Hi all,
You may have seen my original thread "
5uH Aerospace LISN: How dumb would I be to "throw one together"?".
With the most excellent help of a few forum users, we came up with a decent design for an RTCA/DO-160 LISN. For those of you unfamiliar with the standard, it is non-compulsory standard used by the aerospace industry. Avionics module manufacturers may certify their equipment to DO-160 to prove some level of flight-worthiness. While (historically) most aerospace/military companies seem *ahem* okay with spending lots of money, if you are not in that situation, or if aerospace is not your main industry, it may be hard to justify buying all of the standard EMC equipment. The attached design costs around $150USD. Commercial designs are on the order of $2,000USD.
Disclaimers:To be clear, this project is aimed at *PRE-COMPLIANCE TESTING ONLY*. Please do not use this LISN to perform final self-certification of any flight-bound equipment. If you do, you assume all responsibility. In addition, this LISN can potentially be used with mains voltage. Please exercise all usual cautions. It is of utmost importance that the housing is firmly tied to earth to prevent electric shock. Perhaps more importantly, power should be applied to LISN through a galvanically isolated source. A warning label is included in the attached zip that you should print and fix to the housing. I used makestickers.com for a proper durable, laminated label.
Design:
The LISN is built onto a single board designed around the Bud AN-1322-A housing. I chose this housing because it is a reasonable size and has many threaded holes for mounting a PCB. Since a low impedance earth connection is important, this housing is especially good with 12 mounting holes. 3x3.3uF capacitors are used to form the 10uF capacitance and 4x1.2uH inductors are used to form the 5uH inductance. The tolerance of these elements is not incredibly important, as the tolerance of the LISN's impedance (vs. frequency) is actually pretty large.
The 3.3uF capacitors are X2 rated capacitors. In an ideal world, these capacitors would be Y2 rated, since they are connected between line/neutral and earth. However, the size and cost of Y2 capacitors (of these values) is quite prohibitive. They can be anywhere from 3-10x the size of X2, and X2 are already large compared to "normal" film capacitors. Since any mains LISN is a potentially dangerous piece of equipment (due to very high earth leakage current), the rationale against using Y2 capacitors is that the user must already be a skilled operator, and must understand the risks involved with using the equipment.
The 5uH inductance is composed of 4x1.2uH Coilcraft SER2011 inductors in series. Each inductor has a self-resonant frequency over 80MHz, which allows the LISN to maintain good high frequency performance up to the 400MHz requirement. (There is some fluctuation in impedance above 100MHz, but it is insignificant and well within the limits, when measured with a 10dBm signal from a VNA.)
RTCA/DO-160 does not measure EMI by direct voltage measurement (as most EMC standards require), but instead requires the use of an RF current monitoring probe that measures CM currents on the input cable to the DUT. However, an equivalent measurement can be made with the standard voltage measurement, though some adjustments need to be made. (For example, the current monitoring probe measures common-mode current, while a direct voltage measurement measures both differential and common-mode signals. A DM rejection network would be needed.) For this reason, I included SMA connectors on the PCB, but chose not to bring them out to the front panel. Instead, I capped them with 50-ohm terminators. I used Minicircuits ANNE-50+.
Measurement:The S11 impedance was measured using a NanoVNA V2.2, plotting to its PC software. An additional SMA connector was soldered to the PCB to move the reference plane as close as possible to the circuit. Once the circuit is used in conjunction with actual DUT cabling, the effective measured impedance can change significantly, particularly at the high end, but this is expected. I used 10dBm output power to conduct the measurement, which is obviously a considerably lower test signal when compared to true operational currents. Unfortunately, I have not yet had a chance to run a test with a high bias current through the inductors. I will plan to do that as soon as I can. A high bias current may cause some resonant peaking to show up in the impedance plot. In theory, there may be enough parallel loss (with ferrite cores) to keep the peaking relatively controlled, but it cannot be said for certain. If there's an issue, the PCB includes several un-populated resistor and capacitor footprints that may be used to provide damping around the problem frequencies.
Thanks!
Tim
P.S. Regarding the 3.3uF X2 capacitors, unfortunately, as of September 2020, stock of these capacitors is depleted on Digikey and Mouser. There was lots of stock a few months ago. These capacitors are rather downsized compared to some other 3.3uF X2 capacitors, so there may be some difficulty finding direct replacements. Some options are to: A.) Use non-X2 capacitors, but accept that "standard" film types may be less robust or B.) Move the capacitors to the external binding post terminals, using capacitors of your choice. But, be aware that these capacitors will not be protected by the internal fuses and that mains voltage will likely be exposed.
Special thanks to: Jay_Diddy_B,
T3sl4co1l, and
Pitrsek
Home
Help
Search
Login
Register
