• Please review our updated Terms and Rules here

Ibm pc 5150a..... Possibly a pre-release build???

SUMMARY

By using the quoted 'minimum diagnostic configuration', PCFreek has isolated the problem to the motherboard.

Motherboard appears to be 'dead'.

Not a motherboard switch setting problem.

Proven not to be the BIOS chip.

A failure of one or more RAM chips in bank 0 (soldered in) is one cause of the symptom (one cause of many). The piggyback technique was unsuccessful, but the technique is known not to work for certain failure modes in RAM chips. Therefore, failure of bank 0 remains as a possibilty.

Reseating socketed chips did not work (post #40 instructs).

Unknown as to whether or not PCFreek has swapped in the good CPU chip from his other 5150.
 
Something else to try. A 'long shot'. Remove the BASIC ROMs (U29/U30/U31/U32) in case one of those is the cause (e.g. affecting data bus).

On a 5150 motherboard with the 10/27/82 BIOS (as yours is), the motherboard can operate without the BASIC ROMs (but obviously, cassette BASIC is not available).
 
modem7 said:
Reseating socketed chips did not work (post #40 instructs).

Unknown as to whether or not PCFreek has swapped in the good CPU chip from his other 5150.
Click to expand...

I went to bed last night a little bummed that I had not accomplished much in the way of testing... then woke up this morning to find all these helpful posts!!! It's like waking up Christmas morning and finding gifts under the tree!

When reading the posts, I realize two things... #1 you guys are pretty darn smart... #2 I have an awful User ID haha

Yes I reseated all of the socketed chips to no avail.

No, I have not swapped the CPU chip. I will try that when I get home this evening! I can't wait!

modem7 said:
PCFreek has a 5150 motherboard that neither issues a beep at the end of POST, nor displays video. Therefore, cause not related to motherboard switches.
Click to expand...

One correction here... early on, I did not have the later 5150 and had neither monitor nor video card. Now that I have them, I have never hooked the 1981 motherboard to a display. I have the motherboard, speaker, and power supply out of the case and on my coffee table. I hope that this is not dumb to expect beeps. I thought it would beep after failing to detect the video and keyboard.

modem7 said:
Something else to try. A 'long shot'. Remove the BASIC ROMs (U29/U30/U31/U32) in case one of those is the cause (e.g. affecting data bus).

On a 5150 motherboard with the 10/27/82 BIOS (as yours is), the motherboard can operate without the BASIC ROMs (but obviously, cassette BASIC is not available).
Click to expand...

I'll try this too! Thanks!
 
Last edited:
per said:
Has anybody considered if there is a bad cap on the board?
Click to expand...

I have examined the caps individually under magnification (as discussed earlier in this thread... sorry it is getting so long), they all look fine. I did use an ohm meter to make sure that the 3 caps near the power supply connectors have continuity and they appear fine. Since the power supply voltages are good/stable with the motherboard attached, we have determined that caps are not shorted. I have seen many other posts where people use 5150's with several exploded caps as many of them are for filtering purposes only (as I understand it anyway).

If you have a suggestion as to how I can test them with a digital multi-meter (without unsoldering a leg), I will gladly test them all.
 
PCFreek said:
If you have a suggestion as to how I can test them with a digital multi-meter (without unsoldering a leg), I will gladly test them all.
Click to expand...
If the power is good with the board plugged in, then it's unfortunately not the case. The caps are only for filtering as you mention, so bad caps are quite easy to deal with.

Do you have an oscilloscope or logic probe? You can easily do some crude tests of the CPU with one, like test for activity on the address lines and such. Another posibility would be to use a custom boot ROM that basically does something simple like sounding the speaker immediately.
 
And one thing I thought of as I drove to work today...

I have the Olivetti (Xerox PC6060) that has 18 socketed RAM chips. I can replaced the soldered RAM chips in bank 0 with sockets from the Olivetti and then swap RAM til my heart is content.

My one concern about this though... this is a VERY early 5150 s/n 0103XXX... oldest one that I have seen documented anywhere online. I do not want to do anything to diminish its value and would like to keep it as original as possible. I also want to keep as many 1981 date coded RAM chips in the motherboard as possible. While adding sockets will allow me to keep the 1981 chips in place and determine which ones are faulty, it would alter the board. Any thoughts on this??
 
Last edited:
glitch said:
I wouldn't bother...it's likely anyone with a programmer has a pile of 2764s, they're extremely common. I have both a pile of 2764s and an EPROM programmer, so PM me if you can't find anyone local to program one for you.

As to testing/replacing 4116 RAMs, I usually use a modified 4164 RAM. They're more reliable than 4116 RAMs in my experience. I also have a huge pile of 4164s if you need them.
Click to expand...

I have PM'd you regarding the diagnostic ROM.
 
MikeS said:
Ouch! I wonder if the 4164(s) survived 12V Vcc...
Click to expand...

Good question...at least PCFreek can test it with his working 5150 once it's modified.

I wouldn't bother stealing sockets from another board. If it were me, I'd find replacements (4116s if you want to keep the "originality" for value) and test them thoroughly in the working 5150, then solder them in. If you decide to sell it in the future, it should help keep the value up.
 
glitch said:
Good question...at least PCFreek can test it with his working 5150 once it's modified.

I wouldn't bother stealing sockets from another board. If it were me, I'd find replacements (4116s if you want to keep the "originality" for value) and test them thoroughly in the working 5150, then solder them in. If you decide to sell it in the future, it should help keep the value up.
Click to expand...

The working 5150 is a 256k motherboard while this one is 64k. I am under the impression that the RAM chips are not compatible between the 2. As for the 4164's... while I may be somewhat embarassed for doing something dumb, I have about 20 more of them.
 
PCFreek said:
The working 5150 is a 256k motherboard while this one is 64k. I am under the impression that the RAM chips are not compatible between the 2.
Click to expand...

If i can indeed test 9016 chips in my 256k motherboard, I can select 8 good ones and replace bank 0. Please advise if this is possible. I now realize we are dealing with different voltages. I also do not want to unsolder and resolder all the chips in bank 0 of the 256k machine so I am thinking that this is probably not possible due to the mix of voltage requirements. Of course, i would consider soldering sockets into bank 0 of the 256k machine.... hmmmm....

UPDATE: I just purchased 20 NOS 9016 chips... all 1981 date code. I may just start at the parity chip and replace bank 0 one chip at a time. First though, I will try to boot it after switching processors and also after I remove the Basic ROMs.
 
Last edited:
I swapped the CPUs... no beeps...

I removed the Basic ROMs... no beeps...

I piggybacked every RAM chip in bank 0 with 9016 chips and mixed them back and forth a few times with the Basic ROMs removed and the new CPU installed... no beeps... I am more and more tempted to solder in sockets in bank 0. I can always remove them later.

Thanks to glitch, I have a diagnostic ROM in my future and an adapter on order. Let me be sure I understand: with a diagnostic ROM, I should just wait before doing any soldering on bank 0 as this ROM can tell me where the POST fails?
 
Last edited:
PCFreek said:
Thanks to glitch, I have a diagnostic ROM in my future and an adapter on order. Let me be sure I understand: with a diagnostic ROM, I should just wait before doing any soldering on bank 0 as this ROM can tell me where the POST fails?
Click to expand...
The Power On Self Test (POST) sequence of the 5150 is listed [here]. A failure in many of the early tests results in the symptom that you are seeing (motherboard appears dead) because the POST simply halts the CPU.

The diagnostic ROM replaces the IBM POST. One of the first things it does is to initialise the video card (requirement: MDA or CGA only), and thus when it runs its early tests, it can display the results of those on-screen (rather than simply halting the CPU as the IBM POST does).

If the bank 0 RAM is indeed faulty, the diagnostic ROM will show that.

But, as you will be aware, the problem in your motherboard may be something else. And if it's very low level, the diagnostic ROM is not even going to get the chance to run.
 
While waiting on the diagnostic ROM and adapter, I have been cleaning/testing the other components on my newer 1985 5150. I am very pleased that both floppy drives and the floppy drive controller are working perfectly. This gives me hope that the motherboard was also not adversely affected by the storage environment.

So my summary as of now, the following items from the 1981 PC have been found to be fully functional:
Power supply
2 FDD's
FDD controller card
Graphics card
Speaker
BIOS ROM
Processor
Original keyboard with metal plug

I have twenty 1981 date coded 9016 RAM chips on order, a ROM adapter on order, and I believe that glitch is providing me with a diagnostic ROM. Almost there!!
 
OK... finally ran the Diagnostic ROM (thank you so much glitch!!!) and the errors can be seen in the attached photo. I was very happy to see that it ran and completed the diagnostic utility. That's much better than poking around in the dark. Now I would appreciate any feedback and will start researching these errors. As stated earlier in this thread, this motherboard is a 16KB to 64KB very early 1981 5150 that was stored in less than ideal conditions.

This testing was performed with just the motherboard, power supply, keyboard and monitor. Nothing else attached. Not even the speaker. From what I read, U28 should have failed because it is empty... but it passed... hmmm.... I saw 4 failures in the RAM (more registered after this photo was taken) and I believe that I can handle those as it shows an "X" under 0, 3, 5, and 6 with a variety of codes each pass (error at address: 00000, 08000, 08004, 08012, 08052, 08054, 08072). As you can gather from past posts, I am not overly knowledgeable about these PC's, so I apologize if the problems are something obvious that I should know.

P4220268.jpg
 
Last edited:
Seems like the 8253 timer is acting up, and one of the things the timer is responsible for is generating the refresh signal for the DRAM cirquits. If the 8253 is dead, the DRAM won't be properly refreshed, and all you will get from RAM reads will be random data (unless the read is directly following some writes to the same RAM addresses).

since the CPU runs, the 8284 clock generator problably works. The fault is therefore problably either in the clock-divider chip (U26) or most likely in the 8253 itself.
 
You must log in or register to reply here.
Back
Top