Android app for the Classic

[binkino]

Hi

I use
192.168.1.222:502 (48v)
and
192.168.1.223:502 (24v)
worked before with version 2.1.5 and again I am back on 2.1.5

It was after reboot past of midnight


In 2.1.5 I can also set both Classic in each tablet in the network setting, but use only one of the classic each tablet for displaying the datas, no problems, no interfering

[Graham]

worked before with version 2.1.5 and again I am back on 2.1.5
...
In 2.1.5 I can also set both Classic in each tablet in the network setting, but use only one of the classic each tablet for displaying the datas, no problems, no interfering

...
How long will it accumulate data for the screen? App and screen must stay active?

hmm, looks like I'll have to add an option to disable the System Tab that was added in 2.2.0,
In order to display the System power and consumption, the app connects to all classics to sum the power and calculate the consumption by subtracting the whizbang current. This only make sense when multiple classic are charging one battery bank in follow me mode, not when you have independent classics with each their own battery bank with different nominal voltages.


The Live chart plots an entry for voltage, supply and battery current every 2 seconds, after six hour of plotting, the page swiping wasn't very smooth so I might have to limit the amount of data plotted.
If the app is in the background it can be unloaded by android if it needs the memory so you need to keep it in foreground to accumulate data for the graph but it doesn't have to remain on the live chart tab to plot the data.

[Westbranch]

well I managed to do a screen capture of the LIVE CHART and there are some interesting observations.. seen/see-able below I hope...  note how low the Batt V goes! every time the fridge kicks in...

[Westbranch]

this one is clearer

[Westbranch]

Vic asked a question on the other thread to which I will investigate a poor 2V cell, high resistance connection,   etc.. this weekend.

Is anyone else seeing this type of pattern if they use the latest version with Live Chart?

ie is this an artifact of the 2 sec sampling rate, or

[Graham]

well I managed to do a screen capture of the LIVE CHART and there are some interesting observations.. seen/see-able below I hope...  note how low the Batt V goes! every time the fridge kicks in...

It might be invalid readings from the modbus service or a bug in the chart,
the charge state (line at the bottom of the chart) should display as follows;

Resting: Black (code 0)
Absorb: Orange (code 3)
BulkMPPT: Green (code 4)
Float: Blue (code 5)
FloatMPPT: Cyan (code 6)
Equalize:  Pink (code 7)
HyperVOC: Red (code 10)
EQMPPT: Violet (code 18)

you have blank gaps in the green line that indicate it's not one of the valid charge states or the x axis values are out of sequence
I'll let my test system run this weekend to see if I get it to occur


Graham.

[Westbranch]

Thanks Graham, I did note that some of the dips were when the fridge kicked in, not watching constantly but looking over when the fridge started +- 1 min. or so, but I see on full screen , not 7 inch, that there is one group of 2 full dips that has 1 short and one long break in the green line...

I will try to replicate the graphs this weekend...

[Graham]

Did catch one this morning, 
I'll have to look through the android logcat data to try figure out whats going on,
I suspect I'm plotting a reading for the whizbang modbus register (4360...) while not getting a response for the volt & supply current (register 4100 ...)


found a use for my chromecast device, here is the live chart on the big screen

[Westbranch]

found a use for my chromecast device, here is the live chart on the big screen

Great Idea....makes for easy viewing compared to a 7" screen...

[Graham]

I have an updated beta apk in the dropbox location (https://www.dropbox.com/sh/z04q8hpw5ibpwfd/AACbRXpllos-IuLR6Z4pGWska?dl=0) if you would like to help me test it.

The live chart drops to zero where due to a modbus network disconnect, the apps service thread running in the background sends out a modbus read request on the network, when it gets a reply, it broadcasts the data to the listeners in the app (gauges etc), when it doesn't get a reply, it broadcast a charge state code of -1 to clear the displays etc. This was causing the gap in the green line on the graph.

The latest beta will not chart the disconnects if a reconnect succeeds, if it does get a SOC code of -1, it will display it a white segment.

Andy,
I also added a new setting to disable the system view when each classic is independent (not in follow me mode, separate batteries...). When in this mode, the SID for PVOutput is assigned to the current selected controller, allowing you to have a PVOutput for each CC.
When the 'System View' is not checked, the app should only connect to one CC at a time. Make sure you restart the app after you uncheck the setting to make sure it disconnects from the other CC.
There was a lot of changes to the code to support this, especially the PVOutput uploader, it will need a fair amount of testing before it can be published to google play.

Let me know if this setting works for you.

I still have to address some performance issues with the chart.

If you see anything weird with the chart, please don't assume there is something wrong with your system since this is a beta version of the app.



Graham.

[Westbranch]

Hi Graham, well loaded the new version, 10 minutes on , looks good ...  will advise if I see any drops on the screen or   ...  waiting for the clouds to part this morning, ......it's coming slowly..

[Westbranch]

OOPS!            

Perhaps I mispake myself Too  SOON..... the white bits are showing up in the SoC line and the Batt Voltage drops to zero...

[Graham]

OOPS!            

Perhaps I mispake myself Too  SOON..... the white bits are showing up in the SoC line and the Batt Voltage drops to zero...

oops, think I might have put the 'if' statement in the wrong spot,

another update (2.4.2) is now in the dropbox location (https://www.dropbox.com/sh/z04q8hpw5ibpwfd/AACbRXpllos-IuLR6Z4pGWska?dl=0)

[BioPower]

Appears to working nicely now, thanks Graham.

Is there a how to somewhere to show what to do to be able to use the app via the net?

[Graham]

I got it working with noip.com but it took a few hours for the dns entry to replicate to the dns servers.

the first thing I did was to setup and test port forwarding;
- I added a port forwarding rule with the local IP address of the classic (192.168.0.22),  the local port (502) and the public port (I used 39502). Then saved the changes and rebooted the router.
- Once it rebooted I made note of the wan IP address my router was assigned by my service provider.
- I then disabled wifi on my phone to force it to use the cell network, opened the android app and added an new device entry using 39502 for the port and my wan ip for the address and it connected to the classic ok.

DDNS setup
- back to the router page, selected the ddns tab
- entered my noip.com username/pw and hostname as per the attachement
- a few hours later I was able to ping the hostname (skyetracker.ddns.com)
- disabled wifi on the phone again, created an entry using the hostname and 39502 for the port and it connected to the classic.

My noip hostname is skyetracker.ddns.com if you would like to try connecting to it, it's a classic that's just connected to a car battery for test purposes, I'll leave it running for a couple of days.


Graham.

The router I have is from nexxt solutions. But I found all the settings so far as described in the local app manual. But am still no able to access over the internet. I'm not sure what Imy missing. The iPad dress of my provider seems to be dynamic since I have a new one everything I reconnect. So I opened the account with noip.Com which my router even showed in the drop down menu in dynamic dns.

Here is a previous post related to accessing your classic remotely,

Step 1 is to setup port forwarding
Step 2 is to use a ddns service to map your wan IP to a network name


Google port forwarding with the make of your router