Timing chips? what's so special?

WBahn

Joined Mar 31, 2012
29,976
Why are these apparently special, surely timers like this are a dime a dozen, what gives?
Did you read the article? Much of it is devoted to explaining what needs these chips serve. A bit more detail is contained in SiTime's press release that is linked in the first paragraph.

What are some of the "dime a dozen" timing ICs that meet these needs going forward?
 

Ian0

Joined Aug 7, 2020
9,667
I can see the need for frequency accuracy in things like LIDAR, RADAR and radio comms, but most microcontrollers these days appear to be able to do the same job with a single crystal.
CAN requires accurate timing, but I wonder how many crystals could have been saved had CAN been Manchester coded or some other protocol from which the clock could be recovered. Did we need CAN when we already had MIL1553?
 

Thread Starter

ApacheKid

Joined Jan 12, 2015
1,533
Did you read the article? Much of it is devoted to explaining what needs these chips serve. A bit more detail is contained in SiTime's press release that is linked in the first paragraph.

What are some of the "dime a dozen" timing ICs that meet these needs going forward?
Well it is described as being a "low power oscillator" the term "timer" or "timing" does not even appear in the device's datasheet. So it is an oscillator and these are a dime a dozen, NE555? so that's what I'm driving at, what exactly makes these devices so special? I'm not saying they aren't, only that I have no idea what it is that actually makes them noteworthy.
 

Ian0

Joined Aug 7, 2020
9,667
I don't quite think a 555 will manage phase jitter of less than 750 femto-seconds, and the frequency won't be stable within 50ppm over the full operating range, because an ordinary ruthenium oxide surface mount resistor only manages 100ppm/°C; but I still think that the calculated distance between the car and the one it's about to crash into could vary by >0.1% without making much difference to its ability to avoid the crash.
 

Thread Starter

ApacheKid

Joined Jan 12, 2015
1,533
I don't quite think a 555 will manage phase jitter of less than 750 femto-seconds, and the frequency won't be stable within 50ppm over the full operating range, because an ordinary ruthenium oxide surface mount resistor only manages 100ppm/°C; but I still think that the calculated distance between the car and the one it's about to crash into could vary by >0.1% without making much difference to its ability to avoid the crash.
OK I see, so these are actually very high stability oscillators, not "timing chips", that's where my confusion arose.
 

AnalogKid

Joined Aug 1, 2013
10,986
100 parts-per-million, which is 0.01 %, is not good enough for automotive applications. Yet, it is over 1000x better than a 555 circuit with 1% resistors and capacitor. A capacitor that can hold its value to within 0.01% over a brutally wide temperature range plus aging is, in round numbers, impossible; and that's at half the needed performance. You can debate whether or not a 50 ppm oscillator is required, but once that grand proclamation is made, *all* R-C circuits are out.

Next up, crystals. Good enough accuracy and temperature performance, but at heart a thin sheet of glass that is way less reliable than just about any other non-vacuum-tube electronic component. The original IBM PC had a CPU clock of 4.772727 MHz because IBM refused to put two different crystal oscillators in the system. In the world of electronics reliability, everybody hates crystals.

Maxim and others have been developing and refining the "silicon oscillator" idea for decades; got, Got, GOT to be a reason.

ak
 

Thread Starter

ApacheKid

Joined Jan 12, 2015
1,533
I'd like to know what's so unusual, unique, about car systems, that such high stability is deemed essential anyway.
 

debe

Joined Sep 21, 2010
1,389
I'd like to know what's so unusual, unique, about car systems, that such high stability is deemed essential anyway.
Well in relativly cheap cars such as a Hyundai I30 they have active braking when in cruise. From experience if a vehicle in front of me stops suddenly, so does mine. They will also self steer betwean lane line markings. So yes there is need for accuracy in the electronics.
 

WBahn

Joined Mar 31, 2012
29,976
Well it is described as being a "low power oscillator" the term "timer" or "timing" does not even appear in the device's datasheet. So it is an oscillator and these are a dime a dozen, NE555? so that's what I'm driving at, what exactly makes these devices so special? I'm not saying they aren't, only that I have no idea what it is that actually makes them noteworthy.
A 555-based oscillator would be nowhere near good enough for many of these applications.

As stated in the article, the four key specs that this oscillator is focused on are: jitter, power consumption, reliability, and stability.

Starting off, an automotive electrical system is a very harsh environment with lots of noise and extreme temperature differences. You oscillators have to be sufficiently accurate over that entire temperature range. They also have to have a very, very low failure rate since they must be expected to last the entire life of the vehicle -- and not just one of them in the car, but all several dozen of them.

As far as just what the actual specs that need to be met for modern automotive oscillators, that would require some digging to ferret out if you aren't in that industry, but some pretty good guesses for some of them can be had by considering some of the protocols, such as CAN, that are in use as well as looking at some of the various systems that are being incorporated into vehicles. Again, it must be kept in mind that you aren't looking for a chip that meets the spec at 25°C, but that meets it over the entire automotive temperature range (at least).
 

Ian0

Joined Aug 7, 2020
9,667
In the world of electronics reliability, everybody hates crystals.
I never knew that - it doesn't ring true (dreadful pun) with my own experience. Semiconductors fail more than MLCC capacitors (which part company with their metallisation), and MLCC capacitors fail more than crystals.
Although they consist of a think sheet of silicon dioxide, that's not a lot different from a semiconductor wafer, it's just not oxide all the way through.
 

AnalogKid

Joined Aug 1, 2013
10,986
I got a mini-lecture from one of the big defense contractors' reliability dude about 10-15 years ago. It is entirely possible that manufacturing techniques have changed for the better since then, or that he was burned by a batch of crystals way back when and I just stepped into the remnants of an old mess. Either way, he was crystal clear.

ak
 
Top