Teensey 4.1 Schematic For Small Project

c172jeff

Well-known member
Hi,
I am looking for an eagle download of the Teensey 4.1 Schematic as well as versions of the cpu and supporting chips in SMD size. After buying an SMD Hot Air station to fix some broken PCB's, I wish to design a circuit of interest and make a PCB to fit select components. I see people doing this with the Arduino/Atmegs328, I want to try it with the Teensey. I was thinking of adding a GPS, SDCard, Ham radio transmitter, maybe an amp... I really don't know yet.

I was also wondering if the cpu, an MIMXRT1052 , comes in a slightly larger size with pads that can more easily be soldered by the hot air station around its perimeter.

I was also wondering if, after I find slightly larger versions of the chips with exposed pads, the firmware can be easily flashed and I can a custom app. of course I will need to add a usb port to the board.

I understand that the atmega 328 is fairly cheap and easy to put on a board, just wondering about trying this with the Teensey hardware.
Jeff
 
The Teensy 4.0 and 4.1 only come in BGA; although, they do have 2 variants with different pitch. These are both using the IMXRT1062. The IMXRT1052 was only used for the Teensy 4 beta boards.

The Teensy 3.6 (MK66FX1M0) does come in an LQFP footprint in addition to the BGA footprint used for the Teensy 3.6, but good luck finding stock beyond engineering samples.
 
Ok, Thank you. I'm wondering how tough/risky it is to solder a BGA chip to a PCB with a hot air gun? I do not have any way to XRAY the board.
 
There are a lot of examples of reballing BGA chips on YT, but this tutorial seems to be a good one for the initial
soldering of a BGA package on a new PCB:
https://www.youtube.com/watch?v=15RFI9wKHq8

Disclaimer: I've not done this myself.... Practice with cheaper chips would seem a sensible first step (for instance some
jelly-bean logic chips are available in BGA (74LVC245 for instance) which would be easy to verify functionally as well as cheap.)

BTW you need 4 layers at least to route out all but the tiniest BGA packages, 6+ layers for more complex ones.
 
Thank you. It does seem promising. That was a good video, especially the part that showed the testing of connections via the internal diodes of the soldered chip.
 
This is the schematic for the Teensey 4.1
https://www.pjrc.com/teensy/schematic.html

I am wondering if anyone has this in an electronic format that I can work with/modify. Yes, I know I can simply enter it into Eagle and go from there. I was hoping to cheat a little and just import something I know works and add/delete as necessary. I use Eagle and LTSpice. I am not expert in these tools, but have made a board and simulated a circuit.

Of course, at some point, I will just enter the linked schematic if no one has an electronic version. That seems like an exercise in editing, which is probably better spent on the other added stuff.


I have a Teensey 4.1 (2 of them). I like how it works, it will probably replace my arduino based amateur radio perfectly. I envision modifying the base Teensey 4.1 schematic to be larger and more accommodating. I would include include a VHF chip that functions as the radio, maybe an amp, a low pass filter, and connections for a display. I might even add a gps. The current arduino mega version is a mish mosh of wires and supporting bread boards.

It is my understanding that simply sourcing a pcb based on the linked schematic and soldering the components will yield a working Teensey that can be programmed like any other, at a much higher cost (hopefully no blown chips).
 
Thanks.. Well, I do like a good challenge. That said, I sometimes get over my head.

I do currently have a working radio using an arduino, dr818 vhf chip, low pass filter , etc ... it does work.
I have an older version of my project here....
https://www.youtube.com/watch?v=WrztebNYmOA&t=1108s

It has a touch screen display. The screen is slow to update. I feel that with a teensey 4.1, the screen should be extremely snappy.
If I could move everything onto one board, I would be very happy. Mostly for the speed increase and the clean up of wires.
Simple transfer of old proj to new, but replace the arduino with a teensey.


That said, I was poking around and saw some reference to using a 4 layer board/not breaking out all the IMXRT1062 chip pins in order to stay at a reasonable pcb cost from jlpcb ($10 or so).

I have to do some more research. At the moment, I would just see what I can add/remove and not break the bank in making a single pcb. It appears that there might be some complexity in routing the traces around the pcb in a way that minimizes cost. Hopefully, they can be routed arbitrarily with no regard to stray capacitance or unexpected interactions (only on the teensey side).

I think the hardest thing is to get the teensey made/working. I figure making a board 2x4 inches. I have to look, it does not have to be as small as the teensey 4.1.
 
FYI, the sparkfun micromod Teensy processor is an EAGLE design with 4 layers and has all the essential components to make a basic Teensy 4.x. I think that design will be my personal step 1 towards a 4-layer board where high-speed signals are needed. I've not taken that step yet ;)

So the following is just what I've read online. I'm doing this hoping that my mistakes will be caught and corrected as we can both learn from the experts here ;) I would actually worry about capacitance if I wanted some level of consistency and integrity of data on USB. The impedance of the trace (per inch) relies on the board material (substrate and copper) and prepreg of the board and the thickness of them in the stack. So even if you send that sparkfun design to a board house that does 4-layer boards, the outcome may not be working boards due to the differences of materials and thicknesses. You must calculate based on the board house's specs.
 
Thank you, I will look more about that SparkFun design. I will download their Eagle files and investigate. I don't really think I have high speed devices in my design. I communicate at 9600 baud with the dr818. The VHF signals are separate from the Teensey code. The USB would be programming the Teensey I guess at high speed. Thanks, I will dig in.
 
I don't really think I have high speed devices in my design.

The device USB port that you will need to program your board runs at USB 2.0 (480MHz) so that is the part that needs some careful routing. I was able to route them on 2-layer boards that embed Teensy 4.x boards on them either with short traces or with very few other things nearby so they work fine. But if you route the entire processor with other high speed pins such as your QSPI for FLASH (required) and possibly PSRAM (optional), that's when things will get harder on a 2 layer board if not impossible. Distance makes things easy because of the inverse square law of irradiation power (if not inverse cube for dipole-like differential wires).
 
I found this helpful link about stray capacitance

https://www.edn.com/pcb-signal-coupling-can-be-a-problem/

In practice, I am wondering if people actually calculate the capacitance where traces are close together along with their downstream effects. Same thing with inductances in traces that loop. It seems easier to carefully route and add a ground plane where appropriate (if ground planes are easy fixes) based on best practices/shortcuts/glaring pitfalls. I can't imaging designing a circuit, as a hobbyist, that takes advantage of stray capacitance and inductance to do useful things as a function of careful/clever board layout generating these effects.

I would prefer stray capacitance and inductance didn't exist at this point.


Anyways, I looked at that sparkfun schematic for the micromod. It seems the micromod connector breaks out everything that is needed to connect to usb and access aa bunch of pins on the chip.

Just to confirm, if I built (or bought) a microrod device and simply connected up the usb connector and provided power, would I be able to program this just like a normal Teensey 4.1? I am only looking a the pins on the Micromod connector...primarily the usb ones.

I like using the IDE normally used for Arduino (with its suite of libraries), but tweaked for Teensey. I think this is the only IDE anyway.
Of course, my board would have a preprogrammed IC_MKL02Z32_T4_QFN16 Bootloader Chip on it (other stuff too).
 
Just to confirm, if I built (or bought) a microrod device and simply connected up the usb connector and provided power, would I be able to program this just like a normal Teensey 4.1? .... Of course, my board would have a preprogrammed IC_MKL02Z32_T4_QFN16 Bootloader Chip on it (other stuff too).

Yes. The purpose of the IC_MKL02Z32_T4_QFN16 chip is to give you a way to build your own DIY circuit board which is fully compatible with Teensy 4.0 or 4.1. Like any custom PCB design, there are many technical details you must get right for it to work properly. Details are on that bootloader chip page. Pay careful attention to the power supply startup sequence, especially delivering power to VDD_SNVS_IN first. The chip will not start up if you apply power to the other pins before VDD_SNVS_IN and VDD_SNVS_CAP stabilizes.
 
I was hoping @Paul would chime in at some point :D Thanks! So if I used the micromod, then both IC_MKL02Z and the correct power sequence are on the micromod, right? I was looking at the schematic. Sparkfun has 3.3V regulator on the carrier board but the PMIC_ON_REQ pin controls a P-MOSFET to feed power to the rest of the processor while both MKL02Z and power sequencing are on the micromod:

Screenshot 2022-04-20 105252.png

So if I make a carrier board, the design workload is really a lot less since the micromod has all the routing done, correct?

About impedance matching, I have no practical experience but it's important when you have high speed signals routed next to one another. I have a colleague who does analysis for high-speed routers. He says signal integrity engineers are well-paid and highly-sought out. He does simulations on things like reflection of EM waves due to vias etc. So I imagine if your only high-speed signals are the USB D+ and D- (already in differential pairs, please route as diff pair in EAGLE) while other signals are routed away from them and stay orthogonal from them, then your chances of issues are lowered.
 
I think I can make a board the way I like. I copied the sparkfun schematic and am rekeying it into a new one (with additions). It was a learning experience, no obstacle though.

I can order a board and a stencil for not too much money. I was thinking I can buy the MIMXRT1064DVL6B chip and the other supporting chips it needs (mentioned above) as well as some kapton tape.

Bada bing, bada boom --- I can lay some kapton tape on a board with a ripple in it and edge the bga chip (MIMXRT1064DVL6B) right up against it. I can fire up the hot air station and solder it on and get this thing to work. Same thing for the bootloader and memory... What can go wrong?

Anyways.. I believe the MIMXRT1064DVL6B chip will be impossible to find.

I was only planning on buying 1 chip and try all this.

I figure since most places seem to like to sell large lots of chips , how hard can finding 1 chip be with a little price premium. $20 instead of $16 (whatever)..

Where can this chip be sourced, only one. I hate to make a board only to wait until 2023 for a chip. This is ridiculous, are these chips sitting in ships off the California coast or is there a problem finding copper and silicon.

Where would one find just one of these chips? Maybe 2 if things don't go as planned
 
Is that the MCU for a T_4.0 - not the larger T_4.1?

Would it make sense to rework the PCB to hold a SparkFun MicroMod Teensy Processor?

That $21.50 item is available - would need the M.2 connector - but include a ready to run 1062 MCU with more pins (+/- ) than a T_4.0 with larger 16MB Flash and Bootloader installed and tested, without having to use up precious kapton tape :)


@Paul has historically ended up with a few extra MCU's after doing a Teensy run and offered them at 'normal' retail pricing with purchase of the accompanying bootloader to support DIY building.
 
The 1064 chip is not supported. Only 1062 works, and only with specific Winbond flash chips.

We do have some extra 1062 chips here, but only a couple dozen of them. They're the larger 12mm size, as used on Teensy 4.1. If you can't find the chip and need just a few to get a prototype made, email me directly to check if we still have them and how we can get you these parts.
 
I was hoping @Paul would chime in at some point

Ok, I'm chiming in, though I'm not sure which parts of your message were meant to be a question for me. Well, apart from this:


So if I make a carrier board, the design workload is really a lot less since the micromod has all the routing done, correct?

Yes, that's pretty much the idea behind MicroMod. Sure, Sparkfun talks about swapping modules between carrier boards, and you can do that to some degree, but the real value is you can make your custom PCB with just the M.2 socket and peripheral circuitry. It saves a lot of BGA layout work and the hassle of BGA soldering. But the really useful part is managing cash flow if you're trying to launch a low volume commercial product. Your total BOM cost will end up slightly higher than if you had put everything on your own PCB, so it certainly is a trade-off, but you also don't need to take the financial investment and risk in all the parts up-front. You can make PCBs ahead of time with just the M.2 socket and then buy the MicroMod at the very end when you know you have sales for your product.


About impedance matching, I have no practical experience but it's important when you have high speed signals routed next to one another. I have a colleague who does analysis for high-speed routers. He says signal integrity engineers are well-paid and highly-sought out. He does simulations on things like reflection of EM waves due to vias etc. So I imagine if your only high-speed signals are the USB D+ and D- (already in differential pairs, please route as diff pair in EAGLE) while other signals are routed away from them and stay orthogonal from them, then your chances of issues are lowered.

Indeed this can be deep rabbit hole.

The main points are to route the signals with a ground plane or well decoupled to ground power plane underneath, to route them together, and keep the length equal.
 
Thanks for the reply. My bad. I did mean to say the 1062 chip. It was a copy/paste from a prior google search I made.

I have a MIMXRT1062DVJ6B in my schematic (based on the MicroMod from SparkFun). It is 12 mm x 12 mm x 1.37 mm. So, I guess that is the right one for a Teensey 4.1 and I will be emailing Paul soon.

I did not want to use the SpackFun device as I wanted the most minimalist solution that functioned as a Teensey 4.1 and let me use the Arduino libraries. Minimalist in the sense of no M.2 connector/separate SparkFun order.

I do realize that $21.50 + an M.2 connector is a bargain, bargain of the century when considering the risk reduction of making a board and soldering a 12mm chip. Not sure how much a foot of kapton tape is?

This is also an exercise in BGA soldering, PCB replication/design, sourcing parts, and circuit enhancements. Well, no circuit enhancements on the 1062 processor implementation side at least.

Ok, let me carry on with my schematic. I probably should order the chips before I get too far down the road.
I'm hoping to get a little more confident with this low level design process.
 
Thanks @Paul. I think you answered my questions. I may be interested in a few 1062 and the smaller chips with bootloader. It's for one of the projects I've been bugging you guys with weird USB questions for. Should I PM you for pricing?
 
Bada bing, bada boom ---

I wish I had your optimism or skill level. I've been designing boards for over a decade but on 2-layer boards so far. As I mentioned, I only know some things in theory. Hope you can get some prototype boards working. I'll have this on my to-do list. I use a toaster oven with heating elements on top and bottom so I can get better results on my batches.
 
Back
Top