Re: Re: Simat 1 wire Hubs

[Jon Payne <jgpayne@xxxxxxxxxxxxxx>]

(wow, just found this in my 'drafts' folder, obviously been sitting
there a while so apologies for those who wonder what this is about!)

Hi Frank,

I made this change (start temperature conversion of all devices on a branch
and then go back later and read the temperature) and it has reduced the
time to take temperature readings.

Unfortunately although all my temperature sensors are DS18B20's, some are
family code 0x10 and some are family code 0x28!
Only the 0x10's work with the SkipRrom/ConvertT method, so I have to
explicitely select and ConvertT on the 0x28 devices.Even the combination of
this, accessing all devices on a branch together and then going back later
to actually read the temperatures saves quite a bit of time and more
improtantly doesn't tie up the bus for anywhere near as long (because the
bus can be used whilst the temperature conversion's are being done).

I've also added support for Louis Swart(sp?) LCDs (with xAP OSD) and
reading of Thermocron's...

Cheers,
Jon


----- Original Message ----
From: Frank Mc Alinden <frankmc@xxxxxxx>
To: ukha_d@xxxxxxx
Sent: Wednesday, 23 May, 2007 9:31:42 AM
Subject: Re: [ukha_d] Re: Simat 1 wire Hubs


Hi Jon

> I'm going to look at optimising the temperature gathering now by
grouping
> sensors by path(location on branches), which will also apply to polled
> sensors in general, and by starting >temperature conversion off
> simultaneously (which I've only just realised can be done) for all
> >sensors on that path, then fetching their readings

Let us know how this goes ....

Frank
----- Original Message -----
From: "Jon Payne" <jgpayne@xxxxxxx>
To: <ukha_d@xxxxxxx>
Sent: Wednesday, May 23, 2007 5:21 PM
Subject: Re: [ukha_d] Re: Simat 1 wire Hubs


> Great to know others are using 1-wire on UKHA!
>
> I have now speeded up my network scanning a little by using the DS2409
> SmartOn feature - this lets you determine whether there are /any/
devices
> on a branch without having to actually scan it which obviously saves
time
> if a branch is unused or used for touch contact ibutton applications.
>
> I'm going to look at optimising the temperature gathering now by
grouping
> sensors by path (location on branches), which will also apply to
polled
> sensors in general, and by starting temperature conversion off
> simultaneously (which I've only just realised can be done) for all
sensors
> on that path, then fetching their readings.
>
> cheers,
> jon
>
>
> ----- Original Message ----
> From: Frank Mc Alinden <frankmc@xxxxxxx>
> To: ukha_d@xxxxxxx
> Sent: Wednesday, 23 May, 2007 7:49:12 AM
> Subject: Re: [ukha_d] Re: Simat 1 wire Hubs
>
>
> Hi Mark / Jon
>
>> Sort of - as I understand it, it does a startup scan to detect new
>> devices and adds them to the database of connected devices. It
will
>> then scan for devices in the database so if a device has been
>> connected at any startup it should then be scanned for. However, I
>> think the new hub code breaks this in that the hub branch number
is
>> remembered at startup and the devices are only searched for on
that
>> branch.
>
> That appears to be the case from my observations as well....
>
>> Nice to see some 1-wire discussion here, I'm just getting into it
and
>> going through the opnode sorce code has helped a lot but it's good
to
>> hear of other ways of doing things.
>
> Keep us posted on your progress
>
> Frank
>
>
>
>
>