How To Kill Off Future Sales

Infineon appears to be methodically killing Cypress, in my opinion.

As you know, I am a fan of the PSOC5 with the PSOC Creator, using VSCode to code/debug on the Mac after design on PSOC Creator, with KitProg or KitProg2 (see previous posts. That introduces some clunky, but with much more speed.). The PSOC3, PSOC4, and PSOC6 all have their places, but no one has created the mixed signal and FPGA combination that is in the PSOC5.

Unfortunately, PSOC5 appears headed for the graveyard. Why do I say this? Certain iterations of the PSOC5 will be available from distributors in 93 weeks if ordered now. 52 weeks are in a year. If you add other timings into the system, 93 weeks is 2 years out.

Unfortunately, in today’s design world, 2 years means the parts will not be put into future designs. So, future sales take a very big hit. Very big. Huge. Crippling.

Other manufacturers are in much the same boat. Renesas NEW MCU’s will be available in 2025. But Infineon is worse. Infineon’s PSOC current MCU’s are 2 years out. Atmel CPU’s are in stock, including some that have Comparators, A/D and DAC, and similar peripherals to PSOC. But no UDB/LUT.

What Happened To Cause This?

Simply put, chasing today’s profit, ignoring tomorrow’s disaster. Combine that with governmental agencies that do not want manufacturing plants in the U.S.A. (This governmental hostility started mainly due to past malfeasance by those manufacturers regarding the environment.) This combination prevents the building of facilities in the U.S.A., which would have prevented most of the current set of problems.

For instance, need cheaper gasoline? Easy, refine more! Nope. Go green is the woke agenda, so refineries are being shut down by the government and not allowed to open. Easy fix, build new ones and shut them later. No deal. It takes about 10 years to get the required waivers and permits, and even that does not guarantee approval to open. To be fair, refineries STINK! I would not live next to one.

Bottom line: Higher fuel prices for the foreseeable future. It is so pervasive that even city governments are in on this new American Exceptionalism.

San Francisco is the poster child of city governmental agencies running without oversight. It takes 10 agency approvals and $30,000+ dollars to open a restaurant in that city. According to this article (, one man spent $200,000 and still could not open an ice cream shop. Although they are promising to “look at a new law” to help fix that. Yeah. Right. It is the government we are discussing. Won’t happen.

Back to the PSOC5 Problem

Fabless companies have shifted the majority of their production to overseas companies that are controlled by governments who want to be the only game in town. Rather than supporting manufacturing in the U.S.A., these companies were/are having parts manufactured in countries controlled by CCP, FSB, or some other unfriendly governmental iteration. Fast forward to today, and chips that provide $30,000 of revenue per 1,000 units are put on the back of the line at plants no one has control over. Mouser has at least $250,000 worth of PSOC5 parts ordered. Some of those won’t be delivered for 93 weeks. Mouser is small potatoes to Infineon.

What Can We Do As Normal People?

In brief, nothing. It is what it is, and many companies are dying without recovery. The sad thing is they don’t know they are dying, they think things will be fine in the future. We, the normal people, have to deal with the shrapnel from their self immolation as sparks are emitted during the bonfire collapse.

I am the founder of a small company. I have wrangled a way so the company has around 6 to 8 months to make serious decisions. I have to face the fact that Infineon does not want our business, or the business of the Arduino class subset of customers. (Although Arduino in total is a *lot* of processors being shipped annually. Think about it, Infineon. Why not grow that market and become the next Atmel processor? Deaf ears.) I have to find alternatives.

The PSOC6 is not an alternative. Too few UDB’s kills it in the versatility department. It also has an uncertain ability to work above 105C (210F). Incompatible voltages; although that can be dealt with. In addition, it is only available in BGA style packages that make it almost impossible for a custom manufacturing facility or hobbyiests to create small production runs.

In some cases the PSOC3 could be used. However, the 8051 is extremely limited. During tests, I was unable to fit some of my more complex code into that processor system. Not enough room. Ugh. Building FreeRTOS into the 8051 is a no-go.

The Ideal Solution (Assuming PSOC5 Destruction By Infineon Continues)

Three pieces to this puzzle: 1) Arm processor SOC with 256k flash and 64k ram, 2) Analog blocks with DACs, A/D’s, OpAmps, Comparators, and 3) small FPGA with around 256 LUT’s (or 24 UDB’s).

This all has to fit on a 4) 68-QFN or 5) TQFP 100 package width, within about 2 inches in length. And #6) operates at 300F.

Atmel SOC’s meets #1, #2, and #4/#5, sometimes #6. And in stock….Hmm. But not #3.

Renesas looks like a step in the right direction, but that processor is too far out. However, when I contacted them, they said they would work with us to bring that in. Hmm…

GreenPak for the Analog might be a possibility, but they are having shortages and their development board setups do not make sense to the beginner. Odd shortages of things make it impossible to put a development environment together, and there is no quick and simple guide that says: 1) buy these 3 things, 2) download this, 3) do this and test your project.

CPLD devices from Atmel are in stock with about the same number of logic blocks as PSOC5. Hmm. GoWin has small FPGA’s, and they are in stock. TinyFPGA might be a possibility, but it is out of stock.

Nothing matches PSOC5 for sheer versatility. But with 2 to 3 parts, I can come close. The trade-off is the development environments for Atmel/GoWin/Etc. are clunky, similar to Modus Toolbox, the epitome of clunk.

Why Am I Considering Out Of Stock Items?

I am considering these things because their stated delivery time is within this year. Not 93 weeks. Just 10 or 12 weeks. Any one at Infineon notice this difference? Does anyone really care?



Add a Comment

Your email address will not be published.