[GTALUG] mysterious restarts

[Alvin Starr]

On 06/22/2016 07:39 PM, Michael Galea wrote:
> On 06/22/16 16:35, Alvin Starr via talk wrote:
>> On 06/22/2016 01:25 PM, Jamon Camisso via talk wrote:
>>> On 2016-06-22 17:46, D. Hugh Redelmeier via talk wrote:
>>>> Remember, he truncated the values (grep does not round).
>>>>
>>>> My best guess:  the average of the readings would be from a
>>>> distribution centred on 59.996435 (0.005 larger than Lennart
>>>> calculated).  It could be as low as 59.991435 or as high as 60.991434,
>>>> assuming six digits of precision in the fraction.  I know nothing
>>>> about accuracy of the device.
>>>>
[snip]
>>
>>
> No, the device is not a UPS. It is a PLC that measures voltages, 
> currents, real and reactive power of a 3-phase service.  It provides 
> me the results as IEEE754 32-bit floating point numbers. It's me that 
> truncates to 2 digits.
>
The data will be collected with an analogue to digital converter.
There are a number of different A-D converter technologies that can be 
used but each technique will affect your accuracy.
It is possible that the A-D converters are 32 bit but more likely they 
are somewhere between 8 and 24 bits.

All the A-D converters I know of will give an integer as the result so 
the floating point comes from the post processing.

Then there is the sampling frequency.
This will affect the reliability of the frequency if the A-D conversion 
is used to calculate the frequency.

So the accuracy of the measuring system needs to be taken into account 
when you are using it to measure some other system.


> You know I'm sorry I ever implied that the grid wasn't long term 
> accurate at 60 Hz!  What I should have said was that in the short term 
> the grid could be quite off 60 Hz but over the longer term it could 
> average out.

The problem is that the grid needs to be short term synchronized because 
if 2 power sources are providing power and they get slightly out of sync 
then VERY BAD things happen.
So a 0.01 drift from 60Hz will after 50 cycles  be out of phase and at 
all the power from 1 source will sucked up by the other source and the 
grid will collapse. 50 cycles is just about 1 second of drift but long 
before that second elapses the grid would have collapsed.

The point is that the power sources need to be VERY accurate over the 
short term.
So Pickering needs to be exactly in phase with Bruce and with Niagara 
and so on.
Over the long term they can slightly adjust their phase to make sure 
they are in sync but if the frequency drifts in the short term then the 
grid ends up losing power to generator and power-line losses.

That being said any single cycle can be wildly off from 60Hz but that 
will be because of local loading or switching transients.

>
> If I really needed to know the answer to the question of "how much 
> out", I would get ask the PLC to just count cycles and send me the 
> count with a timestamp. That would tell.
>

-- 
Alvin Starr                   ||   voice: (905)513-7688
Netvel Inc.                   ||   Cell:  (416)806-0133
alvin at netvel.net              ||

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gtalug.org/pipermail/talk/attachments/20160623/1868bb12/attachment.html>