OJ’s projects…usually breaking stuff.

[olivier.jan]

Never trust a notch! And I know it very well having multiple STs where some sockets are (factory) soldered in the wrong direction.
Still I managed to burn a perfectly good shifter today :headbang:

[exxos]

Never trust a notch! And I know it very well having multiple STs where some sockets are (factory) soldered in the wrong direction.

Wasn't aware of that but I am now !

[coonsgm]

Yeah, I had one where the TOS ROM sockets were soldered with the notches the wrong direction....

[olivier.jan]

Yeah, I had one where the TOS ROM sockets were soldered with the notches the wrong direction....

Yes, it’s usually the ROMs which have this issue, first time it happens to me on a shifter. The silk screen is Ok, but the socket is not. I should review all my STs and mark these errors with a Sharpie.

[olivier.jan]

Seems these shifters are really resilient: After a good night and cooling down properly, it works like a charm this morning.
I was planning to change it with a spare, but I just tested before changing it and good surprise, it works fine !!!
:girldance:

PS: Don't know if it will last a long time though...

[rubber_jonnie]

My OSI 600d Replica has all of the silk screen the wrong way around, once you realise, it's OK, but you have to pay close attention to the build docs and any photos.

However the 610 RAM/FDD controller board has them all the correct way around, with one exception, the socket for the 6821...

One dead 6821 later I went back and looked at the photos for the 610 and sure enough the photo is correct, but the silk screen for that single IC is not.

Very annoying.

[olivier.jan]

Quick rant tonight after wasting way too much time trying to fix a Roland MC-50.
I had this lying around for nearly a year, with errors when accessing the floppy.
To troubleshoot it :

1. Unsoldered the 1772 and tested it in an ST => OK
I also installed a socket for the ROM so I can upgrade it to MKII once fixed.
And while the desoldering gun was hot, I also replaced the floppy cable with a proper 34pin connector.

2. Tested the floppy, a JU-253, with a GreaseWeazle : Nothing, no sign of life.
3. Found an old Amiga JU-253 lying in a drawer, labelled "Broken", tried it with the GreaseWeazle: Nothing
4. Tested some SMD-380 with the GreaseWeazle: Working fine :)
5. Checked some video on YouTube on JU-253 repairs and decided to replace the 2 electrolytic caps. => Nothing
6. While checking the service manual, I noticed the 12v rail was actually connected on the JU-253, so I built a 5v/12v adapter to power the drive instead of the usual 5v adapter I use with the GreaseWeazle => The Amiga JU-253 sprang to life, while the MC-50 didn't :(
7. Decided to use the Amiga JU-253 in the MC-50, cleaned it, changed the faceplate and installed it in the MC-50 => Nothing

That's when I noticed the MC-50 floppy power doesn't have 12v connected, which is normal as there is no 12v in the device. The two JU-253 have different references and I guess the original one from the MC-50 is designed to work with 5v only.

The original drive from the MC-50 gives me random errors, once in a while it complains about Track 0 but most of the time it just doesn't even respond.

So finally, I decided to installed a Gotek I had ordered just in case: Stock firmware doesn't work with MC-50 as it must be 720k and not 1.44. So tried to flash FlashFloppy firmware, but... I don't have the proper USB-A to USB-A cable....

Ordered the cable, we'll see tomorrow...

Oh, forgot to mention the multiple failed attempts to replace the original barrel jack with a USB-C PD module which ended destroying 2 of these modules....

Very bad day when everything goes wrong, a 2hr operation finally ended up taking most of the day for no result (Ok, I learnt JU-253 needs 12v...).

Time for a few beers and a cigarette I guess...

[olivier.jan]

I received the cable today and programmed the Gotek properly to use FlashFloppy.
Sadly the MC-50 refuses to format any image or to read any disk, same symptoms as with the original JU-253.
As the 1772 is tested OK, I guess this is going to be a long troubleshooting, I hope no Roland Gate Array are damaged as this would be nearly impossible to fix, and not with it anyway I guess.

[olivier.jan]

Not too much time to spend on Atari or other electronic stuff these days. :(
Still I managed to find some time over the weekend to fix my last remaining 1050 floppy drive for the 8-bits.
I actually have 3 drive, but 2 are broken and I never found enough time to troubleshoot.
A few weeks back I plugged the only working one to read a few floppies and realized it wasn’t working anymore. I haven’t used it for a few years with Fujinet and other SIO2SD doing the job, but before going to storage it had been completely cleaned and tested :(

So, I cleaned it again, greased the motors and moving parts and recapped the board, didn’t change a thing.
I finally noticed with the drive open that the floppy rotation was strange (speed was not constant): The belt was not gripping anymore.
I changed it with a belt from a broken drive and it works again.

I just ordered new belts as I suspect these 40 years old belt won’t last long.
Anyway, at least it works now without going for a full chipset troubleshooting.

[olivier.jan]

Back to Atari this weekend, trying to get my USB-C PD PSU working and finally got some success.
Finally got 5 boards working, obviously for a v1, it’s a bodge wire edition.
Still one little issue to fix, it takes ages to discharge when the power is tuned off. Not a big deal, but if turned on before it’s discharged enough, I got weird behaviour. I will try to add a bleeder this week.

Anyway, the ripple is very low (<200mV peak to peak), the 5V is at 4.95V and the 12V is around 12.3V.
It accepts around 3.5A on both rails and provides an USB-A socket to feed an UltraSatan or ACSI2STM.

Once the discharge issue is fixed, I will start working on a 3D model to attach it correctly inside the ST (on the picture below it’s fixed with Blue Tack).

No more 230V near my fingers and plenty of free space inside my STF and STEs.

IMG_2826.jpeg

IMG_2827.jpeg