Teensy 3.2 End Of Life

So sorry to see T3.2 go, it has been, and still is, great.
But i fully understand it has to go.
 
Not such a big deal, @Paul Thanks for all the efforts on the chip, I really learned a huge amount.
I too have been hit hard by NXP not getting all sorts of MCU's that I use.
Not entirely sure what their problem is, if it is some sort of effort to kill the entire lower end line.
Either way it is a shame, because in a lot of circumstances, the lower end MCU is all you need, e.g. battery life, price, etc.
Even so, I have my own bootloader not halfkay, but, that doesn't matter.
 
Not such a big deal, @Paul Thanks for all the efforts on the chip, I really learned a huge amount.
I too have been hit hard by NXP not getting all sorts of MCU's that I use.
Not entirely sure what their problem is, if it is some sort of effort to kill the entire lower end line.
Either way it is a shame, because in a lot of circumstances, the lower end MCU is all you need, e.g. battery life, price, etc.
Even so, I have my own bootloader not halfkay, but, that doesn't matter.
I'm not going to speculate, but I'm sure profit has everything to do with its obsolescence, and I think the real loss was the LC as the 4.0 is the same price and way more powerful. On a positive note, how's the 4.1 replacement project for OC going, and will we ever see anything like the LC ever again?
 
I suspect during the chip shortage they prioritized the chips that were required by major industries like car makers and white-goods, no big mystery really.
 
I suspect during the chip shortage they prioritized the chips that were required by major industries like car makers and white-goods, no big mystery really.
I've been actively tracking the whole thing for a couple of years now.
It's pretty much exactly like the capacitor famine, which, we are actually at the tail end of.

You can blame a couple of things for the whole mess, which are very similar to the capacitor famine.
  1. Of course, the big guys like nvidia and amd got dibs on anything, which eat tons of silicon wafers, and those aren't exactly quickly produced.
  2. Then you have the fact that the node that they are on is getting more and more obsolete, so fabs tend to reduce the machines for that, and in place of those, install the newer process node machines.
  3. On top of all of that, any huge volume customers get preference, since it is a guaranteed sale.
There are a lot of other factors, but those are the main three, pretty much the same deal as the capacitor market famine.
 
Mouser even has 15K+ pieces on stock.
I would buy Teensy 3.2 if they are available - I have actually never needed the processing power of the T4 [and the related power consumption].

Paul
 
There are various posts where people would buy lots of T3.2 instead of moving their designs to T4.0.
Financially, there is a small risk that when T3.2 would be available again, everybody already moved to T4.0.

A better idea would be to design a small pcb, same size than T4.0, to minimize the pinout difference between T3.2 and T4.0. (mainly in the bottom side).
It will be inserted between the T4.0 and the T3.2 PCB. It will redirect the T4.0 pins to T3.2 pinout to allow minimal software changes.

Angelo
 
Teensy 3.2 is not coming back.

Angelo's comment about the financial risk is spot-on, exact for the "small" part. A word like "certain" or "inevitable" would be better.

Just to explain clearly, Teensy 3.2 was already a mature product when the chip shortages hit in mid-2021 and Teensy 3.2 stock ran out by January 2022. Had NXP delivered our orders on time in 2022 (most were already past the 52 week lead time by January 2022, the rest were late by June 2022), or even had they delivered more substantial quantity many months late, we probably could have kept Teensy 3.2 going for several more years. It is the normal life cycle of a tech product under normal circumstance of continuously supply, where "mature" gradually turns into "long tail" as customers with established applications continue buying for many years.

But that's not how things went. NXP delivered pathetic quantities only sporadically from late-2022 throughout all of 2023. Robin and I put a *lot* of work into allocation. Eventually that work turned into direct assistance to help customers migrate to Teensy 4.0, which freed up more of the extremely limited chip supply for those who hadn't migrated, and for occasional retail sale with low quantity limits (we really do care about helping makers). In hindsight, we probably should have put that work into Teensy 4.0 development and just discontinued Teensy 3.2 about 1 year earlier than we did. But that's with the benefit of hindsight. Throughout 2022 and 2023 people at NXP kept promising improvements which didn't come until 2024, which was far too late.

By mid-2023, most customers who had been using Teensy 3.2 in significant volume have moved on to Teensy 4.0 or their own custom PCB using the T3 bootloader chip or (sadly) other non-Teensy products. Several customers buying less regularly usually at lower volume were left, and some pretty widely used open source projects like Ornament & Crime were still tightly tied to the Teensy 3.2 hardware, but the long-term commercial damage to Teensy 3.2 was irreversible by early-2023.

I know this is hard to hear. I poured many years of work in Teensy 3.2 (and 3.1 & 3.0 which is evolved from), so believe me I know the pain! It was the first Arduino ecosystem microcontroller to support large non-blocking addressable LED projects in early 2013, and then DMA-based audio in 2015. I truly am sad to see Teensy 3.x go. It was as still is (would be) good on a technical level.

But PJRC is a small company which can't afford to make large financial mistakes, especially at a time when we're still recovering from having gone all-in to keep Teensy 4.0 & 4.1 (mostly) available during the shortages. The sad reality is Teensy 3.2 could have had many more years of "long tail" product lifecycle, had its life not been cut short at the "mature" phase by 2 years of shortages. The simple reality is nearly all customers who probably would have continued buying (in volume needed to keep Teensy 3.2 a viable product) for years to come were forced to move on during those 18 months. I know many people still want small quantities, and even some would probably use it in modest volume, but it's nowhere near enough bring back an already-mature product that has lost virtually all of its sustaining customer base.

Teensy 3.2 is absolutely not coming back.
 
Last edited:
I didn't checked if they are already available anywhere, but PJRC could maybe provide production files (gerber, silk, placement, BOM,...).
Maybe by removing reference to PJRC for "PJRC would not be responsible for any damage....."
JLCpcb or PCBmain could produce some batches at low cost. But obviously not tested.

Angelo
 
Back
Top