Whats the actual difference? 85 vs 105 degC parts?

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

Some questions around the max temp that an AVR can run at:

 

1) Atmega's come in three temp ratings,  industrial (up to 85degC) extended industrial (up to 105 degC) and Automotive (up to 125 degC).  But what actually is the difference in terms of what's in the IC?  Is there ANY difference?  or are the EOL qualification tests just different?

 

2) When you 'overheat' an IC what actually happens? Assuming you stay under the temp the solder melts at, so it doesn't fall off the pcb ;-)  what happens to the physics of the thing that results in it's temperature rating?

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

Many of the parameters are temperature dependant. Depending on the batch, some may stay in spec better than the others over the temperature range. So, yes, the chips are all the 'same' but the higher spec ones are guaranteed. Temperature has an effect on electron mobility, so this affects the performance of the device. Timing changes as well as threshold voltages. At some point the silicon stops working as a useful semiconductor.

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

Don't rely upon my reply, but I think it all has to do with levels of internal leakage currents.

Ross McKenzie ValuSoft Melbourne Australia

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

Are we talking about factors that lead to a slow failure, ie a reduction of useful life, or simply factor that result in errors when the temperature exceeds a specific value?

 

Ie the EEPROM can't write bytes with a low enough error probability above 100degC for example?

 

 

Is there any official info from Atmel as to what the differences are and what test procedures are used to bin/sample/rate the production devices?

 

 

Has anyone put an avr in an oven an tried to test what happens when it gets toasted?  Maybe write a little program to exercise the devices hardware (uart,eeprom, timers etc) and stick an old Arduino or something in my toaster oven??  Any suggestions?

 

 

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

max_torque_2008 wrote:
Is there any official info from Atmel as to what the differences are and what test procedures are used to bin/sample/rate the production devices?

I've given links before to the Atmel "quarterly reliability reports" that can be found with a determined Google search.  Related docs describe the test methodology.

 

E.g. http://www.atmel.com/Images/relm...

From http://www.atmel.com/images/atme...

14
Any material returned for credit or replacement must be processed through the appropriate Atmel sales office or sales 
representative using our RMA/FA system. Standard request forms are used to ensure completeness of information 
and quick response. For technical returns, authorization will not be granted unless the initial failure analysis verification has 
been completed and validated. Both forms are available on our web site:
(
http://www2.atmel.com/about/quality/default.aspx
, select“RMA & FA Information”) along with shipping instructions.
Change Notification and Product Obsolescence
Atmel relies heavily on continuous improvements to enhance our products. In our fast moving technology, Atmel must also 
stay cost competitive. These factors result in changes to our products after they are released for mass production.
Atmel has a comprehensive system to manage such changes including inventory management and customer notification 
when required. Customers are notified through a change notice prior to the implementation of a major change.
In some cases, change notices will result in product obsolescence. In these instances, specific details on the last time 
buy (LTB) date, last time ship (LTS), and replacement part information will be widely distributed. The Atmel Quality Web 
Site includes a product obsolescence query screen that allows customers to easily search our database: 
(
http://www2.atmel.com/about/quality/default.aspx
, select “Obsolescence Information”).
Atmel Quality and Reliability Web Site
Atmel has made a focused effort to utilize our web site as the primary vehicle for quality and reliability data dissemination.  
It is our intent to continuously improve and enhance the site so that customers can have immediate access to all relevant 
quality and reliability information. Items that are available at: 
http://www2.atmel.com/about/quality/default.aspx
 include:
• 
Quality Inquiry
• 
Quality Handbook
• 
Green Packaging
• 
Environmental Information
• 
Quality and Environmental Certificates
• 
RMA and FA Information
• 
Reliability Information
• 
Obsolescence Information
• 
FAQ Response Summa

http://www.atmel.com/about/quali...

 

 

But if you want to dig into it, I'd look at the automotive documents and the certifications and such.  Should be a lot of info there.

 

 

 

 

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

Last Edited: Tue. Jun 13, 2017 - 05:58 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

A lot of the electronics websites appear to have (interminable) discussion about all this stuff and the effects of heat etc.

 

As usual with the internet the problem is not knowing the provenance of the information given. Some of the answers may be from total experts who are dyed in the wool and know the stuff inside out. While some maybe from the completely ill informed simply guessing at possible reasons without any basis in fact. Having said that you can usually spot the more "professional" looking answers I guess.

 

This kind of thing...

 

https://electronics.stackexchang...

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

ok, so that leaves the question:

 

are the 85degC parts and higher rated parts (105 & 125 degC) actually different silicon?

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

I wouldn't think so.  Perhaps the package material might be different. Just a guess.

 

OK, now you need to tell us why it is important. ;)  Sending a rocket to Uranus?

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

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

The old databooks of the 70s and 80s usually had a section on reliability, the tests they performed and what the test was meant to prove.
In terms of the silicon itself, measuring key parameters over voltage and temperature is the main test. This is what the company is guaranteeing. Temperature and temperature cycling also uncover defects like contamination and mechanical problems like die attach and bond wire attach. The military grade product was subjected to more extensive and harsh testing along with statistical measures- ie if x% died in testing, the batch would be rejected and so on. Similarly, automotive grade have been subjected to a more rigourous test. Specific packages might only be offered in that grade due to mechanical reasons.

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

I have an application, that in certain, extreme conditions could result in an ambient temp of my processor up towards 100degC.  These conditions are rare, and the system is not safety critical.  Although production system will use either the 105/125 degC part, proto systems (just 2 of) due to sourcing constraints, are going out the door with 85degC rated processors in.  I am trying to understand the possible implications of that.  I am planning to oven test a third unit, but that cannot be done before the proto's have to be shipped.....

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

Possible implications? The device might not work as you expect. The reality will probably be that it will work fine. What about the rest of the circuit?

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

I guess this almost is like over clock, we don't know.

 

But to hot, I will guess max speed get reduced, and don't write to EEprom (and program Flash), at to high temperature.

 

And then my guess is that the used plastic of the house might be different (die the same), so the bonds internal stays connected after many temperature cycles.