Beautiful machine, it's like something designed in the 60s.

dominusprog wrote on 2026-09-17, 11:43:

Beautiful machine, it's like something designed in the 60s.

Thanks! 😀

I appear to have solved the Boot Problem!

It was just one of the dip switches defining the systems memory configuration:

How it was initially set:

How it’s set now:

I’m still going to re cap the power supply when I get my rework station!

Oh good old times with half of the ICs made in Italy and the other one in Portugal, not Malaysia, Taiwan, Singapore....

This is my first look at this machine and the issues I've encountered so far:

https://youtu.be/75koPEHmiDQ

So I watched the video, and you should use contact cleaner instead of IPA for RAM sockets. And for the battery situation, you can use a rechargeable 3.7V 18650 battery if you can find an empty space for it.

https://en.wikipedia.org/wiki/18650_battery

dominusprog wrote on Yesterday, 11:11:

So I watched the video, and you should use contact cleaner instead of IPA for RAM sockets. And for the battery situation, you can use a rechargeable 3.7V 18650 battery if you can find an empty space for it.

https://en.wikipedia.org/wiki/18650_battery

Overkill and potentially dangerous, given the charging circuit is tailored for NI/CD, not Lithium.

It's plenty of batteries with the original specs, or very close.

https://www.amazon.co.jp/NI-CD-%E9%9B%BB%E6%B … 0/dp/B0DXDLD3XZ

dominusprog wrote on Yesterday, 11:11:

So I watched the video, and you should use contact cleaner instead of IPA for RAM sockets.

I still intend using DeoxIT on all of the socketed IC's. It's unfortunately not sold locally, so I'll have to order some from DigiKey at considerable expense!

I'm still experiencing instability in the form of the M21's lack of willingness to post with the jumpers set for 640K, so I've pulled all of the chips and as a future investment I'll get myself one of these Retro Chip Tester Professionals to test the memory chips. (Jeeez these things are expensive!)

I've checked all of the voltages and they're good, so it's not a Power Supply issue.

The Solutor wrote on Yesterday, 12:35:

Overkill and potentially dangerous, given the charging circuit is tailored for NI/CD, not Lithium. […]
Show full quote

dominusprog wrote on Yesterday, 11:11:

So I watched the video, and you should use contact cleaner instead of IPA for RAM sockets. And for the battery situation, you can use a rechargeable 3.7V 18650 battery if you can find an empty space for it.

https://en.wikipedia.org/wiki/18650_battery

Overkill and potentially dangerous, given the charging circuit is tailored for NI/CD, not Lithium.

It's plenty of batteries with the original specs, or very close.

https://www.amazon.co.jp/NI-CD-%E9%9B%BB%E6%B … 0/dp/B0DXDLD3XZ

I don't know about the overkill part, but I connected one of these to my socket 3 board last year and it works just fine.

dominusprog wrote on Yesterday, 14:11:

I don't know about the overkill part

If replacing something that is 60~80mAh with something else rated 1800 mAh isn't overkill, what is ?

, but I connected one of these to my socket 3 board last year and it works just fine.

It's not matter of working, it's matter of safety, 18650 batteries, especially unprotected ones, are ticking bombs that must be managed by people who know what they are doing.

Now, given the original battery was so smaller, the wrong charging circuit is unlikely to have enough force to start a fire or an explosion, and you're probably on the safe side, but is better to rely on competence rather than on luck.

A not so lucky case....
https://en.wikipedia.org/wiki/UPS_Airlines_Flight_6

For now here are some pics of the machine: (It's in great physical condition 😀 )

The Solutor wrote on Yesterday, 16:33:

If replacing something that is 60~80mAh with something else rated 1800 mAh isn't overkill, what is ? […]
Show full quote

dominusprog wrote on Yesterday, 14:11:

I don't know about the overkill part

If replacing something that is 60~80mAh with something else rated 1800 mAh isn't overkill, what is ?

, but I connected one of these to my socket 3 board last year and it works just fine.

It's not matter of working, it's matter of safety, 18650 batteries, especially unprotected ones, are ticking bombs that must be managed by people who know what they are doing.

Now, given the original battery was so smaller, the wrong charging circuit is unlikely to have enough force to start a fire or an explosion, and you're probably on the safe side, but is better to rely on competence rather than on luck.

A not so lucky case....
https://en.wikipedia.org/wiki/UPS_Airlines_Flight_6

Fair enough, they have higher capacity. But turns out I do know what I'm doing because the board I connected the battery to doesn't charge the battery. Maybe some boards do that I don't know. I charge the battery with a USB charger and put it in the battery holder.

dominusprog wrote on Today, 09:18:

Fair enough, they have higher capacity. But turns out I do know what I'm doing because the board I connected the battery to doesn't charge the battery.

While this machine was running I measured the voltage across the pads where the battery was mounted and measured a voltage of +-3.3V, so there would be a voltage applied to the battery.

Irinikus wrote on Today, 09:59:

While this machine was running I measured the voltage across the pads where the battery was mounted and measured a voltage of +-3.3V, so there would be a voltage applied to the battery.

Of course there is. The original Is a rechargeable battery, perhaps a type that would self discharge in weeks or few months, if no one charges it, there is no hope it keeps doing its function.

dominusprog wrote on Today, 09:18:

Fair enough, they have higher capacity. But turns out I do know what I'm doing because the board I connected the battery to doesn't charge the battery.

There are two options

1) you didn't measure correctly

2) the original charging circuit was broken.

That said such kind of batteries used to be charged with something like 1/10 C, so something like 8~10 mA in our specific case, something that hardly has the force to harm a way bigger battery.

But still I can't see why use the wrong approach when easier and cheaper solutions exist. The most obvious is using the original battery, but a 3xAA /AAA battery holder, filled with 3 AA/AAA NI/CD or NI/MH cells, would have been a good and easy mod.

If the option 2 above was the correct one, likely a Lithium non rechargeable battery like a current CR2032 would have been an option to take in account.

Irinikus wrote on Today, 09:59:

dominusprog wrote on Today, 09:18:

Fair enough, they have higher capacity. But turns out I do know what I'm doing because the board I connected the battery to doesn't charge the battery.

While this machine was running I measured the voltage across the pads where the battery was mounted and measured a voltage of +-3.3V, so there would be a voltage applied to the battery.

Well, you can connect a diode to the battery. In other hand you can use one of those CR2032 battery holders, but it too needs a diode connected to the positive wire.

https://www.amazon.com/Alinan-Button-Battery- … 89904767&sr=8-1

Wow, what a wonderful find! 😍

If the M21 has the same internal RTC as the M24, you might not want to bother getting it to work. As you can find at many resources on-line (as well as in Usenet archives of comp.sys.att for the US version of the M24, the AT&T 6300), the RTC on these systems was notorious for running no further than 1990, and requiring a device driver to work around that limitation.

Also it was non-standard and required either custom MS-DOS versions to work out of the box, or external RTC device drivers or TSRs and such.

If you want a real-time clock on that thing, try to get an expansion card that has one. Some Expanded Memory cards also had an RTC that would work better than the one in the Olivetti machines. Or use something like a PicoMEM 2, which can also provide an RTC, if you install an optional CR1220 Lithium battery in it.

There is an open source ISA riser card project for the M21 on GitHub.

Interestingly enough, when it boots I'm able to enter the current date! (So the BIOS from 1985 seems to be Y2K compliant!)

Of course it needs to be entered every time you boot the system, as there's currently no battery fitted.

If this is due to software patches that are currently installed, I'll have to properly sort out the disk when I reinstall everything once the machine's properly working.

For now I'm going to go through the process of troubleshooting the board (as this is what the hobby's about), but if worst comes to worst, I'll order in an M24 board as a replacement.

dominusprog wrote on 2026-09-17, 11:43:

Beautiful machine, it's like something designed in the 60s.

Molto bello for sure. Great find Irinikus!

Many Bothans wrote on Today, 13:50:

Molto bello for sure. Great find Irinikus!

Thanks! 😀 It's indeed a beautiful design!