Victron voltage metering accuracy?

[tanglewood]

I have very limited experience with Victron, and am wondering how accurate their voltage metering is. Let's assume a Multi or Quattro inverter/charger utilizing remote battery sense. And lets assume an MPPT with short, fat cable runs to the batteries. In other words, let's leave out voltage drops through cable runs.


Does a Multi/Quattro report voltage that is accurate to within what? 10mv? 20mv? 50mv? 100mv? And what about the MPPTs? Within what range of difference is it reasonable to expect all devices to report the voltage?



Obviously with LFP, voltage accuracy matters when it comes to charger voltage settings, and the settings are only as good as the metered voltage. I think 100mv is a bare minimum, with 50mv or less preferable.


For those of you with Victron equipment, what have you found in practice?


I'm asking because now that I have my LFP system on line, I am finding that my Schneider Conext equipment has shirt for metering accuracy, making it very difficult to get the charger settings correct for a variety of different charging devices, potentially working in different combinations. I would say it's unsuitable for use with LFP from what I've seen so far. Or at least requires a lot of trial and error experimentation to get settings that work.

[CatNewBee]

It is sufficient accurate for LFP.
Just tweak the settings to the conservative side.

[_andi_]

some folks did some measuring...



https://community.victronenergy.com/...-accuracy.html

[tanglewood]

Originally Posted by CatNewBee

It is sufficient accurate for LFP.
Just tweak the settings to the conservative side.

How accurate are you seeing? 100mv? 50mv? 10mv?

[Jim Cate]

Tangle, you may have seen this thread recently:

http://www.cruisersforum.com/forums/...nt-210853.html

I had similar questions, and it generated a lot of flak... and I still baven't found a good voltage standard to check my meter against.

Jim

[CatNewBee]

Originally Posted by tanglewood

How accurate are you seeing? 100mv? 50mv? 10mv?

Sufficient, because your voltage loss across the terminals, the wires, the fuse, the solenoid / relay contacts etc. from the controller to the battery cell under load exceeds this range by far, we are talking about 0.01V, that is not critical to a LFP battery. Even 100mV or 0.1V on the 12V side boils down to 0.025v at the cell. It makes no difference if you cut off charging at 3.650V and the battery has 3.625V (at 100mV spread under load), when the safe range is up to 3.8V per cell. Also the measurement is at the controller, so the real voltage is always lower at the cells than what the controller measures, the measurement is consistent and stable and in fact the error is +/- 10mV.

Same applies for any other charge source, the battery voltage under load (charge current) is always lower then the voltage at the terminals of the charger.

Also the voltage spread between 3.625V to 3.650V is a small difference of maybe 3% of the SOC, so you end up at 97% SOC in contrast to 100%, not bad either.

We can of course discuss the wiring and implement the connection via helium cooled supra-conductors to mitigate the voltage drop.

Technically in the real world the accuracy of the voltage meter inside the controller is sufficient to charge the battery and cut off at a safe level without overcharging the cells when set properly.

It is also not necessary, that all charge sources cut off exactly at the same voltage. You may than end up charging witch shore power at 96% and on solar at 97%, it does not matter where the electrons come from.

BUT what is VERY IMPORTANT is to disable all smart features, battery maintenance features like EQUALIZATION and also TEMP SENSOR of the controller, they are meant to adjust the voltages to the temperature for FLA charge profiles, trickle charge to fight sulfation etc. by magnitudes of 100mV, that can kill a LFP for sure on the long run.

It sounds stupid, but a smart charger is a dumb thing when it comes to LFP.

[noelex 77]

I have three Victron MPPT controllers.

They are acurate (on the battery side. I have not checked the input voltage) to within +/- 90mv, but note I have not tested them in hot weather and the error is a little variable with temperature. Also note this is 24v system.

This is a good result especially given the cost of the units. The very fine 0.01v voltage set point adjustment allows systematic error and even some variable temperature error to be dialed out. This is particularly useful to balance the results of multiple charge sources.

I also have a small cheap Victron charger and this is much less accurate although most of this is consistent over reading error of arround 140mv so it likely to be a calibration issue with my particular unit. I dont think this is likely to be representative of the larger Victron chargers.

Of course these are readings directly at the terminals voltage drop in the wiring will add to these errors.

[CatNewBee]

It really does not matter, the BMS is the last level of defense if a charger goes rogue anyway.

In this forum many user vote for conservative settings to stop charge at much lower SOC, but it is not a necessity.

From 3.55V to 4.0V cell voltage (that is within the limits) we are on battery level between 14.2V and 16V.

3.75V should be the level when at least the BMS stops the charge bus as over voltage protection, that translates to 15V battery voltage.

Stop charge should be set to 14.5V or below, that is 3.625V at the cell, while Winston considers 3.65V as 100% (14.6V - this is also the GEL battery set point for absorption)

From 3.55V to 4.0V we are already on the shoulder of a LFP battery, with 3.65 on a steep increase. That said, not much Energy is charged between this 2 set points and it is OK to stop charging somewhere in between without loosing too much SOC capacity (around 5..10% depending on the charge current, the lower the current, the fuller the cell)

That said, the battery could be disconnected at 14.2V to 14.6V, this are 400mV difference with not much effect on the battery and not much harm.

Precision of 10mV or 100mV really do not matter in this equation, more important is the time you stay at this levels and the current flowing.

So why not stop charge at 14.2V you may ask.

Well this depends on your BMS and his requirements. Most BMS start balancing at 3.6V (14.4) and use the Absorption time to equalize the cells as part of the battery maintenance at the end of the cycle. If you never go above the 14.3V on a slightly imbalanced system, the cells will not be treated and you may end up with cells lets say at 60% SOC (around 3.48V or so) and a full cell with 3.6V. If you charge further with low current, it will be balanced, if you charge with high current you may trip over voltage protection at a SOC of 70%,on the other side you may hit the bottom when using 60% of the capacity and consider your bank aged and damaged. In fact your cells are not that bad, but just unbalanced.

A healthy bank is one with balanced cells.

Another option could be to start balancing earlier - like at 3.5V (14V for balanced cells), but most cheap cell modules do not allow you to tweak the parameters. A good BMS decides if balancing is necessary based on many parameters, like balancing start voltage, charge current, voltage difference of the cells etc. Cheap cell module BMS just measure the single cell and start burning energy at the set point. They just heat with 1 or 2 Amp the compartment when the cell reaches the Voltage independent of the other cells. That said, at the end of the charge all 4 cells burn 1..2A or heat up the compartment with 15..30W. If your charger only delivers 1..2A output at 14.4v and is set to cut off at 14.6 for instance, he will never hit that point and the BMS will heat up the compartment for ever. A good BMS will quit balancing as soon as the cell difference is less than 10mV, no energy will be burned and the 14.6V set point will be reached easily.

What you really want is reproducible cut points independent of ambient temperature, no fancy stuff like trickle charge with high voltage impulses, no equalization programs that run once a week or once a month.

[noelex 77]

It is good question r the forum. It is easy to develop a false impression of accuracy especially on instruments with a resolution down to 0.01v.

For comparison the good mid level (not excellent) Fluke 77 multimeter has a DC voltage accuracy of +/- 0.3% (plus 1 count).

So measuring a 24v battery the specified accuracy is about +/- 70 -80mv.

Often good quality equipment will do better than their specifications, which are a worst case, but a meter like this is designed to nothing other than measure parameters such as voltage as accurately as possible (while keeping the cost reasonable). You can purchase higher accuracy (and more expensive meters). 0.05% is not unusual for the top quality handheld meters, but expecting this sort of accuracy from a solar regulator or battery charger is optimistic.

[CatNewBee]

This is also a big problem with "cheap" drop in LFP batteries. They have cheap BMS with cell modules inside that monitor only the cell and burn energy from a set point, also disconnect a daisy chain for OVP and LVP with some relays inside. The whole thing is in a sealed enclosure where heat dissipation is not optimal. The 15..30 W can heat up the battery when charging near full and shorten the life of the cells, not because they do not like high cell voltages, but because of the stress by high temperatures.

Better BMS do not burn excess energy for balancing but do it actively by voltage converter, it remains cooler. Even better set ups are, where the battery and BMS are not in the same enclosure, so no unnecessary heat at the cells and cooling can be done optimally.

[tanglewood]

Originally Posted by noelex 77

I have three Victron MPPT controllers.

They are acurate (on the battery side. I have not checked the input voltage) to within +/- 90mv, but note I have not tested them in hot weather and the error is a little variable with temperature. Also note this is 24v system.

This is a good result especially given the cost of the units. The very fine 0.01v voltage set point adjustment allows systematic error and even some variable temperature error to be dialed out. This is particularly useful to balance the results of multiple charge sources.

I also have a small cheap Victron charger and this is much less accurate although most of this is consistent over reading error of arround 140mv so it likely to be a calibration issue with my particular unit. I dont think this is likely to be representative of the larger Victron chargers.

Of course these are readings directly at the terminals voltage drop in the wiring will add to these errors.

Thanks for the concrete answer. Just what I was looking for. The system I'm dealing with now has a metering spread of about .8V between the solar chargers, mains charger, and battery actual. And to make matters worse, the spread varies over the course of a charge, presumably due to heating in the active charger. It's annoying to have to do a bunch of mental math to calibrate each device, and adjust settings to accommodate their meter error. And a double pain when it's a moving target with unknown factors that make it move.


Re the whole question of "how much is good enough", I'm thinking 100mv is a good number. Accurate enough to be a respectable implementation, and accurate to 25mvpc for a 12V bank. A means of programming in an offset would be nice too.


Can anyone speak to a Quattro, or Multi, or Skylla?

[rgleason]

Originally Posted by CatNewBee

Heavily Condensed and Paraphrased ..BMS is the last level of defense if a charger goes rogue.

3.55vpc-4.0vpc (14.2V-16V) LFP Limits, already on the high shoulder (Energy stored -5-10%C capacity depending on charge current. Lower charge rate results in > SOC)

3.75vpc (15v) BMS stops, over voltage protection
3.65vpc is on steep shoulder.
3.625vpc (14.5V) Stop charge -Winston considers 3.65V as 100% (14.6V)
3.60vpc (14.2V) Most BMS start balancing, absorption time used to equalize cells.
3.57cpv (14.3V) If charging never goes above this on an imbalanced system, it can leave cells at roughly 60% SOC (about 3.48V) and others at 100% SOC (3.6V).

..If you charge further with low current, it will be balanced, if you charge with high current you may trip over voltage protection at a SOC of 70%,on the other (low) side you may hit the bottom when using 60% of the capacity and consider your bank aged and damaged. In fact your cells are not that bad, but just unbalanced.

A healthy bank is one with balanced cells.

Another option could be to start balancing earlier - like at 3.5V (14V for balanced cells), but most cheap cell modules do not allow you to tweak the parameters. A good BMS decides if balancing is necessary based on many parameters, like balancing start voltage, charge current, voltage difference of the cells etc....

CNB

What do you think of operating in a more conservative range for the charger and only occasionally top balancing to 3.60vpc (14.2V) ?


How often do you think LFP should be top balanced? Bottom Balanced?


How do you llke the OrionJr BMS now?

[CatNewBee]

Originally Posted by rgleason

CNB

What do you think of operating in a more conservative range for the charger and only occasionally top balancing to 3.60vpc (14.2V) ?


How often do you think LFP should be top balanced? Bottom Balanced?


How do you llke the OrionJr BMS now?

It depends, when living on board, you have a decent cycle each day, at night the battery discharges by lights, entertainment, navigation, fridges, warm water, watermaker.... you name it. At daylight solar re-charges. Sometimes he can charge to full, sometimes not.

There is no reason to not charge to full at this scenario, so settings at 14.4...14.5 V are ok, but you can cut off at 14 if you want without losing much capacity.

Very different on storage with very little discharge, like AIS bilge pump, alarm system, web cam, whatever you want to run when not on board. You must re-charge to not go too low, but you dont want to stay at full. You can set charge end to 13.5V...13.6V or something similar, so the battery stays / charges to aronnd 50..80%.

Regarding balancing, I am pro balancing, but with a good BMS. I have set my thresholds accordingly. My BMS supports top and bottom balancing and has lots of parameters to tweak. Check the thread below, i discussed it in detail there, top balancing is the one when staying at the upper end of SOC, bottom balancing prevents long cut offs when one cell is low and distributes available energy across the cells if possible.

Over time balancing causes very few energy transfers and keeps the cells in synch, what is very desirable..

I have no experiance with the Orion Jr BMS, never seen one yet.

[rgleason]

Originally Posted by CatNewBee

It depends, when living on board, you have a decent cycle each day, at night the battery discharges by lights, entertainment, navigation, fridges, warm water, watermaker.... you name it. At daylight solar re-charges. Sometimes he can charge to full, sometimes not.

There is no reason to not charge to full at this scenario, so settings at 14.4...14.5 V are ok, but you can cut off at 14 if you want without losing much capacity.

Very different on storage with very little discharge, like AIS bilge pump, alarm system, web cam, whatever you want to run when not on board. You must re-charge to not go too low, but you dont want to stay at full. You can set charge end to 13.5V...13.6V or something similar, so the battery stays / charges to aronnd 50..80%.

Regarding balancing, I am pro balancing, but with a good BMS. I have set my thresholds accordingly. My BMS supports top and bottom balancing and has lots of parameters to tweak. Check the thread below, i discussed it in detail there, top balancing is the one when staying at the upper end of SOC, bottom balancing prevents long cut offs when one cell is low and distributes available energy across the cells if possible.

Over time balancing causes very few energy transfers and keeps the cells in synch, what is very desirable..

I have no experiance with the Orion Jr BMS, never seen one yet.

That's right, you wrote:

I chose the REC ABMS with custom programming, optional second temp sensor, serial touch display and the programming software package. Great BMS, active balancing, highly configurable and 4 programmable outputs - 2 NO/NC relay contacts and 2 opto-coupler.

http://www.cruisersforum.com/forums/...ml#post2639756