DALI light level control issues

Good grief I've just noticed that I have been made a 'senior member' (so hopefully I know what I am talking about!). Thanks to the forum pixies for recognising my contribution to the World of C-bus.


We have just finished being involved with a job where the 5753PIRL unit is being used to switch on (PIR function) then control the LUX (LL Function) of rooms and open areas of an office block. The Fluro Lights are being controlled by DALI ballasts and getting the system to work correctly was not as straight forward as it might have first appeared.

1. The 5753PIRL issues 'ramp up' or 'ramp down' commands in response to the ambient light level being measured by the 5753PIRL as being lower or higher respectivly than the target level programmed in to the 5753PIRL.

2. The ramp rate set in the PIRL (and the 5031PE) is of the order of 60sec from 0-100% or 100-0%

3. The PIRL has a 4 second buffer (slug factor) built in to that it ignores any change in measured light level for less that period of time (ie a cars headlights sweeping across the room or a worker in a white shirt walking under the sensor and reflecting up more light than the, say, gray carpet does etc)

4. The DALI ballast can be programmed (using WINDIM or equivalent from Tridonic of other DALI manfacturers) to either turn on/off or ramp up/down not both and the default ramp rate is 8secs. If you want to control the brightness of the load you need to select the ramp option and note that it doesn't matter that the ramp rate from the PIRL is 60sec the DALI sees the initial 'on' command and ignores the 'over 60secs' bit.

Think about that little lot for a moment.


So the situation is, you walk in the PIR detects you sends out an on command but the DALI ballast can only have been set to ramp (see 4) the light ramps to 100% over 8 seconds but probably will pass the target set in the LL. But the 4sec delay means that the LL will not react until the light level is way above the target so a ramp down command is issued and the light gets dimmed down and again shoots past the target - it finally settles down after a few excursions.

Well there are two options: reduce the 4 sec 'slug time' in the PIRL (not currently available as a variable) but this would mean that the loads would dim up and down as people in white shirts walked around the room/office. So not a real option

or

increase the DALI ballast ramp time so that the 4sec slug becomes less significant (ie set the ramp rate to ~60sec and everything is sweet.

But (!) because the DALI can only be set to switch on/off OR ramp up down this would mean that as you walk in to the room the PIR send out an on command and the DALI can only then ramp up from 0% to the target (or 100% - which ever is the lower) over the 60secs so you are left standing there as the lights slowly come up around you (in fact you could walk in and out of the office a couple of times and the lights may still not have reached the target brightness) this is not only inconvenient but also may have Ocupational, Health and Safety implications.

There is no happy ending. Untill the DALI guys allow their Ballasts to be turned on and then dimmed down to a target level (which is what a totally C-bus sollution would allow if you were using incandescent lights or fluros run by DSI or 0-10Vdc ballasts) to give you a bright room that then goes to the set level making the room conveniently bright as you walk in and safe so you won't trip over the wast paper basket the only option is to use as short a DALI ramp time as is possible (given the over and under shoot issue) while still bringing the lights up quick enough to met any OH&S issues.

If anyone has any better solutions (ie if there is some way of setting the DALI ballast to on/off then ramp up/down) please post it here.

 

slug factor

Hi Uncle Dick!

I'm not sure how helpful this is, but the "slug factor" only applies once the PEIRL has reached the target level. When the PEIRL is NOT ramping, it requires that the averaged measured light level is outside of the margin for at least 4 seconds continuously before it will take any action.

The sensor measures the light level every 16ms, and averages it (over 2 measurements for early models, and over 16 measurements for most recent models).

When the sensor takes action to correct the light level, it will issue a ramp-to-level-at-rate command, with the rate set to whatever is chosen in the GUI. While it is ramping, the PEIRL keeps a model of the expected dimmer behaviour, and ramps its internal model up at the rate of the command, ie. over 60 seconds for your example. As soon as the average reaches the target, the unit will issue an instant ramp to the current level as perceived by the internal timer associated with the PEIRL ramping.

Now if the DALI ballast is ramping more quickly, it will see a command to ramp back to the level that the PEIRL thinks it aught to be at, which is likely to take the light level out of the controlled range again, and will result in another ramp up command. Apart from the disco effect, it will evenually reach the target.

The best way to fix this would be to make the DALI ballast ramp at the same rate as the PEIRL (which is what a C-Bus dimmer would do). Second best would be to make the PEIRL rate match that of the ballast, unless this rate is too fast to achieve stability. If stability is not possible at (or near) the DALI ballast ramping speed, I would tend to err by setting the PEIRL to a slightly slower ramp than that of the ballast. If the corrections are small and the margin is set high enough, it should behave itself

Don

 

Thanks for the info Don, once we had taylored the settings (along the lines you detailed - we found out by educated suck it and see) we got the best fit that we could. We were fortunate enough to be able to set up a pilot system back in the office and quickly trial a few options - if we had tried to do that on site (which was live with office workers tring to settle in to their new home) it would have been more error than trial.

The ramp rate of the PIRL (although it is selectable unlike the 5031PE) was of limited use as once it issues a Ramp to 100% over X seconds the DALI Ballast only sees either an instant on command or a ramp to 100% over whatever seconds (default 8sec) that it is set to

If the lights were to turn on once in the morning and off at night it wouldnt be such an issue, the minute to minute changes in the ambient light level could be coped with.

The spider in the fruit basket was that there was PIR control as well so the expectation was that rooms or areas would be turned off (ie reasonably short time out periods) if no one was occupying the room/area and then when someone walked in the light HAD to come on quickly enough for safety reasons but slow enough to limit the 'disco' effect. During the day there was enough ambient light from windows or adjascent areas for this not to be too bad but we also had to cope with the 'what if it is late at night and I walk into this room' scenario

 

DALI Commands

Hi,

The quick turn on then slow ramping could be acheived if the DALI Recall Max command was used to turn the ballasts on. This command doesn't depend on the Fade Time.
Unfortunately, this command is incompatible with C-Bus messaging. Making it happen would require a dirty, dirty hack.

Cheers,
Mark.