10 things to check when evaluating second source silicon/chips (in desperation)
Many OEMs are scrambling to find second sources (i.e. equivalent chips from other suppliers) to the chips in their BOM for which they have a shortage issue. This is causing companies to go to new suppliers they’ve never worked with, and who may not even have been around for very long.
Here are ten technical things to check when deciding whether to use a second source chip or not:
- Ask for a datasheet that indicates how each parameter in the electrical characteristics table is guaranteed. Each spec in a datasheet’s EC tables has min, typ and/or max. For parameters with min/max you want to check if every part is tested at ATE (automatic test equipment) to verify it meets the spec OR if the spec is guaranteed by a sample of units measured on the bench from a number of different lots when it was being characterized at the development stage OR if the limits are guaranteed by design/simulation. It’s important to understand this since you may be relying on a min/max for your system design that may not be ATE tested on every unit. For important aspects of the chips that are not listed as parameters, you need to understand how they are tested at ATE and how they were verified in simulation.
- Ask for the evaluation reports taken by the supplier over the operating conditions stated in the datasheet. Every supplier should have thorough evaluation reports available for their parts. These reports go well beyond what is on the datasheet, and while these reports cannot be used as guaranteed specifications they should show how thoroughly this supplier’s characterization process is. Make sure the reports are for the part you are buying.
- Determine any safety critical aspects of this chip to your system, and request verification (i.e. sims) and validation (measurements) methodology and test results/reports. Many safety critical loops are implemented as hardware state machines, so it’s vital that you check how thoroughly these state machines have been checked in simulation and in measurement, and whether they are checked at ATE for every unit before it ships. A classic example is a charger state machine that gets stuck in the charge state randomly, this could lead to a battery explosion event, so better check as much as possible because this could be one of those hard to repro failure modes. Also check if registers for safety critical things have any type of protection against erroneous (or malicious) software writes.
- Ask for the JEDEC qualification report. Check how much of the testing was done using this part and how much is based on QBS data from other parts. Dig into the logic of why QBS is applicable and ask to see the data for the QBS parts being referenced. Double check that the testing meets the applicable JEDEC standard test conditions, sample size, # of lots, all applicable tests performed, etc…
- Ask for evaluation kits. Order some evaluation kits and test the parts as much as possible to your operating conditions. It is valuable to also perform temperature testing when doing this evaluation. Use an oscilloscope to check for stability issues, and weird waveforms that could indicate the circuit is not behaving as expected even though still within specs.
- Order a few loose parts from a distributor. Some suppliers will cherry pick the parts they put on evaluation boards, but then you test randomly selected parts and they may not behave as nicely as the cherry picked ones. So if you can first test the eval kits and check performance there, and then after replace the part with one from the loose parts you got from the distributor, then you can repeat your testing and see if these loose parts look like the one that came with the eval kit. If it looks like it has a lot of odd behaviors, and you’re certain you didn’t mess up the soldering when replacing the eval kit part, then you may have reasons to worry about selecting this second source.
- Ask for process corner parts and use as a configuration in your system build. It depends what type of chip you’re second sourcing. For digital chips asking for the digital process corners is usually good enough: slow NMOS, slow PMOS (SS), slow NMOS, fast PMOS (SF), fast NMOS, slow PMOS (FS), and fast NMOS, fast PMOS (FF). Ask for 100 of each corner and build it. For analog chips, you can decide which chip parameters are the ones that most matter for your system, and then ask for corner parts in the distribution for those parameters then create a configuration for your system build with these parts.
- Double check your BOM for incorrectly selected higher grade parts. Sometimes in the interest of focusing the volume on a few parts per supplier a chip with automotive or industrial grade specs is used in a part of the system that would be fine with the lower grade consumer electronics version of the part you’re second sourcing. Check if this is true in your system, and whether the consumer electronics part is available more readily.
- Unleash your quality team on the supplier. Your quality team should be able to assess whether the supplier follows good practices or not.
- Ask for the errata. It may not be publicly available, but most/all chips once released into the field have a set of (hopefully) little bugs that customers have been reporting. Ask for a summary of these bugs so you are aware of them and can assess whether they matter to you or not. To be clear, this is not the datasheet errata of typos, this is a list of known bugs in the chip.
If you want to go deeper, you could start digging into what EDA tools are used (are they legit commercial tools?), dig through the supplier’s methodologies for verification (simulations) and validation (measured data), and even review tape out verification for the chip that interests you to ensure it has been thoroughly simulated. In reality, this is a lot of work and most suppliers will not provide this unless you are a special OEM to them. But it’s worth doing it if you can get it. My list only covers technical things, but you should of course perform other commercial checks as well.
In the end, field issues can be very costly and could lead to safety issues for your customers. So when evaluating a new chip supplier you need to make sure that you not only trust, but verify too.
CustomSilicon.com is the leading consulting firm in the custom silicon strategy and project management space for AR/VR, automotive, mobile, server, crypto, sensors, security, medical, space and more.
Raul has 20 years of combined experience in the system electronics and silicon industries. He is currently responsible for major system company’s custom silicon and sensor projects. Raul was the directly responsible silicon manager for 18 chips ramped to mass production at Apple for iPhone and iPad, and 23 total chips ramped to mass production counting projects where he was an expert reviewer. Raul was directly responsible for the development of mobile processor System PMICs for the iPad2, New iPad, iPad mini, iPad 4 and iPhone 5s. Other silicon included, backlight/display power for iPhone 5 and iPhone 5s, lightning connector silicon and video buffers. He managed supplier teams across the Globe.
Our network of experts provide our clients with an A+ silicon management team from day one.
Photo by NeONBRAND on Unsplash
Trust, but verify. How to catch peanut butter engineering before it spreads into your system — Part 1: Validation.
Trust, but verify. How to catch peanut butter engineering before it spreads into your system — Part 2: Verification.
Originally published at https://customsilicon.com.