Re: [Development] Lights functions required

["Dr John Tankard" <john@xxxxxxx>]

--- In ukha_d@y..., "Keith Doxey" <ukha@d...> wrote:
> Here I go, answering many questions in one go.
>=20
> Channels Per Dimmer Pack/Packs Per Room...
> 8 channels per pack, possibly 2x4 channel boards each with one PIC=20
per
> board. That allows a 4 channel pack to be built and also makes the=20
PCB
> cheaper. Granted you would need 2 PCB's but you are buying more so=20
the price
> per board comes down. Also limits failure.

Keith, are you thinking that the dimmers are on the channel boards? I=20
was thinking we would have a motherboard which the rabbit plugs into.=20
The pic and optos on the same board. then upto 8 small drive boards=20
which would connect to the motherboard.

I have no problem with either route, its up to you. At the moment I=20
have been running the pic at 10Mhz and 8 channels are no problem. But=20
if we go down your route we might be able to drop down to 4mhz if you=20
are talking two pics.
>=20
> Channel Assignment.
> Some form of Dimmer/Address mapping. Lutron assign dimmers to=20
channels. This
> allows several dimmers to have the same identity so you can handle=20
a 20A
> load as 4x5A.

Thats an idea. how bright do you want your lights !! or is ths to run=20
the pizza oven :-)
>=20
> Dimming Curve/Dipless Crossfade....
> I dont believe anyone would notice the difference unless it was=20
pointed out
> to them. For dipless fading my idea is this. If the level is 0 to=20
100 and
> you have 2 channels that require to do the following
> A 25 > 75
> B 100 > 0
> with a fade time of 10 seconds you need to do the following.=20
Increase A by
> 50 steps and decrease B by 100 steps. You have 10 seconds to=20
complete this
> so that gives A needing to step 50 times in 10 second =3D increase
1=20
step
> every 200mS, B needing to step 100 times in 10 seconds =3D decrease
1=20
step
> every 100mS. The only calculation needed it to work out how many=20
steps in
> the time period.

Yep, at the moment the prototype fades at a constant rate across all=20
channels, which i agree is not good enough, so if one channel needs=20
to go one step and the second channel needs to go 50 steps, the fi4st=20
channel will get there in the first time period. I will fix it.

>=20
> Fault detection.
> Subject to having spare pins available on the PIC you can detect
> No Output - fuse or relay blown
> Continuous Output - Triac blown
> No Load - All lamps disconnected (NOT INDIVIDUAL BULBS unless 1=20
bulb per
> dimmer)

In my case i have wired every lamp back to the packs so it is=20
possible to check each bulb. (sorry my LV units are in groups),=20
however i suspect most people will not have every lamp isolated so=20
thats not a option.

Since we have a protection relay so if the triac goes short circuit=20
it will still go off when we go to off mode.

I dont realy know which would be best. Do you know what the comercial=20
systems do ?
>=20
> Cost.
> Prices quoted by others seem reasonable at about =A315 per channel,=20
possibly
> =A320 with fault monitoring but that can be optional. The expensive=20
bit is the
> choke. By keeping the dimmer rating low - say 3A per channel ~720W=20
the
> chokes are cheaper. for heavy loads you use more than one channel=20
as per
> channel assignment above.

I like that idea, perhaps if the drive boards were seperate we could=20
do different triacs. From my personal view we need to go over the top=20
on triac size, pain in the ass to swap when a bulb goes. As a side=20
note i have some channels on ssr's where i dont need dimming and out=20
of 10 channels I have had two go in the last three years even though=20
the are 5 amp rated

John





____________________________________
Automated Home UK
http://www.automatedhome.co.uk
____________________________________=20

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/=20