What would you like to see in The Next Chip?

[Deouss]

I think you're taking about some another chip and IMHO this topic is not about, but about next version of ESP32.

K210 and ESP are designed to do different tasks. They couldn't compete. K210 SoC is rather a CPU than MCU, is very fast, have plenty of RAM, rich audio/video processing skills, but doesn't have Wi-Fi and Blue. Evaluation boards with K210 have... ESP32 for Wi-Fi - what's a shame. ESP32 doesn't have so much computing power, but provides connectivity with 3 lines of code. And have Flash memory in-package, where is the one in K210?

To summarize: it will be very nice and very competing to another industry players if Espressif make a SoC/SoM/SBC with 64-bit multi-core based high computing power and built-in wireless connectivity. But it shouldn't be a replacement for ESP32 - such thing will be bigger, more power hungry and most important: will be more expensive. ESP32 really need new version, but it should be version, not totally new chip. Doubling the bits also double the price. I don' think it also double users satisfaction and Espressif income.

I think you are completely wrong. You probably haven't even looked at K210 specs. The design is very similar to ESP but much improved.
Also K210 has onboard esp8285 module specifically for WiFi so you are wrong again. There are non-wifi modules too
The chip max power is only 250mW when fully operated so it is not so power hungry as you claim here.
Sipeed-M1w
Read specs first please.

[dmaxben]

Whatever Espressif has planned for the future chip...please do NOT remove the CAN bus controller!!

[CollinK]

Whatever Espressif has planned for the future chip...please do NOT remove the CAN bus controller!!

I agree. Though, I would not mind if the CAN controller were upgraded. Say, to a CAN-FD compatible interface. I really, really like the MCP2517FD chip. Now, I know that can't just be put verbatim into an ESP32 but something with those sort of features would be awesome. It's probably not practical though. At the least it would be nice to retain the existing CAN hardware. Though, perhaps figure out the oddities inherent with the 64 byte buffer. I'm told if you fill it up to, say, 60 bytes with frames and then another one comes in it'll get written but only the first 4 bytes then it is plain truncated. But, it's possible to still "receive" the truncated frame but it'll contain all sorts of random garbage. Maybe don't store partial frames and make it look like you can grab a frame there even though you can't/shouldn't.

[jrakus]

I think you are completely wrong. You probably haven't even looked at K210 specs. The design is very similar to ESP but much improved.
Also K210 has onboard esp8285 module specifically for WiFi so you are wrong again. There are non-wifi modules too
The chip max power is only 250mW when fully operated so it is not so power hungry as you claim here.
Sipeed-M1w
Read specs first please.

You really know what K210 is? This is the datasheet of K210 CPU: https://s3.cn-north-1.amazonaws.com.cn/ ... 248_en.pdf
Show us where are the wireless things. K210 doesn't have any of these, it's just a fast CPU with accelerators and some basic peripherals.

The development board from Your link have some kind of SoM module (MAIX-I) with K210 as a CPU and other chips to provide memory and connectivity. Quote from the linked page:

MAIX-I module, or called M1, integrate K210, 3-channel DC-DC power, 8MB/16MB/128MB Flash (M1w module add wifi chip esp8285 on it)

So you're taking about K210 or modules with it? Feel the difference between on-chip and on-board?

[Deouss]

Maybe it should be referred as KPU210 or MAix module similar to ESP module which is not cpu.
And WiFi of external chip is not a bad thing because it wouldn't consume main CPU power or interfere with program tasks.
Anyways it's a different module and future more high end esp module could have some better peripherals and faster memory/cpu
Right now there is some kind of gap between modules like ESP and R-Pi where many projects don't require monstrous boards
with power hungry and complex PC designs like Pi boards but still require a bit more than what current esp module offers

[rudi ;-)]

cheers together,
how looks the roadmap in 2020 ?

i got several answeres to my questions in the past, here and there with espressifer's and direct conntact with espressif (2018),
one was from 6 Dez 2018 that 5 GHZ comes soonish's, also few other goodies and tech news.
i thinked and believe it, that 5GHz comes 2018 QIII-QIV but latest 2019 Q1,
cause we go now from 2019 to 2020 and we talked this in 2018, can you confirm again, that 5 GHz comes or
does it be on(ly at) the List (for) later?.

I think, 5 GHz comes with USB OTG on ESP32-S2 / S3.. but not in SOC itself?
Does we have to handle 5GHz over 3thrd hands? I hope not.

I can imagine that with "pwnd" 3-6-9 months everything threw everything behind. from the beginning ESP and IoT I had opted for ATECCxxxx, ( Olimex, Analoglamb ) so I am not very surprised at the use with ATECC608A and ESP32 in the new ESP32-WROOM-32SE. Unfortunately, glitching is a sensitive topic - and that was also the case with ESP8266 and ESP32 - I hope espressif learned the lessons from it. I talked this theme in the past to espressif but there was no ear for this theme.

In what i am have the most interesst is USB Device ( OTG is fine ) and 5 GHz compatible with 2.4..
Is there progress in the 5 GHz direction? Has it been put on hold? Unfortunately, it was already said in January 2018 that it would be available soon. So far, there has been no report. Unfortunately. And in a few days it will be January 2020, two years ago.

IPO was very successful, maybe too successful? You hear and read very little of people at Espressif with whom you have always been in contact. The "Twitter relationship" with some was getting thinner and then it was a long time coming, unfortunately. There was a time in 2015 when Espressif could be reached. Unfortunately, today you wait in vain for answers and you feel a bit "used".

I understand that here and there company positions have been filled and new tasks are waiting for the previous contacts that were known, but I also have to say that I have not had any of the management levels for over a year (more than 1 1/2 years) heard of Espressif - let alone - that they answered my questions. Too bad. Angus or Jeroen were mostly the only ones who sporadically gave answers here in the forum. Sure Ivan is the Top in few technicals "tricks" with his answeres and helps allways.

I hope that it has nothing to do with the domestic political situation at the moment and that Espressif will always be free from political influences. But I'm afraid that exactly the cut happened here at times like IPO and investment.
Be that as it may - I wish Espressif all the best for the future and would be happy to continue to cultivate and maintain old contacts or that Espressif simply preserves and appreciates such contacts.

Espressif will appear as a CO exhibitor at eMbeddedWorld 2020. After the successful IPO, of course for 2020 I would like Espressif to come as a direct exhibitor. But that's ok, I'm happy that Espressif is coming again. Maybe this will result in some interesting conversations.
Unfortunately, I wasn't lucky enough to test a new ESP32-S2 or MARLIN B3. Anyway - many thought - but no - I got nothing this time

I hope that many suggestions from this thread will be used here, and I also hope that this or that will flow into the new chips. When it will be or what roadmap MileStones it will be, unfortunately you can not see.

Since 2020 will bring some changes, I think there will also be changes here. I hope that we can all survive and use them well.

I hope from espressif and for the future in the other chips from espressif, and from our all suggestions here in this thread, that espressif will return to the beginning as it all started.
also an ear has for the community, more opensource and more open, communicated more directly with it and shows more accessibility and more contact as before.

I am a little disappointed with how it is currently going and how it went.
Company structures change - times change but technologies also change. I am not sure that we can hear everything that Espressif develops and publishes in the world. We watch many channels every day and read some political concerns that are also worrying. Espressif is currently very closed.
Nothing penetrates outside, nothing is speculated.

The future list here is fine and beautiful. But if it only serves hope and not implementation, it will have no future. People watch that too.
Like me - 5GHz since Oct 2017, Jan 2018 soonish's and now two years have passed and all other manufacturers are passing where Espressif was once a pioneer and figurehead.

I am sorry for the word and difficult for anyone who knows me.
I hope that changes and actions will follow soon.

missing the roadmap and the milestones from this thread.


best wishes
rudi

[ESP_Sprite]

Hi Rudi! Let me see if I can reply to some of your points, where I know the answers. Not sure how satisfactory they'll be to you, but here goes.

First of all: the roadmap. For 5GHz, all I can say is that it's still on our roadmap. Analog stuff is hard and a chip is the end result of many different development processes; all I can say is that we're getting there, but we don't have anything to announce at this point. If any, the next chips (ESP32-S2 and -S3) will not have it yet.

Wrt USB OTG: It's likely that this will be included in any chip that we make that's not on the very bottom end. I can't give any guarantees of speed (-S2 only has low/full-speed, for instance), but I don't see the USB in the -S2 being an one-off.

I can't say anything about closed-ness of Espressif, apart from that we don't have anyone specialized in handling social media working for us: both replying on this forum as well as the Twitters etc are all things someone does next to their main job. It may just be that what you're seeing is a result of them being preoccupied with other things.

If any, when it comes to this topic, we still take everyones ideas into account, and some of them made it into the USB32-S2 already. If any, I can be pretty confident in saying that if people didn't heap on the request for USB here, it may not have made it into the end chip design. The thing with making chips is that it's a long process. The ESP32-S2 was an exception, but as an indication, USB was requested as a feature 2.5 years ago. No such thing as instant gratification here, I fear.

Wrt roadmap: not sure what I can say that hasn't been said before. To recap, and with the usual disclaimer that what I say here is what I know and may not play out like this: we're expecting the ESP32-S2 non-beta version to get here Any Moment Now (say, Q1/2020) and that chip is supposed to be a somewhat cut-down version of the ESP32. We're not ignoring the fact that there's also a need for more powerful chips than the S2 as well, so there will be something slightly more relevant there coming after that. Also, you can take the new ULP in the -S2 as a sign of things to come: maybe not immediately, but certainly soon.

[bvovag]

what lauros ...
I thing you need at first not a H/W update, but release a properly working & correctly documented IDF.

[ESP_Sprite]

Noted, but as stated before, that is not the subject of the topic. For one, if we don't develop chips in parallel with optimizing ESP-IDF, we'll have an office full of chip designers and layouters twiddling their thumbs, as you probably don't want to see them to try writing embedded software. Furthermore, if you have issues with ESP-IDF, I'd invite you to document them as issues on our Github rather than posting here.

[tvoneicken]

After spending a couple of days banging my head against auto-light-sleep I wish the next round of chips have a bit more thought put into low power operation. Specifically, I find the fact that most peripherals stop working properly in light sleep to be abysmal, the uart being the first issue that must just hit everybody. In this specific case I expect the uart and its buffers to continue operating like normal unless I choose to switch the device off. In the STM32L world there are registers to power each peripheral on/off and a table that shows typ power consumption for each one. This way one can easily decide what to leave powered on and what to cut together with the CPU. Yes, the esp32 is much, much better than the esp8266, so great progress, but sadly you haven't arrived yet.
I think also more details into the functioning of the peripherals would be a good investment as it should carry forward to new processor designs, e.g. an ADC that delivers accurate results. My fear is that such fixes don't make for great PR headlines and thus get left by the wayside...