question

ulfertg avatar image
ulfertg asked

Voltage difference between MPPT 100/20 & Smart Battery Sense

Hi all,

I noticed that the MPPT 100/20 (joined with a Smart Battery Sense) reports a lower voltage than the Smart Battery Sense.

It seems to be lower as a function of the current delivered to the battery like this:

V_MPPT100/20 = V_SmartBattery_Sense - 0.current(A)

Example:

Smart Battery Sense reported voltage: 13,50V (and most likely actual V, as verified with two different measuring devices).

MPPT 100/20 reported voltage: 13,40V

Current to battery at this moment: 10A


Could this be some sort of cable loss compensation in the MPPT 100/20 (which probably shouldn't kick in while the Smart Battery Sense is connected and reporting and could hence be a bug?)?

Can it be turned off, so that the MPPT 100/20 reads the exact voltage which is reported by the Smart Battery Sense?


I bought both devices a couple of days ago and they updated their firmware when I first connected them, so I assume they are on the latest FW.

bug report
2 comments
2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

christian-bell avatar image christian-bell commented ·

How is the battery sense connected to the charge controller? Are they bluetoothed together?

If there was voltage drop the mppt would see a lower voltage than the voltage measured directly at the battery i think. Sorry i don't have answers just trying to work through it.

0 Likes 0 ·
ulfertg avatar image ulfertg christian-bell commented ·

Yes, they are connected in a VE.Smart Network via BT.

The MPPT should receive the voltage from the SmartBatterySense and work with it, but apparently it calculates a current dependant voltage drop into it.

0 Likes 0 ·
4 Answers
ulfertg avatar image
ulfertg answered ·

To sum it up: today I received a beta Firmware which eliminates the bug.


While I'm happy that I received a solution, I must say it was a very time consuming and frustrating journey to get there.

During the last couple of weeks I wrote mails back and forth with support.

I explained the issue over and over again.

Wish there was an easier way to report simple bugs like this.

4 comments
2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

rpetrovski-forum avatar image rpetrovski-forum commented ·

what firmware version did you get? For which gizmo? I'm having similar troubles with 4.03 BMV712 and 1.50 MPPT100/20. At the moment they're supposedly talking via DVCC but I think there was the same issue when I used VE.Smart via bluetooth group.

Also, Despite being configured to go into absorbtion (see time markers) at 14.1V, MPPT did that at 14.2V which at the time was 14.3V on BMV-712.

I guess I can live with it, but looks extremely lame on Victron side.

0 Likes 0 ·
1600955865813.png (233.6 KiB)
ulfertg avatar image ulfertg rpetrovski-forum commented ·

I got 1.52_A05F for MPPT 100/20.

I guess / hope that the bugfix will be incorporated in future FW releases.

0 Likes 0 ·
rpetrovski-forum avatar image rpetrovski-forum ulfertg commented ·

FYI, the 1.53 rolls back that change. No idea why, but once you upgrade Victron Connect, you won't be able to change settings unless you update the firmware. Current is 1.54


v1.53 – 06 October 2020

Released only for all models.


Changes:

  • Remove improvement, added on v1.52, of the remote voltage sense compensation to speed up filtering. That improvement is postponed for now.


Bug-fix:

  • Text messages now always sent every 2 seconds and no longer interrupted when a hex message is received. That could reduce the risk of the MPPT control to get unresponsive when on the “connecting”, as well as the “fetching history” screens.
  • History total data now always announced together with the history day0 data reducing the risk of the MPPT control to get unresponsive when on the “fetching history” screens.


v1.52 – 15 September 2020

Released only for all models.


Changes:

  • Include low power mode for the small models. Idle current consumption (when PV is around 0V) is reduced by around 70%.
  • Improve remote voltage sense compensation to speed up filtering when necessary and improve voltage control response.


Bug-fix:

  • Fix remote voltage compensation for small models. The voltage compensation of small MPPTs did not consider some current depended voltage drops, causing the MPPT to have a higher voltage setpoint, up to 200mV, depending on the MPPT current. This FW version fix that.


Add new models:

  • Include support to BlueSolar 150/45 rev3 and SmartSolar 150/45 rev3.
0 Likes 0 ·
ulfertg avatar image ulfertg rpetrovski-forum commented ·

I just checked my devices (MPPT 100/20 FW1.54 & Smart Sense), while charging with 8A they both report the same voltage (previously the MPPT would report an 80mV lower voltage while charging with 8A)

As far as I understand the change-log Victron removed:

Changes:

  • Improve remote voltage sense compensation to speed up filtering when necessary and improve voltage control response.

in FW1.53, but not:

Bug-fix:

  • Fix remote voltage compensation for small models. The voltage compensation of small MPPTs did not consider some current depended voltage drops, causing the MPPT to have a higher voltage setpoint, up to 200mV, depending on the MPPT current. This FW version fix that.

I have to admit though that their description of the bug fix seems misleading to me. It would make more sense to me if they had written something like: "The voltage compensation of small MPPT's did falsely calculate some current depended voltages drops while receiving the accurate voltage readings remotely ".


Side note: thank you for making me aware that there is a Victron Professional sub section which provides change logs and other useful information.

Didn't discover that previously.

0 Likes 0 ·
Kevin Windrem avatar image
Kevin Windrem answered ·

If the voltage drop varies with current then it is most likely due to voltage drop in the wiring from the MPPT. 0.1 volts @ 10 amps means there is 0.01 ohms in that circuit. You can minimize this drop by either shortening the distance between the MPPT and battery or increase the wire size. 10 feet of #12 wire (5' distance to the MPPT) is 0.016 ohms.Upping this to 10 gauge drops resistance to 0.0 ohms (50% less resistance). With 8 gauge, resistance is only 0.006 ohms.

5 comments
2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

ulfertg avatar image ulfertg commented ·

But this is exactly the reason why there is a Smart Battery Sense in the network: to read the exact voltage directly at the battery, without any drop due to the resistance of the wiring from battery to MPPT.

If it was the issue you describe then the voltage reported by the MPPT should be higher than the voltage at the battery (and without the Smart Battery Sense and thin/long cable it is). In my scenario it's the exactly the opposite, the MPPT reports a lower voltage than the Smart Battery Sense reads directly at the battery.

0 Likes 0 ·
Alexandra avatar image Alexandra ♦ ulfertg commented ·

@ulfertg

The difference is the battery sense is not carrying amperage it is just reading. Resistance increases under load and voltage drop happens that is why the mppt will read differently. Also, as was mentioned cable length and size, as well as strand count matter.

In any case is 0.1v a huge enough difference to cause an issue?

Also the mppt is also not a measuring instrument. So the difference may even be a rounding issue. Battery sense rounding up and the mppt rounding down.

The information given will be a pretty good indicator of what the system is doing for most users.

Have you measure the terminals yourself at both points?

Most managed batteries have systems in place to prevent any problems in any case, in my experience. This drop has not been an issue for charging there are many other factors with greater impact on system performance

0 Likes 0 ·
ulfertg avatar image ulfertg Alexandra ♦ commented ·

@Alexandra

Well, either I still haven't managed to explain the issue properly or I'm completely missing the forest through the trees.

In case it's the former I'll try to explain it again:

I have a MPPT 100/20 connected to solar and a battery.

I have a Smart Battery Sense connected directly to the battery's terminals.

The MPPT 100/20 and Smart Battery Sense are connected wirelessly via Bluetooth in a VE.Smart Network.

The MPPT reads the voltage and the temperature wirelessly from the Battery Sense.

It applies a voltage correction to the reported voltage of the Battery Sense as a function of the electric current delivered to the battery (V_MPPT100/20 = V_SmartBattery_Sense - 0.current(A)).

By doing this it renders one of the two functions of the Battery Sense useless: the reading of the voltage directly at the battery and hence bypassing any potential cable losses of the MPPT.


In case I'm mistaken: What am I missing?

0 Likes 0 ·
Alexandra avatar image Alexandra ♦ ulfertg commented ·

@ulfertg

I think I also misunderstood what you were questioning there as well. I will assume that you know why there is a difference in the voltage readings.

Do I understand correctly that you are saying that the mppt is not using the voltage reading from the smart battery sense for its charging calculations? Even though it has been smart networked?

Or that it is using it and now the mppt is charging the battery incorrectly?

My other question is which way around are the terminations on the battery, as i also know this makes a difference with resistance etc. Do you have the terminals for the battery sense under or over the mppt terminals?

I would verify what the data is showing with an actual reading with a meter. Graph reading and data are dependent on sample rates. Sometimes a meter is just better.

Just another piece of info. Most batteries with battery management prevent over charging in any case. For instance a well known brand we use has winston cells and they can be charged up to 3.85v (according to their datasheet) a cell but the bms will limit them to about 3.35v per cell. The bms will say it is 100% but in reality it is not. (I appreciate that lithium are not meant to be kept or charged to 100% - but this is off topic). But for the consumer it is 100% capacity. What I am really saying is the difference is not a problem for a well self managed bank. And if the bank is still drawing charge from the mppt then it is probably trying to internally balance and needs the power to do so, whether passively or actively balanced, so mppt's algorithm is also based on what the battery is taking the amps or not among other things not just voltage readings, or at least that is what I understand at this point.

0 Likes 0 ·
ulfertg avatar image ulfertg Alexandra ♦ commented ·

@Alexandra

The thing is: I believe there shouldn't be any difference in voltage reading in the first place, that's what the Smart Battery Sense is for in the system.

It should read the Voltage directly at the battery's terminal and send it via Bluetooth to the MPPT, so any potential cable loss should be left out of the equation.

But nevertheless the MPPT assumes a lower voltage (while connected with the Battery Sense) than is actually present at the terminals.


If the issue was in any way related to cable losses and resistance, the voltage reported by the MPPT should be higher than the voltage at the battery's terminal (during charging of course).

But it's lower.


And this is indeed a problem in my opinion. Of course certain BMS (not all of them) can prevent an unhealthy overcharging of LFP batteries, but I understand that to be a fail safe rather than a regular function.


And no, I don't know why there is this difference. At this point and without looking at the code I can only guess. And my guess is, that it is a software bug (you could look at my reasoning in the opening post).


Of course it still is possible that I'm completely mistaken and I'm open to this option.

But then it would be very helpful and appreciated if someone could explain the behaviour of the MPPT, which to me doesn't make sense at all.

0 Likes 0 ·
ulfertg avatar image
ulfertg answered ·

Is there any way to directly contact Victron about it?

It really looks like a software bug to me.

The way it's implemented now makes voltage analysis worse with the Smart Battery Sense in the system.

3 comments
2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Leslieanne avatar image Leslieanne commented ·

Hi, I can confirm what you are seeing and have reported exactly the same issue with mppt100/20 and BMV in smart network configuration. I have described it in great detail here:

https://community.victronenergy.com/questions/33706/smartsolar-10020-showing-incorrect-low-battery-vol.html

This problem appeared with the major update around 1.39 firmware. Before that the mppt would switch to absorb too early, after the update would switch at too high a voltage even though in the app it was reporting the correct voltage.....so unless you looked at the reported BMV voltage at the same time carefully and/or measured things you would not notice!!!!

Definitely a bug. No response from Victron at the time, may have just gone under the radar. I simply do not use the smart network anymore for the sake of my lithium batteries!!

0 Likes 0 ·
Leslieanne avatar image Leslieanne Leslieanne commented ·

https://community.victronenergy.com/questions/33706/smartsolar-10020-showing-incorrect-low-battery-vol.html?childToView=33732#comment-33732

Last link seemed not to work. This was from 11 Nov 2019

0 Likes 0 ·
ulfertg avatar image ulfertg Leslieanne commented ·

Thank you for confirming this issue.

And also helping me understand why there are not much more reports about it - it must be much harder to notice when the MPPT is hooked to solar.

I connected it to a programmable power supply for testing purpose.

Because of the stable power input to the MPPT it was quite easy to see the difference between the reported voltages. And to understand the linear dependancy of the reported voltage drop to the amount of electric current flowing to the battery.


Now we only need a way to make Victron aware of it.

If my theory is correct this is probably easy to fix: the assumed "voltage drop due to cable loss compensation" subroutine should not be called if there is a voltage reporting device present in the network which reads directly from the battery.

Victron, can you hear us?

0 Likes 0 ·
wkirby avatar image
wkirby answered ·

Hi

I have a 100/20 and a Smart Battery Sense. All have latest firmware.
I did manage to replicate your problem in my setup. Sorry for the poor photography.

After some time this does settle to the correct Voltage I have noticed.

Victron are already aware of this issue.


img-4129.jpg (1.0 MiB)
2 comments
2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

ulfertg avatar image ulfertg commented ·

Thank you for confirming this issue.

You wrote that it did settle to the correct voltage after some time. Could it be that the battery was full then and there was no significant current to the battery anymore? This would confirm my theory of the erroneously applied current depending voltage compensation.

Since when is Victron aware of this issue? I'm asking because I'm still in the return window of my Smart Battery Sense unit and I'm trying to figure out if I better send it back. Since it makes the charging process worse the way it is programmed now.

If there is any chance that Victron is going to fix it in about the next couple of weeks I'd love to keep it though.

0 Likes 0 ·
Show more comments

Related Resources

Additional resources still need to be added for this topic