What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

New RFI(?) issue without electrical changes

Acenels

Active Member
This one has me scratching my head.

About 2 years ago I completed an avionics upgrade. I flew a couple hundred hours since then. No issues. On all the new avionics I used a home run forest of tabs insulated from surrounding surfaces and connected to the battery. It was a very quiet system. (Edit: the FOT is connected to the firewall pass through on the passenger side of the grounding strap, and from there a cable goes to the battery)

I took my plane out of service in September for conditional. I rebuilt the airbox, executed a SB for tail cracks, replaces some gaskets, swapped the engine mounts, and complied with all conditional items.

Upon flying again in January, I now have a PTT issue that causes my EI oil temp/pressure sensor to drop to zero when transmitting on about 121-123, and gets interference outside of that but not quite as significant. This hold true on either radio (gtn650 or gtr). This happens in air or on the ground. I get very minor egt/cht interference, too, but only around 5 degrees.

Here are the various things I’ve tried and the result (yes means it still interferes):

PTT from passenger stick: yes, either radio
PTT with radio off: No
Disconnected pilots stick grip completely (tosteen) and use passenger PTT: yes
Audio panel off: yes
Antenna disconnected: no, not for disconnected antenna but does still happen on the other radio
Every piece of electrical equipment off and can pulled: yes

I have run the grounds on the instrument, and replaces a crimped connector. It ohms out good.
I disconnected each antenna and I stuck my fluke on the bnc connectors while transmitting and in that configuration I get interference (with the circuit competed through my fluke).
I “jiggled” the cables including the antenna cables behind the panel while transmitting to see if there was some strange new antenna RFI between the coax and the gauge—no change.
I checked my Forest of tabs for any loose connectors—none found.
Both antennnas are CI-122.

I’m just so puzzled because I wasn’t working on the electronics at all.

Next up is jump the avionics master CB to see if that’s gone bad, but I’m just not understanding what could have gone bad from the plane sitting.

And great ideas? It seems like it must be a ground or similar, but can’t figure it.
 
Last edited:
Radio interference

When transmitting on the VHF radio, it throws out 6-10 W of power in the coax cable to the antenna.
This creates a strong magnetic field around the coax.
Some sensors does not have shielded wires and they typically use very low voltage for the sensors, from 50 mV -14 V.
The lower the sensor voltage is, the more sensitive wiring is for strong magetic fields close to the wires.
If the wires are running in parallel for any lenght it´s getting worse.
There are lots of wires behind the panel, shielding all unshielded wires is a lot of work. I have put a tinned copper tubular braid on the outside of the RG400 coax cable going to the antenna.
It is grounded at the BNC connector behind the radio. Length is about 3 ft.
When the coax no longer is parallel to sensor cables, the braid is not
required.
https://www.westshorecontrols.com/product/ttce-tinned-copper-tubular-braid/

Good luck
 
This one has me scratching my head.

On all the new avionics I used a home run forest of tabs insulated from surrounding surfaces and connected to the battery. It was a very quiet system.

.

While a forest of tabs is a good idea, I am not sure that isolating this from the airframe is. The goal is to provide a low resistance ground path and NOT to isolate from the frame ground. The frame ground I think should be there as a redundant backup as well as to keep everything in harmony. Maybe a EE can weigh in with an opinion, but my understanding is that the dedicate grounds should be on top of the frame ground and not in isolation from the frame ground. How does errant static electricity in the airframe get back to the battery if that battery is not connected to the airframe? Not my area of expertise, but doesn't quite sound right to me.

Larry
 
Last edited:
You want the forest of tabs bonded to the airframe. You're trying to have a single point ground, not isolated.
This is very likely associated with the coax to the antennas. (when you key the radio, a bunch of RF comes out where you don't want it and couples into the sensor wires or acquisition board (and makes an inadvertent AM RF detector))
Don't intentionally transmit into an unmatched load! i.e. your Fluke. That is asking for trouble. (as in burning out the radio tx amplifier)
If you have a VSWR meter, hook it up inline at the radio if possible. A mismatch (bad coax connection) will show up as a high VSWR.

Basically, take a real hard look at all your coax connectors and crimps. Sometimes you can put in a new (quality) test cable, bypassing the installed stuff, for a quick test. I (cough, cough) ran my xpndr like that for years until I finally stood on my head and re-terminated the installed cable connector.
 
Basically, take a real hard look at all your coax connectors and crimps. Sometimes you can put in a new (quality) test cable, bypassing the installed stuff, for a quick test. I (cough, cough) ran my xpndr like that for years until I finally stood on my head and re-terminated the installed cable connector.

This^^^^^. While you’re at it check the braid to bnc connector to antenna connector to airframe ground connections. What you’re describing are symptoms of RF in the cockpit.
 
You want the forest of tabs bonded to the airframe. You're trying to have a single point ground, not isolated.
This is very likely associated with the coax to the antennas. (when you key the radio, a bunch of RF comes out where you don't want it and couples into the sensor wires or acquisition board (and makes an inadvertent AM RF detector))
Don't intentionally transmit into an unmatched load! i.e. your Fluke. That is asking for trouble. (as in burning out the radio tx amplifier)
If you have a VSWR meter, hook it up inline at the radio if possible. A mismatch (bad coax connection) will show up as a high VSWR.

Basically, take a real hard look at all your coax connectors and crimps. Sometimes you can put in a new (quality) test cable, bypassing the installed stuff, for a quick test. I (cough, cough) ran my xpndr like that for years until I finally stood on my head and re-terminated the installed cable connector.

Just to make sure I have it right: this is the co-ax from the radio to the antenna (as the likely culprits)? Is there some way another cable might be a likely spot? I can’t figure it (e.g. the transponder antenna causing this interference when transmitting.

I will give that a shot. It seems weird that they would both act up at the exact same time, but then again I could have not noticed on the number 2 radio as I generally use it as a rx radio or talk in a higher band.

Thank you for the idea.
 
You want the forest of tabs bonded to the airframe. You're trying to have a single point ground, not isolated.

Not sure if this changes the comment but the way it is set up is the FOT is on the sub-panel, and then I run the "big" ground to the battery return.

I did this because when I was researching this the VAF posts said this might be better to avoid electrons following a different path back. If this is wrong, it would be super east to just remove the rubber isolator from the sub panel. I was just worried about it coming back through two separate paths causing a problem. I am no EE by any stretch :) CLEARLY.

I can remove the rubber isolation from the FOT if that would be "better."
 
Yes.
You did use good quality crimped BNCs, correct? (not the screw on type).

Yes, crimped BNC's. I used the "old" ones when I did the Avionics install, though. I didn't replace those as the all looked to be in great shape.

The only one I built myself was a new one for the transponder, also crimped BNC.
 
Ground Bus

It is important to have a single Ground bus bar for all avionics.
The Ground bus should be mounted to the air frame or have a short wire to the
air frame. The Ground bus bar should be connected to the battery negative side. This Ground wire should be sized to carry the full current of all installed avionics.

Good luck
 
Update:

Got a RG400 and BNC tester cable to test.
With all Nav/Comm/GPS antennas removed from the back of the GTN650 and the GTR, I ran the cable straight to the antenna. I tried both antennas and both radios and still had the interference (oil temp dropped to 0 from about 60, the hangar temp).

I then removed the antenna from the airframe and tried it. Interestingly, if I pointed it at a 45 degree angle I got little to no interference when it was grounded to the airframe.

I cleaned the mounting surface and reattached, checked the ground, and tried again (optimistically) and still have the RFI. Oil pressure drops to 0.

In flight the “big” problem is the Oil pressure dropping and I need turn the plane on to do that, so I’ll test that tomorrow in case the oil pressure is now “fixed” from my tinkering and re-crimping the ground to the pressure sensor.

I’m not optimistic, but we’ll see.

Part of me wonders if I should shorten the cables to the oil pressure sensor (they are long and looped in places), but part of me says that makes no sense as they have been like that for 2 years with no issue.

Man, this one has really got me…and I love a good electrical challenge, but I’m over my skis.
 
The RG400 cable works as an ANTENNA (HF) when you transmit on the GTN.
Yes it's shielded but not 100 % of the RF stays inside the RG400.
The oil pressure wires works as a receiving ANTENNA when you transmit, if it's not shielded.
More shielding may be required, either on the outside of the RG400 or the oil pressure wires.

Good luck
 
Grounding and bonding

Looking back at your original posting you noted the things you had worked on before you had the problem. Changing engine mounts was one item that you noted. The engine sensors are generally all grounded to the engine inherent to their construction. I would start by ensuring that the ground strap between engine block and airframe is at least a minimum of 4 AWG ( preferably 2AWG). The crimps on both ends are properly made and the attachments on both the airframe and engine are secure and tight. The “forest of tabs” grounding block is best mounted on the airframe close- or at the point of the engine strap mounting. Without a detailed wiring diagram and a good look at the physical wiring to suggest any other recommended changes than just following best practices would be chasing your tail. You really want a single point where the airframe, engine and electrical wiring can be ground referenced to take measurements and minimize ground loops. That is generally done on the firewall that is convenient to get to for troubleshooting.
Another approach to figuring out the source of your interference is to purchase a 50 ohm dummy load and the necessary adaptors (BNC to TNC etc), remove the Tx coax at the radio, insert the dummy load and do a short transmit and check for interference. If you still have a problem then the problem is likely a conducted interference issue - if the problem goes away then it is likely a radiated RF problem and one step closer to a robust solution.

KT
 
Last edited:
The RG400 cable works as an ANTENNA (HF) when you transmit on the GTN.
Yes it's shielded but not 100 % of the RF stays inside the RG400.
The oil pressure wires works as a receiving ANTENNA when you transmit, if it's not shielded.
More shielding may be required, either on the outside of the RG400 or the oil pressure wires.

Good luck


RG400 is double screened. If the screen is terminated correctly in the BNC or TNC connector at both ends of the RG400 then there should be no measurable radiation or conduction outside of the RG400. The oil pressure sensors generally have two wire connection and recommend a twisted wire configuration. The input to the sensor instrumentation is generally a differential input circuit that has a very low impedance to RF and has a moderately high common mode rejection ratio to deal with generator and other lower frequency conducted radiation interference. If the common mode range is exceeded (as it would appear to be in this case) then the problem is more likely to be ground loops, poor grounding or bonding issues.

KT
 
Ground strap bond

Did you disrupt the ground path from the engine to the airframe when you 'changed the mounts'?
There should be a braided ground strap between the airframe and the engine case, to create and electrical bond.
Your symptoms point to resistance in ground somewhere.

Looks like KiethT were reading this at the same time..
 
If the instrument is an Electronics International, the PT-100GA Pressure Transducer is used. This Transducer has four wires, +5V, GND, signal - and signal +. This indicate that the transducer is isolated from engine GND.
As I understand it, the problem exists only when the radio transmits.
To me this would indicate an interference from the RG400.
This is provided that the wiring from the oil pressure Transducer is located
within 1-5 inches from the RG400.
The Transducer signal is probably a ratiometric one, 0,5 - 4,5V.

Good luck
 
If the instrument is an Electronics International, the PT-100GA Pressure Transducer is used. This Transducer has four wires, +5V, GND, signal - and signal +. This indicate that the transducer is isolated from engine GND.
As I understand it, the problem exists only when the radio transmits.
To me this would indicate an interference from the RG400.
This is provided that the wiring from the oil pressure Transducer is located
within 1-5 inches from the RG400.
The Transducer signal is probably a ratiometric one, 0,5 - 4,5V.

Good luck

If the pressure transducer is an EI PT 1000 type and the display instrument is an EI offering then the likely configuration is a ratiometric differential DC output which is likely referenced at about half the supply voltage to the sensor (5 to 13 volts according to the spec sheet). The likely instrument input would also be differential ratiometric too with a very low bandwidth to suppress any interference picked up between the sensor and display instrument. The wiring configuration for power and ground between the instrument and sensor is important to minimize noise and conducted interference. Power and ground connections for the sensor should be directly connected to the display instrument head. Grounding the sensor at any point other than that is inviting a problem.
Just to get the order of magnitude of RF voltages for a 10 Watt transmitter into perspective. Power is voltage squared divided by resistance (assume we have a VSWR of 1 so are purely resistive). Resistance is 50 ohms and 22.36 vrms will give a power of 10 watts (peak voltage of 31.6 volts). Peak current is 0.63 amps. The voltages are between the coax center conductor and the screen and the currents are flowing in the outer screen (which is the area of interest). The coax screen will have a resistance of milliohms at most so the voltage developed over a length of coax outer screen is millivolts at most and more likely microvolts.
If you are using RG400 and have correctly terminated the connectors then the problem is not likely to be in coupling between the coax and sensor wiring but more likely to be in inadequate general grounding and bonding or incorrect wiring connections.
Trying to troubleshoot some of these electrical/electronic/RF problems can be very challenging and frustrating especial without access to the specialized test equipment (even for those of us that have spent a professional career doing this type of work). The best advice is follow the guidelines laid out in the industry standards and manufacturers installation documentation.

KT
 
Two antennas?
If both radios are doing this, there has to be a common issue. Power return, ground, etc.
Not to be a wise guy, but are the radios fully seated in the trays? (Were they removed recently?). Bezels hanging up on the panel?
Start with the basics. I learned that lesson recently!
 
Two antennas?
If both radios are doing this, there has to be a common issue. Power return, ground, etc.
Not to be a wise guy, but are the radios fully seated in the trays? (Were they removed recently?). Bezels hanging up on the panel?
Start with the basics. I learned that lesson recently!

Antenna grd bonded to airframe and radio grd bonded to battery neg and isolated from fame grd. hummm. Doesn't sound right to me. I would first emove the isolation from grd bus to airframe. Possible that ground path is through sensor ground wires and messing things up. Nothing wrong with a grd bus, but it should be connected to a frame ground also.
 
Last edited:
Wow--I really appreciate all these responses. Super helpful stuff, folks. Hopefully it isn't too much if I pull out a few below with some responses and things I've since tried. I've been heading straight from work to the hangar everyday, so trying to get this sucker figured out!

Quick note on the grounding: I think I complicated this by describing my ground point in a poor way; all the avionics are to a forest of Tabs, and that is connected directly (via 8ga) to the firewall at the point the grounding strap comes through from the engine. There is then a cable from there to the negative on the battery. I didn't ground the FOT through the airframe where it is mounted (subpanel) as I didn't want to risk any weirdness as the subpanel was painted and primed and the touchpoints were not super well grounded. The FOT is wired via 8ga directly to the firewall passthrough for the engine grounding strap.

New things I have tried: I disconnected the instrument completely from the plane's power source and ran wires straight from an external battery into the connector via a new CPC connector I fashioned. No Change.

I then removed the instrument completely from the panel and tried running it with the oil sensor wires (4 wires) disconnected from the connector. I ran all these "tests" referencing the Oil Temp gauge is a bit easier to test in the hangar with the plane off :)

So now I have it completely decoupled the instrument from any Aircraft avionics, which I believe narrows it down as the problem persisted :(

No difference.

In this configuration it allowed me to also physically move the instrument around in he cockpit to check for variations based on position, and vary it did, from dropping 20 degrees (60 to 40) to dropping to 0.

It impacted by BOTH moving the instrument position and by moving the the position of the cables.

I also took aluminum foil and wrapped the wires the whole way back to the firewall and wrapped the instrument (OPT-1). That helped a bit. Instead from ~60 degrees to 0, it went from 60 degrees to between 25 and 50 degrees, depending on the position I was holding the instrument and which radio I was transmitting on. A little better! And probably more in line with where it was. I watched my test flight videos from my avionics upgrade and realized that unbeknownst to me, it WAS dropping about 20 degrees, and maybe 5 PSI, I just never noticed....so it was impacted, just not as much as it is now. Wild that I never noticed. Glad I have that cockpit video! It certainly wasn't dropping like it has been, though, so there is a change of some type.

Strangely, at one point I had the foil-covered cable in an "S" where it overlapped itself and the temperature actually went up instead of down.

Removing the aluminum shielding from the instrument itself results in the temp dropping an additional 10 degrees compared to when I have it wrapped in aluminum foil.

I have what I believe to be quality antennas (CI122) so I am assuming that is unlikely to be the culprit? Is it still worth a VSWR check? I am due for a new tool...
Can antennas go bad in some way?

Here is something weird, that may be completely inconsequential: If transmitting on radio 2 to the "right" antenna, and I remove the radio 1 (left) antenna's BNC at the antenna, the interference is ever so slightly reduced (about a 3-4 degree swing). The opposite happens if I do the opposite (transmit on radio 1 and remove the BNC on radio 2). This may just be something with the field, so I don't want to over think it...but...I think I'm beyond overthinking things here.

Twisting wire questions: Does it make sense to twist the K20 thrmorcouple wired with the 4 Oil Pressure sensor cables? Currently, none are twisted groups.

My next step plans are
1) To shield the instrument with faraday tape
2) Shield the oil pressure sensor wires and oil temp thermocouple and with tin shielding up to the firewall. (twist them?)
3) Replace the current unknown type coax with RG400 and fresh BNCs.
4) Run a VSWR test on the antennas
5) Say a little prayer to the RFI Gods.





This^^^^^. While you’re at it check the braid to bnc connector to antenna connector to airframe ground connections. What you’re describing are symptoms of RF in the cockpit.

I think I am about to ask a dumb question, but here we go: The antenna to airframe ground should be accomplished from the antenna touching bare metal under the mount, and through the screws to the doubler, right? I ohmed those good, and took of the antennas and cleaned the surfaces. Making sure I am not missing something here.




Did you disrupt the ground path from the engine to the airframe when you 'changed the mounts'?
There should be a braided ground strap between the airframe and the engine case, to create and electrical bond.
Your symptoms point to resistance in ground somewhere.

Looks like KiethT were reading this at the same time..

Good call. I removed the lower cowl and checked this. All is still good there. I have a 4ga grounding strap from the engine tot he firewall.


RG400 is double screened. If the screen is terminated correctly in the BNC or TNC connector at both ends of the RG400 then there should be no measurable radiation or conduction outside of the RG400. The oil pressure sensors generally have two wire connection and recommend a twisted wire configuration. The input to the sensor instrumentation is generally a differential input circuit that has a very low impedance to RF and has a moderately high common mode rejection ratio to deal with generator and other lower frequency conducted radiation interference. If the common mode range is exceeded (as it would appear to be in this case) then the problem is more likely to be ground loops, poor grounding or bonding issues.

KT

I tested with RG400, but the coax in the plane is not RG400. I am thinking I should rebuild these connectors (maybe?) with RG400 just as a better practice.

The Oil Pressure sensor is 4 wires, and they were not twisted. However....tonight I tried twisting them, but that didn't appear to make an impact, but ask above if I shoudl still twist them as a best practice.


If the instrument is an Electronics International, the PT-100GA Pressure Transducer is used. This Transducer has four wires, +5V, GND, signal - and signal +. This indicate that the transducer is isolated from engine GND.
As I understand it, the problem exists only when the radio transmits.
To me this would indicate an interference from the RG400.
This is provided that the wiring from the oil pressure Transducer is located
within 1-5 inches from the RG400.
The Transducer signal is probably a ratiometric one, 0,5 - 4,5V.

Good luck

The coax is not RG400 currently (it is black coax, I wasn't the builder so not sure on the type...but I am going to redo it in RG400 with good BNC connectors. Order placed tonight--haha!

Two antennas?
If both radios are doing this, there has to be a common issue. Power return, ground, etc.
Not to be a wise guy, but are the radios fully seated in the trays? (Were they removed recently?). Bezels hanging up on the panel?
Start with the basics. I learned that lesson recently!

I'll double check this tonight. They haven't been removed, but can't hurt to check this!
 
Quick update for anyone interested, or for a future person with the same challenges...

I re-did all cables with RG400 (previously RG58). I did a SWR test, both antennas were under 1.5, even with the CI-122 whip antennas on the belly.

I twisted the oil pressure sensors. I also added shielding to both the temp and the pressure sensor cables, and grounded the shielding. I added faraday tape to the OPT-1 sensor itself, as well as the cables from the instrument to the circular connector.

It was only deviating about 5 degrees down on the temp (definitely acceptable), so I reinstalled it.

HOWEVER....then when I changed the output frequency to the higher 130s, instead of dropping, it increased about 60 (yes, 6-0) degrees, from ambient of 50 to as high as 109. YIKES!

So I grabbed another OPT1 from another plane (my late father in law's) and installed it. Guess what? NO issues.

So after all of this, it was in fact the instrument. It must be some soft of noise reduction failure or something. This is apparently a very rare failure, but I also learned the instrument, which I purchased on the secondary market, is nearing antique status.

The instrument I tested with was built in 2004, per EI. Mine is much older than that. I don't know the born on date (I didn't ask), but the one I tested with was S/N 75527, vs mine is S/N 21675 as a reference point. Mine actually uses the older, pre PT100GA pressure sensor natively, and requires an interface to work with the PT100GA.

The solution? I called EI, who is ALWAYS great to work with, and they are mailing a new one at the repair cost. What a great company.

I'll install it mid week here and report back, but I suspect this will fix all the issues, and now I have a meticulously shielded, grounded set of wires under the instrument panel :)
 
Annnnd new instrument sent from the always-friendly folks at EI, and there is zero deviation on the oil pressure and the temp.

Problem solved!

Occam’s razor after all 🤓
 
Back
Top