Modern WOW! factor chips.

[VK3DRB]

A few days ago I mentioned to a colleague at work, a WOW! factor chip called the LTC2984 which I discovered a few years ago when Digikey put out a bulletin about them. Oddly enough, the colleague whom I have only work with for a few months, also got the big WOW! factor feeling about this chip when it was announced at the same time. That chip was a game changer from its integration level, the number of inputs, its sheer versatility and accuracy.

http://cds.linear.com/docs/en/datasheet/2984fb.pdf

What other WOW! factor chips do you know about? Forget ye olde 555's, 741's, XR2206 etc. - most of us know about them and they are old news. Please post any relatively new chip that you find that might be an order of magnitude better than anything else out there, or are very unique for an application. In other words, it has has to spark a feeling of WOW! when you first discovered them.

I think such a list will not only be interesting to the readers, but also very useful.

[@rt]

I was surprised to see the YouTube video about Potato chips.

1GHz 7400 series logic.

[technix]

Intel Xeon E5-2670 two years ago. The WOW! factor is its second hand market price. Back when it was released in 2012 it was a $1552 MSRP processor with 8 hyperthreading cores at up to 3.3GHz, 2 more than the best desktop chip available at that time i7-3970X. Fast forward to mid 2016 the price of this chip plummeted HARD to as low as $70. I actually stepped the game up a bit and bought a pair of E5-2680s for less than $150, now living in my HPC workstation.

[KE5FX]

For the Redditors among us, also check out http://www.reddit.com/r/nicechips .

[NiHaoMike]

LTC3871 is a very interesting bidirectional DC/DC converter control chip. It unfortunately doesn't quite meet what I want for a project I'm working on, so it looks like I'll be designing a DC/DC controller out of multiple chips anyways.

Then there's the SM74611 which is a drop in replacement for a blocking diode in low voltage PV systems. What's notable is that I was using a Schottky diode in my solar setup prior to that and going from a surprising amount of heat to almost none was way better than I expected. That solar setup, BTW, is likely going to be the most special solar setup I'll ever own.

[PartialDischarge]

From that reddit post, LMX2594 10Mhz to 15GHz PLL
Looks like a GHz is not much anymore...

[ikrase]

The low end (<80 USD) Apex high voltage high current opamps are pretty cool, especially from my perspective (since I do not know how to design discrete amps)

On the more normal end of things, there is a 140V Linear op amp.

[Circlotron]

LT1166 - Power Output Stage Automatic Bias System
http://www.linear.com/product/LT1166

[GeorgeOfTheJungle]

The greenArrays 144 cores on a chip chip http://www.greenarraychips.com/ IDK what for but I'd like to play with that.

[David Hess]

What other WOW! factor chips do you know about?

The Linear Technology LT1008/LT1012/LT1097 operational amplifiers have low input bias current combined with precision.  Being bipolar parts, their input bias current does not double for every 10C rise in temperature so unlike less precise and noisier JFET and MOSFET parts, they can maintain low input bias current at elevated temperatures.  The now discontinued National LM11 which is an improved LM308 used to fit into this niche.

The National LMC6081 operational amplifier can be graded for an input bias current of 2 femtoamps.  This part was later released as the much more expensive pre-graded LMC6001.

The National LM301A is almost uniquely useful for having an input common mode voltage range which includes the positive supply for easy high side current sensing and a compensation pin which can be used to clamp the output preventing windup.  Despite details searches, I have not found a modern replacement for this ancient part.

[gibbled]

Silabs si4836 wowed me.  A versatile multiband digital receiver on a chip!

https://www.silabs.com/products/audio-and-radio/multi-band-radios/si4820-24-25-31-35-36-radio-receivers

[KE5FX]

Silabs si4836 wowed me.  A versatile multiband digital receiver on a chip!

https://www.silabs.com/products/audio-and-radio/multi-band-radios/si4820-24-25-31-35-36-radio-receivers

Interesting, I didn't know anything like that was (still) being built.  I wonder if they use varactors internally or fast V-to-F converters, or just an ADC whose output is used to tune a conventional LC-based PLL. 

[Alex Eisenhut]

[VK3DRB]

SiLego. Truly impressive programmable silicon dust.

Yep, very impressive; as appears to be the IDE software. And my guess if Dialog might be a good company to invest in providing this technology is covered by patents from SiLego.

[tggzzz]

The greenArrays 144 cores on a chip chip http://www.greenarraychips.com/ IDK what for but I'd like to play with that.

You would be better off playing with the XMOS chips, since they are hard realtime, have FPGA-like I/O, have a good IDE, up to 32-cores, many variants - and are commercially successful.

[1design]

These:
https://www.infineon.com/cms/en/product/rf-wireless-control/mmwave-mmic-transceivers-24-86-ghz/mmw-backhaul-and-fronthaul/

[TK]

ESP8266 and ESP32 WiFi SoC

[daqq]

Well, the TimerBlox ( http://www.linear.com/parametric/timerblox ) from LT are very useful.

[Someone]

Not a bad chip all said, would buy again. Head over the ditch and try some exotic foods too:

[Someone]

What other WOW! factor chips do you know about?

The National LMC6081 operational amplifier can be graded for an input bias current of 2 femtoamps.  This part was later released as the much more expensive pre-graded LMC6001.

Fun times trying to get the leakage low enough to push the limits of the '6001

For recent advances TI have pushed out a big range of chopper stabilised opamps, trimmed for offset and drift in various prices/powers/speeds

[David Hess]

The National LMC6081 operational amplifier can be graded for an input bias current of 2 femtoamps.  This part was later released as the much more expensive pre-graded LMC6001.

Fun times trying to get the leakage low enough to push the limits of the '6001

When we were using the LMC6081, the LMC6001 had not been released yet.  I preferred the higher power LMC6081 for its lower noise compared to the lower power LMC6061.

We used the DIP packaged LMC6081 and bent the inverting input pin up into the air to avoid leakage through the printed circuit board.  The integration capacitor was a polypropylene part qualified for low leakage at high temperature and the switched input was through a similarly qualified 2N3904 collector-base junction which was clamped to prevent high reverse voltage and charge pumping.  Total leakage current was below about *calculating from memory* 0.5 picoamps and probably on the order of 0.05 picoamps or 50 femtoamps.

For recent advances TI have pushed out a big range of chopper stabilised opamps, trimmed for offset and drift in various prices/powers/speeds

Why would a chopper stabilized operational amplifier require trimming?

[Someone]

For recent advances TI have pushed out a big range of chopper stabilised opamps, trimmed for offset and drift in various prices/powers/speeds

Why would a chopper stabilized operational amplifier require trimming?

To flip that around on you, why don't chopper amplifiers have zero offset and drift? take a look at the TI parts, they clearly have been trimmed at the factory to get the offsets in a tight distribution but now include 10+ MHz GBW parts that are very universal.

[John Heath]

LTC2984 with a 24 bit A/D ! that seems outside the range of possible. The highest bit would have to be within + or - .0000007. I imagine their process engineers are as busy as a one armed paper hanger trying to live up to that spec.

[David Hess]

For recent advances TI have pushed out a big range of chopper stabilised opamps, trimmed for offset and drift in various prices/powers/speeds

Why would a chopper stabilized operational amplifier require trimming?

To flip that around on you, why don't chopper amplifiers have zero offset and drift? take a look at the TI parts, they clearly have been trimmed at the factory to get the offsets in a tight distribution but now include 10+ MHz GBW parts that are very universal.

You will have to link an example datasheet or part number.

Also note that chopper amplifiers are *not* the same thing as chopper stabilized amplifiers.

The residual offset and drift of a chopper stabilized amplifier and as well as the best non-chopper stabilized precision amplifiers is limited by thermocouple effects which are outside the bounds of trimming.  There is no expectation that they will not have close to a Gaussian distribution in various parameters and there is nothing unusual about a 10+ MHz chopper stabilized amplifier; that kind of speed would be unusual in a *chopper amplifier* which is not the same thing.

There is actually a good reason chopper stabilized amplifiers are often relatively slow; high speed CMOS amplifiers are usually noisy.  Some TI parts are incredibly noisy but I assume this does not apply to their chopper stabilized amplifiers.  Higher speed is also usually accompanied by higher heating which leads to larger errors from thermocouple effects.  Precision is often associated with low power operation for this reason.

Chopper stabilized amplifier actually should be faster than precision bipolar precision parts because of built in transconductance reduction from using an FET input stage but this applies to any FET input stage.

[ogden]

Those who are not yet familiar with Cypress PSoC series chips, especially 5LP, can have their WOW moment.