General > General Technical Chat
Semiconductor Marketing Tips (Through the Use of Misleading Specs)
niconiconi:
1. Make sure to use a different test methodology or standard to specify the performance of your part in the datasheet, so it's impossible for designers to make a direct comparison against your competitors.
2. When your device has a good parameter, make sure the first page of the datasheet shows its performance tested by a method that gives the best-looking numbers. Bury the other numbers tested under more realistic settings in the least visible page in the Electrical Specification section.
3. When one parameter of your component is far from ideal, never even list that in the datasheet, just pretend it doesn't exist!
4. If your device performance starts degrading seriously beyond a point, just cut off the X-axis early in your graphs before that happens. For example, if the frequency response is really bad beyond 500 MHz, just stop at 450 MHz, so people will never notice this deficiency in the datasheet!
5. Always advertise the best theoretical performance, even if it's not practical or literally impossible to achieve in real-world applications. For example, assume input impedance is perfectly matched to a non-standard value. Or, when the package limits the maximum MOSFET current to 100 A RMS, still say your MOSFET can handle 200 A in large boldface font!
6. Always boast your device using its "Typical" specs and claim it's the latest and greatest thing ever. What? Guaranteed worst-case performance is 30% lower than "Typical"? You're just unlucky! Better luck next time in your next silicon lottery!
7. When one of your device parameters has small unwanted variations, make sure they're unreadable from the datasheet charts. With a scale large enough or logarithmic, only the extreme ends are distinguishable, everything looks like a straight line! Remember the saying of the great seismologist Charles F. Richter, "logarithmic plots are a device of the devil."
8. Remember to list all prices of your parts on their product pages to increase sales. But DO NOT list them in the parametric search list, so it's impossible for designers to compare your part's price/performance ratio, and hopefully will pick a more expensive part by accident.
After finished writing this, I later learned there's even a special word for these marketing practices, called specsmanship.
--- Quote ---Specsmanship is the use of specifications or measurement results to suggest [...] superiority [...], especially when it is inappropriate.
--- End quote ---
Anyone has more to add? :-DD
dbctronic:
Claim that the part is 'equivalent to' a part it just plain isn't.
I've seen a Chinese transistor datasheet claiming that their transistor is a replacement for the 2N2222. Casual inspection of the operating parameters shows that several are just plain NOT 2N2222 equivalent, at all--the minimum hFE is 20, not 35! It's a general purpose NPN signal transistor, and that's about it. 'Replacement' my sore butt.
I didn't bother to check whether the pinout is the same... :rant:
niconiconi:
--- Quote from: dbctronic on April 12, 2022, 12:19:54 am ---Claim that the part is 'equivalent to' a part it just plain isn't.
I've seen a Chinese transistor datasheet claiming that their transistor is a replacement for the 2N2222. Casual inspection of the operating parameters shows that several are just plain NOT 2N2222 equivalent, at all--the minimum hFE is 20, not 35! It's a general purpose NPN signal transistor, and that's about it. 'Replacement' my sore butt.
I didn't bother to check whether the pinout is the same... :rant:
--- End quote ---
An example I've seen. Nexperia has a "PMEG" series of low-Vf Schottky diode, useful for squeezing the last half percent of efficiency out of a power supply. The PMEG3020ER for example, has 365 mV typical, 420 mV max at 25 degrees. Meanwhile, a quick LCSC search finds another Taiwanese company, Shikues, sells the same part. But the datasheet says the Vf is 500 mV typical, 550 mV max. :-- Yes, definitely beware.
But speaking of your 2N2222 example, I think these classic parts are all JEDEC-registered standard parts. If I get it right, the JEDEC's requirement for 2N2222 only says the DC current gain at 1 mA should be 50, and nothing more... So, if this part meets that bare-minimum spec, it's a fair game to make the claim... But if not, then it would be not just a case of "marketing" but plain dishonesty.
Tomorokoshi:
Why bother measuring anything? Just plagiarize someone else's datasheet!
SpacedCowboy:
The most egregious example of this I ever saw was Silicon Image - I was writing some firmware to translate MHL to HDMI using one of their chips. It. Did. Not. Work.
I called up the rep, and he told me, quite serenely that the public data sheet was wrong, and he'd send over the correct one under NDA. That man was lucky not to be in my immediate vicinity.
I couldn't do much about it then, time pressure was a thing, but I told my boss I would never work with their chips again, and if he wanted a project with their chips in it, to get someone else to work on it. To my knowledge we never bought one of their chips thereafter.
Navigation
[0] Message Index
[#] Next page
Go to full version