5150 Gen1 and lo-tech XT Memory Board question

[BarryM]

Hey all. Looking for some advice on how best to proceed. I’ve got a 5150 from 1981 with a fully populated 16-64k motherboard, the 10/27/82 BIOS, V20 CPU, Paradise VGA, 2 * 256k memory boards, and the 5161. It’s all been reasonably well stored over the years and I got the urge to fire it up again recently to see what was on the old MFM drives. I was able to boot with the Lo-tech XT-IDE board and was going to run some more tests, but ran into a bit of a snag.

Upon power up, I’m getting POST codes that says one of the removable 4116 RAM chips has developed a fault and the system halts. Since I wasn’t sure about the overall health of the rest of the memory in general, I ordered up one of the Lo-tech memory cards to see how well it works.

When it arrives, my plan is I’ll configure the Lo-tech board to provide the base 0000h page and disable 0-16k.

Am I correct to assume I’ll have to also physically remove the motherboard’s socketed 4116 chips?

Is it also correct to assume that with the Lo-tech board installed I should not change any of the motherboard SW1 settings to reflect this removal?

As an aside, if I wanted to try to fix the 4116 issue with a new chip, are there any known issues when using 150ns memory with this old system? Couldn’t find a good local source of speed matched stuff. I’m assuming if I replace one chip with higher speed stuff that I’ll have to do the entire bank with the same rating.

 

[Ruud]

Looking at the schematic of the 256 KB version, the RAM memory is selected by a PAL. It activates a 74245 bi-directional buffer for reading and writing the DRAM. My conclusion is that you HAVE to use onboard DRAM. IMHO it will be the same for the 64 KB board with 4116 DRAMs. So I'm afraid you have to find the culprit and replace it. And you have to populate the whole board with DRAM.

 

[modem7]

Upon power up, I’m getting POST codes that says one of the removable 4116 RAM chips has developed a fault and the system halts. Since I wasn’t sure about the overall health of the rest of the memory in general, I ordered up one of the Lo-tech memory cards to see how well it works.

"Removable", so the second, third, or fourth bank.

And "suggests" is more appropriate than "says". For example, this could be as simple as the RAM chip needing to be re-seated in its socket.

Am I correct to assume I’ll have to also physically remove the motherboard’s socketed 4116 chips?

Ruud is right. On the IBM 5150:
- All four banks of motherboard RAM are permanently enabled.
- All the RAM related motherboard switches do, is inform the power-on self test (POST) of RAM amount.
- Removing the RAM chips from a motherboard bank does not disable the bank.

And you have to populate the whole board with DRAM.

A requirement of bugs in the 10/27/82 BIOS.

 

[Malc]

Never tried it myself but this thread may be of interest... https://forum.vcfed.org/index.php?threads/is-it-possible-to-use-only-expansion-ram-on-ibm-pc.49627/

 

[twolazy]

Not to steer too much off topic, but to save creating a semi duplicate question, is the Compaq Portable in the same boat and motherboard dram (populated or empty) before expansion on isa bus? So cant use a 1mb card to fill in first 512?640?

Example be say an AST Six pack. Im a little confused...

 

[BarryM]

Never tried it myself but this thread may be of interest... https://forum.vcfed.org/index.php?threads/is-it-possible-to-use-only-expansion-ram-on-ibm-pc.49627/

Interesting read that thread - thanks for digging it up. There’s probably a similar modification that could be done to this Gen1 PC I guess, but I’m going to try avoiding hardware hacks for now.

My gut feel was that I wasn’t going to be lucky, so I ordered up a full set of 4116 chips, but am still waiting for them to arrive. At least the reported bad chip moved when I swapped it with another chip, so looks like it’s just the IC itself.

 

[modem7]

Not to steer too much off topic, but to save creating a semi duplicate question, is the Compaq Portable in the same boat and motherboard dram (populated or empty) before expansion on isa bus? So cant use a 1mb card to fill in first 512?640?

The original Portable, sometimes referred to by users as the Portable 1 (to help distinguish it from later Portables). I have just now taken a look at the circuit diagram for the Portable 1. Compaq have duplicated what is done in the IBM 5150, that is, switches 3 and 4 on SW1 do not enable/disable motherboard RAM banks. And I cannot see any other bank enable/disable mechanisms. So, all motherboard RAM banks are permanently enabled. And like the IBM 5150, there is a transceiver between the data bus of the RAM chips and the motherboard's data bus, and so removing the chips from a RAM bank will not disable the bank.

Example be say an AST Six pack. Im a little confused...

If I am an operator in a particular nuclear power plant, and the operations manual says, "Do not turn on switch 6 unless switch 13 is in the on position", then I do what the manual instructs. I do not need to know the technical reason why. If I'm curious, then I need to become educated in the technical workings of the particular nuclear power plant. Nuclear power plants come in different designs, and so I understand that at nuclear power plants of different design, the switches 6 and 13 rule may not apply.

In the Portable 1, Compaq have functionally duplicated the IBM 5150. Therefore, the RAM rules/requirements are the same:
1. All motherboard RAM banks are permanently enabled. Removing chips from a bank does not disable the bank.
2. Set motherboard switch banks SW1 and SW2 to inform the POST/BIOS of how many of the motherboard RAM banks are populated.
3. All motherboard RAM banks need to be populated before a RAM card (conventional memory) can be added. Adjust SW2 accordingly.

As for point 3, an IBM 5150 example is at [here].

( One thing that I doubt that Compaq duplicated, is the RAM related bugs in the IBM 5150's third motherboard BIOS (10/27/82), bugs that require that all motherboard RAM banks be populated. )

 

[jafir]

I would imagine that even if Compaq did duplicate the bugs somehow, they had more chances to fix them, because there are significantly more revisions of Portable BIOS than PC BIOS.

 

[modem7]

( One thing that I doubt that Compaq duplicated, is the RAM related bugs in the IBM 5150's third motherboard BIOS (10/27/82), bugs that require that all motherboard RAM banks be populated. )

I would imagine that even if Compaq did duplicate the bugs somehow, they had more chances to fix them, because there are significantly more revisions of Portable BIOS than PC BIOS.

Something else in common with the IBM 5150: Per [here], users need to be aware that a BIOS upgrade may require different switch settings.

 

[BarryM]

My gut feel was that I wasn’t going to be lucky, so I ordered up a full set of 4116 chips, but am still waiting for them to arrive. At least the reported bad chip moved when I swapped it with another chip, so looks like it’s just the IC itself.

The replacement 4116 chips finally arrived on Saturday. The replacement of the one bad AMD 250ns chip with the 150ns TI seems to work fine and let me get past that problem.

This also exposed a further POST memory problem with the 256 kb boards that I remembered from ages ago. I was planning to swap them both out with the lotech board anyway to get the full 640 kb and a free slot. That also worked just fine too. Maybe I’ll order up a few 4164s and fix them as well.

Now to have some fun with ISA USB and Intel 8/16 Ethernet!

Thanks for the info and pointers everyone!