KCC's Quizzes AQQ305 about a mutual exponentiation equation
1. Quote of the month: "A train station is where a train stops. A bus station is where a bus stops. I have a work station on my desk, so..." - unknown
2. Quiz AQQ305 about finding a mutual exponential equation
Good luck, and try to be among the
View All
September Monthly Blog Quiz! Read the blog, take the quiz, and you could win a gift card
Quiz: How Well Do You Know Adaptive Power Switching?
Is your I/O module overheating? The AD74416H's Adaptive Power Switching (APS) can cut heat dissipation by up to 40% — but how well do you really understand how it works?
Read the blog first: Cooling
View All
ADM3251E
Production
The ADM3251E is a high speed, 2.5 kV fully isolated, single-
channel RS-232/V.28 transceiver device that operates from a
single 5 V power supply. Due to...
Datasheet
ADM3251E on Analog.com
While conducting ESD testing on a circuit using the ADM3251E, there are numerous cases where the transceiver will latch-up. When this occurs, the current consumption goes to max (150mA), the Viso rail drops to around 1.2V and the ability to transmit or receive anything is lost. The only ways we've found to reset this is to power cycle the transceiver or to momentarily short the Viso output to GNDiso.
We've tried various modifications to the circuit to try and isolate the root cause of this but nothing has provided a complete or repeatable solution. The behavior has been observed from as low as 4kV (level 2) in most cases, and as high as 8kV (level 4) with some potential fixes implemented. The ESD strikes have been on multiple pins of our output connector, even with some of the IC pins lifted and not attached to the PCB.
Is this a known behavior of the IC and/or has anyone else observed this? If so, are there any recommended fixes or mitigation techniques?
Thanks.
Snip of schematic showing the transceiver... DB9 shell is EARTH

Hi Brian,
The ADM3251E has ESD protection built-in only for HBM (human body model) ESD. I presume your testing is to IEC 61000-4-2, which is a much harsher system level test. I do note the external protection/filter circuit (D4, FB3, FB4, C18, C19). The exact components chosen here will affect whether ADM3251E is successfully protected from disruption.
Another consideration is the effect of discharge routed to GNDISO vs. the other side GND. ADM3251E is rated to minimum common-mode transient immunity of 25 kV/us, but it is possible the ESD discharge could exceed this.
One option sometimes used for ESD protection on isolated designs is to place high voltage capacitors from GNDISO to GND. C17 might be helping here, but it might be possible to connect D4 to GND rather than GNDISO via a capacitor. Or high voltage capacitors could be used for C18 and C19, to connect to GND vs. GNDISO.
One fundamental consideration too is where the equipment is earthed - for example is GND earth referenced, or is GNDiso earth referenced (GND floating/isolated). This is important in terms of immunity offered by any cable shield as well.
In general the system design will need to confer the ESD immunity if testing to IEC 61000-4-2. The built-in HBM ESD protection is really just for handling in a controlled environment for PCB assembly/component stocking/handling.
Best regards,
Conal
Hi Brian,
As Conal mentions, your exact protection circuit component selections will affect the success of this circuit function. This is why I'd suggest you may want to select a different TVS diode to protect the TxD line.
The ADM3251's Tout pin's ABSMAX is +/-15V. The TVS diode in the schematic (STS232120B301) has a 19-25V clamping voltage depending on the peak transient current (19V @ 1A, 25V @ 12A). I suspect that this TVS's stretch allows the Tout pin to go high enough to trigger the latch up.
Eric
Hi Conal, Eric,
Thanks for the suggestions. We were able to find a solution by populating both R2 & R3 as well as using some capacitors instead of GDTs.