“All models are wrong, but some are useful” Part I

The original business case that David Tatchell and Harvey Rosten put forward for the formation of (what was to become) Flomerics had the following quote on its front page:

All models are wrong, but some are useful

G. E. Box

This sentiment has underpinned the approach my colleagues and I have always taken in the packaging up of computational fluid dynamics technology for the thermal simulation of electronic systems. It still does so as we now continue to work as the ‘Mechanical Analysis’ division within Mentor Graphics. The quote is our pragmatic dogma.  This and the following few blog entries will explore the issues associated with the usefulness, accuracy and cost of CFD based electronics cooling simulations.

Back in the day (and sometimes even now) the issue of the predictive accuracy of FloTHERM was front and center of each sale. “How accurate is FoTHERM?” was answered in a number of ways:

“Accurate enough to be useful”

“It depends on the accuracy of your input data”

“It depends on the accuracy of your experimental measurements”

And finally, when push came to shove:

“Usually ~10% on dT predictions”

With the following response: “O, really, err, ok then”.  Such a response I put down to expectations that are more often than not laid down and set in academia. A wonderful environment, free and unfettered from the restrictions of time (=$$) and singularly minded in the pursuit of accuracy at all costs.

Should we ever be happy with 10%? Yes, and in addition we should be somewhat surprised with <5% and unbelieving of <1%.

Why? Why be accepting of something that could be smaller, could allow FloTHERM users to have an even greater confidence of the thermal compliance of their yet to be made electronics? There’s a long nasty line up of suspects, the ring leader of which is power dissipation.

Power, Q (Watts -> J/s ->Nm/s (speed, hmm, more on that some other time…))

Power, it’s why you need tools like FloTHERM, that and the fact that solid materials tend to grumble by getting hot when you first dissipate heat in, then force heat to pass through, them.

From_hot_to_cold

Electronics cooling is getting the cold in to quench where the power is being dissipated as quickly/easily as possible. Conversely it is trying to get the heat out as quickly/easily as possible to stop the temperature building up at the power source. Either way, power’s a main culprit.

So what form do power dissipation errors take?…

Usually, in an electronics cooling simulation, you’ll be assuming a steady state of your system, meaning in this case that the power dissipation in the die is unchanging in time and therefore an time averaged value should be sought and used. Alternatively if you are performing a transient simulation you might be assuming that the time averaged value changes from one value to another. For many packages and operating behaviours such assumptions themselves will contribute to the overall simulation error.

There was a study once (whose reference alludes me) of taking two ‘identical’ packages from a production line, powering them up and measuring their actual consumption. The consumption that should have been identical varied by ~10% between the two.

If you are using a detailed model of a package in your simulation, where the important internal 3D geometric structures are explicitly represented (e.g. lead frame, tie bars, heat slug etc.) likely you’ll be assuming that the die is a single piece of Si with the power dissipation uniformly spread over it. This assumption can also carry an error. (Note that in the V8.1 of FloTHERM you can use the Die SmartPart to specify a non-uniform collection of discrete areas with differing power dissipation over the die, this goes a long way to shed the inaccurate assumption of overall die power dissipation uniformity).

So what is the relationship between power inaccuracy and resulting thermal predictive accuracy? Hey, good question. I did a little study in FloTHERM, took a detailed model of an 84 lead PLCC, stuck it on a 2S2P test board in the standard JEDEC still air test environment, set the ambient temperature to 45degC and solved at 1W, 0.9W and 1.1W (~10 minutes start to finish, man, I love FloTHERM!). Looking at the resulting Tj values I found that if power dissipation is known to within +/- 10% then the resulting dTj (Tj rise above ambient) is also +/-10% accurate. (Hey, thermal experts, under what situations would it not be a proportional relationship when Q+/-10%?).

plcc84_power_vs_tj_and_graph1

The point about this power inaccuracy issue is that there will be inaccuracies that you will have to carry and suffer not because you are a poor modeler, but because of assumptions you have to make due to the utter lack of good quality thermal data. Getting even the average power dissipation for the package as a whole is difficult enough, let alone  knowing how to account for manufacturing variations in the package construction! Experience has shown that in a design world of unknowns and unknowables a +/- 10% accuracy of an electronics thermal simulation is to be expected, not feared.  If you believe your design requires predictions of even tighter margins then I’d be more worried about your design itself rather than its simulation.

There are suspects in the line up of modeling errors that you as a modeler will have control over. Whereas power dissipation inaccuracy is something you have to suffer, there are others that you could be responsible for….. more on that in Part II

Robin Bornoff

Hampton Court

May 12 2009

Post Author

Posted May 12th, 2009, by

Post Tags

, , ,

Post Comments

6 Comments

About Robin Bornoff's blog

Views and insights into the concepts behind electronics cooling with a specific focus on the application of FloTHERM to the thermal simulation of electronic systems. Investigations into the application of FloVENT to HVAC simulation. Plus the odd foray into CFD, non-linear dynamic systems and cider making. Robin Bornoff's blog

Comments

6 comments on this post | ↓ Add Your Own

[…] of a package. A 3D object representing the die, with the correct material properties, with the power being set to dissipate on its surface. All the other important internal objects such as die attach, bond wires or […]

Commented on June 23, 2009 at 5:15 am
By Chris Hill

If your ring-leader for model inaccuracy is power dissipation, then I would like to offer “material properties” as a close-run second. Sure, for common materials such as copper the thermal conductivity, specific heat, density, etc., have been tested to death and are well known and readily available. But what about, say, the plastic encapsulant used in that Far-Eastern QFP package? If you’re lucky you *might* get a figure for thermal conductivity – but specific heat? Good luck with that… Also, most leadframe materials are not pure copper anyway and usually have some small percentage of other metals, which can have a profound effect on thermal conductivity. I would say that material properties often provide me with the biggest headache because many of the other factors are at least partially under my control.

Commented on June 24, 2009 at 5:50 am
By Robin Bornoff

http://www.matweb.com/ is the most extensive materials database. It’s not free (well, not all of it) but is the best source for troublesome rho and Cp.

Commented on June 28, 2009 at 11:13 pm
By Chris Hill

Indeed I was fortunate enough to discover matweb some years ago and have used it often. I particularly like the ability to search for alloys based on percentage composition. However, my experience is that even matweb often cannot help me with density and SHC figures.

[…] the fact that all models are wrong a detailed model representation of a package is least wrong and comparatively the best. Each […]

[…] them to his/her advantage. A about a year ago to the day I ran a series of blogs based on the quote “All models are wrong, but some are useful”. In the same way that art is never reality, just an interpretation of it, then a computer model will […]

Add Your Comment

Archives

November 2014
  • If You’re Going to Lose it, You Might as Well Use it!
  • October 2014
  • Thermal Bottlenecks. This is Hot, This is Why.
  • Blue LEDs. Since When is Improvement Invention?
  • Leg Hair? What a Drag
  • The Electronics Cooling Metaphorical Drinking Game
  • September 2014
  • Xilinx Patent for Critical Tj Prediction
  • Dell Precision – Spot on Thermal Design
  • Top 10 FloTHERM V10 Features – #11: Odds and Sods
  • Top 10 FloTHERM V10 Features – #10: Improved Solar Calculator
  • Top 10 FloTHERM V10 Features – #9: Data Center Simulation
  • August 2014
  • Top 10 FloTHERM V10 Features – #8: Thermostatic Control with Hysteresis
  • July 2014
  • Top 10 FloTHERM V10 Features – #7: Super-fast Parallel CFD Solver
  • June 2014
  • Top 10 FloTHERM V10 Features – #6: Integrated Summary Columns
  • Top 10 FloTHERM V10 Features – #5: FloSCRIPT
  • Top 10 FloTHERM V10 Features – #4: Updated CAD
  • Top 10 FloTHERM V10 Features – #3: FEA Interfacing
  • February 2014
  • Top 10 FloTHERM V10 Features – #2: Advanced Find
  • Top 10 FloTHERM V10 Features – #1: New GUI
  • January 2014
  • Come and Learn about the Latest Release of FloTHERM, V10
  • Heat Your Home Office for 8p a Day. Part 5 – Putting it All Together
  • December 2013
  • Heat Your Home Office for 8p a Day. Part 4 – Comfort Temperature
  • Heat Your Home Office for 8p a Day. Part 3a – Was Dave Right?
  • November 2013
  • Heat Your Home Office for 8p a Day. Part 3 – It Takes Time
  • Heat Your Home Office for 8p a Day. Part 2 – Thermal Interception
  • Heat Your Home Office for 8p a Day. Part 1 – Really?
  • Happy 25th Birthday FloTHERM !
  • July 2013
  • Why Not Just Shove a Heatsink on Top of it? Part 3: Pads, Vias and Undersinking
  • May 2013
  • Why Not Just Shove a Heatsink on Top of it? Part 2: Heat Flow Budgets
  • Why Not Just Shove a Heatsink on Top of it? Part 1
  • April 2013
  • Experiment vs. Simulation, Part 5: Detailed IC Package Model Calibration Methodology
  • CFD – Colourful Friday Distractions
  • Experiment vs. Simulation, Part 4: Compact Thermal Models
  • February 2013
  • Experiment vs. Simulation, Part 3: JESD51-14
  • January 2013
  • Experiment vs. Simulation, Part 2: TIM Thermal Conductivity
  • Experiment vs. Simulation, Part 1: Them and Us.
  • September 2012
  • “Why Cartesian Grids Are Good”
  • August 2012
  • Where’s the Best Place to Put a Radiator in a Room? Part 5: Get a Job
  • Where’s the Best Place to Put a Radiator in a Room? Part 4: Premature Simulation
  • July 2012
  • Where’s the Best Place to Put a Radiator in a Room? Part 3: 13% Better
  • Where’s the Best Place to Put a Radiator in a Room? Part 2: PMV and other TLAs
  • Where’s the Best Place to Put a Radiator in a Room. Part 1: Such Things are Important
  • May 2012
  • Agile software development practices in the Mechanical Analysis Division
  • A Little Goes A Long Way (But A Lot Doesn’t Go Much Further)
  • April 2012
  • More Than Two Decades and Still Going Strong; FloTHERM and FloVENT V9.3 Now Released
  • Simulation Software So Simple Even Teenagers Can Use It
  • February 2012
  • Bottlenecks and Interface Materials; Part 3 – Relieving Thermal Bottlenecks Reduce Temperatures
  • January 2012
  • Bottlenecks and Interface Materials; Part 2 – When TIMs Go Bad
  • Bridging the Simulation Supply Chain; NXP Semiconductors, a Case in Point
  • Bottlenecks and Interface Materials; Part 1 – Great Thermal Bedfellows
  • Emails, more Emails and Jeff Bridges
  • LEDs; The future’s bright and hot.
  • December 2011
  • From Megawatts to Milliwatts; sub-micron scale thermal modelling with FloTHERM
  • November 2011
  • What! All that just for that? The bonkers world of CPU cooling.
  • October 2011
  • Ho, Ho, Ho! Facebook moves to Lapland
  • All Detailed Thermal IC Package Models are Wrong… Probably
  • Underfloor Electric Heating. Part III – Penny wise, pound foolish.
  • August 2011
  • Underfloor Electric Heating. Part II – Infrared Thermography
  • Underfloor Electric Heating. Part I: In by Christmas
  • June 2011
  • Come, meet FloTHERM/VENT/EFD users, learn and enjoy!
  • PC Overclocking and Aftermarket Modding. Part III – Power vs. Frequency?
  • PC Overclocking and Aftermarket Modding. Part II – Liquid Nitrogen Overclocking, How Cool is That?
  • May 2011
  • PC Overclocking and Aftermarket Modding. Part 1 – When Colour Matters.
  • April 2011
  • Desktop PC with Integrated Toaster – As if!
  • Thermal Design Perfection Starts with the use of FloTHERM PACK
  • We Love FloTHERM V9.2
  • Desktop PC with Integrated Toaster – the Future is Now
  • March 2011
  • Do you know the way to San Jose?
  • February 2011
  • Beer Fridge – A Case Study in Thermal Design. Part 6 – Baffles and Bottlenecks
  • January 2011
  • FloEFD HVAC Module – Taking Built Environment CFD Simulation to the Next Level
  • Beer Fridge – A Case Study in Thermal Design. Part 5 – Time for a FloBEER
  • Beer Fridge – A Case Study in Thermal Design. Part 4 – FloBEER
  • Beer Fridge – A Case Study in Thermal Design. Part 3 – Side Up or Upside Down?
  • December 2010
  • Beer Fridge – A Case Study in Thermal Design. Part 2 – TEC Effect
  • November 2010
  • Beer Fridge – A Case Study in Thermal Design. Part 1 – A Gift
  • What Can You Learn When You Turn It On?
  • We Love FloTHERM – 8 Reasons to Upgrade to V9.1
  • October 2010
  • On the Vilification of Smokers
  • Identifying Thermal Bottlenecks and Shortcut Opportunities – Taking Simulation to the Next Level
  • August 2010
  • How many frogs does a horse have?
  • It’s a wireless world! No it isn’t.
  • July 2010
  • Are you using ‘Smart’ in a way I am not familiar with?
  • An Interview With… Clemens Lasance
  • I was led to believe we’d have flying cars by now
  • Red Hot Electronic Thermal Analysis?
  • June 2010
  • The art of modelling using CFD. Part VI – Peripheral Boundary Conditions
  • The art of modelling using CFD. Part V – Grid
  • May 2010
  • The art of modelling using CFD. Part IV – Fans
  • The art of modelling using CFD. Part III – TIGs
  • The art of modelling using CFD. Part II – Grilles
  • The art of modelling using CFD. Part I – What happens if you cross art with science?
  • April 2010
  • How much do ‘U-Value’ good thermal insulation? Part VII – “Ooo, shut that door”
  • “A Faster Horse” – Mentor ‘IDEAS for Mechanical’ driving product development
  • March 2010
  • How much do ‘U-Value’ good thermal insulation? Part VI – revenge of the radiative heat flux
  • IC package representation is central to Electronics Cooling
  • How much do ‘U-Value’ good thermal insulation? Part V
  • February 2010
  • How much do ‘U-Value’ good thermal insulation? Part IV
  • How much do ‘U-Value’ good thermal insulation? Part III
  • How much do ‘U-Value’ good thermal insulation? Part II
  • January 2010
  • How much do ‘U-Value’ good thermal insulation? Part I
  • Keeping the caveman warm – HVAC blog
  • FloVIZ, the free FloTHERM/FloVENT CFD results viewer, try it, it’s free
  • ‘Heat Trees’ – taking a leaf out of natures book
  • The Most Extreme CFD Model Ever Ever – Explained
  • FloTHERM and its new XML neutral file format
  • The Most Extreme CFD Model Ever Ever
  • So, you want to predict component temperatures do you? Part VII
  • December 2009
  • So, you want to predict component temperatures do you? Part VI
  • So, you want to predict component temperatures do you? Part V
  • November 2009
  • A trip to MPH and Top Gear Live
  • So, you want to predict component temperatures do you? Part IV
  • So, you want to predict component temperatures do you? Part III
  • October 2009
  • So, you want to predict component temperatures do you? Part II
  • So, you want to predict component temperatures do you? Part I
  • Underfloor Thermal Insulation; Why? Part III
  • September 2009
  • Underfloor Thermal Insulation; Why? Part II
  • Underfloor Thermal Insulation; Why? Part I
  • Is all Software Rubbish?
  • August 2009
  • Thermatronic Stagnation (nothing to do with male deers)
  • Fractals: Gods Artwork, Part III
  • Thermatrons Must Leave
  • July 2009
  • At the Speed of Heat
  • A Load of HVAC TLAs
  • How-to: Invert your thermal model to good effect
  • Clogged cooling fins, a cautionary tale
  • Invert your thermal model to good effect
  • “I work with computers”
  • Fractals: Gods Artwork, Part II
  • Fractals: Gods Artwork, Part I
  • “All models are wrong, but some are useful” Part V
  • June 2009
  • 3D Electronics Cooling CFD, with FloTHERM, in Pictures
  • Spend some time with FlyGuy
  • “All models are wrong, but some are useful” Part IV
  • Flying
  • “All models are wrong, but some are useful” Part III
  • May 2009
  • “All models are wrong, but some are useful” Part II
  • “All models are wrong, but some are useful” Part I
  • Welcome along!