autoexec.bat file. Otherwise I have to manually set the date/time on reboots. So is there another card/way I can achieve setting the system date time on reboots without cmos?

Other note: A few days ago I received the original 5150 keyboard, and it must have had 20+ years worth of gunk/grime/dust in it. The surface under the keys seem to have rusted off some of the black paint. Hope it won't get worse over the years. I cleaned it real well; works like a champ, and I absolutely love the keyboard. Wow it takes some getting use to compared to our modern keyboards.

Thank you very much for your time; I appreciate it!!
 
Last edited:
Thank you Robin4 for your information!!! VERY interesting. I guess licensing of products by the manufacturers and resellers have always worked that way.
 
I'll take off the battery but will miss its feature; I use the real time clock on this board to set the 5150 system time using the setclock.com in the autoexec.bat file. Otherwise I have to manually set the date/time on reboots. So is there another card/way I can achieve setting the system date time on reboots without cmos?
I presume that the battery is a 3.6V NiCad, standard for that application.
If so, you could solder in a 3.6V NiMH replacement (Oz eBay example [here]).
 
You can always do what I did and move the battery off-board. Its a much safer option and it eliminates the risk of further damage in the future.

This can be done either with alkaline batteries or the factory 3.6V NiCd/NiMH. I go into a little bit of detail in the thread here: http://www.vintage-computer.com/vcf...-My-first-(hand-assembled)-computer-486DX2-80

Read the thread right to the end - I ended up with a lot more elegant/working solution
 
Back
Top