Re: z wave energy meter DSB28

Posted on
Sun Apr 27, 2014 10:55 am
matt (support) offline
Site Admin
User avatar
Posts: 18732
Joined: Jan 27, 2003
Location: Texas

Aeon Labs Home Energy Monitor (DSB28)

This thread is for discussing the Aeon Labs (Aeotec) Mult Sensor.

Device details
How to use with Indigo

Image

Posted on
Sun Apr 27, 2014 11:14 am
matt (support) offline
Site Admin
User avatar
Posts: 18732
Joined: Jan 27, 2003
Location: Texas

Re: z wave energy meter DSB28

Hi Karl,

Interesting. Have you only noticed these types of bad values from the DSB28? I'm not sure what is causing it (bad module, firmware buglet in DSB28 or Z-Stick, etc.).

I will definitely add sanity checking to the values in Indigo before it posts the state changes.

Image

Posted on
Sun Apr 27, 2014 12:04 pm
kw123 offline
User avatar
Posts: 6004
Joined: May 12, 2013
Location: Dallas, TX

Re: z wave energy meter DSB28

Could not find the right forum for this device, used the only "energy meter" device type to post

I get (very few) entries like this with very large + or - numbers from the energy meter:
id ts accumenergytimedelta accumenergytotal curenergylevel accumenergytimedelta_ui accumenergytotal_ui curenergylevel_ui
---------- ------------------- -------------------- ---------------- -------------- ----------------------- ------------------- -----------------
352841 2014-04-27 05:44:21 8213836 8906.994 743.385 8213836 seconds 8906.994 kWh 743.385 W
352842 2014-04-27 05:54:15 8214430 8906.994 4263.132 8214430 seconds 8906.994 kWh 4263.132 W
352843 2014-04-27 05:54:16 8214431 8907.569 4263.132 8214431 seconds 8907.569 kWh 4263.132 W
352844 2014-04-27 06:04:16 8215031 8907.569 4042.169 8215031 seconds 8907.569 kWh 4042.169 W
352845 2014-04-27 06:04:16 8215031 8907.976 4042.169 8215031 seconds 8907.976 kWh 4042.169 W
352846 2014-04-27 06:04:17 8215031 8907.976 -2660064050.7 8215031 seconds 8907.976 kWh -2660064050.70 W
352847 2014-04-27 06:14:16 8215631 8907.976 328.893 8215631 seconds 8907.976 kWh 328.893 W
352848 2014-04-27 06:14:16 8215631 8908.322 328.893 8215631 seconds 8908.322 kWh 328.893 W
352849 2014-04-27 06:24:16 8216231 8908.322 1062.27 8216231 seconds 8908.322 kWh 1062.270 W
352850 2014-04-27 06:24:16 8216231 8908.594 1062.27 8216231 seconds 8908.594 kWh 1062.270 W
352851 2014-04-27 06:34:16 8216831 8908.594 327.758 8216831 seconds 8908.594 kWh 327.758 W


274063 2014-02-04 18:40:35 2203898 15652.36 2648.801 2203898 seconds 15652.360 kWh 2648.801 W
274064 2014-02-04 18:42:31 2204014 15652.36 2649.46 2204014 seconds 15652.360 kWh 2649.460 W
274065 2014-02-04 18:44:32 2204135 15652.536 2649.46 2204135 seconds 15652.536 kWh 2649.460 W
274066 2014-02-04 18:44:32 2204135 -2140219.72 2649.46 2204135 seconds -2140219.720 kWh 2649.460 W
274067 2014-02-04 18:44:32 2204135 -2140219.72 2650.801 2204135 seconds -2140219.720 kWh 2650.801 W
274068 2014-02-04 18:46:36 2204259 15652.659 2650.801 2204259 seconds 15652.659 kWh 2650.801 W
274069 2014-02-04 18:46:36 2204259 15652.659 4109.2 2204259 seconds 15652.659 kWh 4109.200 W
274070 2014-02-04 18:48:34 2204377 15652.794 4109.2 2204377 seconds 15652.794 kWh 4109.200 W

There is no log file entry around that time.

That kind of data creates problems for plots i.e. smoothing and the KWH(accumenergytotal) seem not to be impacted.

(a) Any idea how to avoid this
(b) how to fix the data
(c) is this a broken device?


Jay,Mat could you reject negative KW's or anything bigger than e.g. 50KW ?


thanks

Karl

MODERATOR NOTE: See support reply to this post ABOVE. I had some problems moving this post to the correct topic and the timestamps became off, resulting in abnormal sort order here.

Posted on
Sun Apr 27, 2014 5:18 pm
kw123 offline
User avatar
Posts: 6004
Joined: May 12, 2013
Location: Dallas, TX

Re: z wave energy meter DSB28

I had 6 data points*) in 3 months for 2 modules (don't have more data in sql)
Not really bad, but one or two of these can have bad effects on plot.

thanks

I have added ignore (min/max) parameters for the plots - can be set to e.g. -500 / +50,000 well outside any number I can come up with for temp, distance, energy,humidity, windspeed, direction, pressure rain amount ...

Karl

*) these are in Watts (well beyond the copper cable capacity going to my house)
274,195.581
246,974.3
301,662,164.701
1,264,353.6
268,973.578
-2,660,064,050.7

Posted on
Sun May 18, 2014 10:43 am
kw123 offline
User avatar
Posts: 6004
Joined: May 12, 2013
Location: Dallas, TX

Re: z wave energy meter DSB28

some more info on bad data from the device

The devices (I have 3) are delivering the accumulated energy consumed. By definition the next number should be equal or higher than the one before.
In about 0.3%*) of the cases that is not correct. Sometimes the values are just a little bit lower but some times you get
besides the obviously wrong numbers like -121231234.87 (should never be negative) or 345,675,346.99 (that is in KWH .. more than a nuclear power plant can deliver)

you find:
first value: 12345.89 OK
second value 0.1234589 bad
third value 12345.99 OK

or a minor problem:
12345,89 ok
12345,80 bad
12345,92 ok

Anyone seeing effects like this?

I am trying to filter these things out, but the pattern are difficult to program and there is one situation where the next number can be less than the one before, i.e. when you reset the accumulated energy, and that looks exactly like one of the bad examples.

As I have no access to the raw data, I was wondering if this is due to the z-wave network, the device or some issue in indigo.

Karl


*) 0.2% sounds small but its is ~ 1 bad value every 3 days -- if you look at the value over time in a graph these things really pop out.

Posted on
Wed May 21, 2014 8:50 pm
matt (support) offline
Site Admin
User avatar
Posts: 18732
Joined: Jan 27, 2003
Location: Texas

Re: z wave energy meter DSB28

I added (after your original report above) to 6.0.12, or might have been 6.0.11, logic that should ignore negative W and kWh values. In the Event Log it should still show the values but report a "value out-of-bounds" error. Is that occurring for the negative values? I didn't add anything to drop really large values or cases where kWh drops below the previously read value (that will be harder to detect).

Image

Posted on
Thu May 22, 2014 3:51 am
kw123 offline
User avatar
Posts: 6004
Joined: May 12, 2013
Location: Dallas, TX

Re: z wave energy meter DSB28

don't recall when I upgraded to 6.11, so the negative numbers could be old.

and for accum Energy next value less than previous value can be correct after a reset.

BUT anyway. The error rate is too big. Instead of trying to chase some bad numbers, can we fix the root cause? Should I get in touch with AEON and send them the data? I found the same issues with 3 units.

The very big numbers and negative numbers could be bad transmission although is z-wave not using a checksum? The small error ie 12345.66, next value 12345.55 seems to be a measurement problem. But they make an error correcting algorithm very difficult ie a lot of IF's and highly tuned to not throw away good data.

Karl

Posted on
Thu May 22, 2014 6:45 am
matt (support) offline
Site Admin
User avatar
Posts: 18732
Joined: Jan 27, 2003
Location: Texas

Re: z wave energy meter DSB28

I'm not sure where the problem is: firmware bug in the energy module, firmware bug in the Z-Stick, firmware bug in the Z-Wave communication stack (caused by collisions or noise), or a USB level or driver problem. I will check how Indigo is handling the checksum for that particular packet to make sure there isn't something Indigo can do to easily identify the bad packets.

Image

Posted on
Tue Jun 24, 2014 1:34 pm
bschollnick2 offline
Posts: 1324
Joined: Oct 17, 2004
Location: Rochester, Ny

Re: z wave energy meter DSB28

matt (support) wrote:
I'm not sure where the problem is: firmware bug in the energy module, firmware bug in the Z-Stick, firmware bug in the Z-Wave communication stack (caused by collisions or noise), or a USB level or driver problem. I will check how Indigo is handling the checksum for that particular packet to make sure there isn't something Indigo can do to easily identify the bad packets.


Whatever happened with this? I'm considering picking up the Zwave DSB28 energy meter as well, since our electric bill was outrageous a few months back.. I'm wondering if there was someway I could of actively detected it, or countering RGE's claims...

Does the data just go into the Indigo log? Or does it accumulate in a state in the device?

Does anyone know how accurate this is? Is the data reliable?

------
My Plugins for Indigo (v4, v5, and v6) - http://bit.ly/U8XxPG

Security Script for v4 - http://bit.ly/QTgclf
for v5 - http://bit.ly/T6WBKu

Support Forum(s) - http://www.perceptiveautomation.com/userforum/viewforum.php?f=33

Posted on
Tue Jun 24, 2014 3:13 pm
kw123 offline
User avatar
Posts: 6004
Joined: May 12, 2013
Location: Dallas, TX

Re: z wave energy meter DSB28

I have 3 of these. They are ok. But sometimes they send bad data ie negative numbers or realllllu large numbers. In about 0.3% of the cases. The following number is normally ok.


Sent from my iPhone using Tapatalk

Posted on
Tue Jun 24, 2014 3:30 pm
kw123 offline
User avatar
Posts: 6004
Joined: May 12, 2013
Location: Dallas, TX

Re: z wave energy meter DSB28

The send watt and KWh. KWh can be reset. The installations can be done by experts in the fuse box.


Sent from my iPhone using Tapatalk

Posted on
Thu Jul 10, 2014 12:51 am
haavarda offline
User avatar
Posts: 640
Joined: Aug 18, 2012
Location: Norway

Re: z wave energy meter DSB28

I am considering a DSB28 to measure the wattage to 1 circuit. This circuit is charging a electrical car so I am dependent on clamp measurement due to the high current. It runs three wires to the charger. Live, neutral and ground. Do I have to measure on Live and neutral to get this device to work? In other words: do I need 1 or 2 clamp version? How is this device displayed in Indigo (child devices for each clamp)? And is the wattage updated instant if the circuit is turned on/off?

Hope someone that has this device can share some light on this :)

Håvard

Håvard

Posted on
Thu Jul 10, 2014 8:11 am
kw123 offline
User avatar
Posts: 6004
Joined: May 12, 2013
Location: Dallas, TX

Re: z wave energy meter DSB28

only live.
Always only one wire through one of the clamps (if you run live and neutral through ONE clamp the magnetic fields would cancel each other.) I believe they sell a two and three clamp version

you could also put a clamp around the neutral wire, but that will measure the same. The total would be then twice the Watt or KWh. As you have access to both values in indigo you could use this to check if both measurements show the same. They sometimes show bad measurements like negative numbers of realllllllly large numbers This way you could kick out bad values.

As you have AC
in the up phase the current runs from your fuse box through the live wire to your car and back through the neutral wire.
and the down phase the other way around

the AC current creates a magnetic field around the wire and the clamp measure that. if the net current is zero i.e. if you have both wires in one clamp there would be no magnetic field. And it only works with AC not DC.

hope that helps

Karl

Posted on
Thu Jul 10, 2014 9:37 am
RogueProeliator offline
User avatar
Posts: 2099
Joined: Nov 13, 2012
Location: Baton Rouge, LA

Re: z wave energy meter DSB28

First, I would double check the wiring to be sure that you actually have a live, neutral and ground on 120V. Never hurts to verify -- I only say that because many car chargers are 240V, so you could have two live wires going there; code does not require a new wire color, so some electricians re-use the white wire as a live load wire in these cases. With electricity, never hurts to verify everything twice before proceeding!

The energy monitor does not seem to react to events (such as no current flowing), it simply reports on a schedule. However, that interval is user-settable (I assume there are limits, but you can change it). Given that it is obviously a powered device, more frequent polling should be ok.

You define the device in Indigo and it will create child/pseudo/whatever you call them devices for each clamp.

That doesn't answer all your questions, but hopefully will help a bit.

Adam

Posted on
Thu Jul 10, 2014 11:47 am
haavarda offline
User avatar
Posts: 640
Joined: Aug 18, 2012
Location: Norway

Re: z wave energy meter DSB28

Well, fist of all thanks for the physics lesson Karl :wink:
You confirmed what I already know. Since this device seams to be available in so many variants I was unsure which type I should use. I have used a similar non z-wave device were you were supposed to clamp on all the clamps, and the device calculated the power based on this. I guess like an average or something like that.

Here in Norway we use 230 V, and I am pretty sure that it is a live, neutral and ground wire. This is because it is connected by a normal schuko plug.
http://en.wikipedia.org/wiki/Schuko

Thanks for all your inputs.

Håvard

Håvard

Who is online

Users browsing this forum: No registered users and 1 guest