Dimmer Neutral Wiring

Thanks for the advice cabra, the lights that this happens to are incandescent, but I do have some cheap LV transformers for some lights that are connected to one of the dimmers (I don't dim those lights though, always . I've had no issues with those lights flashing and this "flash event" occurs on two dimmers that I have and the LV txfmrs are only connected to one.

Thats what I'm beginning to conclude. I started the install on my house at around the same time a new subdivision occurred over my back fence. The new homes are all fed from the txfrmr that supplies my house. As houses have gradually been added, I firstly experienced water hammer issues (which might now be flash events on the relay unit activating my sprinkler solenoids) and now the flash events on the dimmers. Hence, your observation of a local influene could fit, as does Nobes' observation of it being a larger influence on the whole supply system. I'm now headed down the path of installing a filter and seeing if that works... Thanks again for the suggestions, Cheers, cc&c.

 

Thanks for the advice cabra, the lights that this happens to are incandescent, but I do have some cheap LV transformers for some lights that are connected to one of the dimmers (I don't dim those lights though, always 0% or 100%, haven't made a direct correlation yet with those being the source of the problem, they could well be though). I've also had no issues with the lights on LV txfmrs flashing and as this "flash event" occurs on two dimmers that I have not just the one that the LV txfmrs are connected to.

Thats what I'm beginning to conclude. I started the install on my house at around the same time a new subdivision occurred over my back fence. The new homes are all fed from the txfrmr that supplies my house. As houses have gradually been added, I firstly experienced water hammer issues (which might now be flash events on the relay unit activating my sprinkler solenoids) and now the flash events on the dimmers. Hence, your observation of a local influene could fit, as does Nobes' observation of it being a larger influence on the whole supply system. I'm now headed down the path of installing a filter and seeing if that works... Thanks again for the suggestions, Cheers, cc&c.

 

This is very interesting stuff, although I notice that the CIS people aren't saying much...

Anyway, if what Nobes says is correct then how about checking / changing the settings of the dimmers (and relays) in Toolkit for a test... edit it so that they restore to ZERO LEVEL upon power failure.

That way, when the (whatever causes it) event occurs they will turn off rather than "flash" on then restore back to the dimmed setting.

I am thinking that using such a method will help determine / confirm if it's a brownout that's causing it. Whether my thinking is correct or not is highly debatable because I am making this up as I go along... but it's worth trying, even if only as an attempt to get more info to feed back to CIS about it.

Anyone else got any ideas about my idea ?

 

Recovery

Perhaps CIS could clarify how C-Bus recovers from a power failure or brown out? I have to admit that I have never noticed or had knowledge of a system check which would entail a momentary full-brightness setting, followed by whatever setting it should recover to.

 

Just quickly ...

My observation is that this isn't a C-Bus (Software) event, it is a hardware issue. I take point the point that setting the restore to zero level will help to confirm that this is the case. Yes I've noted the lack of CIS prescence, I appreciate they can't do everything. I'm sure they've had many times the philosophical discussions about providing support by this forum. "Contact Technical Support" must be wearing a bit thin with some people, along with this is "CIS Engineering". To most people this is the CBUS and it is the easiest way for people to help themselves in resolving issues and gain a quick resolution from experts.

Clipsal need to look at the internal silos that obviously exist and present a united one stop web prescence. They need to move with the times, there are plenty of other companies providing support in this way and we're dealing with a technology product.

Now that I've finally got that off my chest, I'm headed down a path of installing a filter for my mains supply. Cheers, cc&c.

 

What a beautiful story Ashleigh

It had begining, a middle and an end - top shelf.

 

Ashleigh

Your reply was excellent and also was considerably more helpful than all of us people speculating about the possible causes of the problem. At the end of the day, you guys know a HELL of a lot more about the products than we users do... we simply don't have the technical knowledge to troubleshoot wierd problems like this.

I don't think that CC&C was attacking anyone, in fact there was actually defence of the engineering team that visits here at the forum.

 

Ashleigh,

First up, thanks for the detailed response. As we all experience all too regularly these days (through emails and forums like this), the tone behind written words can often be mis-intrepeted. (I guess that's why smilies were invented).

I've obviously got the wrong impression about why this forum was set up and the purpose of it. I'll create a different thread (under general discussion) so as to avoid confusion with the purpose of this post and potentially assist you in getting the message across to others. I trust you'll feel free to discuss.

Transformer. Automatic tap changer will be on HV side at the HV sub-station (in my area). You've got an excellent point with a tap change being the possible cause.

I'm in the process of trying to beg, borrow or steal.

The Australian Standard is 230 +10%/-6% and specification quotes 220-240V so are you saying that CIS won't guarantee the dimmer meets this standard for voltage supply? or compliance with the standard is inferred by quoting 220-240V. From what you are saying, I suspect the later is correct.

Completely randomn. No flashes on the dimmers the last few nights, 6-7 one night last week. Had 4 x water hammer events in 5 minutes last night but I've previously said that this isn't necessarily a C-Bus problem, but it is certainly convienient for me to assume that it is.

Observation about the cold is that the flash events appear to be more prevelant on cold nights. This could be related to the larger loads on cold nights and hence possibly invoking more "tap change" events as you've suggested.

I'd suggest "conniption" as possibly a more apt description. Certainly agree that the solenoid wouldn't activate for an event of only a few cycles. The flash events on the dimmer and a possible flash event on the relay unit don't have a high correlation, hence it might be a completely different flash event (i.e frequency instead of voltage) or I might be unfairly blaming C-Bus at the moment. As I appear to have two completely randomn events happening and the common element is C-bus, well .....

Will do.

Ever experienced a perturbation or conniption?

No to both.

After all the help that I've received (from everyone), I certainly agree that the mostly likely cause is voltage variations and yes your post has assisted greatly. Cheers, cc&c.

 

I run here on a 10KW generator-interactive inverter. Pure-sine, low distortion, frequency when on batteries is about .005%

When the genset is running, because of the computer-controlled air/fuel mixture, it tends to vary in frequency slightly. I've set it to a nominal 50Hz (counting cycles over several hours to fine-tune it), but as the AFR varies I see the frequency vary from 49.9 to 50.1Hz cyclically with about a 2 second period. I've not experienced any "flashing" or other sync issues with any of the 108 odd channels of dimmers, so they're obviously not *TOO* sensitive to change

How does a tap change result in a frequency change? *puzzled look*

I pricked my ears up at this one.

While building (in the late stages thereof) I had most of the cbus stuff all set up (although no key units - all the light fittings were in, all the output units were in - we could turn stuff on and off at the output units). I also had a small pressure-pump (a little 10 litres/minute toy). The supplier said it had a pressure-switch in it and was safe to leave it run. Silly me, took him at his word. The pump was mounted outside in the pump pit (actually about half a metre above the ground) and hardwired to the power circuit through an RCD.

Whenever someone used a little water (one of the trades flushed a loo, or the painters were rinsing brushes etc), all the lights in two rooms only would intermittently "sparkle". (They're all Megawatt 240V/11W CFLs). They would strike and immediately go out. Not all in unisen. This bank would fire, then another bank in a seemingly random pattern.

In the end, the "internal pressure switch" gave up and burned out (it was never a proper pressure switch, it was an overpressure safety!) and I replaced the whole thing with the pump we'd intended to use but didn't have (a variable-frequency drive with pressure transducer that only runs as hard as it needs to to supply the volume of water required at the time) - and the sparkles "went away". Only one of the 9 dimmer modules did it. Never did get to the bottom of it. We *occasionally* see a flick on those lights when they're off. I'm guessing this is related to noise on the mains resulting in false triggering of the triacs (since it used to happen with NO key inputs and NO programming!)

RossW

 

All triacs will turn on due to a fast transient

There is a parameter called "critical rate of rise of off-state voltage" which all triacs can be characterized by (typically a few volts per microsecond). If the triac is OFF and the applied voltage changes quickly enough; the triac will turn ON. The triac will remain conducting only until the next zero-crossing of the current through the device at mains frequency. The turn-on can only be avoided by protecting the triac with bulky filter components (rated to the full voltage and current rating of the product, of course). While there are some filter components included in ALL C-Bus triac-based dimmers, these are kept small to allow a compact product, and are not sufficient to avoid triggering by very rapid voltage changes.

When the dimmers are tested for immunity to surges and fast transients to meet the European requirements for EMC, the test standards stipulate that the loads can turn on briefly (because few dimmer products would otherwise meet the requirements), as long as they recover to the intended state (OFF) at the end of the tests.

This behaviour is a characteristic of the triacs used to switch the load. In products based on other technologies (like trailing edge dimmers), this behaviour may not be seen.

If you really require a load to be OFF at all times and you have a very noisy mains, you could wire that load from a dimmer as well as in series with a voltage-free relay. The relay could be set up as a threshold switch so that it is closed for any level other than OFF.

Don

 

Don, the original posted CC&C was experiencing problems of a slightly different nature, and not experiencing flashing when the level is set to zero :

So, it's not flashing from "off", it's flashing from "nearly off" - would the triac characteristic you describe cause that as well? You stated that :

... so are you saying that the triac will turn on for a full 1/2 cycle (causing a flash), where at a very low dimmed level it should only being switching "on" for a small proportion of the 1/2 cycle?

I ask for clarification because some people's interpretation of the term "off" will be different to others - since in this case the Triac is off then on then off then on again continuously (at 100Hz) whilst it is operating as a dimmer... I hope that makes sense ?

Also, I am having a bit of a problem understanding the ramifications of that event from a lighting point of view. An incandescent filament is a REALLY slow device - I doubt you'd ever notice a 1/2 cycle flash, as the filament simply couldn't heat up and cool down that fast. If it was a Xenon Flash tube, then yes... but is it even possible that such a short duration (10mS) of 100% power could produce a noticeable increase in apparant brightness in an incandescent lamp ?

That said, we are really getting somewhere - an excellent thread, and shows all the great minds working together (and mine isn't one of them) !

 

Ooooh - what technology is the new "Universal Dimmer" using? Would that new dimmer be less susceptable to these mains-bourne annoyances?

 

Yup, the old "dv/dt" problem. Well aware of it.

Again, yes. Triacs (and SCRs, SCSs etc) will turn off when current=0 (or falls below their threshold which may be a non-trivial amount for larger devices).

I wasn't disputing this happens - I know that it happens and why (my background was in hardware, then hardware/software integration and development, then in industrial/commercial/military R&D, troubleshooting, etc)

I was more puzzled why different devices in the same dimmer module fired at seemingly different times, while other modules appeared to not exhibit the problem *at all*. I'd have thought that if one discrete component was slightly more sensitive than others, it'd have turned on first, as the noise got worse the next-most sensitive one would fire *ALSO* and so on. The observed events didn't fit.

Contemplated doing that, but it would require more switchboard space than I have (I split everything over 4 seperate "areas" throught the house, each one is 3 x 24M boxes except the aviary which only has one 24M box)

 

To achieve this kind of dimming, the triac is turned on somewhere *BEFORE* the end of a half-cycle. Since it will always turn *OFF* when I~=0, the earlier it is triggered, the longer it will remain on, so yes if the dimmer was at a low level (triggering at say 120 degrees in the cycle) and something was inducing sufficiently fast risetime noise, the triac could trigger at (say) 10 degrees and remain on in effect for the whole half-cycle.

If the noise is in a burst going over several half-cycles, the triac could be retriggered nearly continuously, resulting in an (apparant) 100% setting for the duration of that noise.

 

I don't know if this is the mechanism of your observed fault or not, but thought it was worth clarifying what is accepted as "normal" behaviour for a dimmer under conditions of fast transients on the mains. From what I've read, it could still be the culprit.

For JohnC, if the load is dimmed to a low level, the triac should be OFF most of the time, so a sudden surge to full brightness from "dimmed" also fits the picture, as the triac itself is in the OFF state and susceptible to dv / dt triggering for a percentage of each mains half cycle.

Note that most sources of fast transients like switch contacts, etc. do not produce a single event, but produce a burst lasting typically several 10s of milliseconds, so in a noisy environment you definitely do see incandescent loads lighting up from time to time.

If there are several loads connected to different channels of a C-Bus dimmer, and only some loads are flashing, I would suspect that if dv / dt triggering is the issue, that the ones that flash are coupled more closely to the source of the disturbance! It certainly is possible that some other mechanism is at play here, but It sounds to me like the triacs are just being triacs.

 

CC&C,

Sorry for not getting back to you, been in the INT2 C-Bus course.

I have a friend in Aurora that works in the fault investigation area. He was saying the other day they regularly get people asking for thier supply to be monitored. From what I could gather, you probably shouldn't have too much trouble getting them to come out. I think they put monitoring equipment at your board and the sub that it comes off.

As far as a relay pulsing on and off, I have never see or heard of this happening, I probably think the water hammer is something else. Still worth you running a toolkit log for a night or two, or touch screen if you have one. Or just secure that knocking pipe and forget about it...

 

Thanks, I've managed to track down a monitor. Because this happens so randomnly Aurora could have the gear here for a couple of weeks and they might not record anything. But when it does happen it generally sticks around for the night. I'll then be able to move the monitor from the main board to the c-bus cabinets to upstream and downstream of the dimmer.

Observation was based on that I've got two randomn events occurring and they're both connected to C-Bus. The monitoring will confirm or deny. Cheers, cc&c.