Next-gen and mini Axoloti hardware discussion

#22

How Pin compadible is this poing to be with the older axolot?i I'm currently working on expansion board for the orginal that adds fair amount of IO (button, leds, function keys, screens, pots etc.) and would like to open source it for people not interested in building a fully custom interface for the axo. Is it likely to need a substantual revison to work with this?

#23

Compatible pins will be available for the core functions; I'm going to make as much IO available as possible in the space constraints we have to work with. Pins might end up in physically different locations on the board though; the board itself is going to have different dimensions. So an expansion that depends on the actual shape of the original board is unlikely to work directly with the new board. You'd need to wire it up manually. There is going to be a standard header somewhat like the Raspberry Pi that will allow for stacking expansion boards.

#24

Can't really contribute to the technical discussion, but I can guarantee I'll get at least another 2 Axoloti's and force all my friends to buy one if you slap on some more SRAM and CPU power :smiley:

#25

If you design it to fit in Eurorack you would probably make a lot of people happy. Could even add some CV I/O (or head unit) and a Eurorack faceplate to mount it in if you wanna go all out. Not important to me personally, but could help sales.

#26

I actually think that an axoloti fork that would be very successful in general would be that one that includes a minimal interface in a very small package. something a bit like a pocket operator but with way more possibilities. Since @SmashedTransistors provided the oled object, now it became essential in order to be able to switch between different patches in the sdcard and actually knowing what you are loading... (among other things) doing a 4x4 matrix of pushbuttons (+ extras) with its respective LEds shouldn't be super costly and then just adding maybe 4 cheap pots to tweak parameters and one encoder to load patches from Sdcard or other scrolling purposes would be really cool. Specially because. "custom" sequencers and patches that are native for this interface will start showing up in the forum and will work perfectly because of using the same configuration.

#27

Fwiw, new teensy 4.0 with an M7


Maybe an inspiration

#28

I just saw these STM32 controlled (Analog)FPGA boards going for $99. cant find the link. grr.

#29

adafruit is selling their own cortexM4(SAMD51) boards too with teensy specs & more form options- just ordered one. [edit] changed my order to the teensy4.0- thanks!

#30

You’re probably referring to zrna.org,
Designed and sold by @urklang

#31

thats it. zrna/urklang. seems i've seen those names somewhere recently... lol

#32

Thanks for the mention @reaganry. Funny you should mention Adafruit as well. Phil and Limor (of Adafruit) bought a Zrna board from me actually and are planning to do some coverage. Stay tuned for possible future collaborations.

What is particularly amazing to me about Adafruit is that they have a manufacturing facility in NYC so they can produce hardware in house. Getting more FPGA/FPAA/DSP/Audio kinds of applications into the Adafruit ecosystem could be huge and would likely help us be able to bring prices way down.

#33

Quick progress update: stuck taking care of some Zrna-related stuff at the moment. Planning to devote this weekend to the mini Axo layout so we can get it in for a quote early next week.

#34

good stuff. just for clarity, we are talking about a 600mhz processor vs. the existing 168mhz, right? that would be a huge boost in performance...

#35

H7 goes to 480Mhz. Yes, I’d expect a big performance boost.

#36

To clarify: a bump in the clock speed sets us up for performance increases, but there may be gotchas and other bottlenecks lurking in the existing software stack. I'm going to have to put in some work on the software to make sure that we're really taking full advantage of the new hardware. The full benefits may not be there on day one. Like I was saying above my approach is going to be to work for stability and compatibility first and performance later, so we may have extra hardware headroom that isn't fully unlocked right away. On the other hand we may get lucky and be able to unlock right away. I'm not going to know for sure until I have the hardware prototype in front of me.

#37

sorry i was reading about the M7 of the teensy 4.0 :slight_smile:

#38

The Teensy uses an NXP chip based on an M7 core. The STM32H7 is also an M7, just a different implementation in silicon. Interesting whitepaper from NXP: https://www.nxp.com/docs/en/white-paper/I.MXRT1050WP.pdf

Would need to look more closely to see how they stack up in terms of peripherals, etc. What jumps out at me right away is that it only has a 128K boot rom, no large internal flash like the ST chips. I think the Teensy is probably using an external flash chip but I'm not seeing exactly where it is at first glance.

#39

i don't think so. from the pictures https://www.pjrc.com/store/teensy40.html and from the price point (19.95$) it seems it is "just" the one chip actually. there are two other "MCUs" on the board but i can't read them.

#40

As far as I can tell, the flash setup is similar to teensy 3.x, meaning : external flash, and a small MCU used only for programming the flash.

Note : MCU means Micro Controller Unit.

#41

yes. that is why i wrote "MCU"