I know that. I have a lot of experience with these things.
50 year career in this type of design. Lumped constant filters are not
Missing, just seem to not be that effective in knocking down very LARGE
Current spikes from high voltage high current inverter. Working outdoors
In Florida this time of year is no picnic with the heat and humidity.
I don???t want to drag out the expensive test equipment into the field just
Yet.
Tom Mills:
Yes that could also be an issue. I had to do that with the old system.
It used a conventional capacitor start motor for the compressor. These
Are sealed units. The new system uses 3 phase inverter driven compressor
Motor also sealed unit. The ECM motor in my old system was pretty quiet
But being the picky person I am, I decided to add the additional filtering
Right at the air handler blower motor, and also added shielding for the
Thermostat connection from the air handler to the thermostat just below
It in the house.
The longer run out to the compressor used un-shielded wiring and I put
A type 31 toroid there just as a safety measure. Overkill is wonderful.
No noise with the old system.
Now with the new so called ???state of the art??? system the noise sources
Are much different. Eventually I will get this solved. Just takes a bit
Of time. Especially since I know where the source is.
Karin Anne Johnson P.E. K3UU
Palm Harbor, FL
_____
From: David Eckhardt [mailto:davearea51a@gmail.com]
Sent: Saturday, September 05, 2020 3:57 PM
To: Karin Johnson
Cc: Rfi List
Subject: Re: [RFI] Rheem AC RFI
You don't look at the 6.5 kHz, specifically! Look for the noise on
40-meters or wherever the problem is occurring. Once you find that noise,
look for tyne spacing of 6.5 kHz which should be characteristic of the
advertised switching frequency. This is typical of switchers. From
experience as an EMC/RFI engineer, when a common mode choke is left out of
the AC line input (a common practice from China and other CM's), I usually
see a large hump of broadband noise between 40 and 70 MHz. Looking at that
closer, I usually am able to pick out the switching frequency by observing
the tyne spacing. Do the same for your interference. First find the
broadband noise at the same frequency you are receiving it on the spectrum
analyzer. Then look closer-in where the noise is strongest to observe the
tyne spacing. You will likely need narrower RBW and span to view the tynes
once you find the noise at frequency.
Dave - W??LEV
On Sat, Sep 5, 2020 at 6:41 PM Karin Johnson <karinann@tampabay.rr.com>
wrote:
David:
Thank you for the reply.
My spectrum analyzer is a Tektronix 7L12. Although the lower
Frequency limit is DC, it is difficult to resolve meaningful signals
At very low frequencies due to the Zero marker leak thru masking
Any signals.
I will say that at HF frequencies above 3 MHz below 8 MHz is where most
Of the RFI is showing up. Using the spectrum analyzer in zero span mode
There is a very complex switching waveform which is hard to separate out
any 6500 Hz components. Viewing the spectrum it occupies the above
frequency
range. Below about 3 MHz there is very little energy showing. The AM
broadcast
Band is not affected at all.
Using some basic DF techniques I am pretty sure that the source is the
inverter
driving the compressor motor.
Just to clarify. The COMPRESSOR motor is a three phase motor as shown in
the
Rheem documentation. It is not self contained ECM motor, nor is it a
capacitor
Start type motor. It is truly a three phase motor. I have not had the
external
Metal surrounding all of the outside electronics off just yet, as I don't
want to
Go too far down the rabbit hole with regard to voiding any warranty.
I thought posting to this list might show another user with similar issues
with a
similar unit. I suspect there will be several solutions applied in the
final accounting
to solve this issue.
Right now the installers used the original un-shielded control wiring from
the inside
Thermostat to the outside electronics. One part of the solution may be
having to
Replace the un-shielded control wiring with shielded control wiring.
This system is sort of a state of the art control system, as it uses what
Rheem calls
ECONET. The thermostat does not send a solid control signal generated by a
closed
Or open relay in the thermostat. It uses a binary data stream coming from
the Thermostat
to the outside compressor unit and also the Air handler in the attic. Sort
of like a CAN bus
communication system. I'm not sure what the protocol and physical layer
component are.
I am speculating at this point, but this may be as simple as re-routing some
of the internal
Wiring in the outside unit to prevent coupling of the high power switching
components from
being coupled on to low voltage wiring. One cannot simply put a low pass
filter on the data
Lines coming from the thermostat as it would kill the waveform logic high
and low.
Common mode choke would be good here. When they installed the new unit
they
Took off a common mode choke I had installed on the old thermostat control
wiring at the
outside unit. The installers did not leave enough slack in the control
cable to install another
Choke.
Plus one other horrible thought. If the high voltage inverter switching
signals are coupling to
The actual high side and low side Freon copper piping that presents a
difficult problem to solve.
Right now I'm just familiarizing myself with the system components. So that
I can drag
The scope and current probes outside and do some more measurements.
Karin Anne Johnson P.E. K3UU
Palm Harbor, FL
_______________________________________________
RFI mailing list
RFI@contesting.com
http://lists.contesting.com/mailman/listinfo/rfi