Any Plans for a Teensy Board with Bottom Pads on Top?

Status
Not open for further replies.
Seeing a trend here ... Wingsy - the Teensy that can fly.
Umm, that would have to be the Weengsy and not the Wingsy, although it would be pronounced the exact same way, at least in America.

I have a number of projects where I've soldered R-A headers into Weengsy and other boards. They even sell 2- and 3-row R-A headers for guys who build pcbs the way I do. Great for low-profile boxes.
https://www.ebay.com/itm/1pcs-90-De...Right-Angle-Pin-Header-Connector/252492439694

wungsy.jpg
 
This is a difficult design problem. On how to get out the most pins from the CPU while keeping the module footprint small and cost low and make everyone happy. With all this back and forth on extra pins, what is PRJC's position so far?

What is currently the most popular Teensy board?
 
dual row for header pins could also mean something like using IDE ribbon cables to bring out the data channels :p
 
I tend to agree, even Teensy 3.5 & 3.6 having 42 breadboard accessible I/O signals is a bit much. Standard breadboards have 63 columns. With only 39 left over, you can't really build up circuitry that uses so many signals.
And a lot of breadboards are the 1/2 size (30x2 pins), which using the 3.5/3.6 only leaves you with 6x2 pins on the breadboard.

I find myself going back to the Teensy 3.2 a lot, to keep the size small. I hope there will continue to be something still in that size factor. If the shortage of MK20DX256VLH7 boards is not temporary, perhaps retrofitting a MK64FX512 (Teensy 3.5) into the 14x2 DIP size would help those of us that want the classic Teensy layout, and don't need all of the pins.
 
Now, it would be interesting to know what the motivation of the Teensy 3.5/3.6 buyers is... Did they buy these mainly because of the increased speed/performance/memory, or would a Teensy 3.2 or LC have been sufficient in terms of horsepower and they "upgraded" just because of the number of available pins?

Contacting T3.5/3.6 buyers and inviting them to an online survey might give answers.

BTW: I am currently developing a project around T3.2, but I'd really like to do the T3.6 upgrade because of the better performance and the FPU, but I unfortunately can't in terms of available physical space. While most users tend to ask for more pins, I'm most probably the only one to wish a T3.6 in 14x2 pin format (no SD card needed)...
 
I have a pretty strong impression, admittedly all from anecdotal evidence, that most people who choose 3.6 do so for the extra speed or memory. Especially audio synth projects benefit from the extra performance. Some people go for the extra speed "just in case". USB host might also be a factor.

Teensy 3.5, which sells the least of all 32 bit boards, is harder to figure out. It seems to be a popular upgrade path from Arduino Mega. The larger number of I/O pins may indeed be a factor. But even with 3.5, a pretty consistent stream of feedback is Mega just doesn't have enough RAM when several modern Arduino libraries are used together.
 
Teensy 3.5, which sells the least of all 32 bit boards, is harder to figure out.

Speaking for myself, it's the combination of 5V tolerance and high pin count. The 3.2 is my preferred device for projects that don't need so many pins. As for processing power, all the 3.x have been massive overkill for my projects so far, but it is nice not to have to worry about it.

--
Best regards, Kári.
 
Teensy 3.5, which sells the least of all 32 bit boards, is harder to figure out.

For me, I upgraded to the T3.5 from the mega for the combination of all the factors mentioned, 5V tolerance, pin count, performance (memory, speed, FPU, etc). It was also the form factor that attracted me, I could breadboard some project elements before final mounting. It also gave me the ability to make custom breakout boards for my projects. To be honest the only reason I haven't used the T3.6 is because the pins aren't 5V tolerant, just very nervous about that one. Have to get over this one :).

Think the other reason I moved away from the Arduino boards is this forum and the support that everyone gives as well as the turn time to update the core or unique libraries if there are issues. Think Paul has mentioned the decision by committee that goes on before.

Respectfully,
Mike
 
For me (a retired Software person), who still enjoys tinkering... I have most recently done most of my tinkering with a 3.6.... Why? It is nice and Shiny ;)

I used to be a lot more concerned about 5v tolerance, which back then is why I waited until the 3.1 came out before I tried a Teensy. But now days I am not as concerned as more and more of the devices/sensors are now 3.3v (or lower)

For just about anything I do, the 3.2 is more than sufficient. But I do like the 3.5/6 as it does give you lots of options (more Usarts, SPI, I2C). Also for those things that need a reasonable number of IO pins, having more on the top row is convenient.
 
Since we are doing anecdotal evidence :D

3.6 for the SD card
LC for neopixel projects with that 5v signal pin
3.2 for everything else

(I know, I am a lazy amateur)
 
Paul: As a placeholder on your napkin drawing did the T_4_? start with the T_3.6 form factor with SD? Hopefully the SD/SPI errata cleared up and will allow it to run full speed? Seems the K664/K64 had errata that stopped ideal SD usage?

T_3.2 size was too good - anything else seems big - unplugging a pinned T_3.6 seems scary and painful. Working with FrankB's Teensy64/C64 needs the T_3.6 so those are on my desk just now.

Hey look - the perfect sized 1.4"x0.7" for $10? Brand new and 2KB RAM and 28K FLASH!
Adafruit Itsy Bitsy 32u4 - 5V 16MHz: What's smaller than a Feather but larger than a Trinket? It's an Itsy Bitsy! Small, powerful, Arduino-compatible - this microcontroller board is perfect when you want something very compact, but still with a bunch of pins.
 
Paul has the great fun of trying to be all things to all people, :).

- I use T3.1,3.2 for most projects, and rarely need all of the available I/O pins as it is, although some of my robots do use most of the I/O pins, due to having a lot of sensors and 5V tolerance is good for this. The greatest limitation of T3.1,3.2 for me is the single SPI port.

- I went to T3.6 for my Arducam-Mini vision project, due to speed, amount of RAM (it can hold a couple of image frames), SD card, and multiple SPI ports, so I can more easily interface camera, LCD, and external SPI RAM chips (23LC1024). I will also be using T3.6 on future expansions of my robots, as the current programs are >100K and I will be adding more high-level, abstract kinds of processing that can use the higher clock speeds.

A lot of people seem to like the smaller 28-pin modules, which is a very convenient size, but also express desire for there being a T3.6 chip on this size pcb. However, I would hate to see the SD card socket go away. So my "favorite" idea would be a T3.6 chip on a 28-pin module form-factor, but expanded to include "2" extra rows of through-hole pads, top and bottom in parallel with the current rows. So, 0.9" tall and 1.5" long. I would remove the end-pads and put an SD card socket there, so it might have to actually be 0.2-0.3" longer. Then, a few leftover I/O pins could go on smt pads on the bottom, like at present, although no one might ever need them.

Too bad for breadboard people, however.
 
defragster said:
Adafruit Itsy Bitsy 32u4 - 5V 16MHz: What's smaller than a Feather but larger than a Trinket? It's an Itsy Bitsy! Small, powerful, Arduino-compatible - this microcontroller board is perfect when you want something very compact, but still with a bunch of pins.

Maybe Paul should trademark "Eeensy Weensy Teensy" before it's too late. Also, FYI:
https://www.kididdles.com/lyrics/e001.html
 
I mainly buy 3.6 modules - for the processing power, speed, RAM, memory, pins, 14bit ADC and price. I do find the 3.3V limitation annoying because many of my sensors are 5V, but it's nice that I can swap out the 3.6 for a 3.5. The pin compatibility between 3.2-3.6 is good.

It's surprising that Teensy 3.5 & 3.6 popularity is on equal footing with the LC, which is second to the 3.2. I suspect the future Teensy will go in this direction. I hope it features a more powerful CPU, but also some smart and easy way to access more pins. I mean, it's a shame to throw away so many CPU pins. Bottom pads are a reasonable attempt to capture some, but bottom is difficult. Maybe something like this instead:

https://www.adafruit.com/product/2772

That pin matrix is for prototyping only and for Feather options. I'd rather have them as active CPU pins. Then I could easily solder-in top side headers. And still be breadboard friendly.

Too bad for breadboard people, however.

That is risky for the customer base. A big problem for me with Arduino was always having two big separate parts. The Due and breadboard. Wires being pulled apart. SPI and I2C transmission line problems. Etc. I'm seeing a trend on breadboard friendly modules. Arduino is adopting these. Adafruit began with this form factor. MBED is still in business.
 
Last edited:
https://www.adafruit.com/product/2772

The Feather board, that people keep bringing up, is like comparing a 1918 Model-A Ford to a 2018 Audi. Feather basically has the power of a T-LC, and only 20 general I/O pins. Not hardly a Teensy 3.6. Lady Ada is about 5-years behind the curve, Paul is the one pushing the curve.
 
... comparing a 1918 Model-A Ford to a 2018 Audi. Feather basically has the power of a T-LC, and only 20 general I/O pins. Not hardly a Teensy 3.6. Lady Ada is about 5-years behind the curve, Paul is the one pushing the curve.

That was the direction of my post #63 above - this is a NEW product with 8 bit cpu - at least they got the size and format right - except using 5V ... all of which Paul starting pushing with his 8 bit Teensy work some years back - with some MYSTERIOUS HALF KAY bootloader.
 
What is PJRC's thought on microcontroller for the future Teensy? The Kinetis K66 is already at the top of the K-series performance class. Are you considering a new direction, like the i.MX processor?
 
Regarding castellated edges, the ESP12 model (And I guess Edison) is kind of neat. There would be an adapter board for bread board that fits the current footprint exactly. It does double the number of things that need to be printed and tested, so I can completely see why that's a no-go.

Regarding T4, I was wondering what https://github.com/PaulStoffregen/cores/tree/master/teensy4 was all about ;)

Just please one reverse feature request. Please never put Linux or other bloated disaster on these boards. I don't want to ever deal with an OS or scheduler and all that jazz. Thanks!
 
I think one of the challenges with the new design is how much of the chip capabilities you want to break out for users. If you look at the i.mx rt 1050 block diagram (https://www.nxp.com/docs/en/white-paper/I.MXRT1050WP.pdf) and the 196 pin bga form factor for the processor it going to be design challenge to keep it to a T3.2 size if you want access to those features think you may have to do some creative layout Maybe a T3.5/3.6 form with double row or a little wider with pins or header at one end. Just thinking out loud here.
 
How much info to share, and how early, and how much public conversation about future features to have is a difficult decision. While I'd like to be as open as possible, and I want to consider as much feedback as I can, there certainly are some risks. Likewise for publishing early code, hardware details, and shipping beta test boards.

Arduino went down this path in early 2011 with Leonardo. The shared everything, believing the community would help. No meaningful contributions came. A few hobbyists made boards over the summer, selling just a few. By late 2011, Adafruit was shipping a clone (but quickly reverted to LUFA and stopped making them when customers had trouble), and in early 2012 Sparkfun was shipping a huge volume, all using the buggy beta bootloader Arduino had published, all before Arduino sold their first Leonardo board in mid 2012. It was Arduino's first product requiring significant new bootloader firmware development and substantial changes within the IDE. Painful lessons were learned.

Arduino also went through relearning the same painful lesson of inflating user expectations by sharing early info. Both Arduino Due and Yun were "released" in announcements at Maker Faire in San Mateo, but weren't actually completed or ready to ship. As happens with new products, both were repeatedly delayed. Arduino suffered quite a lot of negative publicity as a result, on top of the growing frustration and resentment from quite a few very vocal users, naturally disappointed they couldn't actually buy those boards. Massimo instituted a policy of not talking about new products until they were actually ready to ship. But that was just about the time Federico Musto leveraged his position with the development of Yun and a buyout of Martino's manufacturing operation to try to take over all of Arduino. Musto repeated the same well worn mistakes, with a lot of hype about Arduino Star Otto and Primo. Even now, there's quite a number of people unfamiliar with the Musto drama who hold up the failure of Star Otto to mean Arduino has lost its way (despite the fact they're recently released several new products, but focused in a different direction).

So please understand I'm trying to balance openness with the risk of over-inflated expectations of vaporware.

Delays are the norm with new products. There's a very good chance any sort of Teensy4 might not finally ship until sometime in 2019. I can tell you both Teensy LC and 3.5/3.6 were in design here for nearly a year, and I worked on Teensy 3.0 for nearly 2 years before release. While I do need to be careful about iMX-RT details which may still be under NDA (but most info is now online), if you read carefully through the public info NXP has published, you'll see mention of version 1.0 and 1.1 of the silicon buried in a few places. I can tell you we're waiting on version 1.1. PJRC will not ship any boards with the 1.0 version of the chip. I'm not at liberty to discuss why, but trust me it will be worth the wait.

Having just said all this, I will risk mentioning I'm leaning towards two possible paths on the form factor. One way would look like 2 different products, mimicking the Teensy 3.2 form factor as closely as possible with the cost as low as we can manage, with the second higher cost board having a "long" form factor (maybe Teensy 3.6 pinout, maybe longer) and provisions built in for all sorts of I/O. The other general idea I'm considering is adopting either Adafruit Feather or Arduino MKR form factor, probably with some sort of extension for the many extra I/O pins, since both of those provide relatively little I/O. The general idea there would be compatibility with their shields, and future Teensy shields being compatible with their boards.

Again, this is all still *very* early. Expect this all to take a very long time. It will feel like vapor! With other new products, I tried to be open about details, but we asked everyone to please keep everything only on this form, and as betas were shipped, we asked every to not post any photos or realistic diagrams. PJRC & Teensy aren't anywhere on the scale of Arduino. So far sharing on lengthy forum threads seems to work, and I really hope that can continue. I am reading every comment here and trying to consider every piece of feedback, even if I don't manage to directly reply to every message.
 
Thanks @Paul

Let me know if there is anything I can help with (or any new shiny things :D )
 
Status
Not open for further replies.
Back
Top