Apple eMate 300 Expansion Module (Open Source / Open Hardware Project)

[splorp]

Are there options for the eMate 300 to communicate with modern computers?

Newton Connection for Mac OS X (NCX) works over serial and ethernet under Snow Leopard and later. I can probably dig up an older version that work under Tiger and Leopard. I use an FTDI serial to USB adapter cable for both my eMates and MP2x00 devices.

 

[splorp]

Oops. Just saw Mikeyy00’s reply regarding NCX.

I’ve recently taken over management of the Newton Research site from the developer (Simon Bell) and have access to his archive of old software releases and source.

 

[JustG]

This is great news! Been following your progress @Androda. Are you still experiencing issues where installing the expansion module is a one way process where the system is copied over to the card?

 

[Androda]

This is great news! Been following your progress @Androda. Are you still experiencing issues where installing the expansion module is a one way process where the system is copied over to the card?

I think it's intentionally a one way process when installing the flash upgrade modules. The system RAM upgrades can be removed with no issues.

My guess is that there's something in the firmware which "integrates" the extra flash into the system as a whole. Remember that the Newton is based on "soups", so adding more flash is like getting a bigger pot - it all just flows together and can't be separated afterward.

So, you can't remove the flash upgrade module after installation. It'll do funky things to your eMate. I've been looking for a way to undo the bad state that happens afterward, and it might be fixable with the original eMate system update - that update mentions calibration data. Will try that once I'm back in town.

 

[JustG]

Thank you for the reply, I am interested in one of your DRAM-plus-flash modules but am a little wary of the potentially permanent nature of the upgrade.

 

[retr01]

@Stephen and @Androda, I would be interested in this memory module to expand more memory for my eMate 300. PCIMIA cards for the eMate can give 20 or 32 MB of flash memory. It would be nice to have that much in the module to be able to use the eMate's PCIMIA slot for other things

I started a new thread on the eMate 300 covering battery updates, missing items from the original box, documentation, etc.

 

[Fizzbinn]

News on this project: I have designs which work, both DRAM-only and DRAM-plus-flash.

This project has been thoroughly weird. As mentioned before, only one of the 10 I assembled (dram plus flash) last time worked at all (the first one, which gave me a false sense of hope). I revised that design and tried again, with no success. Tried one last time, completely reworking where the chips are on the PCB, and just tested that design this morning.

Poof, it works. I have assembled and installed four DRAM-plus-flash modules in my eMate, and all of them have reported the increased storage space and System RAM. My DRAM-only design also works, five of those assembled and reporting only the increased System RAM.

I haven't tested installing programs (not sure if you even get to select where they end up, on eMate internal or the upgrade stick), but this is certainly a good sign and step in the right direction.

If you're interested in testing out one of these modules, DM me. Three of the ram plus flash upgrades are available (leaving one for me to test with), and five DRAM-only are available.

I got some time this evening to do some testing with your DRAM-plus-flash module, spoiler, it works!

• I made sure to do a Newton device Backup using the Newton Connection Utilities app I have installed on one of my PowerBooks
• Next I removed the bottom panel and disconnected my rebuilt battery (which was fully charged, so no AC adapter connected either)
• Then popped the module in, reconnected the battery, and screwed the panel back on (tempting fate)
• ...and pressed the power button. dark screen. nothing
• I tried pressing the reset button on the bottom for 10 seconds (a guess) then the power button and still nothing
• As I got up to find the eMate docs on resetting I heard the start up chime! Came up factory defaulted
• I did a Newton device restore using the Newton Connection Utilities app and I'm back in business but now with 4MB of RAM (instead of 1) and 4MB of Flash storage (instead of 2) as reported by Extras app Memory Info

Before I started I did do some wayback machine searches for Newer Tech docs that I was hoping might have upgrade instructions for their eMate upgrade module. The best I could find was these pages:

Newton eMate Memory Modules

PowerBook Memory Guide

...which do reinforce the need to backup and restore before and after the install respectively. Some nice old school skull graphics drive home their point.

 

[Fizzbinn]

I think it's intentionally a one way process when installing the flash upgrade modules. The system RAM upgrades can be removed with no issues.

My guess is that there's something in the firmware which "integrates" the extra flash into the system as a whole. Remember that the Newton is based on "soups", so adding more flash is like getting a bigger pot - it all just flows together and can't be separated afterward.

So, you can't remove the flash upgrade module after installation. It'll do funky things to your eMate. I've been looking for a way to undo the bad state that happens afterward, and it might be fixable with the original eMate system update - that update mentions calibration data. Will try that once I'm back in town.

Did you try both Soft and Hard resets after removing the module?

https://tim.id.au/laptops/apple/misc/emate.pdf

It seem like a Hard reset might be needed to get the Flash upgrade changes undone. I might try this later but stopping while ahead tonight!

 

[retr01]

Most likely, it was the hard reset to factory defaults to get it set up with the new memory module. This is what I have from the user guide:

 

[Androda]

Did you try both Soft and Hard resets after removing the module?

https://tim.id.au/laptops/apple/misc/emate.pdf

It seem like a Hard reset might be needed to get the Flash upgrade changes undone. I might try this later but stopping while ahead tonight!

I have now tried a 'hard reset' on my eMate which gives the factory calibration error. The error persists after hard reset.

According to this post on NewtonTalk, calibration data is stored in the very first block of flash with a backup in the second block (blocks 0 and 1). A hard reset does not delete or rewrite those blocks.

This leads me to believe that calibration data is either copied to the flash upgrade or somehow 'merged' into the flash upgrade, meaning that indeed you cannot go back after installing the flash upgrade. Your eMate will function (other than battery charging), and you will have to dismiss that 'factory calibration' error every time the eMate's display is powered on.

Makes me wonder whether this calibration data can be read out and backed up with a special program prior to installing the upgrade, so it can be rewritten if the upgrade is ever removed. In any case I plan to perform a factory hard reset on another one of my eMates, and then pull the internal flash chip off the board to back up its factory calibration data. That will allow me to apply it over any eMate that has this issue, with the annoyance of needing to desolder and resolder the flash chip itself.

I'm starting to think that the flash upgrade might behave similarly to the DRAM upgrade. What if adding that second flash chip is expanding the flash bus width to 32 bits from 16? That would require a pull and merge of data from internal flash onto the upgrade.

 

[retr01]

How was it successful with Newer Technology or MicroTech SO DIMM upgrade for the eMate back in the day?

 

[Androda]

How was it successful with Newer Technology or MicroTech SO DIMM upgrade for the eMate back in the day?

View attachment 7949 View attachment 7950​

I don't understand the question. Two test modules were installed, by Fizzbinn and another tester. They haven't reported any issues.

Installing the module *and then removing it* is what caused the 'factory calibration lost' error for me. And I can't think of a good reason to remove the module once it's been installed. It increases RAM, internal storage space, and system speed (thanks to the bus widening effect).

 

[retr01]

I don't understand the question.

I asked what procedures to make it successful and avoid NOS (Newton OS) errors. So, I wondered about what the instructions were back in the day. What if, back in the day, it was needed to roll back to the original memory module? How was that done?

Unfortunately, it isn't easy to find those original instructions from Newer Technology or other third parties that offered the memory upgrade for the Newton eMate. Maybe someone like @splorp would have that handy?

Installing the module *and then removing it* is what caused the 'factory calibration lost' error for me.

Was the eMate wholly powered off, including removing the rechargeable battery pack and the PRAM battery?

And I can't think of a good reason to remove the module once it's been installed. It increases RAM, internal storage space, and system speed (thanks to the bus widening effect).

Oh yes, definitely! Here is what Newer Technology had to say back in May 1998:

The eMate 300 comes with 1MB of DRAM, and 2MB of Flash RAM soldered to the motherboard. It has a single memory expansion slot. That slot accepts a memory card with either DRAM, Flash RAM, or both. The DRAM compliment on the card may be 0MB, 1MB, or 4MBs. The Flash Ram may be either 0MB or 2MB. The most common version of the card is 4MB of DRAM + 2MB of Flash RAM.​

​

Adding 2MB of Flash RAM is beneficial for several reasons. Aside from the obvious benefit of doubling your internal storage space, there is a far more significant advantage. The 2MBs of Flash RAM soldered in the eMate, sit on a 16-bit wide data bus. When you add 2MBs of additional Flash RAM, the eMate reconfigures itself, and all the Flash will sit on a 32-bit wide bus. Now you have twice the storage space, accessed twice as fast.​

 

[Androda]

I asked what procedures to make it successful and avoid NOS (Newton OS) errors. So, I wondered about what the instructions were back in the day. What if, back in the day, it was needed to roll back to the original memory module? How was that done?

I doubt that they provided instructions on how to remove the module. Based on what I'm seeing it's a one way process, unless a tool can be written to back up and restore those device configuration blocks.

A tool to do the backup and restore might be useful in general. Other people have reported getting this factory calibration error when doing things which are completely unrelated to the RAM/Flash expansion modules.

 

[retr01]

I doubt that they provided instructions on how to remove the module. Based on what I'm seeing it's a one way process, unless a tool can be written to back up and restore those device configuration blocks.

Ah. Makes sense.

A tool to do the backup and restore might be useful in general. Other people have reported getting this factory calibration error when doing things which are completely unrelated to the RAM/Flash expansion modules.

That would be good. I wonder, is there a log that would clarify why the factory calibration error happened? It seems to me that several factors could cause that error. Why is Newton ambiguous with that compared to the SAD Mac codes, which are more specific?

 

[Androda]

Talked to my friend who has one of the original dram+flash upgrade modules from 'back in the day'. He says that installing and removing it has not caused a factory calibration error to appear. Just a full data wipe every time.

This makes me question whether it's really a one-way upgrade. And it also definitely makes me wonder why some eMates throw the factory calibration error and others don't.

This factory calibration error is of a fundamentally different type than a sad mac code. Sad mac codes mean that the ROM has determined your hardware has a fault that makes running the computer impossible. Bad RAM, etc. This Newton error message does not prevent the system from booting and being usable. Calibration data for the battery charge system has been lost, so the battery won't charge. But other than that the hardware is fine.

 

[retr01]

Perhaps the eMate can get confused when the reconfiguration to the 32-bit data path does not work when the memory module is gone. It's like it thought it became twice as bright and twice as fast, then it loses its marbles and does not remember what it was like being half bright and half as quickly.

In other words, I wonder if removing the memory module and starting up the eMate expects the module to be there?

About charging, can replace the battery pack throw off the charging calibration?

 

[Androda]

Something in the eMate's ROM must be involved here to configure the CL-PS7010 go-between chip properly in terms of which memory and flash driver banks to use. It's a strange architecture, with the CPU not directly connected to RAM or flash (maybe not even ROM). The 7010 supports many banks of RAM and flash and stuff, but it has to be configured to know which ones to use. This is how the eMate can switch between a 16 bit memory mode (default) and 32 bit mode with the expansion installed. Probably defaults to 16 bit mode and then swaps to 32 bit later in the ROM code after scanning determines that the wider bank is available.

The calibration data which is being lost is likely related to the MAX471 current sensor chip. It has a scalable voltage output which varies depending on the current flow. So each eMate is individually calibrated to minimize the difference in output based on resistor tolerances (5% or something like that). Makes sense, though the degree of precision required here is hard to tell without the original design spec.

So more or less, we don't have a good answer as to why the calibration data is being lost when a standard hard reset theoretically doesn't touch that block of flash. And we should try to back it up for the future inevitability that the flash cells simply lose their value (because they have a finite storage lifespan).

 

[Androda]

Found some more information about what's in the blocks of 'calibration data' from NewtonTalk. This post talks about what's actually *in* the calibration section a bit. Not all the information, but at least some of it. And the 'factory calibration lost' message probably has nothing at all to do with the battery charge IC, it's the temperature sensor - this makes sense given that all thermocouples and similar temperature sensing devices are variable and need some sort of standard to check against.

 

[Fizzbinn]

Found some more information about what's in the blocks of 'calibration data' from NewtonTalk. This post talks about what's actually *in* the calibration section a bit. Not all the information, but at least some of it. And the 'factory calibration lost' message probably has nothing at all to do with the battery charge IC, it's the temperature sensor - this makes sense given that all thermocouples and similar temperature sensing devices are variable and need some sort of standard to check against.

This can't be the temperature sensor in the battery pack though, correct? I mean wouldn't there have to be some documented procedure on doing some sort of recalibration when you got a new battery pack?

Oh, scratch that. Should have looked at the manual first, page 244:

Replacing the battery
If for some reason, you think you need to replace the battery, contact an Apple- authorized service provider.

I didn't realize the battery wasn't meant to be user replaceable or how important using the same sensor was in rebuilding a battery pack...