registerOne
New member
I think it is rather classy you do this. Makes me want to support _you_ rather than any corp/company.
open-source teensy-compatible - what features do you want?
- Thread starter ptorrone
- Start date
- Status
I think allowing people to dig their own hole is the right move here. You don't need to say a word, it's all pretty clear to us
I thought the OP would have come back after they'd cooled off and asked for the thread to be removed, because it definitely doesn't reflect well on them. But it seems they've doubled down and are directing people here from their own website/blog.
The response to their post on the eevblog forum was pretty much the same as here ("this isn't Teensy-compatible, it's just another RP2350") and they haven't bothered to follow-up there either, so I hope nobody holds their breath waiting for this product to come to market given the reception it's gotten.
I've been playing with the C++ SDK for the Pico for a while now and I'm really not a fan of the dev workflow particularly writing new firmware. It takes a heck of a lot of work to get your dev environment setup and working well. Also a lot of the abstractions seem far away, and much less well documented for nuance of their physical interactions like all the work that has gone into the Teensy libs. PIO is attractive until you have to actually write some complex control code like a servo loop and then just wish you had the raw FLOPS the teensy gives you... A lot of work is towards circuit python and micropython which unfortunately any time I try to use for something important I just hit bugs and weird slow constructs buried in 5 architectural choices for "ease of use" - this unfortunately is not great for fast PIDs or FFT where the Teensy shines.
All in all I agree with most here comparing the RP2350 to the NXP chips is like apples and potatoes. Would be great if companies like NXP eased up the NDA/IP restrictions which actually make open sourcing the Teensy loader problematic but that's not the world we live in.
Shirley Marquez
New member
Those are all features of the microcontroller used in some of the 3.X series. (The inputs on the Teensy 3.6 are not 5V tolerant.) The microcontroller in the 4.X series, which is otherwise a solid choice and has much more processor power than the 3.X series, does not have them.
Microcontrollers with 5V tolerance are becoming rare as the industry moves to smaller process nodes. I don't think anybody makes one that offers the performance of the chip in the 4.X series. Microchip and Renesas make some 5V-tolerant microcontrollers; both companies are big in automotive applications, where the higher noise margins of 5V remain valuable.
The Audio Shield can serve as a replacement for the DAC and ADC in some applications. Its bit depth and low noise are great, but its low sample rate is a limitation that make it unsuitable for some purposes.
mjs513
Senior Member+
Shirley Marquez
New member
Adafruit is committed to making all of its own designs open source hardware, which includes all associated firmware. That's a stance, and one that I applaud them for taking, but it means there are products they can't make. (I'm not saying that everybody has to make that choice, but I'm glad that somebody is.) NDA restrictions on the chip used in Teensy 4.x make it impossible for them to use that chip, so they had to choose a different route.
Their "Freensy" product, which probably won't be marketed with that name (PJRC would rightly see that as trademark infringement), is in the same ballpark of processing power as the Teensy 3.X series, and it has the 5V tolerance and ADC (but not the DAC) that are missing from the Teensy 4.x series. There may be a market for that, especially at the expected price of $10 or less. (It's much easier to hit that price with a $1 microcontroller.) They will have a lot of software work to do to make it a real Teensy competitor; time will tell. They will be able to lean on what the Arduino RP2040 team has already done; presumably they will add support for it as they have done for other RP2040 and RP2350 boards.
PJRC has thus far chosen not to develop a new product that is comparable to the discontinued Teensy 3.x series, which they had to kill because of inability to get the microcontrollers. They're a small company, and likely felt that their resources were better spent on developing a higher performance product rather than serving a niche that's likely to become more of a commodity space. I'm happy that we have the Teensy 4.x series; right now there isn't anything else that fills the same needs at an affordable price. The manufacturing partnership with Sparkfun may enable them to broaden their product line in the future.
KurtE
Senior Member+
Curious, are you saying that they will not make any boards for example that have an IMXRT chip on it, as NXP have NDA to find out everything
about these chips? Especially when several others have made their own Teensy boards? I understand they would not use Paul's bootloader, but
they could I use a different approach?
Personally I keep wondering if this thread is a good or bad thing. I really do wish Adafruit and Sparkfun could settle their differences as I have
purchased things from both and like both!
And I can imagine that Adafruit, still has to decide on their response. They could simply move on/or they could try to build a Teensy 4.x
replacement or with an IMXRT106x or maybe less likely 5.x with something like 1170. Maybe it has same footprint as 4.1? Maybe it grows it
slightly with double row of pins? and brings out enough pins for Cameras and displays, maybe built on SDRAM, PSRAM?, maybe BT/Wifi..., with a
different bootloader which allows debug support, etc. They could fork the Teensy cores and make whatever adjustments they need.
They could probably easily add their variant to Circuit python, maybe they add it to Zephyr and extend it over the minimal T4.x builds...
And/Or maybe they build a version that has a larger form factor like an Arduino UNO or Giga and in more features, etc...
Time will tell!
PaulStoffregen
Well-known member
I'm wondering if some of these recent messages may be chat bot generated?
ETMoody3
Well-known member
I think this is a reasonable response, and I have enough respect for Paul to want to offer another voice of support.
The original post... doesn't exactly make this Adafruit representative look very good. Frankly I don't think they need another dev board but it's their money to spend. The most important comment I feel I need to make is that Adafruit would never have tolerated this thread on their forums.
Shirley Marquez
New member
Yes, I am saying exactly that. They may sell boards from other companies, as they did with the Teensy, but they will not make any.
Yeah i got that too. And they linked to this discussion for the spec's on the new board. I thought that was kind of crappy.
to me the library support means nothing, my firmwares are in rust and i pretty much need to custom write a lot of modules anyway. the only appeal the teensy has for me is the raw power which is useful sometimes. this is just an rp2350 in another form factor (not that i hate it, my FCs are built on the RP235x). but there are already so many boards on the market with the same chip
i would strongly suggest to make this a long term project and use the imxrt
i would strongly suggest to make this a long term project and use the imxrt
Shirley Marquez
New member
It does have a closed-source replacement, the circuit design component of Autodesk Fusion. But I'd like to see them move their future designs to KiCAD.
First, thank you Paul for curating the community here. Although I haven't had a need to create an account until today, I'm sure I speak for many unregistered users when I say that this community has been extremely helpful.
Phillip,
As a longtime Adafruit (and SparkFun) customer, a user of both the RP-series micros (PIO and all) and most Teensy revisions (mostly 4.1), I'm at a loss for a good response to this.
I appreciate that Adafruit values the open source community in the way that you do. That said, Adafruit is a company, and not a small one, nor one that is hurting for revenue, as far as I can tell. Adafruit arguably also has more name recognition, a larger customer base (I've seen boasts online of 8-figure revenues) and more community reach than PJRC. Putting it bluntly, coming to the PJRC forum and posting as such really can only be interpreted as advertising. "Soliciting input", "getting community feedback", "request for comments", whatever you want to call it, the plain truth is that the summary of the thread is "we can't sell PJRC's product anymore, so we're going to make our own". In doing so, what you're really doing is just taking advantage of an actual small business, their good name, and their community in hopes to recoup losses in a petty battle with another company. Nothing I've seen so far would benefit Paul, regardless of his relationship with SparkFun.
Teensy isn't really about the pinout, although an ecosystem has been built around it. You're re-creating your Feather boards in a debatably-new form factor, considering the Adafruit KB2040 Kee Boar was already a stab at Teensy's popularity in the keyboard space (and has a suspiciously similar layout). I don't feel the need to justify what I use the Teensy 4.1 for, or why I carefully chose it even when things like the RP-series micros exist, but suffice to say that to be a Teensy is to be the hardware, chip for chip, and everyone has a different opinion why.
I guess what I'm trying to say is that if you really want to make a new microcontroller form factor and offer some RP*, ESP*, STM*, whatever variants, just do it. You've proven you don't need PJRC for that, so don't come to their forums, posting with your personal account, and advertise as a competitor. It's disreputable, outright insulting to Paul et. al., and nobody benefits from it. If you have company beef, solve it by coming up with fresh ideas.
I expect more from Adafruit.
Colby
Phillip,
As a longtime Adafruit (and SparkFun) customer, a user of both the RP-series micros (PIO and all) and most Teensy revisions (mostly 4.1), I'm at a loss for a good response to this.
I appreciate that Adafruit values the open source community in the way that you do. That said, Adafruit is a company, and not a small one, nor one that is hurting for revenue, as far as I can tell. Adafruit arguably also has more name recognition, a larger customer base (I've seen boasts online of 8-figure revenues) and more community reach than PJRC. Putting it bluntly, coming to the PJRC forum and posting as such really can only be interpreted as advertising. "Soliciting input", "getting community feedback", "request for comments", whatever you want to call it, the plain truth is that the summary of the thread is "we can't sell PJRC's product anymore, so we're going to make our own". In doing so, what you're really doing is just taking advantage of an actual small business, their good name, and their community in hopes to recoup losses in a petty battle with another company. Nothing I've seen so far would benefit Paul, regardless of his relationship with SparkFun.
Teensy isn't really about the pinout, although an ecosystem has been built around it. You're re-creating your Feather boards in a debatably-new form factor, considering the Adafruit KB2040 Kee Boar was already a stab at Teensy's popularity in the keyboard space (and has a suspiciously similar layout). I don't feel the need to justify what I use the Teensy 4.1 for, or why I carefully chose it even when things like the RP-series micros exist, but suffice to say that to be a Teensy is to be the hardware, chip for chip, and everyone has a different opinion why.
I guess what I'm trying to say is that if you really want to make a new microcontroller form factor and offer some RP*, ESP*, STM*, whatever variants, just do it. You've proven you don't need PJRC for that, so don't come to their forums, posting with your personal account, and advertise as a competitor. It's disreputable, outright insulting to Paul et. al., and nobody benefits from it. If you have company beef, solve it by coming up with fresh ideas.
I expect more from Adafruit.
Colby
Here's an interesting thing which I hadn't mentioned but will now:
I said in my post that I only had one bad interaction with AdaFruit over the years. Here is what it was: I was using a BLE adapter that they made. I wanted to create a large number of IDs for various info available from my device. The firmware they had on the BLE module was limited to 30 IDs. So, naturally, them being an open source advocate, I wanted the firmware so I could edit it to vastly increase the limit of IDs. They would not (and could not) do that. You see, the BLE stack was actually under NDA and closed. The firmware they wrote could have been open but would have relied on closed source stuff under NDA and so they didn't even release their own code. There was a bit of an argument related to that whole thing. Ended up dropping the BLE idea and looking elsewhere for ways of presenting the data values.
I believe this is basically the adapter I was using back then: https://www.adafruit.com/product/2479
And, here is the github repo for that device's firmware: https://github.com/adafruit/Adafruit_BluefruitLE_Firmware
You might notice that even today, there are only hex files there. The firmware is NOT open and never was. So, I don't necessarily buy the "we're so open source! We can't stand closed source stuff!" when I first hand experienced that they're plenty willing to go closed source if it's convenient.
ptorrone
Active member
rp2350 dual-core coremark results: first numbers are in
hey folks, just got back from our two-month pediatrician visit and, in between bottles and naps, we had claude code update the coremark repo to support the rp2350 running dual-core, multi-threaded.
results look solid.
for reference, a teensy 4.0 lands around ~2300 coremark. the rp2350, overclocked to 276 mhz, comes in at ~1400. not bad at all, especially given where this chip sits in the ecosystem and what it’s optimized for.
the repo is up, the changes are straightforward, and the benchmarks are now in the readme. if you want to dig in, clone it, tweak clocks, or argue about flags, have at it.
github.com
coremark isn’t everything, but it’s a useful yardstick, and this puts the rp2350 firmly on the ... board!
hey folks, just got back from our two-month pediatrician visit and, in between bottles and naps, we had claude code update the coremark repo to support the rp2350 running dual-core, multi-threaded.
results look solid.
for reference, a teensy 4.0 lands around ~2300 coremark. the rp2350, overclocked to 276 mhz, comes in at ~1400. not bad at all, especially given where this chip sits in the ecosystem and what it’s optimized for.
the repo is up, the changes are straightforward, and the benchmarks are now in the readme. if you want to dig in, clone it, tweak clocks, or argue about flags, have at it.
GitHub - ladyada/CoreMark at dual-core-progress
Benchmark CPU Performance on Arduino Compatible Boards - GitHub - ladyada/CoreMark at dual-core-progress
github.com
coremark isn’t everything, but it’s a useful yardstick, and this puts the rp2350 firmly on the ... board!
ptorrone
Active member
the person of course fails to mention we have been moving to kicad, it takes time, we have over 800 open-source hardware certifications, we donate and support kicad, i totally get that there are new people here that are only here to dunk on me, that's fine - and there are some folks that have great experiences, and some that will have some things to gripe about too, we're here and listening and we'll continue to make and ship hardware, if there is an opportunity to resolve any of the issues sparkfun has with us, we're always open.
Phillip, not trying to be combative, but I have to call your bluff here. You posted a brand new board in the OP that is made in EAGLE. It doesn't make sense to be upset about Teensy being a "closed" platform, while simultaneously starting brand new designs in a tool that is already abandonware and won't be available to download after June 7, 2026.
Fusion for Personal Use, Autodesk's free replacement for EAGLE, only allows for 2-layer designs, meaning "Freensy" won't be editable in it. Users will have to purchase a commercial Fusion license for $680/year to even open the board natively.
None of this holds water.
Fusion for Personal Use, Autodesk's free replacement for EAGLE, only allows for 2-layer designs, meaning "Freensy" won't be editable in it. Users will have to purchase a commercial Fusion license for $680/year to even open the board natively.
None of this holds water.
For the record, I like EagleCAD and still use it. I've got 9.6.2 and a 6.x version installed. You can keep using the old versions of Eagle forever, so long as you keep the files yourself. What stops in the middle of this year is the ability to download it. I'm not sure if it'll still try to phone home after that point and fail. I have a Fusion360 subscription anyway so it may not effect me. And, I can always use my Eagle6 install.
But, I think it's important to delineate between open files and the ability to open them with free tools. It is possible to create open source that requires closed source software. For instance, I could write an open app that compiles with a commercial compiler. It's still open source but not exactly easy or free to compile it yourself. Also, I could call a board open if I were to publish the gerbers and a PDF schematic. From that you can bring it into any CAD package you want. Is it convenient? Not so much. But it's technically open.
KiCAD does a pretty good job of importing Eagle files so if it bothers you that the files are in Eagle, maybe just import them to KiCAD and have at it.
So, while I too question what is going on here, I'm not sure I feel too strongly about this particular issue. I think it's perfectly fine if they use any package they want to design a circuit board. It's up to everyone else whether they want to use the released files after that.
I would still contend that it'd be better to make a drop-in compatible board if they're really insistent on going the totally open route. But, at the end of the day, we are all free to take it or leave it. So, I guess AF can do what they want and see where the chips land.
Collin, appreciate your take. I was also an EAGLE user before switching to KiCAD, and agree that in the grand scheme of things, EAGLE usage in a vacuum is not a huge deal. (I don't believe source-available qualifies as OSHW, but that's a separate topic.)
The reason I'm bringing it up here is to draw attention to the wider context of Adafruit's pattern of "purity testing" other OSHW enthusiasts and organizations. As a company, they often use their platform to browbeat others for not being "open enough". They wield their large OSHWA corpus as a cudgel, but also hide behind it as an excuse not to migrate, while continuing to produce designs in a tool they know future travelers won't have access to, and will be forced to import into KiCAD to edit. EAGLE usage in this specific context is an unfortunate footgun that damages their credibility.
The reason I'm bringing it up here is to draw attention to the wider context of Adafruit's pattern of "purity testing" other OSHW enthusiasts and organizations. As a company, they often use their platform to browbeat others for not being "open enough". They wield their large OSHWA corpus as a cudgel, but also hide behind it as an excuse not to migrate, while continuing to produce designs in a tool they know future travelers won't have access to, and will be forced to import into KiCAD to edit. EAGLE usage in this specific context is an unfortunate footgun that damages their credibility.
Last edited:
hi phil! i’m one of the new people who’s here to dunk on you!
just wanted to say that i had a similar experience with your closed source products as CollinK a few years back. talked to yall a few times online and in person about it but neither of you were very nice or “open source” about it to me, either.
i generally think that people should be able to run their business however they want, but having followed you since the early days of make/hackaday and have seen your style of engagement/open source purity testing in attempts to consolidate power over the years, i’m really enjoying seeing how tilted you are about this!
also am a little bit weirded out about how much you’re talking about limor’s breasts and the details of your home life as a way to to make people sympathetic to your cause, but i guess everyone has a different way of expressing it when they get upset about something.
cheers, get some sleep!
just wanted to say that i had a similar experience with your closed source products as CollinK a few years back. talked to yall a few times online and in person about it but neither of you were very nice or “open source” about it to me, either.
i generally think that people should be able to run their business however they want, but having followed you since the early days of make/hackaday and have seen your style of engagement/open source purity testing in attempts to consolidate power over the years, i’m really enjoying seeing how tilted you are about this!
also am a little bit weirded out about how much you’re talking about limor’s breasts and the details of your home life as a way to to make people sympathetic to your cause, but i guess everyone has a different way of expressing it when they get upset about something.
cheers, get some sleep!
- Status