Not your normal CFL noise

[IndyMike]

Hello Forum members,

I ran across an interesting CFL noise problem last weekend and thought it might help others who are troubleshooting their systems.

 

Setup:

1) ISY 99 running 2.7.7

2) Passive coupler 12' from main panel.

3) PLM plugged into phase A (next to coupler).

4) 1 Accesspoint (RF transceving - all coupling is performed with the passive coupler).

 

Scene test (all devices off)

[GRP-RX ] 02 61 30 13 00 06

[iNST-SRX ] 02 50 09.8C.CD 0C.A8.B4 61 13 30 LTOFFRR(30)

[standard-Cleanup Ack][09.8C.CD-->ISY/PLM Group=0] Max Hops=1, Hops Left=0

[iNST-SRX ] 02 50 0A.E5.EC 0C.A8.B4 61 13 30 LTOFFRR(30)

[standard-Cleanup Ack][0A.E5.EC-->ISY/PLM Group=0] Max Hops=1, Hops Left=0

Edited for brevity

[standard-Cleanup Ack][03.82.89-->ISY/PLM Group=0] Max Hops=1, Hops Left=0

 

----- Link Test Test Results -----

[succeeded] Foyer (3 C5 2B 1]

[succeeded] Mud KPL 3 - 1st Floor (9 8C CD 3)

[succeeded] Mud KPL 4 - 2nd Floor (9 8C CD 4)

[succeeded] Fam Fan (5 55 B2 1)

[succeeded] Mud KPL 1 - Entry Garage (9 8C CD 1)

[succeeded] Master Fan (5 57 66 1)

[succeeded] Foyer KPL 5 - Kitchen Cans (A E5 EC 5)

[succeeded] Foyer KPL 6 - Dinnette (A E5 EC 6)

[succeeded] Foyer KPL 3 - Foyer (A E5 EC 3)

[succeeded] Foyer KPL 4 - Bar Cans (A E5 EC 4)

[succeeded] Foyer KPL 1 - Bar Lamp (A E5 EC 1)

[Failed] Family Room Lamplinc (1 42 D4 1)

[succeeded] Family Room LampLinc (1 42 DC 1)

[succeeded] Matt Switch (3 82 89 1)

[succeeded] Master KPL 2 - Matt (A DB B 2)

[succeeded] Master KPL 1 - Master Overhea (A DB B 1)

[succeeded] Master KPL 4 - David (A DB B 4)

[succeeded] Master KPL 3 - Kath (A DB B 3)

[succeeded] Master KPL 5 - 1st Floor (A DB B 5)

[succeeded] Master KPL 7 - Flood (A DB B 7)

[succeeded] Entry Patio (C 7A 38 1)

[succeeded] BSMT Lamplinc (F 82 A9 1)

[succeeded] David Switch (3 7C 4C 1)

[succeeded] Fireplace Spots (0 35 8A 1)

[succeeded] Fam KPL 1 - Wall (A A9 A5 1)

[succeeded] Entry Porch (3 42 48 1)

[succeeded] Fam KPL 2 - Fireplace Spots (A A9 A5 2)

[succeeded] Bar Cans (3 5F 77 1)

[succeeded] Kitchen Cans (D 4 E2 1)

[succeeded] Kath Icon (3 D3 46 1)

[succeeded] Fam KPL 4 - Fan (A A9 A5 4)

[succeeded] Fam KPL 3 - Lamps (A A9 A5 3)

[succeeded] Fam KPL 6 - Entry (A A9 A5 6)

[succeeded] Fam KPL 5 - 1st Floor (A A9 A5 5)

[succeeded] Entry Deck (C 77 D2 1)

----- Link Test Test Results -----

 

Note: In the above, the Family Room LL failed because the attached lamp was switched off.

 

 

Scene Test with Fam Room Dimmable CFL's on (KPL)

 

[GRP-RX ] 02 61 30 13 00 06

Note: this was not edited. I received 0 responses

----- Link Test Test Results -----

[Failed] Foyer (3 C5 2B 1)

[Failed] Mud KPL 3 - 1st Floor (9 8C CD 3)

[Failed] Mud KPL 4 - 2nd Floor (9 8C CD 4)

[Failed] Fam Fan (5 55 B2 1)

[Failed] Mud KPL 1 - Entry Garage (9 8C CD 1)

[Failed] Master Fan (5 57 66 1)

[Failed] Foyer KPL 5 - Kitchen Cans (A E5 EC 5)

[Failed] Foyer KPL 6 - Dinnette (A E5 EC 6)

[Failed] Foyer KPL 3 - Foyer (A E5 EC 3)

[Failed] Foyer KPL 4 - Bar Cans (A E5 EC 4)

[Failed] Foyer KPL 1 - Bar Lamp (A E5 EC 1)

[Failed] Family Room Lamplinc (1 42 D4 1)

[Failed] Family Room LampLinc (1 42 DC 1)

[Failed] Matt Switch (3 82 89 1)

[Failed] Master KPL 2 - Matt (A DB B 2)

[Failed] Master KPL 1 - Master Overhea (A DB B 1)

[Failed] Master KPL 4 - David (A DB B 4)

[Failed] Master KPL 3 - Kath (A DB B 3)

[Failed] Master KPL 5 - 1st Floor (A DB B 5)

[Failed] Master KPL 7 - Flood (A DB B 7)

[Failed] Entry Patio (C 7A 38 1)

[Failed] BSMT Lamplinc (F 82 A9 1)

[Failed] David Switch (3 7C 4C 1)

[Failed] Fireplace Spots (0 35 8A 1)

[Failed] Fam KPL 1 - Wall (A A9 A5 1)

[Failed] Entry Porch (3 42 48 1)

[Failed] Fam KPL 2 - Fireplace Spots (A A9 A5 2)

[Failed] Bar Cans (3 5F 77 1)

[Failed] Kitchen Cans (D 4 E2 1)

[Failed] Kath Icon (3 D3 46 1)

[Failed] Fam KPL 4 - Fan (A A9 A5 4)

[Failed] Fam KPL 3 - Lamps (A A9 A5 3)

[Failed] Fam KPL 6 - Entry (A A9 A5 6)

[Failed] Fam KPL 5 - 1st Floor (A A9 A5 5)

[Failed] Entry Deck (C 77 D2 1)

----- Link Test Test Results -----

 

Observations/questions on the above -

1) How in blazes were the CFL's able to corrupt communication to devices scattered throughout the house and on both phases?

2) The first activity the scene test performs is an "off" command to the entire scene. With the CFL's off, why didn't some of the units communicate?

3) Curiously, if I turned the lamps on and quickly performed the scene test, everything would pass. If I turned the CFLs on and waited for 10 seconds, every device would consistently fail. Similarly, if I manually turned the CFL's off, then quickly ran the Scene test everything would fail (persistence).

4) There was 0 noise in the 120 Khz range as monitored by my TesterLinc and ELK-ESM1.

5) Query's of the devices always return 100% (CFL's on or off).

 

The fact that there was a time constant associated with CFL's gave me the clue that I needed. I tromped down to the basement and found that the CFL's were on the same phase as my PLM (roughly 50' of wire between). I moved the PLM to phase B and the interference disappeared.

 

Analysis:

1) Since my testers were not reporting noise in the 120 Khz range, the PLM was reacting to "out of band" noise. By moving the PLM to phase B, the noise was blocked by the passive coupler.

2) Theory - the PLM is adjusting it's AGC (automatic gain control) in response to the out of band noise. The time constant noted above is associated with how quickly the PLM recognizes noise and adjusts the AGC circuit.

 

Noise Plots at the PLM

 

CFL Off -

Yellow : 60 Hz Powerline

Violet: Noise triggered at 70 mv

 

CFL On -

Yellow : 60 Hz Powerline

Violet: Noise triggered at 260 mv

 

 

Noise Spectrum

Out of band noise @67.871 KHz

 

 

Notes on the above:

1) I'm using a ScopeTest II adapter to capture the powerline noise. This is essentially a transformer tuned to 120 Khz (CP000). As a result, the amplitude of the out of band noise will be attenuated by the coupler (actual noise amplitude is much worse than shown).

2) The change in amplitude (CFL's on vs off) should be accurate in the above. With the CFL's on, noise increase by a factor of 3.7.

3) Frequencies shown are accurate (amplitude is not).

4) The fact that there are no noise spikes in the 120 - 130 KHz range is significant. The CP000 coupler is tuned to this range and should pick up noise 1:1. This also agrees with the ELK and Testerlinc results.

 

Take Aways

1) This is the first time that I've seen the PLM adjusting AGC due to out of band noise. This is very difficult to detect using traditional equipment (ELK, testerlinc). Clues are the "widespread" failures and the persistence (time constant) of the noise.

2) My PLM is a V 2.75. Since the AGC is largely controlled through the device firmware, other PLM versions may respond differently.

 

IM

[Brian H]

Thanks for the detailed information.

 

I use a JV Engineering XTBM and it has a frequency read out that starts around 110KHz and ends at about 140KHz. It also has a voltage readout that goes to 9.99 Volts PP.

 

I have found a few CCFLs and CFLs that actually walk up from the lower display edge to the higher display edge as they warm up. I have one brand CCFL that makes NO power line noise but spews RF noise in the 120KHz. band and actually if close enough. Gets into the XTBMs filters and give false readings.

 

Yours kind of looks like it changes as warming up and is clobbering the AGC to change the gain. Bummer. One more thing to add to possible problem things.

[IndyMike]

Hi Brian,

 

I had thought this was a warm-up phenomena as well. When I put the oscilloscope on it things were pretty stable in the 66 - 68 KHz range (nothing in 120- 130 KHz).

 

I'm a little disappointed in the AGC implementation. I don't have any hard data, but I don't believe that my Leviton (X10) units would be affected by this out of band noise.

 

I do not have a good way of filtering this noise. Traditional filters (band stop) won't touch the 67 KHz. High pass filters are typically too large to fit in a J-box.

 

Thanks for the detailed information.

 

I use a JV Engineering XTBM and it has a frequency read out that starts around 110KHz and ends at about 140KHz. It also has a voltage readout that goes to 9.99 Volts PP.

 

I have found a few CCFLs and CFLs that actually walk up from the lower display edge to the higher display edge as they warm up. I have one brand CCFL that makes NO power line noise but spews RF noise in the 120KHz. band and actually if close enough. Gets into the XTBMs filters and give false readings.

 

Yours kind of looks like it changes as warming up and is clobbering the AGC to change the gain. Bummer. One more thing to add to possible problem things.