COMPARISON BULLETIN NVE Calls “Shenanigans” on ADI  “To call shenanigans”: to label something as officially deceitful, improper, or otherwise incorrect.  It’s come to our attention that Analog Devices is circulating a misleading bulletin comparing NVE’s award-winning GMR isolators to their “iCouplers.” They cherry-pick specifications, compare apples to oranges, and cite conditions that do not occur in the real world. Here’s what they say: ADI Claim Facts “Data transmission triggered by input logic transitions” Actually true, at least for some NVE isolators. IL200/IL700 Series isolators have unique edge-triggered inputs, which provide better real-world noise performance, because it takes an edge—not just a level shift—to switch the device. Level shifts can come from power supply changes, load changes, and other real-world problems. Additionally, ADI transceivers have to non-deterministically encode data with encoding, which leads to occasional errors. NVE’s technology does not require encoding, so there is no data error mechanism. And finally, because they use significant power only on transitions, NVE parts have lower power consumption and less self-heating.   “dc correctness not ensured” DC correctness is assured. Unless you somehow get rising edges without falling edges.   “Correct output upon start-up not guaranteed” ADI parts don’t guarantee the output on start-up, you have to wait. NVE isolators have no start-up time. IL200/IL700 Series outputs follow the inputs as soon as there is a transition. Other than that, they tend to stay put, even when the power is removed. If you really want a state-of-the-art isolator that emulates optocoupler behavior, use the NVE IL500 or IL600-Series isolators.   “Susceptible to false triggering by input noise” ADI tested IL710s with 8 ns pulses and noted that NVE parts switched and ADI parts didn’t. That’s not “noise immunity.” That shows how slow ADI parts are. Missing pulses is certainly not an advantage. Then they tested with 500 ns input rise times and said NVE parts didn’t always switch. Why would you want them to? Digital signals have rise times on the order of 3 ns, and for best noise performance everything else should be rejected. IL200/ IL700 Series isolators are insensitive to slow level changes, because your system electronics don’t output 500 ns edges. The best measure of noise input immunity is common-mode transient immunity, and NVE isolators are 25% better—30 kV/µs versus 25 kV/µs.  “Retains last state upon loss of input power” Some of our customers like making ultra-reliable systems that don’t skip a beat if they lose power.   “Sensitive to external dc magnetic field” NVE actually has higher magnetic immunity than ADI depending on frequency. ADI tries to make a comparison at DC magnetic fields, which is an unusual condition. NVE devices pass all relevant EMC standards, including dc magnetic immunity. There are dc immunity specifications in our data sheets, and you are very unlikely to see those field levels. Immunity is just a small part of the EMC picture. The other half is emissions, where NVE has a big advantage because there are RF EMI-generating pulse trains in the ADI devices. NVE isolators are used in the world’s most demanding applications, including life-support medical devices, military electronics, and avionics because they are rugged. They have barrier lives of 44000 years. Versions operate to 125C. And they withstand radiation levels that would kill you in a nanosecond.   “Output ringing/overshoot” ADI compares an output trace from their device to what appears to be an unterminated IL710 and touts less ringing. That’s because their part is slower and has a higher output impedance. Our parts are designed to drive electronics, not oscilloscope probes. NVE’s fast rise and fall times ensure crisp timing, and low output impedance means speed isn’t degraded by fanout or capacitive loads.   So why is ADI so worried about NVE? Because NVE IL700-Series Isolators are half the size of ADI’s, with less signal distortion, higher speed, less EMI, lower power, and longer life. Here’s a comparison:      Smallest Package  Max.  Prop.  Delay Max. Ch- Ch Skew Typ. Jitter Power Consump- tion(2) Isolation Est.  Product  Life   NVE IL700V Series  MSOP-8  18 ns 3 ns  100 ps  188 mW 6 kV  44000 yrs   ADuM3 Series SOIC-8 45 ns(1) 16 ns(1) Not  Specified  1375 mw 5 kV 50 yrs (1) The fastest grade of ADI’s fastest two-channel bidirectional isolator. (2) ADI and NVE's fastest four-channel, unidirectional parts; total of both power supplies at 5 V, 75 MHz, 150 Mbps. Half the Size (or Less) ADI’s agency-compliant four-channel parts are only available in a 0.3" SOIC-16; two-channel parts only in SOICs. IL700-Series four-channel devices come in elegant 0.15" SOIC-16 packages. If that’s still too big, use two NVE two-channel MSOP-8s and reduce PCB footprint by another 50% compared to the ADI parts. ADI can’t match NVE’s two-channel MSOPs for miniaturization, or PDIPs for pin spacing. Much Less Signal Distortion As shown in the table above, the IL712/IL721 maximum propagation delay is less than half ADI’s fastest bidirectional parts and channel-to-channel skew is five times better. Jitter is critical for isolating high speed DACs and ADCs, since any clock jitter directly limits dynamic range and precision. NVE’s IL700s offer a remarkable 100 ps typical jitter; ADI doesn’t even have a specification. Usable Baud Rate ADI claims a 150 Mbps data rate on some parts that have propagation delays of 35 ns or more. That’s more than five data cycles of propagation delay at 150 Mbps, so the usable baud rate is much less. Also, there’s a minimum pulse width spec of 6.67 ns, so you need exactly 50% duty cycle with zero rise and fall time. NVE’s speed specifications support practical use at high speeds, including propagation delay, jitter, and channel-to-channel skew. Less EMI NVE isolators have no RF carriers or pulse trains. The result is NVE isolators have virtually undetectable radiated emissions. ADI's transformers and pulse trains are like radio stations on your boards, and their own literature describes ADI devices as “efficient dipole radiators” (i.e., antennas). As a result, ADI isolators exceed EN55022-B and FCC B even with no other system components. They have a 20 page application note describing what you have to do. We have customers who came to us desperate because with ADI parts they couldn’t pass their EMI tests. Don’t wait until then. Design in NVE isolators from the start. One-Seventh the Power Consumption Worst-case power consumption is a ridiculous 1375 mW for ADI’s four-channel part at maximum frequency, compared to 188 mW for NVE’s four-channel part at the same speed. So NVE simplifies power supplies and reduces cost. Barrier a Hundred Times Better A unique ceramic/polymer barrier gives NVE a best-in-class 6 kV isolation rating and a remarkable barrier resistance of 1014 ohms, at least 100 times better than ADI. A better barrier also means NVE isolators have a predicted failure rate of one part in 44000 years. That’s practically forever. ADI specifies a working life of 50 years. That doesn’t sound bad until you consider the statistical chance of a particular ADI device failing much sooner. Broader Product Line In addition to more package types, NVE offers a broad product line for a wide range of applications, including IL600-Series isolators to emulate optocouplers.