Another stupid flashing cable question HP-30b / HP-15C CE

[AnnoyedOne]

Here's my theory. It is just a theory based on my experience. I don't own any AT91SAM7L128 based calculators, cables, software. Nothing.

To start with I have to explain my own experience with flash upgradable firmware. This is unrelated to HP calculators. And it may be long.

Over a decade ago I developed (hardware/firmware/software) for a product. I used an ARM based SoC part for it. Similar to the ones used in HP calculators but not the same. The product had a two line alphanumeric display, for messages, a few buttons but no keyboard.

As part of the project I wrote a totally custom bootloader that allowed the main application to be updated in the field. The SoC part I used did not have one built-in although the manufacturer may've had an off-the-shelf one (with source code). I don't remember.

That bootloader performed two checksum (32-bit CRC's--the SoC had a hardware generator) checks before starting the main application.

Every "application" had a "header". It had one checksum. The actual code had another. If one or both were invalid the bootloader would not start the application and the product was "dead" (aka "bricked"). Over time I changed the header which meant a new bootloader/application. A pair. That was fine because I had control of both parts (source code, tools, etc).

However there was problem. The product was already in production and old bootloaders were in the field. Because the application header had changed an updated application would not be recognised. So I wrote an "application" that flashed the new bootloader. Updating this older design was a two step process.

Load the bootloader update (reflash) application
Load the new application

Because the display showed the bootloader version/date on the display this was fine.

To my theory. HP may've done something similar. With a custom bootloader all bets are off. It may've also changed over time. Any firmware update may've required a new bootloader as well as an application. If so that probably wasn't known outside of HP. Any 3rd party developer may not even know about it.

Again all this is a theory only. But it matches what I've read. Good cable, software, reflashing but still no-go.

The solution would be

Finding an application image that matches the bootloader. That may be very hard/impossible.
Update the bootloader to one that that matches the application. Less hard but risky. If reflashing fails the product is totally dead.

As an aside this is the issue with 3rd party updates. They know only what they know. The manufacturer knows everything.

A1

PS: I once performed a firmware update (beta) on a consumer product I'd bought. It failed and was "bricked". I sent it to the manufacturer for repair. No issues after that. However with HP and these "old" calculators that can't be done. HP may not even know how to fix them.

PPS: The HP-15C CE and 2015+ HP-12C+ use a different SoC and contain a generic bootloader from the manufacturer. None of the above applies to them.