Monday, April 20, 2009

Common-Mode Current Choke (Ver 2)

There are two major changes in Version 2! The first change is purely a mechanical one- no enclosure anymore - the choke is not as compact as it is in the original design (Ver 1). It is not a "one-box" device but rather an in-line jumper-like device. Before, the enclosure was keeping the coaxial from damage by garden critters and the brittle ferrite rings from mechanical damage. Now, with the open design, one should be more careful while handling/installing the choke. On the bright side - there is one less M/F Type-N connector pair (less insertion loss and one less connection to waterproof) - now the balun has one Male and one Female connector on each side, oppose to the two female bulk-head connectors mounted on the box. Because the antenna side pigtail (/w Male Type-N connector) is longer, there is no need for the short Male-to-Male jumper to the antenna. There are only two connections to waterproof. The open construction also allows for better cooling during Legal Limit power levels.

The second change is electrical - since the space is not limited anymore, now the choke has more ferrite rings (24 pcs vs 18 in the old one) and less coaxial cable turns (2.5 turns in the new version vs 3.5 turns in the old) through the binocular core. This improves the upper frequency impedance (speaking of which, I have a couple more mix 43 single turn ferrite beads on the coax pigtails, for a total of 3)!  More ferrite rings in the core - more inductance (to make up for the lower number of turns). Less turns on the other hand means a decreased capacitive coupling between the turns on higher frequencies. My MFJ-259B is limited in measuring the complex impedance (Z) up to 650 Ohms. The choke was measuring >650 Ohms up to about 20 Mhz. The impedance is around 400-500 Ohms at 30 MHz, mainly due to the capacitive coupling between the turns. I have installed an additional plastic "jacket" over the coaxial in attempt to increase the space/dielectric between the turns and decrease the capacitive coupling but can't really tell if this has any real effect! I am using the same LMR-240 coaxial (RG-8X type with a solid center conductor). The choke should handle well power levels up to 1.5 kW but I am not planing to use it with more than 1 - 1.1 kW anyway (ACOM 1000 amplifier).
For steady 1.5 kW use or with antennas, exhibiting very high SWR, I'll recommend the use of high-temp cable such as RG-142 (silver/teflon) for the balun. This cable has a solid teflon center insulator and there is no danger for the center conductor to migrate to the shield and short it, if substantial heating occurs - something that is possible with the polyethylene-foam based center insulator of the RG-8X / LMR-240.
The ferrite toroids  I used are made by FairRite (Part# 2643801002) - mix 43, OD 28.5mm, ID 18.5 mm, H 7.5mm. Price is about $6.10 for 10 pcs - Mouser Part #: 623-2643801002.

The new Common-Mode Current Choke takes a little more linear space but this is not a big deal. I am planing to put extra mechanical support (another plastic stake) for the choke itself.
Update: I tested the current balun with my VNA and here are some plots:

Insertion loss on HF <0.1 dB (due to connectors and coax cable insertion loss).

Complex impedance (Z) for Common-Mode current is 2200 - 2500 Ohms on the 80m -40m bands and drops down to ~300 Ohms on 6 meters (just as expected). On the 20m band chocking is still > 1kOhm.
Overall, for a current balun of this size and cost, the result is excellent and it completely serves the purpose. I wanted to have it primarily for the lower bands - 80m and 40m but get some good values on 20m as well. If mix #31 is used for the ferrite,  the optimization will be even better in the lower HF spectrum.
2-3 kOhm Chocking impedance is completely adequate for a tunable antenna like BigIR where the SWR never exceed 2:1 (given that the number of radials is sufficient and they are properly sized). I actually get SWR of 1.1:1 to 1.2:1 on 80 meters.
The VNA test just gives me a nice confirmation that I have achieved my design goal when building the choke.
If more chocking impedance is needed, more of these in series can be built on the same piece of coax and folded to save linear space.

A transmission scan shows attenuation >30 dB on the lower bands but this is a less relevant test to a real-world application as the measurement of attenuation is done in a true 50 ohm system. In reality, the antenna feedpoint (Z) is never exactly 50 ohms.

Sunday, April 12, 2009

Common-Mode Current Choke Disaster!

It is Spirng time and Spring means rain - lots of rain! Yesterday I wanted to use my SteppIR vertical antenna after a couple of months rest (I've been using my dipoles meanwhile). When I did, the SWR was very high on all bands regardless of the SteppIR tuning. It was 4.3:1 even when the element was completely retracted and the same thing with completely extended element - 40/80 meters tuning - a sure sign of trouble. No matter what I was doing the SWR was not changing and very strong signals were extremely attenuated - about 1/3 of what the dipole was receiving! MFJ-259B was showing the same thing - the impedance was changing a bit on different bands but the SWR stayed very high. Next thing to try was the DVM - aha!!! - DC resistance between ground and antenna was 6-7 kOhm and constantly changing up and down (ideally it should be just a few ohms of DC feedline resistance as the 80m coil has a built-in balun showing as 0 ohm for DC)! After going to the antenna site, everything looked normal until I started the troubleshooting by inspecting first the common-mode current choke/balun in the base of the antenna!
Here is what I have found upon removing the cover of the balun enclosure! Totally flooded - just missing a few toads inside! Not only that! Note the left antenna connector, between the two top screws of the flange connector, where the coaxial is soldered to the center pin receptacle of the female N connector - the pin receptacle, together with the white dielectric insulator (visible between the bolts) were (!) pulled out of the N connector housing and were almost entirely inside the enclosure! As a matter of fact - they were so far out of the connector's barrel - the center pin of the male antenna connector was not making contact - not even close! The solder connection of the coaxial shield was broken off too! I have no idea what happened there but looks like some serious force was in play - I can't do such a thing with my fingers and even with a tool will be a chore! One speculation is ice. In a solid block of ice this might be possible during freezing / melting cycles during the winter but none of  the ferrite toroids were broken! Another possibility is a miniature steam explosion. When I used the antenna last time, was with 1kW according to my logbook, water inside the tightly coupled male-female N connectors could have vaporized and the resulting steam could have pushed out the dielectric inside the box and ripped off the shield's solder joint- the middle insulator is the only part that can easily relief such pressure. Who knows... Bottom line - the whole balun is badly damaged and I'll be making Version 2 once the parts arrive!
The reason for this disaster is very simple - having too much faith in labels such as "Weather resistant electrical box" (aka the balun enclosure) from Home Depot. By "Weather resistant" they actually mean - "water resistant", forgetting that Sun exposure is part of the "Weather". The seal around the cover failed. It is a ring type gasket and either the gasket material changed over time or it was no good in first place. After removing the gasket, the seemingly dry gasket produced a lot of water just by squeezing it with my fingers. Seems to me that the gasket is made of some sort rubberized open-cell foam material (?!? huh?) and it was SOAKED with water. It is possible that the UV exposure broke down the rubberized coating and changed the properties of the material - the gasket no longer functioned as gasket but as a wick letting water to fill the box.
Looks like Carlon - the manufacturer of the so-called "Weather Resistant Electrical Box" needs to learn more about gaskets and materials or revise the description. (Needles to say - the cover was screwed very tightly so the only thing really to blame is the gasket!). There was also a very small discoloration and deformation of the cover (due to the sun exposure) but that was well within the tolerance of the gasket. I should have removed the original gasket when I built this choke and use some silicon sealant instead but back then I didn't even think they will sell electrical boxes with such poor choice of gasket material - now I know - they do! Lesson learned!
Construction notes for the original choke (Ver. 1.0) are on my antenna site. Before the damage, the choke worked just great! The new one, I am planing to build will be of open type design, with no enclosure.
If anyone decides to built this version - my recommendation is to have a small drain hole on the bottom or use a bead of black RTV sealant around the edge of the gasket, once the box is closed to protect the material from UV damage.

Thursday, April 2, 2009

Hamshack RF Signal Path Diagram

I've updated the diagram to reflect the latest changes in the RF paths / equipment. The main change is the way the output power is measured. In the new configuration I am using one watt-meter for output-to-antenna power and another watt-meter for the output-to-dummy load power. In the old configuration, each amplifier had its own dedicated watt-meter for the output. In addition, the main high-power dummy load is now an oil-filled Bird 8201.