Compact Fluorescent Noise

Sorry about the typos. I swear they weren't there when I proofread that. It must be that CF noise...

I forgot to point out my friend the transient right smack in the middle of the center X10 reception window.

This signal was monitored after the bandpass filter in the XTB-II. So this is what is sent on to the envelope detector.

Jeff

When you posted recently about the CFL noise I was going to ask you to get a picture but wasn't sure you had a way to do it automatically. This is a case where the ISA Digital Sampling Oscilloscope I use would really be handy, allowing you to zoom the horizontal axis to show more detail.

Is the spikey burst just past center screen part of the CFL noise? It resembles a triac switching transient.

How are you determining the frequencies in the noise? Are you just relying on the passband of the ESM1? Since X-10, Insteon and UPB have chosen to do their signalling in the time around ZC, this is sure to be an ongoing problem. It would be interesting to know how well an Insteon lamp module would function in this environment. It's also another factor to consider when considering CFLs. For those into HA the cost of filtering CFLs really changes the economics of switching from incandescents.

BTW, now that I have my web page up, I published several screenshots from my scopecard showing how X-10 does gated AGC in the CM15A. Just click on the CM15A link at the bottom of the page.

snipped-for-privacy@yahoogroups.com

Your postscript and my questions crossed in the ether. I should have realized you were relying on the XTB passband when discussing the frequency content. It would be interesting to see the noise through a wider passband such as through the scope adapter circuit shown in Appendix C (Figure C-3) of Microchip's AN236.

snipped-for-privacy@yahoogroups.com

Hi Dave,

(answering both posts)

Yes, a digital scope would be nice, but this is my old friend Tektronix 7603 that I used for 30 years. I bought it when I left the lab. They were probably happy they didn't have to calibrate it anymore.

The XTB-II bandpass filter is not high-Q to prevent too much ringing, but it does attenuate out-of-band signals. What I think is happening is that the high-frequency chopper in the CF bulb varies its frequency over each half cycle, and it enters the X10/Insteon passband as the voltage approaches zero. Higher frequency noise in the middle of each half cycle would have been rejected by the XTB-II bandpass filter. I'll try to take a look at it with just a high-pass filter to knock out the 60Hz.

Thanks for the po>

Another thing that I've found changes the equation is that none of the CFL's I've purchased have come anywhere near lasting their rated lifetime. It seems that the early "lifetime" guarantees given when CFL's first hit the market have been slowly replaced by ten and then five year guarantees and some packages I've seen have no warranty at all.

FWIW I've experienced far more CFL failures in fixtures where the lamp base is pointed up. Probably the heat from the bulb contributes to the early deaths.

Also annoying is most CFL's inability to be dimmed.

A secondary, but still non-trivial annoyance is the lengthening amount of time it takes for a bulb to reach full intensity. That seems to occur far too early in the life cycle of the bulb and makes it hard to compare costs accurately. My wife insists that they be changed when the delay becomes hazardous, as in stairways, hallways, etc.

One plus: CFL's have never burned out and taken an X-10 switch along with them, something that I can't say for incandescents. We've ended up using CFL's in the summer, where the rates are higher because they are so much cheaper to run. The cooler running means less A/C used, too, so it's a double benefit. In the winter, the waste heat helps warm the house and the electrical rates are lower so the benefits aren't as spectacular. We're also far more likely to use dimmers during the darker winter months so it works out quite well.

-- Bobby G.

Well, I didn't mention that because every time I've mentioned it in the past it has triggered a lengthy harangue from the local village idiot. ;)

It's hard to get reliable statistics but a Google search will turn up enough reports of short lived CFLs from reputable sources to indicate this is a real problem. I've read that the quality of the phosphors available in China are the problem.

I think the economics are marginal but I have to qualify that somewhat because I've recently had a rash of sort-lived incandescents. Either the change in ownership of Cinergy (now Duke) came with a change in power quality or the bulbs I've been getting in the past year are much lower quality than those I got in the previous 8-9 years.

snipped-for-privacy@yahoogroups.com

Why would it do that? It would seem to be more difficult to accomplish than a steady frequency. Ort is the frequency of the chopper related in some way to the sine wave?

snipped-for-privacy@yahoogroups.com

At your suggestion I monitored the line with just a high-pass filter to reject 60Hz.

Unfortunately, there seems to be some sort of audio communication device on the line now because I see apparent audio modulation of a 100mVpp 120KHz carrier that is there whether or not the CF bulbs are on. So it is difficult to determine how much more noise is added by the bulbs.

I did confirm the noise pulsation is due to beating between the bulbs by removing all but one. The CF bulb causing this havoc is the TCP ESN18.

Bottom line is that if you use CF bulbs in an automation environment, it is a good idea to isolate those circuits with the Leviton 6287 noise block. We have those on all X10 controlled lighting circuits, and have never had a problem with any of those CF bulbs.

One other thing I noted while looking at the scope photos is that timing of X10 bursts at other than the zero crossing is not well defined. Please look at the following two photos:

Notice how the X10 bursts move from one sample to the next. Also note the zero crossing burst is fixed as is the line transient just after the center zero crossing. That third burst moves just about 1mS, and I don't see how it could hit the reception window in a three-phase system. Signals were transmitted from a RR501 in response to a palmpad.

Jeff

Whoops, that's:

The base of these bulbs is very small, so there isn't much room for an electrolytic. They might just rectify the line and feed it directly into the chopper circuit.

Jeff

I've never seen that and I've looked at a lot of scope traces from various X-10 devices. In fact, I've always been rather impressed that the 120kHz is really 120kHz and other timings are near right on using 30 year old LC technology. For some examples, take a look at the screenshots in this link...

My pseudo-scope measures frequency and amplitude of a signal, making some measurements fairly easy.

I suspect either that you have an erratic RR501 or there's something amiss in your setup or that your scope has rheumatism.

It sounds like a neigbor has an intercom or baby monitor.

snipped-for-privacy@yahoogroups.com

If you could gate your AGC to measure only between the 1st & 2nd X-10 windows, as I suggested before, it should let you deal with the continuous noise.

snipped-for-privacy@yahoogroups.com

I didn't realize we had selected a leader among the many highly qualified candidates for LVI. ;-)

It's a real problem for me, despite any chewing out, declamation, diatribe, jeremiad, philippic, screed, sermon, spouting, tirade or harangue to the contrary. I write the date of installation and the alleged warranty period on each bulb base. While I don't keep actual "hours used" statistics, I know that I've replaced a lot of "ten year" bulbs in less than three years of ordinary use. The worst offenders are the ones mounted base-up that allow the bulb's heat to rise into the bulb's electronics. That's a failure mode quite similar to the small lamp socket modules that X-10 sells.

I've noticed the same thing and went Googling a while back when you first reported premature failures on bulbs connected to an LM14. We're not the only ones to notice lightbulb longevity ain't what it used to be. Still, an incandescent bulb costs me a quarter and most CFL's cost at least 10 times that amount. Whether it's bad voltage spikes, bad manufacturing processes, bad QC or just plain ol' bad luck I don't know.

Claims that: "The average lifetime of incandescent light bulbs is about

750?1000 hours. It would take at least 6-11 incandescent bulbs to last as long as one compact fluorescent, which have an average lifetime between 11,250 and 15,000 hours. This causes an additional total cost of using incandescent bulbs. Another additional (potential) cost may be incurred if the bulbs are not in a readily accessible location and special equipment (e.g., cherry picker) and/or personnel are needed to replace it."

That and a number of other paragraphs illustrate that the tradeoffs between the two types of bulbs involve a number of disparate factors and that no one solution fits all situations.

-- Bobby G.

Since replacing the LM14A with a Smarthome 2000STW, I've had two more failures. Both came after the lamp had been on for a few hours. Most of my other failures came when a lamp was first turned. It may be that the LM14A going loco was not a factor in the premature lamp failure but was merely coincidental.

snipped-for-privacy@yahoogroups.com

There are now compact fluorescent lamps (CFL) that are designed to be used base up in recessed fixtures. Using lamps intended to be used base down will not provide optimum longevity. See my discussion below of just such an application.

ROTFL. I am on the edge of my chair waiting to learn whether CFLs make sense globally and universally based on a future assessment of whether the quality of AC power delivered to Dave's apartment has changed ;-)

But wait! The global economics of CFLs may instead depend on the quality of the incandescent light bulbs that Dave buys. Thar's shore a whole lot riding worldwide on the quality of them thar bulbs what Dave gits ...

I replaced the seven Philips Halogena (Halogen) 90watt flood lamps ($6 at Home depot) in our kitchen ceiling cans with Philips R40 CFL Marathon Classic 85 (930 lumens 2700K CRI=82 $12 at Home Depot).

Neglecting entirely the fact that Philips rates the CFL as lasting four times longer than the halogen, and even at the low cost of $0.9/kWh, the savings after the first year (4 hrs/day) is $22.

After 3 years (the bulbs are rated and warranted 100% replacement for six years by Philips) the savings is $151 assuming that the halogens and the CFL both last 3 years. (The halogens most definitely will not, and the CFL's are warranted to last twice that long.)

Assuming that I continue to experience the same burn out rate with the halogens as in the past eight years, and assuming that the CFL only last 2/3 as long as Philips claims, the savings will comes out to ~$257 after four years.

The arithmetic seems to speak for itself for the quantitatively inclined, and the reasons and explanations that Dave gives continue to be the source of amusement. Keep giving us them hoots!

... Marc Marc_F_Hult

I should add that these CFLs are rated by Philips as dimmable and by my actual measurement meet Philip's claim that they dim to 10% of full. They are rated at 2700K which is near optimum for my wife's preference and have the same CRI (83) as the halogens that they replace. Nice, well diffused lamp face and light to boot.

|PLC|NP&cnt_key=CI_Refl+|PLC|150425++++++++++++&t=3&tree=0&scr_md=1111&leftnav=1_1&nav=Null&loc=us_en Marc Marc_F_Hult

That 120KHz is everywhere - even pick it up with the scope probe floating. Sometimes the modulation looks like music. While I had seen it intermittently in the past, it has been on all day today.

Looking further into the wandering pulses, it seems to be data related. I had thought it might be due to the RR501 power supply sagging, but that is not the case. Dim commands are stable where they should be. F-ON or OFF commands from the Palmpad causes misalignment of the bursts, usually in the "ON" or "OFF" portion. The center burst moves about .5 mS, and the third burst jumps 1mS. It doesn't slew, it jumps. It is not an erratic RR501 because both Powerhouse and Leviton versions purchased years apart exhibit the same characteristic. It is not scope triggering because the zero crossing burst is always where it should be, and the background line glitches also remain stable in the trace.

Jeff

I could put the gate anywhere, but between the 1st & second bursts wouldn't work for:

Setting the threshold between the 1st and 2nd windows would cause that burst on the right to be detected as a "1".

There doesn't appear to be an effective way to deal with the randomness of this kind of noise on a cycle-by-cycle basis. But then a longer term average has its own set of problems. Eliminating this kind of noise at the source may be the best answer.

Jeff

Hello Jeff,

Have you tried narrowing the LC filter? Several kHz wide might still be ok from a ringing point of view.

The other trick is what radio engineers refer to as a "noise blanker". If the spike is short enough versus the channel information a riding threshold detector is set. This detects sharp peaks above a slower signal and then notches out that portion of the receive phase. The good ones work remarkably well in muffling irregular spike noise.

While the LC filter should be an easy one the noise blanker does tend to increase uC workload to the point where one has to tap the cost brakes.