I like the Apollo3 in the Sparkfun Artemis module, but yowza, take a look at its new big brother the Apollo4:
- 2M of MRAM (I think that means it is magnetoresistive non-volatile RAM instead of flash for program storage)
- 1.8M of SRAM
- Cortex M4 with floating point running at 96 MHz or 192 MHz processor clock
- USB interface
- Bluetooth LE
- MIPI Display Driver
- Crypto engine
- 3uA/MHz]
More details here: https://ambiq.com/apollo4/
Hey Sparkfun, any plans for an Artemis Module based on the Apollo4?
If the apollo4 clocking mechanisms are like the apollo3, it means that the apollo4 at 96 MHz will draw the same power as an apollo3 at 48 MHz. It would be amazing!
I can’t help myself: long ago (early-ish 1990’s), I worked at HP Labs. At one point, I got a brand new HP PA-RISC workstation. It ran at 99 MHz, and it was unbelievably fast compared to the PCs of the day. The new Apollo4 is roughly as capable a processor, and running at 96 MHz, it will be drawing about the same amount of current as the ‘power on’ indicator LED on that old workstation. Where will we be in another 25 years…
SWEET!
+1 for SFE Apollo4 based board
Any updates on that? Some plans? Should we expect that the code developed for Ambiq 3 would easily be transferable to Amibq 4 or would it be a big jump?
It is hard to say definitively because there is no data sheet released for the Apollo4 yet. I would not expect it to be much different from a programming perspective. Also, the Ambiq HAL has done a pretty good job so far at abstracting away the differences between the other members of the Ambiq processor family, so the HAL should take care of a lot of the changes.
The big difference is for Sparkfun, not us. The Apollo4 comes in a different physical package meaning that Sparkfun needs to redesign their module. The Apollo4 has an extra 23 pins and is a different shape. That said, from a user perspective, there might be a potential advantage if Sparkfun could use the same module pinout as they ship now. That would allow for easy migration from Apollo3 to Apollo 4. On the other hand, that would hobble the Apollo4 module because it would mean that not all of its pins would be available at the module pinout. And from a practical perspective, it looks like Sparkfun had quite a time getting the Apollo3 pins out to the module pins in the first place. If the Apollo 4 pinout is much different, it might be an impossible task to make the module pinout match. My expectation is that there would be a new module. And if that’s the case, then I would vote for Sparkfun to add the 32KHz Xtal and xtal caps to the module instead of making all of us put it on our motherboards.
Apollo3 Blue Plus with twice the speed, half the power, more pins, smaller chip 4.7 X 4.7 and still waiting on the Apollo3 Blue EVB and ambiq ‘Sale’ to release something besides ‘Apollo4 Product Brief’. I’m still trying to get an response on the Apollo3 AMAP31KK-KCR (WLCSP) 3.25 mm x 3.37 mm, 66-pin WLCSP with 37 GPIO.
robin_hodgson:
That said, from a user perspective, there might be a potential advantage if Sparkfun could use the same module pinout as they ship now.
And not tie all the power supplies together. As far as foot print, more pins, make it longer to get all the pins broken out this time.