JezC H5 C5B Build - one STep beyond...?

[JezC]

Well, I ended up with very limited time this evening (won't bore you all with the details of my life....) but did manage to get my H4 board out & prove that is still booting to the desktop & floppy is working ok.

So, I'll gradually swap bits between this working H4 and the sulky H5CB to see if I can find any difference (and post the details of the jumpers fitted on the two boards).

One notable difference is that I've put an Ajax in the H5CB board & it's a normal WD1772 in the H4 (and also in the other awkward H5C1 boards).

No GAL fitted in any of my boards yet, so all jumpered with floppy A as internal & forced to 720k mode.

More tomorrow (hopefully)...

[JezC]

Well, swapping YM2149F, DMA and 1772 between the H4 and H5 showed that this particular DMA chip is in fact a truly BAD bad DMA chip....*

Also found an issue on the floppy power cable connections, so I was starting to hope...especially when on the next boot it reported the 1 floppy drive...but then a couple of bombs when it tried to run something off the floppy in drive A...and then back to reporting 2 floppy drives on the next boot.

Out of time again for tonight but I suppose I have narrowed down the list of potential problems a bit more. Maybe


* The 'interesting' failure mode for this DMA chip was that I couldn't actually open the floppy drive when I double clicked the icon - no errors reported or anything, just no action.

[exxos]

Ive had all in a of odd things happen with bad connections on the 1772 as well. I've got one zif socket for testing 1772 chips and I have to reseat it often. Cheap sockets I guess..

[JezC]

Ive had all in a of odd things happen with bad connections on the 1772 as well. I've got one zif socket for testing 1772 chips and I have to reseat it often. Cheap sockets I guess..

I'll double-check the 1772 is seated properly (and check the pins & socket contacts look clean) - thanks.

I should have enough bits to have the working H4 and the H5CB both set up at the same time & just switch the video between the two systems...so then it will just be having the time to dig a bit deeper

[JezC]

Ok, so I have the H4 and the H5CB both set up and feeding into a VGA switcher to a suitable 15kHz NEC monitor (thanks for that one @PhilC -proving to be very useful for this sort of testing!).

H4 is still booting to desktop & reading floppies OK while the H5CB will boot to the desktop if no floppy is in the drive but it crashes when reading a floppy (either during boot or when at the desktop).

Both systems have a real ST FDD fitted as the internal drive (and selected for the boot drive), neither has the GAL fitted, both are booting TOS 1.04 so I'll start going through all the jumpers on both systems to see if I can find any differences - and I will try removing & cleaning & refitting the 1772 just in case.

In this setup both systems are booting and showing that jut 1 floppy drive is detected on the boot logo (along with the TOS version and the amount of RAM detected). I'm pretty sure this is the upgraded TOS that also supports booting from IDE devices (on both machines).

More news when I get the next bit of time for diagnosing...

[exxos]

Might have been a early patched 104 ROM Putnik did ages ago ? That was just to display info quickly on boot up and nothing else was done in it AFAIK.

[JezC]

I think I got this from another forum member... I'll check who it was and exactly what it was modded to include but pretty sure it's IDE support for booting...

[kodak80]

The dual rom with TOS 1.04 showing drive details and memory is just that. As per Exxos post. It was created to show this information before booting. It did not include the IDE patch.

[JezC]

Ok, I've just compared the image I have (called TOS_104_PIDE_206_BS.IMG) with an older version I ran with TRUDIE on my H4 board a few years back (and that had TOS 1.04 patched for IDE support including auto-booting) & the first 256k is identical.

So, I'm 99% sure that this one does have the patched IDE support in TOS 1.04 (and it might well have come from czietz in the distant past).

Just looking at the jumpers fitted on the H4 and H5CB compared to
viewtopic.php?f=78&t=2785

[JezC]

Well, tried a few things (including fitting the GAL for HD as well as DD support) - with & without HD enabled makes no difference (with HD floppy in HD mode or DD floppy in DD mode) but it all still ends in a pair of bombs if I put the floppy in the drive at boot or when I open drive A from the desktop.

So, as I expected, there must be something fundamentally different between the working H4 (DD drive, no GAL) and H5CB (DD drive + no GAL, DD drive + GAL, HD drive + GAL) still relating to the floppy.

The only jumpers relating to the floppy that I can find are :
- A/B swap (both boards set for A as internal)
- 8/16 MHz clock select (both initially set to force 8MHz, H5CB also tested with GAL fitted & this set in both modes)
- force 720k/1.44M for drive 0 & 1 (H4 has these jumpers open, H5CB fails in the same way whether open or closed)

I didn't try connecting the HD detect wire (didn't have time) but don't think that is related to the issue.

At this rate, I'm going to have to try and write a bit of code to access the floppy registers on demand to see if I can work out what signals (and on what accesses) are giving the problem on all 3 of the H5C boards I now have but not on the H4.

I've seen @exxos using a flavour of basic to access this kind of thing but I'd have to get some storage working on it to be able to load it in the first place...

Maybe fit a 68k DIL socket on this H5CB & use TRUDIE (I could do the same on the H4 as well - that already has the 68k DIP sockets fitted)...