Coping with global chip shortage - any guess what microcontrollers will have best future availability?

Go To Last Post
24 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

What does your crystal ball say about which microcontroller manufacturers, product lines, and/or package types are likely to have the best availability over the next 1-2 years? Would you expect everything to be impacted by the global chip shortage more-or-less equally, or will some kinds of microcontrollers maintain better availability than others?

 

I'm redesigning a product that I hope will sell around 2000 units per year, and if there were no chip shortages I'd probably choose something like a SAMD21 or SAMD51, or a comparable STM32 device. But I'm also looking at xMEGA and AVR-Dx stuff. Availability of AVR-Dx right now seems better than the others, but I'm not sure if that will last. Right now it's hard to find ANY microcontrollers that are available in > 1000 unit quantity, with confidence that they'll still be available in 6 months, a year, two years. What's your best prediction for the future?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

bigmessowires wrote:
What's your best prediction for the future?

Give up electronics until the situation resolves itself. The only parts you can guarantee availability for; are the ones you hold in your stores.

 

We're also seeing some pretty outrageous price gouging going on for Grey Market chips. E.g. Laird doubled the price to make some Bluetooth Modules using grey market Nordic transceivers & ten times the price for grey market PIC32.

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

Use something generic like a mega48

 

Then you can also use an '88 '168 or '328, thereby giving you a lot more options.

 

You can layout you board to take both a TQFP and a QFN  (there are pads symbols where QFN fits inside TQFP--so you can install either. 

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

AVRxxxDA48 and ATmega4809 also share a footprint.  Probably DB48 too, if you're a bit careful with that extra supply.

 

I sold some stocks back in October, based on the theory that some chip companies would be especially hard hit because they relied on 3rd party fabs, rather than having their own, or would otherwise be more strongly affected.

NVDA, XLNX, NXPI, STM.  So far, they've done quite well since then, despite my theories.

 

There are so many issues in play: fabs, raw silicon wafers, shipping issues, other supply-chain issues, hoarding...  It's near impossible to say how things are going to play out.  :-(
Given the "very optimistic" stock market, I predict that some weaker companies will end up being bought.  (not that that hasn't been happening anyway.)

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

bigmessowires wrote:

What does your crystal ball say ...I'm redesigning a product that I hope will sell around 2000 units per year,...

 

The indications are that things will start to improve Q3 next year, with things back to near-normal by the start of 2023.

 

So, if you want to make 2,000 units in that period you need to find a chip, or as said above a family of chips, which you can buy 2,000 parts now and put them on the shelf.

#1 Hardware Problem? https://www.avrfreaks.net/forum/...

#2 Hardware Problem? Read AVR042.

#3 All grounds are not created equal

#4 Have you proved your chip is running at xxMHz?

#5 "If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand."

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

N.Winterbottom wrote:

Give up electronics until the situation resolves itself.

 

Yeah, that's the worst case scenario. Unfortunately that may actually happen...

 

Thanks for the suggestions of designing for multiple part and package types. Designing for AVRxxxDA48 and AVRxxxDB48 in TQFP and QFN would definitely expand the options. How would you create a combo TQFP/QFN footprint? I'm thinking you'd start with a TQFP footprint and just extend the pads further inward toward the center of the chip? Do you have an example combo footprint?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0


Do a web search ..they will pop up easy 

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0


bigmessowires wrote:
What does your crystal ball say about which microcontroller manufacturers, product lines, and/or package types are likely to have the best availability over the next 1-2 years?
Don't have a crystal ball and wouldn't want one (trip, drop the damn thing on a foot, see a podiatrist)

Am greatly more inclined to listen to the semiconductor grey beards (many cycles of boom-bust)

bigmessowires wrote:
Would you expect everything to be impacted by the global chip shortage more-or-less equally, or will some kinds of microcontrollers maintain better availability than others?
No and yes

bigmessowires wrote:
But I'm also looking at xMEGA
XMEGA has been long lead time for a year; AVRxm "may" have lead time reduction sometime in '22 with some stock relief roughly mid-'22.

Guesses on why AVRxm is long lead :

  • not wafer fab'd by Microchip Technology
  • die designs are approximately ten years of age (wafer fabs with a compatible process is a "short" list)
  • unlikely XMEGA will receive die respins

Most XMEGA have features that are unique to AVR; if a design requires XMEGA then buy enough trays and reels given lead time versus manufacturing plan's dates.

bigmessowires wrote:
... and AVR-Dx stuff.
Some AVR Dx have a reasonable lead time and might be considered as an alternative to XMEGA D.

bigmessowires wrote:
Availability of AVR-Dx right now seems better than the others, but I'm not sure if that will last.
AVR Dx lead time likely will hold at reasonable though very package dependent for a while.

bigmessowires wrote:
What's your best prediction for the future?
"Wise" guy : the unknown will occur (what's the follow-on to white swans and black swans?)

Answer : semiconductors are long lead time

Recommendations :

  • weekly evaluation of BOM's lead times
  • weekly calculation of risks
  • weekly copy of risks to managers, directors, and executives (ones who do risk evaluation, risk reduction, and risk mitigation)
  • purchasing leads manufacturing
  • stand ready to respin a PCBA and firmware
  • revert EOL PCBA
  • patch PCBA and firmware
  • empathy and support for the ones in logistics

 


ATMEL parts are running out of stock due to wafer shortage ? | AVR Freaks

 

XMEGA Lead Time, Dec'20 | AVR Freaks

 

Lead Time | World's Largest Inventory of Microchip Products

 

API Registration | Octopart (Pro column, Lead time row)

Search API (Mouser Electronics)

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

N.Winterbottom wrote:
Give up electronics until the situation resolves itself.
The ones at Tesla Motors respun PCBA though they're unique among vehtronics manufacturers.

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

westfw wrote:
ATmega4809
mega4809 is an AVRxt that's wafer fab'd at Microchip Technology.

 

Product Change Notifications | Microchip (enter ATmega4809)

How to search for Microchip PCNs

 

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

a long time ago I complained that all this outsourcing would lead to a huge snowball avalanche of problems and now that we've had the perfect storm it has.... I also do not think that this problem is going to be resolved very quickly so get used to how terrible the supply chain is....  I am seeing supply chain issues with all kinds of material not just microchips...  example: There's a stupid connector pin I'm waiting on right now, the entire world is sold out...  No estimated delivery date from mfg...  Manufacturing anything reliably is a million times harder.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

As earlier commenters have noticed: your only guaranteed option is to find a usable part for which are currently in stock for 2000 piece immediate delivery. It's not too bad a hit if you're budgeting for sales in one year.

 

Didn't a famous US sportsman once point out that prediction is really difficult - particularly about the future?

 

Neil

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 My approach to the shortage of parts problem is to design every project (and subsequent product) on four multiple hardware platforms. And I use the Arduino IDE because I can write to its Hardware Abstraction Layer (that handles stuff like UART, I2C, SPI, TFT, USB,...).
 Then if there is an order for 100 units and there is not 100 pieces of any one module available, I can still order 25 each of the hardware modules,  then I can ship 25 units on each platform.                                                   
 The hardware that I use is: the AVR mega328P UNO/Nano board, the ARM STM32F103 Blue Pill module, the WeMos ESP8266 module, and the ESP32 WRoom module.  Each sells for about $4-5 per unit.  Each has an USB/UART chip that lets it be flashed with the USB on PC.

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

bigmessowires wrote:
What does your crystal ball say ... 

If anyone had such a reliable crystal ball, they would be rich!

Top Tips:

  1. How to properly post source code - see: https://www.avrfreaks.net/comment... - also how to properly include images/pictures
  2. "Garbage" characters on a serial terminal are (almost?) invariably due to wrong baud rate - see: https://learn.sparkfun.com/tutorials/serial-communication
  3. Wrong baud rate is usually due to not running at the speed you thought; check by blinking a LED to see if you get the speed you expected
  4. Difference between a crystal, and a crystal oscillatorhttps://www.avrfreaks.net/comment...
  5. When your question is resolved, mark the solution: https://www.avrfreaks.net/comment...
  6. Beginner's "Getting Started" tips: https://www.avrfreaks.net/comment...
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

Thanks. I couldn't recall the name.

 

Neil

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0


Interesting screenshot I've just taken...

 

 

 

 

 

 

...out of 11,171 8-bit micros that element14 stock, only 885 are on stock.

#1 Hardware Problem? https://www.avrfreaks.net/forum/...

#2 Hardware Problem? Read AVR042.

#3 All grounds are not created equal

#4 Have you proved your chip is running at xxMHz?

#5 "If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand."

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

At this point, I don't think that anyone can predict which specific microcontrollers will have best future availability. Different manufacturers have different preferred processes. But, then, you have the odds of a building fire or tsunami or earthquake (Taiwan is earthquake prone, after all) that nobody will predict. Even if a product is made in several widely scattered fabs, the loss of one will push up the demand and slow things at the others. 

 

So, my thinking is that the best you can do is understand that there IS risk.

 

And, that risk is NOT just with micros. Multilayers ceramic caps have been volatile. LEDs, resistors, switches, connectors, inductors, what ever. Any of those can kill production. They all have risk.

 

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

Last Edited: Thu. Dec 23, 2021 - 08:48 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

barnacle wrote:
It's not too bad a hit if you're budgeting for sales in one year.
Can enter the backlog queue via the Microchip Preferred Supply Program; consider a discussion at the preferred distributor about which MCU they're willing to order a lot of.

PSP Letter

Microchip Technology

February 4, 2021

[in fifth paragraph]

With that in mind we would like to introduce the Microchip Preferred Supply Program (PSP) and offer you the option to receive prioritized capacity in the second half of 2021 and the first half of 2022.

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Simonetta wrote:

 My approach to the shortage of parts problem is to design every project (and subsequent product) on four multiple hardware platforms. And I use the Arduino IDE because I can write to its Hardware Abstraction Layer (that handles stuff like UART, I2C, SPI, TFT, USB,...). ... The hardware that I use is: the AVR mega328P UNO/Nano board, the ARM STM32F103 Blue Pill module, the WeMos ESP8266 module, and the ESP32 WRoom module.

 

That's a very interesting strategy, because those are some very different hardware platforms. Mega328P is a simple 20MHz 8-bitter with 2KB RAM, while ESP32 WROOM is a 240MHz 32-bit MCU with WiFi and Bluetooth and 512KB RAM! My first reaction is that's crazy, but if you can get the hardware for the right price, and the Arduino libs support it, then why not?

 

I have always believed that the best way to design any MCU-based product is to design the bare MCU chip directly into your PCB, and this will result in the lowest cost, smallest size, most flexibility, and most professional-looking result. I would have rolled my eyes at the idea of integrating a whole Arduino board or RPi or similar integrated module into a professional product. But in these crazy times, maybe I need to revisit that thinking? For some of these modules, the retail cost of the module really isn't much different from the cost of the bare MCU plus voltage regulator, crystal, decoupling caps, etc. For example when a whole ESP8266 NodeMCU module can be purchased for $4.50, and a bare Mega328P chip is $5-$6, maybe it's time to change strategy and build my products around these modules instead of bare MCU chips. Does anyone else feel this way?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yes, some popular boards are in shortage or priced doubled!

 

I am becoming very poor in year 2022!  crying

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Chips are available if you look and are prepared to be a bit flexible. Just this afternoon I placed scheduled orders, for delivery next year, for my anticipated AVR requirements. 

#1 Hardware Problem? https://www.avrfreaks.net/forum/...

#2 Hardware Problem? Read AVR042.

#3 All grounds are not created equal

#4 Have you proved your chip is running at xxMHz?

#5 "If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand."

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

N.Winterbottom wrote:
Give up electronics until the situation resolves itself. The only parts you can guarantee availability for; are the ones you hold in your stores.
This related with my marketplace threat.

www.tokopedia.com/madagang .Buy and Donated cheap electronics and manuscripts.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

bigmessowires wrote:

Simonetta wrote:

 My approach to the shortage of parts problem is to design every project (and subsequent product) on four multiple hardware platforms. And I use the Arduino IDE because I can write to its Hardware Abstraction Layer (that handles stuff like UART, I2C, SPI, TFT, USB,...). ... The hardware that I use is: the AVR mega328P UNO/Nano board, the ARM STM32F103 Blue Pill module, the WeMos ESP8266 module, and the ESP32 WRoom module.

 

That's a very interesting strategy, because those are some very different hardware platforms. Mega328P is a simple 20MHz 8-bitter with 2KB RAM, while ESP32 WROOM is a 240MHz 32-bit MCU with WiFi and Bluetooth and 512KB RAM! My first reaction is that's crazy, but if you can get the hardware for the right price, and the Arduino libs support it, then why not?

 

I have always believed that the best way to design any MCU-based product is to design the bare MCU chip directly into your PCB, and this will result in the lowest cost, smallest size, most flexibility, and most professional-looking result. I would have rolled my eyes at the idea of integrating a whole Arduino board or RPi or similar integrated module into a professional product. But in these crazy times, maybe I need to revisit that thinking? For some of these modules, the retail cost of the module really isn't much different from the cost of the bare MCU plus voltage regulator, crystal, decoupling caps, etc. For example when a whole ESP8266 NodeMCU module can be purchased for $4.50, and a bare Mega328P chip is $5-$6, maybe it's time to change strategy and build my products around these modules instead of bare MCU chips. Does anyone else feel this way?

2000 pieces a Year ? Why didn’t hack the ready made goods ? Their products almost a million or more in products on market.

www.tokopedia.com/madagang .Buy and Donated cheap electronics and manuscripts.