Why does the Teensy 4.0 have fewer pins than the 3.6 and some othe stuff?

Status
Not open for further replies.

PAAUX4

Well-known member
Sorry if this has been asked before.

It looks like the focus is more on using protocols etc. rather than messing about with pins directly, almost like a computer. This is like when we lost the parallel port.

Also, somewhat unrelated:
  • What is FlexIO?
  • Are pins D- and D- under the USB port for the first or second USB port?
  • What benefit does the power pin have over a hardware switch? Is it just the ability to use a momentary switch?


Thank you.

Your pal,
PAAUX4
 
Sorry if this has been asked before.

It looks like the focus is more on using protocols etc. rather than messing about with pins directly, almost like a computer. This is like when we lost the parallel port.


Thank you.

Your pal,
PAAUX4

Well only Paul can answer about the tradeoffs, but others can speculate on the answers (bear in mind, my answers are just idle speculation).

I would imagine the main issue is cost. Larger boards cost more to make. The more expensive a board is, the fewer customers will buy it if they don't need all of the features. I suspect for many people the $20 price is a sweet spot for ordering multiple Teensys vs. only one. In addition to just the size of the PCB, you have the additional complexity of routing all of those signals from the chip to the board. IIRC, the Teensy 3.0, 3.1, and 3.2 all used 4 layer boards, while the 3.5, 3.6, and 4.0 used 6 layer boards. Paul has said cramming all of the routing for those 6 layers took an enormous amount of work.

With some of the parts (notably the SD and USB pins), you can have signal reliability problems if the entire wire length is too long.

As somebody who puts boards into costumes and such (or at least plans out such things, even if I don't always finish things), I find the Teensy LC/3.2/4.0 to fit in better than the Teensy 3.5/3.6 in small spaces. For example when using a breadboard, I can use the cheap 170 pin breadboards to hold the Teensy 3.2/4.0, while for the 3.5/3.6, I have to use the longer 300 pin 1/2 size breadboards.

Also, somewhat unrelated:
  • What is FlexIO?
  • Are pins D- and D- under the USB port for the first or second USB port?
  • What benefit does the power pin have over a hardware switch? Is it just the ability to use a momentary switch?

For FlexIO, you will need to do a deep dive into the datasheet.

I believe the D-/D+ pins are for the second USB port (i.e. not the connected to the micro USB socket used for programming the Teensy). Paul has said there were issues with signal strength and running at the full USB 2.x speeds that prevented the D-/D+ pins of the first USB connector from being brought out.

I believe this is just bringing out a pin on the chip that provides this functionality. If you want a hardware switch, you can always remove the solder jumper that connects VIN to VUSB and put the switch there.

Paul has said that he is looking at a larger layout similar to the 3.6 that brings out more pins for a Teensy 4.x follow-on.
 
Sorry I am not sure about the Parallel port... ;)

Some of your questions can sort of be answered by looking at the product page: https://www.pjrc.com/store/teensy40.html

Form factor: The T4 is the same size as the T3.2, T-LC and some of the earlier boards with as close of compatibility of pins to these other boards as close as possible. As this is a complete different processor underneath the hood, it is not possible to get an exact match, but again close...

Why this form factor versus T3.5/6? Same reason why before some people buy the Teensy 2 and others buy the 2++. Mainly not every project needs something like 60 IO pins. And as such projects, they also don't want to take up that much room... Note: I am pretty sure there will be a T4.x which has more pins and my guess is it will have same/similar form factor to T3.6

FlexIO - There is a whole chapter (49) in the IMXRT manual on this. There is an ability to sort of develop your own subsystems using some basic building blocks. That is the T4 has 3 FlexIO controllers, each controller has a set of shift registers, timers and IO pins associated with them. With these basic blocks you can build up things like a Serial port or SPI setup or ... using these components.

I have a github project: https://github.com/KurtE/FlexIO_t4 That maybe at some point we will migrate some of it into the Core and other parts into other things. But with my experiments I have created Serial ports and SPI objects using them. I suspect at some point we might migrate some of this into SoftwareSerial, which currently on Teensy 3.x (and I think T4) only works on actual Hardware Serial ports... With this we can allow many other IO pins to work.

D+/D- - Yes as mentioned on the bottom of the card this is for USB Host (2nd USB hardware port).

Power pin - Uses the built in capability of the processor. See section 12.5 for ONOFF. It can be configured, maybe the RTC is still running even if no battery backup? Others might be able to give more complete answer.
 
Thanks for the response. I really don't think sale cost is a factor; PCB is cheap and adding more pins is just a matter of more traces. The initial cost of the time spent sorting it out is a good point though, especially with a new chip.

Shame about the USB pins. I really don't like Micro USB because they are weak like a piece of dust.
 
Initially, the original asthmatic Arduino UNO with its huge PCB was sold for 29€ ($32) in Europe. From which I conclude that neither price nor PCB surface barriers exist. I admit that the Teensy’s breadboard compatible format has its advantages but I can’t believe that adding holes for additional 10 pins on each side, making 48 pins instead of 28, which makes the board roughly 1” longer would add more than 1$ to the final price. And IMHO, there is no reason to sell a 600MHz MCU cheaper than a 180MHz MCU...

Instead, people who want to explore more of the already restricted pins (the iMX RT1062 has a total of 196!) have to fiddle with breakout boards.

But who am I to criticize Paul’s commercial policy? Thus, please don’t take me old fart not too serious!
 
IIRC the Teesy 4 is a 6-layer board. I've only done 6 layer PCBs at work, not as a hobbyist but our 6-layer PCB pricing was definitely proportional to PCB area, for prototypes and production volume also. What fraction that is of the total cost, of course, I don't know. I'm pretty confident that some applications do require a "Teensy" board and some here even want smaller than the current form factor to the extent of cutting off edges, or trying to roll their own, smaller boards.
 
The formula changes by vendor and this was several years ago, but I think it was proportional to PCB area for a fixed number of layers, and some more complicated formula on total layers. I'm pretty sure 2 layer was less than half the cost of same size 4 layer board for example, at least in small quantity. Higher layer count above the basic 2-sided board like Arduino Uno requires not merely more board surface area for mask and etch, but more total processing steps (and a significantly longer total fab time) to laminate the individual plys together, and higher precision processes to achieve the same final dimensional accuracy due to tolerance stackup on via alignment.

Beyond that, anything fancy like unusually tight tolerance, blind/buried vias etc. is a cost multiplier. For board assembly, BGA parts cost more to stuff, and parts on both sides costs more (one reason the Raspberry Pi Zero has no parts on the bottom side). If it means more time on a more expensive machine it means more money. Whether assembly costs are part of the board size/cost tradeoff independent of parts count might depend on the panelization; I'm not sure on that one.
 
Last edited:
Status
Not open for further replies.
Back
Top