I had a 6KW solar energy system installed by Sungevity back in June. This
system is based on LG solar panels, SolarEdge Central Inverter and SolarEdge
Power Optimizers. These are the primary components of the system.
Unfortunately I missed the article by Tony Brock-Fisher, K1KP, in April QST
and quickly came to regret this omission as my system started producing
copious amounts of RFI even before installation was complete. Tony
described the RFI mitigation upgrade to his solar array system which is
based on the same SolarEdge components.
I did review the article at this point and quickly determined that I needed
to either fix the system or remove it if I was to continue daytime ham radio
activities.
Fortunately Sungevity agreed with the first option and we embarked on a
joint effort to implement K1KP's approach.
I met with Tony, a fellow member and president of the Yankee Clipper Contest
Club, and we reviewed how he went about mitigating his RFI problem on his
similar set of hardware. He does an excellent job of describing the
shortcomings of the SolarEdge based design as implemented by SunGevity in
the article. He and Chuck Counselman, W1HIS, did an excellent job of
researching the problem and designing a solution.
I did not feel it necessary to repeat the analysis part of Chuck and Tony's
effort but went directly into the design effort for my system with Tony's
assistance.
My system has two strings of 11 solar panel/power optimizer sets for a total
of 22 sets plus one Central Inverter.
Once I had obtained all of the line filter and ferrite core hardware, I met
with the local Sungevity field manager and installation team. I explained
to them what the problem was and how we were going to mitigate it.
The installation team showed up a few days later and spent 2 days
implementing the upgrades to the originally installed system. There was a
certain amount of learning curve to be climbed and inadvertently introduced
errors to be found and corrected.
The results are amazing!
The initial check was with the Central Inverter off (it must have 240 VAC
from the utility feed to sense utility voltage, frequency and phase to
operate). This meant that only the solar panels and their associated power
optimizers were active.
Before implementing the RFI mitigation upgrade, the power optimizers (which
are, in effect, switching mode power supplies) would produce strong signals
at multiples of the switching frequencies during daylight. This resulted in
numerous large and messy noise spikes across all of the HF ham bands roughly
every 20 KHz. These spikes were, of course, absent at night.
After implementing the RFI mitigation upgrade, the noise spikes have
disappeared to the point of not being detectable in the ambient noise level
here. As I noted previously, these results are amazing!
In order to configure and test the Central Inverter with the line filter
module and ferrite cores that we introduced, the installation team
temporarily installed a dummy production meter which connects the Central
Inverter to the utility line through the line filter complex.
Limited system operation with the Central Inverter active before the RFI
mitigation upgrade did show strong noise spikes at roughly the same
frequency as reported by Tony. These were spaced roughly every 200 KHz but
were difficult for me to discriminate against the Power Optimizer noise
spikes.
Little time was available to investigate the effects of the line filter
complex on Central Inverter noise spike production after the RFI mitigation
upgrade was complete but the previously obvious Central Inverter
contribution to the noise environment was substantially reduced.
At this point in time, the final step in bringing the system to full
operational status is installation of the production meter by my utility,
National Grid. This should take place in a few days.
At this point in time, the solar panels and power optimizers are active
during daylight hours and I am not seeing any of the Power Optimizer
produced noise spikes that were present before the RFI mitigation upgrade.
I am declaring the RFI mitigation upgrade a conditional success.
As far as how I implemented the upgrade, Tony's fine article provides the
bulk of the detail and rationale.
Here is a brief summary:
1. For each Solar Panel we wound its output wires through a 2.4" OD ferrite
core and added a new choke wound on a 2.0" ferrite core before connecting to
its Power Optimizer. Pairs of wires were twisted.
2. For each Power Optimizer we wound its output wires through a 2.4" OD
ferrite core.
3. The DC loop connecting all Power Optimizers in each string was
reconfigured in accordance with Figure 2 in Tony's article. Pairs of wires
were twisted and a 2.4" ferrite core was inserted between Power Optimizers.
Rather than being connected in a long daisy chain forming a loop, the
connections between Power Optimizers were staggered. Figure 2 in the
article shows this arrangement.
4. A filter provided by SolarEdge was installed at the point where wiring
from each string entered the transition from rooftop to side of the house.
We assume that these filters are differential mode filters and cannot
characterize their contribution to overall RFI reduction.
5. My original installation had the wires leading from the rooftop to the
Central Inverter running through EMT conduit so no change was required.
6. At ground level, the Central Inverter was relocated to provide wall
space for mounting the electrical boxes holding the Corcom line filter and
the array of 30 2.0" ferrite cores.
A total of 52 2.0" OD and 70 2.4" OD ferrite cores were installed. Should
anyone walk by my house with a magnetic compass, it will make a shift in the
direction it is pointing.
I used one of my Elecraft K3 radios with its associated P3 panadaptor to
capture 50 KHz wide sections of the HF bands under various operational
states of the solar energy system including day, night, solar panels
installed, solar panels removed, etc. The P3 was calibrated with my
Elecraft XG-50 signal generator and no settings were changed during this
period to ensure reliable comparisons between before and after. This setup
is not a precision test rig but certainly gives me a decent feel for what is
going on.
I will be happy to share any and all screen captures, photographs and
details of the implementation of the RFI mitigation upgrade to my solar
array system.
And I want to express my appreciation for the support provided by Sungevity.
Jim Spears
N1NK
Tiverton, Rhode Island
N1NK@cox.net
_______________________________________________
RFI mailing list
RFI@contesting.com
http://lists.contesting.com/mailman/listinfo/rfi
|