Profile cover photo
Profile photo
electrodacus electronics solar off-grid
electrodacus electronics solar off-grid


Post has attachment
I updated the SBMS user manual needed to make new photos with the SBMS screen as to much has changed from original version and there are links on last page of the manual to newest firmware 3.2h the only version that will support the DMPPT450 as there where some recent changes there also

The DMPPT450 testing works good I'm almost done hope to be done this weekend and just need to cut the silicone pads and assemble the case then do a final shorter test testing also the individual cables then maybe even before the end of the week I can start shipping. Will need to ask for the address on Kickstarter as I will start shipping to those backers first.

I also started working on a DMPPT450 user manual but not sure it will be done before I start shipping hopefully I will have something by the time you start receiving the DMPPT.
Below you can see a simplified diagram of the DMPPT450 (very simplified but I think a good representation of basic functionality and may make you understand better how is all connected). I made this as it will be used in the user manual.
Add a comment...

Post has attachment
I made multiple changes to both DMPPT and SBMS firmware this week and updated the SBMS user manual (I should release that next week).
Today was cloudy with a few sunny breaks and I did a full day test after I update the SBMS firmware last night.

I also update the html page to include the DMPPT450 data (second table) and I even managed to find a solution to display the DMPPT450 graph (the last one in orange).
I was already using almost all the WiFi module buffer of 2kbyte but I use the existing PV2 variable to store the DMPPT450 data and in PV1 variable I include PV1+PV2 and use the DMPPT energy counter to decide if I will be displaying DMPPT data in PV2 variable or PV2 data if the DMPPT energy counter is zero.
This of course required quite a few changes in the SBMS firmware and in the HTML page. This way the new HTML page is still compatible with an SBMS without DMPPT.

I took quite a few screenshot's during the day but before describing what happens in there I should mention the current setup.

There are a total of 22 x 260W PV panels connected to DMPPT450 in the flowing configuration (not ideal but good for testing).

TPV1 - nothing connected.
TPV2 - 3 x 260W
TPV3 - 7 x 260W
TPV4 - nothing connected
TPV5 - 6 x 260W
TPV6 - 6 x 260W

TLD1 - 2x 1.65Ohm in parallel so about 0.83Ohm heating element.
TLD2 - 2.2Ohm heating element
TLD3 - 2x 1.65Ohm in parallel so about 0.83Ohm heating element.

Those 4 x 1.65Ohm are the heating elements in floor heating in the living room still need to install another 8 of this in the rest of the house hopefully before winter :). The 2.2Ohm just heats a 20liter (5 gallon) stainless steel pot full of water.
There are two separate thermal groups controlled by the two thermal sensors one is made out of TLD2,TLD3,TLD5 controlled by Temp235 and the second one made of TLD1,TLD4,TLD6 controlled by the Temp146 but both of this sensors where just measuring the room temperature in this test not connected to thermal mass in any way.

You can notice that I added the time in the top left corner with red so I will refer to that when discussing about the data.

The first image is at 12:31 and what you can see is that while cloudy the battery was already fully charged at around 11:30 in the morning and from that point the battery started discharging as there is a new future when SBMS is used with DMPPT to keep the battery at lower SOC after the first charge in this case setting was at 90% SOC meaning that battery will be maintained in the 85 to 90% range during the day and not fully charged until next day.

There was about 727W load on the battery about 90W for computers and LED lights and my small fridge and from inverter 638W that was constant from a water distiller (I made 4 liters of distiled water with about 3.5kWh so I saved 2CAD :) as that is how much distilled water costs at the store).
As the distiller was inside the house all this 3.5kWh eneded heating the house also.
On the DMPPT450 side the TLD3 heat element was ON with about 32.7V taking around 39A so 1282W that is because it was cloudy just a few minutes before there where 53A so two of the heat elements where ON.
All 22 panels where contributing to this and you can see how much from each TPV 2,3,5,6

The peak was at 14:35 when there where about 50A going to battery (limit was set at 50A) and 96A where going to all heating elements so total 1350W to battery plus 3161W to heating about 4511W total but while more was available from the PV array (notice the 32.9V not quite max power point) there where no additional heating elements so there was no place for the extra energy as mentioned not perfect setup but is what I had now for testing.
The 22 x260W = 5720W of PV panels so at least another 0.83Ohm heating element will have been great to get all the available power in those short periods where it was sunny.

This current configuration is about half of what the DMPPT450 can handle so 40 to 44 panels max

You can also notice that internal DMPPT450 temperature was great with 36C minimum in the morning and 44C max and this indications are a bit high as the temperature sensor is inside the main micro controller so is like CPU temperature not quite the board temperature but the die temperature that is a bit higher. For this reason limit is set at 70C instead of 60C on the SBMS to allow for this higher reading.

By the end of the day 12.8kWh where produced in a sunny day I will expect way more than 2x maybe 3x as much if setup is optimized with additional heating elements. I should get up to around 50kWh with my full setup.

You can also see a screenshot with the row data for those interested in how it has changed.

Not sure what else to say let me know if you have any questions.
8 Photos - View album
Add a comment...

Post has attachment
Just a short update on the DMPPT450 progress.
It was a marathon but the DMPPT450 main board's are build I will just need to test each one but before that I need to make some sort of custom firmware so I can more easily test all the functionality.
Then I need to cut the silicone thermal pads and also punch holes in them for the screws (they are needed between this main boards and the aluminium PCB for better thermal contact) and I also need to build the communication cables needed between SBMS and DMPPT450.
2 Photos - View album
Add a comment...

Post has attachment
I had a lot of work on the software as there where quite a few small "typos" that created a lot of extra work for me understanding the random intermittent behavior.
Today was a full day test did not touch the DMPPT450 and it seemed to perform great with no unexpected issues so I think I'm almost done with the only thing that need's additional testing being the heating digital max power point as currently just a single heating element is connected about 500W that heats (boils) some water outside.
Today it started working around sunrise (I was sleeping quite a few hours after that :)) and the battery was full very fast in just about an hour after sunrise.
There are 26 panels connected but I think there are about two or three of those low cost inline fuses outside that are damaged so maybe around 23 working panels (will need to find better fuse holders before winter).
so about half or so of the PV array that can be supported by the DMPPT450 and of course about 4x more than what will normally be connected to SBMS with this size battery (about 6 panels) thus battery can charge up to 4x faster in early morning or when is cloudy with this larger connected array.
The charge current was limited to 50A and the charge SOC was limited to 90% (this future can only work with DMPPT450) so first charge of the day is done at 100% SOC but then the next charge will only start at 89% and stop at 91% thus keeping the battery around 90% the rest of the day.
You can notice that barely no energy went to battery as there was not much need in this nice sunny summer day with the huge 23x260W (~6000W) or so PV array.
As soon as the battery was fully charged the water heating element started working (red graph in the photo) as the PV array current up to that time was below 50A max accepted by battery so there was no extra for the heating element. After that time the heating element worked for about 10h continues so the entire day (was just for testing had no need for so much hot water) and then as the end of day got closer and available PV array power dropped the heating element stopped while battery was being charged.
So all seems great i2c communication is solid now but I needed to change the firmware on the SBMS also so unfortunately just the new 3.1h version of firmware will support the DMPPT450.
So the only thing remaining to be done is to test with multiple heating elements and see how that work to get max power point and interaction with battery charging.

One thing for those that prepare to setup the SBMS and DMPPT is that DMPPT450 should be on the left of the SBMS and while I only left about 5cm (2") between the two I suggest to leave more space at least double will be comfortable so 10cm (4") as it will be easier to access the 8pin side connector when installing or making modifications.
I think the default communication cable provided will be about 15 to 20cm (6 to 8") so if you need to have them further apart that that then you will need to ask me for a custom longer cable or extend that yourself. Not sure what the limit will be in therms of length but I will try to keep it as short as possible. You also need to consider that the two PV outputs form DMPPT450 are also connected to SBMS120 so is more ideal for them to be close.
My heatsink as you see is made of a aluminium plate about 80cm (32") long around 16cm (6") wide and about 1cm (0.4") thick and spaced about 2cm (1") from the back wall so I can route cables behind that and for better cooling.
I did a test a few days ago where I limited the current to 100A so I was charging around that for about an hour and + the 15A or so heating element so around 110W and the heatsink temperature got to about 38C from the about 26C ambient with the hottest component getting to around +45C so thermals are very good s those current shunts where at about their max limit and less than 10C above heatsink temperature.
That hot spot was the PV1OUT going to SBMS120 while passing 55 to 60A for about half an hour. So excellent as that is the upper limit for that path.
Those are 4x 1mOhm current shunts in parallel so while they see 15A each (60A/4) that will be just 0.225W each (they are 1W rated) but since they are close to each other that group there will dissipate 0.9W plus the copper trace and the near by mosfets ad up a bit but with a great heatsink they will stay very cool and last forever.
I wanted to start manufacturing but there was a lot of forest fire smoke so air quality was bad outside and I could not reflow boards that will make additional toxic smoke and so opening the windows for ventilation will not work.
Things start to improve as air quality was better for the past two days so I think I can start production this weekend.
5 Photos - View album
Add a comment...

Post has attachment
I work on the DMPPT450 software (more work than expected).
I wrote the main software for the DMPPT450 quite some time ago and since then when I wrote the part for the SBMS I made changes to the protocol so I needed to remember what I did there on the DMPPT and implement the protocol changes.
The PV redirect to SBMS seems to work today (needs more testing)
You can see in last photo (just made that one) where limit was set at 50A and max current to SBMS was at a peak of 49.152A
Thee is a small portion where I limited to 10A and it did as expected and there are quite a few power cycles as I was testing software updates.
Is quite difficult to test as is on my main system and I need energy also and since is far from computer I need to use a long USB cable with USB optical isolation to upload new software.
The thing is that I can not test all this on the bench as a power supply can not simulate a PV panel exactly (for SBMS and battery charging is simple to just use a power supply) but DMPPT works quite different and needs real PV panels for testing.
I have not installed the case on the DMPPT as I need to measure signals.
Is also extremely hot outside 38C (101F) today (right now) and tomorrow it will be the same. And I know this is not bad for some of your locations but is unusually hot for Sask Canada.
What is even worse is the smoke from forest fires that is always present for the past few days so I can not produce extra heat inside as I can not open the windows at night to cool down as air quality is very bad outside.
If this smoke goes away I can probably start main production on the DMPPT450.
This weekend I try to finalize most of the software.
4 Photos - View album
Add a comment...

Post has attachment
Time for Battery capacity test.
There is about 1 year since I installed my A123 System battery that powers my entire offgrid house.
The photos are in the order they where taken so I started in the morning charging the battery even added a few more panels on PV2 to charge the battery faster and after the first time it went to 100% fully charged I disconected the panels on PV2 so charge current is lower and added a 25A DC load just to spped the discharge a bit so it will start charging again at around 10 to 12A then one more time before starting the discharge test.

My settings are 3.50V Over voltage so when the highest cell gets there the charging is stopped and End of Charge is set at 3.48V thus the charge re-enable voltage will be 130mV lower (fixed setting inside the ISL94203) so highest cell needed to drop to 3.35V before charging will start again as it seems there is about 2Ah between fully charged and when the charging is re-enabled thus for this battery just a bit over 1% SOC

Now battery is fully charged indication is 99% and PV are physically disconnected from SBMS120 so no more charging so test is started by resting the energy counters.

Now you can see how battery voltage will progress with highest cell voltage and delta to lowest cell as follow:
99% -- 3.458V -- 10mV (1.3A)
89% -- 3.288V -- 7mV (88A)
80% -- 3.286V -- 10mV (87A)
80% -- 3.303V -- 4mV (28A) (this is a few seconds later from above).
70% -- 3.263V -- 12mV (87A)
60% -- 3.263V -- 7mV (50A)
50% -- 3.257V -- 7mV (50A)
39% -- 3.249V -- 7mV (50A)
30% -- 3.230V -- 7mV (50A)
20% -- 3.205V -- 8mV (50A)
10% -- 3.170V -- 9mV (50A)
1% -- 3.108V -- 12mV (50A)
0% -- 2.998V -- 34mV (31A) (point where the test was stopped).

The entire discharge test took just a little bit less than 4h so average discharge current was around 48A (0.25C) with peak at 88A (0.5C) thus absolutely no trouble for this battery.

Now while at 3.5V this battery was actually fully charged (there is no capacity gain to go to 3.6V) on the lower end I stopped the discharge at about 3V so there was still something down to 2.8V but not much and official tests go down to 2.5V but again not much extra capacity to be had going there as voltage will drop fairly fast below 3V

Now to compare this with the initial test after installation made one year ago you need to read this old post here

That test was done a bit differently as after the cells where fully charged I disconnected the PV and continued normal house discharge usage and so the discharge test was about 2 days instead of just 4h for this test so average load was a bit lower so that test had a bit of an advantage over this one but not much.

If I ignore the difference in test procedure and possible small measurement tolerances and just compare the numbers then that test made a year ago resulted in a discharge capacity of 187.89Ah vs this test made today that had resulted in 186.22Ah then delta in capacity is 187.89Ah / 186.22Ah = 0.9% delta
so less than 1% degradation and even if that is true that is an excellent number as that will mean a battery life of 20+ years to get to 20% degradation from original capacity.

Now on last photo you can see how much was the battery used and what my house used.
So over 8544h (that is about 356 days) the house used a total of 1122.349kWh + 373.944kWh = 1496.293kWh that is an average of 4.2kWh/day or 126kWh/month
Out of all that 806.714kWh went trough battery so in average day 2.27kWh went out of the battery and that is about 47% of the battery capacity that is around 4.8kWh
There are of course multiple small cycles each day but it will be equivalent with 356 cycles at 47% DOD (very rough approximation). The idea is that using the battery this way makes no impact on battery degradation and all the degradation is related to battery calendar aging degradation (small chemical reactions inside the battery).
The largest impact by far on battery calendar degradation is battery temperature and as an approximation each 10C higher temperature doubles the rate of degradation.
My battery is always at a nice +18C to +22C for about half the year (winter) and the other half is at +22C to +26C so ideal temperature.

So battery temperature (that for this LiFePO4 is basically the same as ambient as they do not generate almost any self heat in typical offgid usage) is the most important factor in battery life and cycle is completely irrelevant for this type of cells.

There is quite a bit of detail in here so that in a year time I can do this test almost the same in order to remove as many variable as possible.
18 Photos - View album
Add a comment...

Post has attachment
The Toshiba mosfets from DMPPT450 arrived with no more delays and today I build two preproduction board's that I will test the next few days and if all is good I will produce the DMPPT450
I wanted to do this boards yestarday but of I had again problems with the LCD on the pick and place (the resistive touch screen).
I used to think that some sort of calibration was lost over time while not powered ON but I think the touch screen is failing so I opened the pick and place again and took the LCD model number then while searching for a replacement I seen that most adds mention this is used in GPS and then remembered that my GPS has actually a 5" LCD with resistive touch so I opened my GPS removed the LCD and installed on the Pick and place as while LCD model was different it had the same resolution and pinout seemed to be the same and it worked.
It was a bit tricky to get to calibration menu but after the calibration all worked well. Hope the touch was the problem and now is permanently solved.
Now I need to get another GPS :) but the replacement screen had about the same cost as a low cost GPS so no loss and did not needed to wait for the LCD.
A lot of you started receiving the SBMS for the past week so I was busier than normal answering emails.
Maybe tomorrow I will test my house battery capacity degradation after exactly one year since it was installed. I do not expect to see anything relevant as the expected degradation 0.5% to 2% will be hard to measure as the measurement tolerances are already in those ranges.
2 Photos - View album
Add a comment...

Post has attachment
Just a small update to the user manual v1.2 now same link as before

Is mostly the update's of the links in last page with the latest software version 3.0g that has shipped with the last batch of SBMS.
Add a comment...

Post has attachment
Last night we had a severe thunderstorm mostly lighting as there was no rain. The day before it was very hot (at least for here +36C (97F)) and then around midnight there was a bit of wind and this powerful thunderstorm.
At around might night a close lightning strike damaged my SBMS120 (the one powering the house).
Before that the house was powered just DC as we had our computers and some LED lights on top of the fridge and maybe some other small stuff.
The PV was disconnected from circuit breakers but since now I have single pole breakers that means just the positive side of PV is disconnected.
My system is not grounded and the strike was not a direct one just very close to it was about electromagnetic induction in the house wiring.
The SBMS failed in a safe manner as the LCD was white so microcontroller was not working and the ISL94203 disconnected the Load+
It took me about 10 minutes to remove the SBMS all DC breakers where disconnected then I connected the battery+ to the house DC circuit directly and turned ON the LED light breaker when instant magic smoke escaped from the LED driver in my lab so I turned back off and removed that driver from circuit and then all else worked except no light in my lab.
In the mean time I had another SBMS120 that I installed (the one that failed was a beta prototype that I had installed for almost a year).
After I installed the new SBMS and all was back to normal around 00:42 (I know since I set the time and date on the new SBMS) I took some photos of the thunderstorm. Photos where about one and a half hours to two hours after the incident and I used 1.6 to 2.5 second exposure to capture some lightning strikes. I took a few hundred photos most just black as it was night but managed to capture a few nice shots and seen way more live :) It was quite an active storm with lightning almost every few seconds.
While those lightning bots look close in the photo they where not at least compared to the one that took out the SBMS that was at most a few hundred meters away maybe even on my land.
I seen recently a documentary about lightning and was quite fascinating and it seems most of the mechanisms on how it works are not understood.
I think this was the one I seen
Forgot to mention what failed on the SBMS as I repaired it this morning. The current sense amplifier for the external current shunt failed (the internal input diodes) you can see in the photo where the magic smoke escaped and the microcontroller also failed but nothing else. On the microcntroller also an internal diode got damaged but there is no visible physical damage as on the current sense amplifier.
Somehow the GND (common negative) become more positive than the battery positive so the current shunt amplifier had a more positive voltage on the GND pin than on the current sense input leads that where connected to the external current shunt that in turn is connected on the battery positive side.
And internally the current shunt amplifier has some diodes connected between each of the sense input pins and GND and those lost their smoke.
I seen this sort of fail on the PV input current sense amplifiers when some users connected the PV panels with reverse polarity on the SBMS60 and SBMS100 as there the current shunts where first after the PV connectors and on the new models SBMS40 and SBMS120 I had the current shunt moved after the ideal diode.
The microcontroller failed for the same reason an internal protection diode on the 1.8V voltage reference pin that is used for the ADC.
Not sure how I can protect against this sort of fault while not affecting the measurement accuracy.

12 Photos - View album
Add a comment...

Post has attachment
I started shipping this week and up to now I shipped to all of you that had paid the shipping cost. Now I need to start contacting all the others so I can also ship those.
Then I will start working on the DMPPT450 software and testing that until the msfets get here so I can build the DMPPT450 then I can finally ship all the remaining devices.
When all this is done I will do another Kickstarter for those that did not get the chance to order one from this batch and I will be designing a new low cost SBMS0 that will be similar to SBMS40 but without any power electronics and missing the WiFi so it can be small and front panel mount. It will use the same 2.4" LCD and have all the SBMS functionality.
There are quite a few people that wanted a portable power pack and they needed a small front panel mount BMS and it will also work in a power wall type application where it can be integrated with something like a Victron multiplus inverter/charger where SBMS0 can control both the inverter trough EXT IO3 and charger part trough EXT IO4. As some of you know I do not recommend power wall type products where grid is available and reliable as battery cost amortization is still higher than grid cost but in time lower cost higher quality Lithium batteries may become available. As of now nothing has changed and I still recommend LiFePO4 as the lowest cost amortization battery (around $US0.2/kWh best case realistic coat amortization).
Add a comment...
Wait while more posts are being loaded