exxos's Atari, Amiga & retro forum

stephen_usher wrote: 07 May 2022 17:06 Hmm... The signal generator on the Owon 'scope is a bit pants anyway as it only produces by default positive/negative pulses centred on zero, +- 2.5V. The positive voltage only goes up to 2.5V maximum, so I tried changing the lower voltage to 0V, which works, but means that I can only ever get a 2.5V signal. It also seems to partly reset when the power goes off, such as if the power saving kicks in.

The interface is a bit iffy too as if you somehow accidentally change the lowest significant figures of a frequency, accidentally change range and the try to go back, it won't allow you to get down to the lowest values to change them to zero. Daft.

That all sounds a bit annoying!

This weekend I've been doing a bit of retro-computing for work. Basically one of our labs has some equipment which only works if an IBM PS/2 (286) talks to it over serial. Newer machines don't work. So I've been resurrecting the lab's PS/2 Model 50Z.

Actually, I've been looking at it at work on and off over the last fortnight and I thought I'd have to hack it to get it working with a standard floppy drive, which would have required a lot of soldering!

Anyway, I brought it home today and decided to try to get the floppy working again. It chewed disks and couldn't read them anyway.

Pulled the drive apart, fully refurbished it and re-aligned the heads.

I thought it wasn't working properly when I first tried booting from the IBM BIOS configuration floppy (yes, it's one of those types of IBM without a BIOS config screen) as it just sat there. When I tried an MSDOS 6 boot floppy that just crashed the machine with L06 being printed constantly.

Anyway, I'd remembered reading that some PS/2s could only boot from 720K drives without a BIOS setting. The only issue was that the config floppy image was only available as a 1.44MB one!

Thankfully I found that the PC-DOS 4.01 install floppy images are 720K. But this still meant getting my 2004 vintage laptop down from the loft, swapping to my Linux HD for it and writing the floppies on there, as the USB floppy drive doesn't do 720K. *sigh*

Even with that floppy I thought it wasn't working as it seems that PC-DOS doesn't announce that it's booting. The cursor just flashes in the top left corner of the screen. Still I heard the drive seek so kept going. and after about 5 minutes it booted to the install program for PC-DOS. Obviously the PS/2 goes into hobble mode if the BIOS settings are blank.

Anyway, this allowed me to get to a DOS prompt, swap the floppy over to the config one and thankfully that worked, so I could run the config utility SC.EXE. *phew*

Except... it couldn't complete the auto-configuration as it didn't know about how to configure the RAM card and network card in the machine!

At least I was able to manually tell it that it did actually have a floppy drive and it was 1.44MB.

Anyway, save to settings, set the date and time and power off.

After removing the cards and powering up I could boot from the config floppy and finish the configuration. It was only now that it could see the hard disk controller, at last.

I found the config description files for the two cards on-line and copied them onto a floppy. I could now fit the cards, re-run the SC.EXE and load the information, which gets saved in NVRAM along with a driver.

So, that recovered the machine. It booted straight from the hard drive, which has MSDOS 6 on it, so I ran SCANDISK... It only found 4 bad blacks, which isn't bad for a 1988 drive which wasn't parked before it was transported and where I'd had the lid off to see if the heads were moving!

Anyway, here it is.


I'll take it back to work tomorrow.

Nicely done sir 😁

Not much retro this morning, just an attempt to repair the power backplane from the water damaged Aruba core switch which got doused on January.

The contaminated water caused a short on the POE 48V side, which cooked the PCB substrate to the point where it became conductive and shorted even more. A couple of months ago I rectified this by isolating the power planes on the board.

This morning I re-wired it using reasonable heavy duty wire and paper clip wire as buzz bars across the connector pins. Hopefully it'll all fit back in and work.

I was going to run Frontbench on my 520STFM with the TF536 installed (and running the October 2020 firmware) but the machine doesn't like the DMA again, with or without the UltraSATAN if switched to the TF, but works OK with the 68000. *sigh*

I'm now testing the TT-RAM on the TF, just in case, and that passes.

Next to run the diag ROM in ST mode.

The TF was failing to read the ST bus properly... RAM test errors in TOS 2.06.

I had to reseat it several times and "deoxit" the socket as well. I do have turned pin sockets, they have very poor connectivity as there's no active pressure between the pin and the socket. Unfortunately the only sockets you can get to fit the CPU are now turned pin.

I've had a very frustrating afternoon trying to get a usable Raspberry Pi Pico C development environment set up.

I first looked if there was a Windows package I could just download, install and use... erm, nope. Some people have got something working just jerry-rigging and hacking some compilers and then doing handstands whilst whistling hey-Jude backwards and slightly off key, so that's a non-starter.

So, I thought, why not use my Raspberry Pi 3? That should be Easy(tm), after all it's all Raspberry Pi stuff so it should just be an 'apt install" away, surely?

Reading the docs... nope. "But we have a script that you can run..." but first you need the bleeding edge version of Raspberry Pi OS... *sigh*

So, boot up the RP3... It's running Buster, and I need Bullseye. So, change the /etc/apt/sources.list file and run 'apt update' and then 'apt upgrade' and wait. Of course, it doesn't just go away and do it, oh no, this is Debian... you constantly have to keep an eye on the install as at random places it will start asking you questions.

Anyway, a couple of hours later... Done. Reboot.... Pi hangs at the rainbow screen and then screen goes blank!

So, at this point I don't know if it's just a non-bootable SD card or the Pi is toast (even though I have a heatsink on the CPU). I'm now trying to make a new boot SD card but the MacBook Pro's SD card slot seems glacially slow. The maximum write speed I can get is about 1.3MB/s and I need to write 4GB. *sigh*

Wish me luck!

Thankfully it's not the Pi. It looks like the upgrade failed to install a working boot loader for some reason. Anyway, I can try getting the SDK working on the new installation.

@stephen_usher i program the rp2040 using the arduino IDE. It’s still a bit of a pain but at least it works, more than I could say about vscode.

Hmm. I'll have a look at that. Thanks.

Anyway, I've finally got the SDK installed and the Pi booting from the SD card but with root and swap on an external USB connected SSD. Far faster than the SD card and a lot larger. (It also has wear levelling.)

I've not tried uploading any of the examples to the Pico yet, or even connecting it to the Pi via USB.