Battery Temperature ### (in data only)

An area for discussing the OutBack FLEXnet DC product.

Moderators: OutBack, OutBack Moderator Team

Post Reply
Forum Member
Posts: 17
Joined: Fri Aug 04, 2017 6:24 pm
My RE system: Radian, Mate3, FlexNet, FlexMax

Battery Temperature ### (in data only)

Post by SwDoctor » Fri Aug 04, 2017 7:00 pm

I will try to explain properly the issue, please read thru before posting a fix, because it is not the same as I have seen other users post about temperature sensor data...

have a RTS connected to the Radian, have also tried another connected to FlexMax to rule that out.
The temperature IS reported properly to the MATE3. So the "system" is getting the data.

The problem is whenever you logon to the MATE3 "web-server" the temperature ALWAYS displays ###
(the same is true about the data exported from the mate3 to my automation system)

Is there any fix for this?

Mate 3 Firmware: 002.017.000
GS Firmware: 001.004.001
MX60 Firmware: 002.001.000
FNDC Firmware: 001.001.071

This issue has remained since installation (year(s)) thru several FW updates.
I do NOT want to upgrade to a firmware that does not support "OFF-GRID" (no internet) web-serving to devices on the LAN. (if that is a suggested option)

The battery temperature data can be very useful to my automation system programming, it is not just a desire to view the temperature for knowing. I want to make automated decisions about power usage with temp considerations.
- I could hard-wire a new (separate) temp sensor for the automation, but importing the data from MATE3 requires no wiring or expense or additional failures, and it will report the temperature as the system sees it. (if it ever worked.)

I have imported battery state, voltage, shunt data, KW, etc. (in the same manner) to the automation system keeping me from adding a myriad of sensing devices - at least where the power system is concerned. It worked for every other data field, except battery temperature (Which ALWAYS reports "###" even when MATE3 "displays" a valid temperature reading.)

My guess/take:
The data is internally valid and being read from the sensor, but not being exported...
(This would be due to a firmware coding bug that no-one seemed to notice thru several firmware releases.)

NOTE: The "data" exported concerning "Battery Temp" is included within the FlexNet DC data port. This seems odd, but it is true. And this is why I posted this question within the FlexNet DC forum instead of the MATE3 forum. It would seem that the FlexNet DC would have an RTS port on it...

A "sample" data dump: (you can see the issue)
"Dev": "FNDC","Enabled": ["A","B","C"],"Shunt_A_I": -17.0,"Shunt_A_AH": -137,"Shunt_A_kWh": -
'6.830,"Shunt_B_I": -2.0,"Shunt_B_AH": -14,"Shunt_B_kWh": -0.710,"Shunt_C_I": 0.0,"Shunt_C_AH":
'13,"Shunt_C_kWh": 0.670,"SOC": 88,"Min_SOC": 75,"Days_since_full": 0.3,"CHG_parms_met":
'false,"In_AH_today": 518,"Out_AH_today": 468,"In_kWh_today": 28.130,"Out_kWh_today": 23.250,"Net_CFC_AH": -
'138,"Net_CFC_kWh": -6.910,"Batt_V": 49.4,"Batt_temp": "###","Aux_mode": "manual","AUX": "enabled"}]}}

Any information would be greatly appreciated.

User avatar
Forum Czar
Posts: 703
Joined: Mon Nov 28, 2005 3:20 am
Location: Ontario, Canada

Re: Battery Temperature ### (in data only)

Post by intellact » Sat Aug 05, 2017 11:25 am

Hi Rick:

I can't help you with why the webserver data is putting out the battery temperature as "###". It sure sounds like a MATE3 bug to me. ("###" looks like some kind of value-formatting token instead of the actual data.)

I also don't know how you do automated system decisions - perhaps you are quite comfortable doing all of that coding work yourself.

However, I would be remiss if I didn't ask whether or not you had considered third-party software (that does NOT use the internet) that already has very sophisticated OutBack system automation tools that you can easily define in plain-English. :smile:

The Automatic Actions function of WattPlot Monitor PRO supports a large number of system conditions (including battery temperature, voltages, amperages, modes, time-of-day, etc.), that can be combined into multi-condition triggers with trigger delays, state holds, and more. :cool:

(And I can assure you that the data sources that WattPlot Monitor uses (and there a few) will NOT get the same "###".)

See the feature described here.
Andrew Welch, creator of WattPlot

Forum Member
Posts: 17
Joined: Fri Aug 04, 2017 6:24 pm
My RE system: Radian, Mate3, FlexNet, FlexMax

Re: Battery Temperature ### (in data only)

Post by SwDoctor » Sat Aug 05, 2017 7:36 pm

I tried the wattplot trial version, it will not allow LAN connection to mate3, so I am left unknowing, and judging from the software website my scripting does the same thing (plus more), so I am not needing to purchase it for the one data field.

If WattPlot Monitor PRO can PROPERLY get the data from the MATE3, then it should be available using the method I used to grab it. Can anyone confirm that it actually DOES work in wattpro? (or does it display ### with wattplot also).
It probably would not be worth it to integrate with another software package just for that ONE field of data, but it may be revealing to know how WattPlot software gets the data in realtime without "###" being returned.

If you can confirm temp is visible is wattpro, let me know the firmware of your MATE3 and I will get to the bottom of it somehow...

I never knew about wattpro before this, honestly I looked for something to view outback data 2-3 years ago and never found it. But I really went out with the scripting since then.

Heres a sure to be long story about my system and why I would like to use that data;
(and why I am reluctant to switch to anything else)

I use homeseer, in particular the "hometroller" hardware embedded with homeseer software.

The hardware uses only 15watts. That is what sold me on the system...
In fact, with all supporting components and isolated PSU losses, the entire automation system draws under 30w.
(My toughbook laptop consumes that much!)

Homeseer is mostly programmed using its web interface within the LAN (or internet if you have it), it uses VBscript with a very sophisticated event handling that can make extremely complicated decisions based on inputs from anything imaginable, (hardware or virtual software) not just power related, but easy to integrate with doors, lighting, and custom probes. It responds and replies using voice, and a lot of other interactive stuff I don't use much. It is rather impressive actually - it even has built in solar and lunar times of interest, such as "solar noon" solstice etc. Handy for solar decisions (like uncovering production losses you may never know about otherwise.)
An example of its use:
I have recessed LED effect lighting that color changes based on homeseer events, flashes on security events, pulses on system events, and color determines class of event (blue=water related, red=security, yellow=solar, etc,etc,etc).
[example1; flashing two times red with pause and repeating indicates zone 2 security alert]
[example2; gently pulsing blue indicates solar water heater is (over producing) dumping load geothermally]
These lights are in several key locations to be easily seen thru house from anywhere.
-this type of thing is fairly easy using homeseer.

(There is about 15miles of wire JUST for automation system control & sensors, not including AC POWER, DC SYSTEMS, or SOLAR WIRING.)

My first inclination was to use an industrial PLC system to automate the house, but found homeseer was much more cost effective & capable (& web interface too!) - and easily integrated with the outback system with a tiny bit of code (ok, it was a nearly a thousand lines of code, but I had the unfortunate opportunity to sit for a couple months).

With homeseer I am able to assign a "device" (in data/automation device) to be populated with data, say- the "battery voltage" (as an example) whereas the program polls the Mate3, gets the data and populates the "device data" for each individual device. (however many as needed, value and/or string, plus status). Once the data is populated, decisions can be made using the device's data.
For example, if the "battery voltage" device value is below 46.0 then "turn off all lights after being on for more than 15minutes, but only if 'discharging' and whenever 'bay door 3 is closed' for more than 5 minutes and 'well pump has been off for more than 5 minutes'".
Device examples (above) are "Battery Voltage", "Discharging", "Bay door 3".
Whereas each can have "ON/OFF" status, or a numeric value, or a string value. (ie: 46.0, ON, "CLOSED")
(a strange logic example perhaps, but you get the point).

The MATE3 is not the only device on the LAN that homeseer "looks at", it also monitors the solar water heater controller thru its web interface to confirm it is operating properly, and the program can adjust settings on the water heater device without my concern, and compare the results with external sensors to confirm it is accurate and not faulted, and to adjust the target temp depending on weather/occupancy etc.

It all works -flawlessly after 2 years so far, automation certainly has lots of practical uses for off grid use.
Most automation systems are used to change channels or adjust lighting when a movie starts, that's not my thing...
(I don't care about "lazy person automation"... to me this is about power management & security).
An ACTUAL system event example is "turning off kitchen when no one is sensed in the living/kitchen area for >1hr" (killing a solid state relay that feeds the kitchen appliance circuit) this saves ~200watts of phantom loads for microwave, coffee pot, venthood, dishwasher [bosch digital], toaster oven, etc. [all that crap has digital clock/displays, etc.] So if no one is within proximity to use these appliances (and taking into account that the power circuit to the kitchen indicates that no such appliance has been used within the last hour) then it is safe to turn it off until motion is detected again. While away all day, or sleeping for 6-8hrs, this saves a lot of power and no user intervention is required. And no occupant has ever "noticed" that my kitchen "dies" and comes back every so often...
That's just ONE "REAL" example. (there are too many to explain)
BTW: the 3 fridges are on separate dedicated circuits unaffected by the kitchen "countertop" circuits. Fridges only turn off when the food in the target fridge is thawed beyond hope based on temp probe inside each compartment, and only would allow power down if battery is not charging when faulted for duration. All refrigerators are monitored thermally, electrically (for cycle), and door closure, and each has it's own "kill" via SSR. (with fail safe)

This is how I get the house to operate "like on grid" whereas the occupants never need to know it is off grid, and if they leave something running (even plumbing) the computer can isolate the "leak" or faucet left running where there is no occupant, and close a zone valve (or kill a circuit) to stop the power/water waste (and filling up the septic tank) while leaving the rest of the house operational. (works for toilet stuck levers and supposedly water leaks too). The custom "master LCD display" (serial driven LCD display - driven by homeseer also) will indicate the issue in detail, and pressing "reset" button returns everything to normal, clearing the event.

There are hundreds of things it watches for, including security issues like triangulating radio/cellular signal strength to determine if someone is on the property -approaching, and if detected it alerts within (and/or closes security shutters/doors/gates if away/sleeping).
Off grid weather data is on the same system, detecting lightning strikes, rain, wind, temp, baro, humidity, and solar output provide plenty enough data to calculate and predict basic weather, with data understandably used to control automatic doors and ventilation exchangers, but also "counting & rating any clouds" to predict solar day production before solar noon, giving a 'rating' of how much power is estimated to be "expected" to be wasted, so the "waste" can occur (ramping up) as the day progresses instead of "all at once/ALL ON" once the battery is charged, because turning A/C units on high AFTER the battery is full may be too late in the hot day to keep the house comfy, but going full blast too soon will result in battery not getting a full charge because it was wasted on the A/C units too soon!

It works so well at predicting/adjusting the remaining solar "curve" and cloud ratings that I cannot tell when diversion starts by reviewing the graphing harvest charts! The curve is so "exact" all the way until the end of day it looks "textbook". The only exceptions are days when lightning is detected and rated to be increasing, or cloud rating "increase rate" is not steady, whereas charge is prioritized under the premise that a storm/clouds are possible that may obscure a full battery event. Even on those days, the power adjustment is hardly noticeable on the production chart. I really did not expect such favorable results from a 15w computer, but I swear by it now. The first year had some bugs in my program, and it is spot-on for two summers and one winter so far.
In fact, I purchased a spare hometroller just to keep for a spare. It has no fan being an atom processor, so it should not fail easily running on conditioned redundant isolated source supply fed from the 48v battery bank. But I wanted a spare anyway- this system just works too well to not have it.

Another important off-grid task managed by automation is the battery watering cycle.
Homeseer controls a battery watering solenoid valve, and takes input from a flow sensor.
The water supply comes from R.O. system thru D.I. filter, and implements a water quality alarm (PPM ALARM) set at 10ppm. The quality (PPM) alarm trigger is fed to homeseer as an input.
It takes into account the amp-hours in/out of the battery and waters based on usage, (or every 14days- whichever comes first). It keeps records of water used, and adds water when floating when possible. Errors/anomalies are recorded and reported such as short/long flow times (than expected) etc.
-Another effort to make off grid seem like on-grid. No battery watering to worry about and much less corrosion.
(with 72 cells, it would take a better part of a day...)

"out of the box" homeseer is as dumb as the cardboard box it ships in. It requires considerable programming to do anything special, especially to make it pull data from the MATE3 and use it, it takes a bit of scripting/hacking (within homeseer, not the MATE3). But I did most of it when I was forced to sit in a wheelchair recovering from a broken ankle a few years ago, right after bringing the system live. Now that I am mended I don't know how I had the mental focus to write that code when I was swigging that much Jack Daniels - instead of the addictive prescription!

If anyone else needs the code/scripts (or advice), I can share info and the source fragments that you need, but the entire code would require your house to have same EXACT setup, down to the inputs/outputs that would be a clone of my entire house (down to the exact doors, gates, valves, solar config, sensors, and even placement of plumbing!). Otherwise it would need revisions. It would only work on homeseer also, since it is integrated.

Only homeseer or a similar total home automation system could do this level of integrated logic and event based programming, plus- many of the sensors and interfaces were custom made in my on site electronics workshop since they are not commercially available to do the crazy things I wanted it to do.

Now you see how I wanted to use the battery temp data to improve high current times - to give consideration to the battery temperature before allowing high-current loads to switch in, if too hot. Or by turning something else off before switching... (...72 forklift batteries are not cheap enough to ignore the little stuff...)

I do agree, it seems to be a bug that programmers just overlooked for several releases of firmware.
I have a spare hometroller, mate3, and a spare Radian, hub, and RTS, so I can bench test a suggested fix, and have even considered delving into the firmware myself and fixing it- but I try to stay busy doing real work if possible. I have another fence row to complete and am digging a pond, so this is not that important yet.
But I wish it would be addressed by outback before i kill something trying...

User avatar
Forum Czar
Posts: 703
Joined: Mon Nov 28, 2005 3:20 am
Location: Ontario, Canada

Re: Battery Temperature ### (in data only)

Post by intellact » Sat Aug 05, 2017 8:26 pm

Hi Rick:

Sorry that we flew under your radar a few years ago. :wink:

Clearly, you have advanced very far down the path of automation, and I can see why you would not want throw much of that away. Indeed, the kind of automation you have undertaken is somewhat different from the automation built-in to WattPlot. You manage home appliances etc. WattPlot focuses on controlling renewable energy system components. The only external logical connectivity right now would be via component AUX ports being switched on or off.

I'm still not sure that we can help you though. :sad:

WattPlot can connect to a MATE3 via UDP/IP, USB, or Modbus TCP. It does not use the MATE3's web page data. The UDP/IP and USB protocols use the MATE3's inherent network and serial data streams respectively, which are specifically designed for external monitoring software. The more recent Modus TCP interface goes directly into the MATE3 and can read/write any one of the hundreds of settings and measured performance data.

Speaking of which, I can assure you that one of the hundreds of WattPlot installations worldwide would soon have let us know if the battery temperature was coming out as ###. I have never seen that in any WattPlot/MATE3 communications. :wink:

Congratulations on what sounds like a really neat home automation project. Perhaps you could look into acquiring a third-party temperature sensor...


Andrew Welch, creator of WattPlot

Forum Member
Posts: 17
Joined: Fri Aug 04, 2017 6:24 pm
My RE system: Radian, Mate3, FlexNet, FlexMax

Re: Battery Temperature ### (in data only)

Post by SwDoctor » Sat Aug 05, 2017 8:59 pm

You are using streaming data, and the firmware must have gotten that correct concerning temperature.
My method was parsing the dump as a file, and that's the method the web page uses.
(that's where the firmware is fragged.)

I could switch to use streaming, but I only needed to sample by request, and already have the massive program written using that method.

BUT: I do have a totally redundant system (spare parts) whereas I can try to make use of the streaming data instead.
If I ever get time to tackle that for such a menial data element.
(when I could just toss a sensor in the battery room).

I still wish the system was not flawed- That is true, but am happy it is not a more important fields of data...

I left a reply under Wattplott Poll about some ideas for features on your software.
Hope it helps.
And you are certainly a pioneer in the datamining outback, I thought I was the only one needing data that was not on the cloud.
(I dont do cloud for 2 reasons. 1;no internet on the outback 2; I don't like to give ANY data to a computer I DO NOT OWN.) <-- i'm a dying breed there, but does a cloud based home security system seem like a good idea when I ask if the guy working in Ireland at Amazon Cloud Services has a partner in your town who wants to "burgle" your home when his friend can remote control your system or at least report when you are away?
-What about a home automation system being on the cloud? How scary is that crap- Is anyone awake anymore?

"no thanks cloud. I live off solar, I don't like clouds..."

-backups still work, I have not lost ANY data in over 25 years... (clouds not welcomed)


Forum Czar
Posts: 904
Joined: Thu Feb 17, 2011 11:12 am
My RE system: VFX 3648, FM80, 3k panels, Mate3, FNDC (retired), 8ea L16 mixed. Wonderful wife.

Re: Battery Temperature ### (in data only)

Post by JRHill » Mon Aug 07, 2017 7:31 am

SwDoctor wrote:... Can anyone confirm that it actually DOES work in wattpro? ...
Yup. Works fine,

Post Reply