• Please review our updated Terms and Rules here

Request foor help on 5170 Mobo

S

Sabertooth

Member
Joined
Apr 20, 2019
Messages
36
Location
Texas USA
I have purchased an IBM 5170 with the Type 1 Mobo. I knew it didn't work when I got it but I do have some electronics background and thought I could make it work. I have changed out all the tantalum caps as well as a few of the MLCC. I installed the Supersoft BIOS v1.2 and it led me in a direction where I changed both 9517 DMA IC,s, replaced the original memory chips with NOS KM4128-15 the 8254 timer, 8942 keyboard controller and the 146818 RTC. I still get the same results as I have befofre all the replacements. I have listed them below in hopes someone can lead me in the right direction.

Supersoft

Beeps 6/1, 6/2, 6/2, 5/9 which I believe is the video card issue. I also get beeps of 1/4, 1/5, 1/6, 1/7, 1/8, 1/9, 2/1 & 3/8. Then after about a minute I get POST codes on the card of the following with the 1st being the last code given.

01,02,03,04,05,06,07,08,09,0A,0B,0C,0D,0E,0F,10,11,12,13,14,15,16,17,18,19,1A,1B,1C,1D,1E,1F,20,21,22,23,24,25,26,27,28,29,2A,2B,2C,2D,2E,2F,30,31,32,33,34,35,36,37,38,39,3A,3B,3C, END

AMI BIOS

When I install AMI I get 3 quick beeps that repeat (I believe it means bad memory although I have changed it out with several different batches but still can't prove it's good) Then I get POST codes of 10,0C,06,0A,09,08,07,06,05,04,14,03, END


Anyone have any idea of which direction I need to move next? I do not have an ocilliscope but I may be able to borrow one if I have to. Also, is there a memory test unit for the piggyback DRAM where I can see if they are good. I see many for single chips, both factory and home build but never see one for these.

Thanks for all your help in advance.
 
Welcome to these forums.

The manual for the Supersoft/Landmark diagnostic ROM (it is a diagnostic, not a BIOS), does not indicate that the diagnostic ROM outputs codes to a POST card. If the diagnostic does, the codes have not been published in the Supersoft/Landmark manual. Therefore, we do not know what test corresponds to what code. So the codes you see are of no help.

Regarding POST codes from the AMI BIOS. People are often misled, because they go to the Internet and see something like, "List of AMI POST codes", and not realise that AMI/Phoenix/IBM, and the like, changed their POST codes periodically. Unless you have a list that you know for for fact matches the particular BIOS, then you could be being misled.

What symptoms do you observe when the IBM BIOS ROM's are fitted ?


Supersoft/Landmark diagnostic ROM's

Back to the Supersoft/Landmark diagnostic ROM's. A screen shot would be nice. Do you have a suitable card/monitor ?

Sabertooth said:
I also get beeps of 1/4, 1/5, 1/6, 1/7, 1/8, 1/9, 2/1 & 3/8.
Click to expand...
That is quite a few.

temp_8874523530263252.png


At least one of the video RAM tests is expected to fail because you will not have that video card type fitted.

And some test failures are cascading. Generally, you aim to address the first reported failing test, then retest.

Tests such as "8254 Timer channel 0" involve more than the quoted chip. E.g. Missing clock, chip-select logic faulty, corrupted/nil address/data.
 
If this were my 5170, I'd start with the first 16K issue. I think that the 8254 must be operational; after all it's where the beeps are generated.
 
I sure appreciate all your help so far. I do think the Supersoft diagnostic ROM is misleading me for sure. The reason I say this is I just obtained a DTK mobo that I fitted with the same video card I was using in the IBM mobo and I get the same no video beeps on the DTK but I now have a picture on the screen.

You are absolutely correct on the POST codes. I referenced the booklet that came with the card and then I looked on the internet and found BIOSCENTRAL and the post codes there are different.

The original factory IBM BIOS gives me no beep codes and POST codes of 02 and 01.

Also thanks for the info on the 8254, makes sense. I have changed memory out with supposedly new memory but have no idea if it's even good and have swapped to different positions more times than I can remember. I can't fInd a tester that will test it either. I was thinking I could get a standard memory tester and move the RAS to check the second chip.

Again, thanks for your input
 
With the AT version of the Supersoft/Landmark diagnostic ROM's, if one removes all RAM chips on the motherboard, one would expect to see what is at [here].
If you do same and see a test failing before the 16K CRITICAL MEMORY REGION one, then obviously, there is a non-RAM failure to fix before looking at the RAM.
 
More great info. Even though I could get nothing on the screen and could only go by POST beeps but that same video card worked with a clone board, I have another card on it's way to me. When it arrives, that will be my first step to see what I get. Let's say I can't get any video with the card I'll be getting, what would be the cause of no video even when the cards are proven good?
 
Putting your posts together, I deduce that you have not seen the information at [here].

Note the information about some video cards not being suitable for the Supersoft/Landmark ROM's, and the erroneously beeping of the speaker with a "Cannot initialize monitor" pattern.
 
Yes I have seen that but the strange thing is I get the video error on the IBM with no video and I get the same error on the DTK board but the video card works. I would think that same card that works with DTK would also work on the IBM error or no error.

Thanks
 
Sabertooth said:
Yes I have seen that but the strange thing is I get the video error on the IBM with no video and I get the same error on the DTK board but the video card works. I would think that same card that works with DTK would also work on the IBM error or no error. Thanks
Click to expand...
Note that video related errors generated by the Supersoft/Landmark diagnostic, and whether or not video is seen, are two separate things.

Video related errors

Even in my fully functional machines, the Supersoft/Landmark diagnostics (displaying video) always gives me at least one video related error:

* Both the XT and AT versions of the Supersoft/Landmark diagnostics test video RAM. But they test it for all of the video card types of MDA, CGA, and EGA. An example is shown at [here]. I have a CGA card fitted. As a result, the CGA MEMORY test passed, but the MDA MEMORY and EGA MEMORY test failed, as expected.

* The AT version of the Supersoft/Landmark diagnostics always generates a "Cannot initialize monitor" beep pattern.

Whether or not video is seen

Assuming that you are using the AT version of Supersoft/Landmark diagnostics in both machines, your 5170 motherboard problem could be such that it results in the video card's registers not being able to be initialised.
 
Sabertooth said:
The original factory IBM BIOS gives me no beep codes and POST codes of 02 and 01.
Click to expand...
So, that is getting as far as 02.

Refer to [here]. The meaning of 02 depends on the revision of the IBM 5170 BIOS. You indicated a type 1 motherboard, so I will assume (because of your "original factory") that you have the 01/10/84 revision of BIOS. See note 6 at the pointed-to web page. I can imagine someone taking the motherboard out of its case and then powering the motherboard using an XT-class power supply.

Using an AT-class power supply ?
Good to rule out the introduction of a second problem.
 
Re power supplies. Have you verified that the +5V is within tolerance, and is stable? Tried another AT-class power supply? Good to rule improper power out.
 
I am using a 250 watt power supply out of an older 386 unit. Will that be OK or should I use the original AT supply that I do have?
 
Power is rock steady at 5.01 V. I have confirmed that it is an original IBM BIOS with part #6181028 & 29 which is, as you stated, 1/10/84. So I need to check the shutdown byte in the 146818 based on the chart. Thanks for all your help.
 
Sabertooth said:
I am using a 250 watt power supply out of an older 386 unit. Will that be OK ...
Click to expand...
I can't see why not.

Sabertooth said:
I have confirmed that it is an original IBM BIOS with part #6181028 & 29 which is, as you stated, 1/10/84. So I need to check the shutdown byte in the 146818 based on the chart.
Click to expand...
You could either chase down the cause of the timer error shown by the Supersoft/Landmark diagnostics, or chase down the cause of the 'shutdown byte in 146818' error shown by the POST in the IBM BIOS.

Everything has 'pros and cons'. An advantage of using the IBM BIOS is that the source code for it is available. Someone who understands programming code and PC-family hardware can see what the POST is getting up to between the POST codes.

Either way, hopefully, if you find the cause of one error, the fix will remove the other errors.

BTW. The shutdown byte is one of many bytes of RAM in the 146818 chip. The BIOS uses it when the CPU switches from protected mode to real mode. The Supersoft/Landmark diagnostics do not appear to have a dedicated test of it. If the 146818 chip is the only problem on a 5170 motherboard, the Supersoft/Landmark diagnostics run error free through to, and then fail at the PROTECTED MODE CPU TEST.
 
I will look into the timer issue this weekend when I have a chance and see what I come up with. I will let you know what I find. Again, I really appreciate all the info your passing on to me and hopefully it will help others too.
 
Not sure what I did but I am now getting to 05 on the post test with the IBM BIOS. Looking at the source, that is testing the 8254 for all bits of. So now I'll try and run that one down.
 
Looks like I hit a dead end. I can't figure out for the life of me why I'm not passing test 05. Thought I could borrow a oscope but that didn't work out. Maybe just have to go purchase one.
 
The hunt continues. I have looked at all the waveforms that are listed in the Sam's guide and for the most part they looked ok. Maybe a couple that looked funky but not sure what's causing it. Are there anymore resources that show where other waveforms should be taken and what they should look like??

Oh, and for some reason I'm back to stopping at post code 02. So I have to assume I'm back to the RTC CMOS on the shutdown byte. Do I have to make any programming adjustments on this IC or is all that handled by BIOS?
 
Sabertooth said:
Do I have to make any programming adjustments on this IC or is all that handled by BIOS?
Click to expand...
For the IBM 5170 motherboard to start and progress through the POST (or diagnostics), the chip does not need to be initialised in any way.

What happens on a fully functional 5170 motherboard, fitted with an uninitialised chip, is that part way through the POST, error codes like 161/162/163 are displayed.

Sabertooth said:
Are there anymore resources that show where other waveforms should be taken and what they should look like??
Click to expand...
Not that I am aware of.
 
You must log in or register to reply here.
Back
Top