Five Festive 5150 Boards (BOARD #5 Thread)

[VeryVon]

Welcome to Board #5!

[Link to Original Thread / Board #3]

Here's what I know so far:
- Motherboard type 16KB-64KB
- Known good CPU inserted
- Known good RAM populated in Bank0, other banks unpopulated
- RDR RAM Tests Pass (see screenshot below)
- Booting with known good "5700671" BIOS gives continuous beeping, no video
- Booting with known good"1501476" BIOS gives continuous beeping, screen says "109C" (one time it showed "1098")

Basically I'm trying to boot it up with one bank populated, I know I have to use the 5700671 BIOS in order to do that.

 

[VeryVon]

Using TEST5067 (which should suffice) I tested for stuck address bits on the ISA slot (A0-A19) and on the external address bus U15, U16 (XA0-XA12.) Logic probe shows activity on all lines.

 

[VeryVon]

I put a known good rom set in there (5700671 BIOS) I now get some video (see below) and message "10C9"



To summarize:
- Known good ROM set + 5700671 BIOS: Video above + "10C9" Message
- Known good ROM set + 1501476 BIOS: "109C" (sometimes "1098") Message only

 

[modem7]

Here's what I know so far:
- Motherboard type 16KB-64KB
- Known good CPU inserted
- Known good RAM populated in Bank0, other banks unpopulated
- RDR RAM Tests Pass (see screenshot below)
- Booting with known good "5700671" BIOS gives continuous beeping, no video

So, let's look at that particular configuration.

With the IBM 5150, I am always aware that if I change the BIOS, or alter the RAM amount, then both switch banks SW1 and SW2 probably need to be altered.

From the information at [here]:
For a 16KB-64KB motherboard, with 10/19/81 (e.g. stamped 5700671) BIOS, fitted with only 16 KB:
- Switches 3 and 4 need to be set to ON, ON.
- Switch block 2 needs to be set to ON, ON, ON, ON, dc, dc, dc, dc (where 'dc' is: do not care)

The 'PC SW2' area of the RDR screen shot in post #1 shows that your SW2 is not set per above, so that is at least one problem.

I will discuss the "continuous beeping" in a following post.

 

[modem7]

Re the "continuous beeping" (which you hear with both 'BIOS). Is that one continuous tone, or a continuous series of beeps ?
I discuss the former at [here].
In post #3, you no longer mention "continuous beeping", and so I presume the symptom disappeared.

 

[modem7]

- Known good ROM set + 1501476 BIOS: "109C" (sometimes "1098") Message only

Do you have the SW1 and SW2 switches set correctly for that particular configuration ?

( FYI: '1098' is an often [not always] quoted number for the 5150's POST testing non-existent RAM. I even make mention of it in the 'Multiple bit failure' section of [here]. )

 

[VeryVon]

- Switch block 2 needs to be set to ON, ON, ON, ON, dc, dc, dc, dc (where 'dc' is: do not care)

Ok, perfect, all set

In post #3, you no longer mention "continuous beeping", and so I presume the symptom disappeared.

Correct, beeping gone now

RDR Shows switches like this:


Now when I replace RDR with 5700671 and boot up, 109C is gone, however it's in a boot loop that looks like this:

 

[modem7]

Now when I replace RDR with 5700671 and boot up, 109C is gone, however it's in a boot loop that looks like this:

If this was a 5160 (not 5150) motherboard, I would say that switch 1 on SW1 is in the wrong position, but this is a 5150 - that switch has different meaning.

This will be the "known good 5700671 BIOS" EPROM that you wrote about in post #1. How do you know that it is known good? Because you have a proven 16KB-64KB motherboard?

What video card is in use?

 

[modem7]

Something that may be very relevant is what is shown in post #1. All of the 'Check ROM at Fx000' checks are failing. On a 5150 motherboard, only the F4000 one is expected to fail (because there is nothing in the motherboard's U28 socket).

Data problem? Address problem?

 

[VeryVon]

This will be the "known good 5700671 BIOS" EPROM that you wrote about in post #1. How do you know that it is known good? Because you have a proven 16KB-64KB motherboard?

Correct, I'm using the complete ROM set from my working 5150. All original ROMs, no EPROMS.

What video card is in use?

STB “The Chauffeur” - same card I've been using for all troubleshooting up until now. I have another video card here, it's a no-name MDM / Parallel card. I tried that. Slightly different behavior. The cursor comes up, then there's a small 1/2 line flash, then cursor again, repeat in loop.

Something that may be very relevant is what is shown in post #1. All of the 'Check ROM at Fx000' checks are failing.

At that stage I only had the U33 ROM in, no others.

Data problem? Address problem?

Must be. Possibly I/O to the video card, or lower ROMS?

 

[modem7]

STB “The Chauffeur” - same card I've been using for all troubleshooting up until now. I have another video card here, it's a no-name MDM / Parallel card. I tried that. Slightly different behavior. The cursor comes up, then there's a small 1/2 line flash, then cursor again, repeat in loop.

That "small 1/2 line flash" is done on purpose by the IBM POST.

When I look at the video in post #7, it kind of looks like the Chauffeur card is doing some kind of self test.

At that stage I only had the U33 ROM in, no others.

Could this be the problem cause! Be aware of the 'Failure' section of [here]. The symptoms that I have listed there for the 10/19/81 BIOS are for when there is a floppy controller and drive fitted. Maybe what you are seeing is what happens if there is no floppy controller and drive.

 

[VeryVon]

Maybe what you are seeing is what happens if there is no floppy controller and drive.

Yes! But a little different! I enabled floppy drive on SW1-1 and fitted the known good floppy controller and drives from my working 5150. Switches 7 & 8 are set to Two floppy drives. Here's what I'm seeing:

With empty drives:
- Floppy light comes on, head moves to track 39 and back
- Basic comes up (see screenshot below)
- Keyboard doesn't work

With known good boot disk in drive:
- Floppy light comes on, head moves to track 39 and back
- Drive runs continuously, flashing cursor

 

[modem7]

Yes! But a little different! I enabled floppy drive on SW1-1 and fitted the known good floppy controller and drives from my working 5150. Switches 7 & 8 are set to Two floppy drives. Here's what I'm seeing:

With empty drives:
- Floppy light comes on, head moves to track 39 and back
- Basic comes up (see screenshot below)
- Keyboard doesn't work

Cassette BASIC appears! So you changed two variables: added floppy controller+drive, and added Cassette BASIC ROM's.

That symptom table that I pointed to is for when a Cassette BASIC ROM is faulty (or missing).

With known good boot disk in drive:
- Floppy light comes on, head moves to track 39 and back
- Drive runs continuously, flashing cursor

Your computer only has 16 KB of RAM. Not even DOS 2.1 will run in that. See [here].

- Keyboard doesn't work

Does the POST report a 301 error if the keyboard is not present ?

 

[VeryVon]

Your computer only has 16 KB of RAM. Not even DOS 2.1 will run in that. See [here].

Interesting! I pulled out the Greaseweazle and wrote a DOS 1.0 floppy, verified it will boot on my 5160. Doesn't boot on this board, drive just runs. I never had an issue with these floppy drives. They're clean and pretty low mileage.

Does the POST report a 301 error if the keyboard is not present ?

No it doesn't, I've never seen 301 appear with or without the keyboard connected.

I re-ran RDR with the keyboard and floppies present. Floppy passes, keyboard fails. I just noticed that ROM F4000 is failing, I could have sworn that was passing before. I'll run RDR a few times and see if that's intermittent. Oh nevermind, just realized F4000 is the empty ROM socket haha.

 

[modem7]

Your computer only has 16 KB of RAM. Not even DOS 2.1 will run in that. See [here].

Interesting! I pulled out the Greaseweazle and wrote a DOS 1.0 floppy, verified it will boot on my 5160. Doesn't boot on this board, drive just runs. I never had an issue with these floppy drives. They're clean and pretty low mileage.

I re-ran RDR with the keyboard and floppies present. Floppy passes,

The 'Trying to read a floppy"' test in RDR reads track {cylinder=5, head=0} on a formatted floppy. That works. But the test is limited; it doesn't test head 1, nor cylinders other than 5, and if there is a problem in those areas, the problem cause can be either controller or drive.

Since the DOS 1.0 floppy is booting on your 5160's controller+drive, how about temporarily moving that controller+drive over to the 5150 motherboard.

I just noticed that ROM F4000 is failing, I could have sworn that was passing before. I'll run RDR a few times and see if that's intermittent. Oh nevermind, just realized F4000 is the empty ROM socket haha.

I know that with the U28 ROM missing, the test will fail on some 5150 motherboards, and will pass on others. But something in the back of my mind tells me that I have seen varying pass/fail results on the one motherboard.

 

[modem7]

- Keyboard doesn't work

Does the POST report a 301 error if the keyboard is not present ?

No it doesn't, I've never seen 301 appear with or without the keyboard connected.

Obviously, with no keyboard plugged in, we expect the IBM POST to display 301.

The photo in post #12 shows that you are not using an AT2XTKB Keyboard Converter (or equivalent). So we can rule that out as to the cause of no 301 from the IBM POST. See the 'Observations' section of [here].

A possibility is that something is bad in the keyboard interface circuitry (which includes part of the 8255 chip), bad in a way that puts the POST into 'Manufacturing Test' mode. You can see that at step 24 at [here], that the 301 test is skipped if the POST is in 'Manufacturing Test' mode. I think this is the case. At the web page, see 'LOOP POST IF REQUIRED' after step 28. Initially, you had SW1-1 on the ON position, and if the POST is in 'Manufacturing Test' mode, the POST restarts. It explains the 'looping POST' that you saw earlier on. Later you set SW1-1 to the OFF position, and saw a boot to Cassette BASIC happen.

If you put SW1-1 to the ON position, does the 'looping POST' happen again ?

 

[VeryVon]

Since the DOS 1.0 floppy is booting on your 5160's controller+drive, how about temporarily moving that controller+drive over to the 5150 motherboard.

I have my working 5150 sitting here open on the bench, so I just moved the ROMS+controller+drive back into it, replicated the dip switch settings and fired it up.

Without the keyboard I'm seeing the 301's as expected. DOS 1.0 and Cassette BASIC boot. When the keyboard is connected it works as expected.

If you put SW1-1 to the ON position, does the 'looping POST' happen again ?

Yes. Same behavior with or without keyboard connected. If I then switch it OFF and leave the floppy out I get Cassette BASIC.

 

[modem7]

If you put SW1-1 to the ON position, does the 'looping POST' happen again ?

Yes. Same behavior with or without keyboard connected. If I then switch it OFF and leave the floppy out I get Cassette BASIC.

It does very much suggest 'Manufacturing Test' mode.

My 16KB-64KB motherboard has a 10/27/82 BIOS fitted and I cannot replicate the jump into Cassette BASIC (with 'Manufacturing Test' mode forced, SW1-1 off, CGA, floppy controller/drive fitted, no boot floppy).
I will go find/create a 10/19/81 BIOS, because maybe the jump into Cassette BASIC happens with that.

 

[VeryVon]

Finally got a chance to look at all the materials you referenced above: POST Details, Keyboard Startup, 'Manufacturing Test' mode, grounding the keyboard data line and of course the Keyboard Interface Circuitry. Impressive collection. I especially like how you call out the "reference view" looking at the DIN socket. I can't tell you how many connector pinouts I've seen online that leave me wondering if I'm looking at male / female, inside or outside etc. haha but I digress.

It does very much suggest 'Manufacturing Test' mode.

I checked out the resistances between ground and the data line (PIN 2) of the keyboard DIN connector: (bottom-most pin on DIN connector behind the empty math co-processor slot)

Board #1 - 4.9K
Board #2 - 4.9K
Board #3 - 5.7K
Board #4 - 4.9K
Board #5 - 146 Ohms <-- current board
My 5150 - 4.9K

Looking at the schematic, pull up RN1

 

[modem7]

My 16KB-64KB motherboard has a 10/27/82 BIOS fitted and I cannot replicate the jump into Cassette BASIC (with 'Manufacturing Test' mode forced, SW1-1 off, CGA, floppy controller/drive fitted, no boot floppy).
I will go find/create a 10/19/81 BIOS, because maybe the jump into Cassette BASIC happens with that.

Explained. It turned out that I had no power going into the floppy drive. Once, I fixed that, I eventually saw a boot into Cassette BASIC. So now I am absolutely convinced that you are observing 'Manufacturing Test' mode.

I checked out the resistances between ground and the data line (PIN 2) of the keyboard DIN connector ...

I should point out that I used, "grounded" in "the motherboard's keyboard data line ('KBD DATA') is grounded." because that is sure how IBM will be invoking 'Manufacturing Test' mode.
This is TTL logic, and so a LOW will do as well.
But good that you ruled out some kind of direct short.

Reference diagram at [here]. The 8255 and 74LS322 are certainly under suspicion. U80 and U82 have the ability to take the data line LOW. If you power up the motherboard with no keyboard attached, does the data line show as HIGH or LOW ? We expect it to be HIGH (U80 is a disabled state, and RN1 pulling the data line to +5V).