IBM 5160 - Known Problems/Issues
Flawed methodology for determining total conventional memory size
In the 5160's predecessor, the 5150, the Power On Self Test (POST) is informed of the total amount of conventional memory (on motherboard + on expansion cards) by way of the switches on SW2.
In the move to 5160, it seems to me that the designers of the 5160's POST thought that they had come up with a clever alternative to SW2. To establish the total amount of conventional memory, the 5160's POST performs a write/read to the first two addresses of each successive 64K block. As soon as the POST finds a block where the bytes that were read back do not match the bytes that were written, the POST decides that it has encountered the end of conventional memory.
That methodology is flawed. Why? In most cases, if the POST encounters faulty RAM, it will incorrectly assume that it has reached the end of conventional memory.
In my IBM 5160 that has a 64-256KB motherboard, if I pull out a chip from the fourth bank of RAM, and then power on the 5160, I do not see the 201 memory error that I expect. Instead, what happens is that the 5160 counts up to 192 KB of RAM instead of 256 KB. When DOS later boots, DOS thinks that only 192 KB of RAM is fitted.
I want to increase the RAM in my IBM 5160 from 256 KB to 640 KB. I buy a RAM expansion card, configure the card appropriately, and fit the card. But the 5160 still only counts as high as 256 KB. One or more faulty RAM chips in the first bank of RAM on the expansion card is one possible cause of this symptom.
In some circumstances of RAM chip failure, the POST will correctly determine the size of conventional memory and go on to produce a 201 error. It depends on the type of failure of the RAM chip/s. Click here to see examples.
At some point, IBM must have realised that the methodology used in the 5160 was flawed, because in the 5170, IBM reverted to a mechanism where the machine is 'told' by the user/technician as to how much conventional memory is fitted.
Short circuited tantalum capacitors
It is common for tantalum capacitors on the motherboard and on expansion cards to go short circuit, stopping the power supply from working (although the power supply fan might still turn).
It will appear that the motherboard is 'dead', although to be noted is that that symptom has many causes.
If you suspect that your 5160 may have a shorted tantalum capacitor, then try the diagnostic procedure here.
Floppy disk drive problems
Early 5160s were fitted with a full-height floppy drive, which was typically the Tandon TM100-2 (or TM100-2A), or the Micropolis equivalent.
Click here for issues associated with the TM100-2 (or TM100-2A), although most of the content also applies to the Micropolis.
POST cards do not work
Information on this subject is here.
Cards put in expansion slot 8 must be slot 8 compatible
Information on this subject is here.
16-bit ISA cards
Some 16-bit ISA cards advertised as being 8-bit slot compatible, may not work in the IBM 5160.
Here are some possible causes:
• When the maker wrote "8-bit slot", they meant "8-bit slot in an AT-class computer". This kind of requirement is usually found in the card's user manual.
Click here to see an example.
• Some 16-bit cards need to be reconfigured for 8-bit slot operation. That configuration is sometimes done by switches or jumpers. On at least one VGA card, it is done via the card's configuration software.
Possible ROM conflict - hard disk controller and VGA
• In the IBM 5160, IBM reserved the 32KB sized address space of C0000 to C8000 for any BIOS expansion ROM on a video card. Refer here for a diagram.
• The ROM on XT-class hard disk drive controllers typically starts at C8000.
• According to Microsoft's Q63588 article, the ROM on some VGA cards extends past C8000 (quote: "A few cards may extend up to CA7F.")
Possible ROM conflict. See the diagram here.
Maybe none of the offending VGA cards are capable of being used in an IBM 5160. This possible ROM conflict is just something to 'keep in the back of your mind'.
BIOS ROM replacement
Early 5160 motherboards are wired for a Mostek MK38000 series ROM in the BIOS ROM sockets (U18 and U19). On such motherboards, a replacement ROM of type 27256 (or 27C256) usually works, but not always. See here for more information on this subject.
Swapping out the 8088 CPU for a V20 CPU can sometimes cause problems. See here for more information on this subject.
Motherboard BIOS revision of 05/09/86
Use of this BIOS revision may result in third-party floppy controllers being unable to boot from 1.44M diskettes. See here for more information on this subject.
RAM failure in motherboard bank 0
Vintage RAM chips have a relatively high failure rate.
The failure of any chip in the 5160's first bank of RAM (bank 0) results in what appears to be a 'dead' motherboard.
Click here to see a diagram that shows the four RAM banks.
Note that rather than a RAM chip actually failing, what might have instead happened is that a RAM chip has developed poor electrical contact with its socket.
And so the first thing to try is to simply to reseat all of the chips in bank 0 (wiggle each chip in its socket).
For the 5160, one way of diagosing a faulty RAM chip in bank 0 is to fit the bank 1 chips into bank 0 (assumption: the chips in bank 1 are all good).
If your 5160 motherboard is the 64-256KB type, then you have the option of using RAM chips from banks 2 and 3 as well - all 4 banks use chips of the same type (generic type 4164).
If your 5160 motherboard is the 256-640KB type, then you do not have the same option - the chips in banks 2 and 3 are a different type (generic type 4164) to the type used in banks 0 and 1 (generic type 41256).