Smart shunt drift after 16days with no calibration. Who is to blame?

We're charging the battery to 3.45V again and the smart shunt should show the same SOC on the display as 16 days ago. But it shows 2% less .
Is this a problem with the smart shunt accuracy? Have the cells really drifted because of different internal resistance and therefore different heat loses? Or is the SOC inconsistent at 3.45V as we have seen in our test recently (link below)?
Inconsistent charging to same voltage:
• Finding 20% and 80% SO...
Here is the spreadsheet and graphs for this test:
bit.ly/2VYuol1
Please subscribe and join me on my journey of the Off-Grid Garage!
For more information, please check out my website with links to all the batteries, materials, devices and products I have used in my videos: off-grid-garage.com
The 'buy me a beer 🍻' link is here (on the website, scroll all the way down):
off-grid-garage.com Thank you for your support!
You can also click on the Thanks button under the video to support the channel. Thank you, if you have already!
**** These are the batteries you are looking for ****
🔋➡EVE LF280 LiFePO4 cells 3.2V/280Ah: off-grid-garage.com/batteries/
🌞➡Solar, Controllers, Inverters and accessories: off-grid-garage.com/solar/
🔧➡Tools, Analysers and Testers: off-grid-garage.com/equipment/
📈➡ Battery Tests and Results: off-grid-garage.com/battery-d...
🔌➡Batteries, charger and accessories: off-grid-garage.com/batteries-3/
⚡➡Cables, Inverters and Electronics: off-grid-garage.com/electronics/
Get 5% off when buying Bluetti all-in-one Power Storage Products!
off-grid-garage.com/bluetti/
New to AliExpress? Use my referral code below to register and earn US$19 store credit. What a great start! a.aliexpress.com/_mrtR9Ux
Take a live sneak peak at the Off-Grid-Garage in the Victron VRM World:
vrm.victronenergy.com/install...
My Tesla referral code (free 1500 SuperCharger km / 1000mi for you and me)
ts.la/stefan96963

[@kelvincoffey3121]

Considering it’s had 14 cycles and there is a bms and losses in battery, I Am pretty impressed how accurate it is

[@OffGridGarageAustralia]

I thought so too while editing the video. I will leave it as it is now and see how further it will drift or if it even comes closer again.

[@wideawaketotruth5301]

Thanks again Andy, it was necessary for me to rewatch your video a few times. To begin to digest what you were explaining. Andy you never cease to amaze me. Truly I am a novice at best. But you sir are bringing me along at a pace to where I consistently move forward. Cheers and hats off!

[@john_in_phoenix]

All of the above. I also think based on my own measurements that approximately 1% is the least drift you can experience, since that is what I measured for cell efficiency in a much more controlled environment. In other words, if you use 100 amp hours, it takes up to 101 amp hours to recharge the cells. I have actually been measuring slightly under 1%, but you get the picture. Not many real world examples of 99% efficiency, but it really is.

[@OffGridGarageAustralia]

Thanks John. That works with my observation as well. And some of the causes for drifting may even cancel themselves out. Different charge/discharge current, temperature, SOC... this all are variables which could balance each other and make it less and more drift. I have left the shunt uncalibrated now and will see how this progresses over time, say 6 months. I will charge to 3.45V as a defined state (well, this 'low' voltage alone has some drift in it) . Maybe I increase charge voltage to 3.55V afterwards for 6 months to compare. Well, then we hit in the colder season again which causes other... it's getting complicated 😬

[@DMIINC2013]

After extensively playing with smart shunt setting during spring time (such as P.E., charge efficiency, tail current and charged detection time) I have achieved 0.7% accuracy over extended period of time. I think longest was 3 weeks beginning from mid July. Reason smart shunt let's us make those changes is because to achieve the necessary accuracy it can not take into the account your setup. Think of type of wire, length of said wire, gauge of the wire, type of wire terminals, switches, etc.... All of these components have inherent voltage drops across them which in reality relate into energy losses over time. It is worth exploring various settings to achieve best accuracy over prolonged periods of time between being fully charged.

[@charlesrichards5389]

0.7% accuracy is awesome. Wire, switches, and terminals may be a significant factor to the drift if you have inconsistent charge/discharge cycles (in other words, higher load cycles can over-heat conductors, especially if you use cheap stuff). I don't think that's Andy's issue though.

[@aureltanasan1413]

Can you post the settings values that resulted in this small drift? I’m intrigued

[@OffGridGarageAustralia]

Yeah, post these numbers here!!!!

[@DMIINC2013]

@@OffGridGarageAustralia Tail current 1.00% Charged Detection Time 4m P.E. 1.02 Charge Efficiency Factor 98% Nothing special here. Your settings may vary. And let's not forget difference between summer and winter temps effecting capacity. Interesting note. Charge Efficiency factor of 96% for me causes 10% drift over couple of days.

[@DMIINC2013]

@@aureltanasan1413 See response below.

[@JayRSwan]

Great video. I assume you have the settings for the battery set correctly. I am actually surprised it's only 2 %! Really shows how good the smart shunt is.

[@RICHIE26636]

All of the above! It's not a perfect world. LOL Wiring loss, heat is loss of energy. Any heat on the battery side of the smart shunt, the smart shunt will not see. Love your videos. Very entertaining.

[@OffGridGarageAustralia]

Thank you.

[@wayne8113]

Thanks Andy

[@dgriffejoen]

I think it is a combination of all the options, but mostly the difference in charging current.

[@OffGridGarageAustralia]

I think so too. Could some of the options compensate others, cancel the error out, causing less drift under certain circumstances? High current causes more drift but warms the battery which causes less drift, something like this...

[@egnegn123]

Set the parameters mentioned in the manual correctly: • Tail current • Peukert exponent • Charge efficiency • Discharge floor Peukert can be calculated from the results of the discharge tests.

[@doglabdogtraining-gus.8873]

this is one of the best tips on charging and and drift i heard so far , thank you , you just saved me another 50 hours of youtube ha ha ha

[@MrDingaling007]

Temprature differences also affect cell capacity, not to mention the slow degredation after more cycles, that's why they usually produce more AH from brand new.

[@mane46]

Cell: Temperature to internal resistance differences 80% SOC Temp 20°C = 4.2 mΩ 80% SOC Temp 30°C = 3.7 mΩ = 12% < as Temp 20°C 80% SOC Temp 40°C = 3.5 mΩ = 17% < as Temp 20°C

[@craigjameslister290]

Hi Andy, another great and thought provoking video. Could it be that your batteries all have different wear patterns and so you will need to adjust charge settings as they age? BTW I sense a new video about a Daly BMS soon 4:22 ;-)

[@OffGridGarageAustralia]

Thanks Craig! I think the wear on the batteries is fairly even. As they are all in series, the current is the same. The only difference is the internal resistance which varies a tiny bit. That is so small though but over time I think this accumulates and makes the measurement drift. Daly? What? 😂

[@SailingMayhem]

Thanks for a most imressive channel! Prismatic cell concern: I have seen several DIY'ers finding themself in the same situation: One or two of the cells in the pack act differet that the others from the same batch. We know about the expansion when using the full (or near full) chargeprofile at 3.65V. We also know about the more or less fragile terminals on most prismatic cells. So, I'm thinking; since most are using solid bus bars tightening the cells together via the terminal, could that result in making disorder in the terminal inside the cell when there is no "compression" to hold the pack in place? That the combination of busbar, expanion and top craging result in stress on the screws? It has been a concern of mine for a long time that this seems to be a common issue. What you think? Would a connection with a more flexible bus (like a cable) work better?

[@OffGridGarageAustralia]

Thanks for your comment and information. Great question but I would not be too concerned about the forces on the busbars/terminals. I have mine connect in a loose setup with space in between the cells. The contraction/expansion is minimal if there is one at all. In the specs it says up to 0.5mm between 30%-100%SOC. Cables would work better in these occasions but have more resistance due to higher contact resistance.

[@shadowmoon7763]

Hello Andy from Sunny and Hot Florida, would you post your 'shunt' settings to your my-settings page on your website please. Thanks and keep up the outstanding work

[@OffGridGarageAustralia]

Yes, OK, I'll get this video up soon...

[@Pimpasxx]

Hi Andy, my smart shunt has drifted in the opposite direction by about 10% in a week. As a result, I have to reduce the SOC by around 10% every week. Peukert is 1.00 (cannot be lowered), and efficiency is 100% (cannot be increased).

[@Pimpasxx]

:D

[@mdunbar04]

Hey Andy perfect Video idea for OGG. Lithium Solar and Will Prowse have put out videos about cell compression with mixed results and answers. I think with your Scientific approach to this subject is very much needed. Compress? No Compression? How much Compression? How to Compress? Cycle life Compressed? Cycle life Uncompressed?

[@reubenj.cogburn8546]

He has addressed this issue fairly recently. Just checked his video history for your answer.

[@bobfoulke4875]

Awesome video as always. In looking at the data it seems to me that 3.5v capacity curves are better by about 2% than 3.45v in the 5-40amp range tested. Maybe being a little higher up on the curve would provide better long term performance, especially seeing the videos of the wild wattage variations due to the clouds.

[@OffGridGarageAustralia]

Yeah, the higher you charge the less difference it makes how you got there. But remember, charging to 3.45V gets us already to 99.5%SOC. So is it worth chasing these remaining 0.5%SOC? In my opinion, it only stresses the battery for nothing more than 1.4Ah of capacity you could gain.

[@randycarter2001]

Even a slight difference in temperature will cause cell balance to drift over time. Mostly caused by capacity variations. Cell A is 100.001aH, cell B is 100.002aH, cell C is 100.000aH. It adds up over time. You are only looking at it for a few minutes per day while the system is running 24/7.

[@OffGridGarageAustralia]

That's a good point.

[@evil17]

I agree with this, and that there are probably many other inconsistencies & variables throughout the whole system when connected to a solar PV setup 24/7 with temp changes, etc, but this is lab level experimentation too, where we are examining these inconsistencies to try and get some answers for a better understanding of this chemistry. Was it Einstein who said, the “definition of madness, is doing the same tests over & over and expecting to get different results!” Cheers all

[@jornheitmann3206]

Hi Andy, whats going on with the Daly Bms Manual in the background (arround 2:23)?

[@petertulp8806]

Helle Mr Off Grid, I have a quick question about your bms. In case of a low cell voltage cutting off situation, what can you do to fix that? The relais is open, so the battery can not be charged right? Or does the bms detect a charging voltage and closes the relais again for the battery to charge? Thanks

[@Hstruik01]

What is your charging efficiency set at in the shunt . You may have to experiment with that value based on your average charging/discharging current.

[@OffGridGarageAustralia]

It sits on 99% and the Peukert Exponent on 1

[@chuxxsss]

Morning Andy, have you gone the shelves done mate.

[@edwardvanhazendonk]

Isn't the smart shunt telling us it still needs 7.4Ah to reach 100%, hence it knows it's not full even when the cells are 3.45V. Also in your capacity tests you (as far as I remember) didn't charge with a absorption, so you stopped at a certain voltage or current, this will give you other results as well. Also a battery charged with 40A will be a bit warmer than charging with 5A. So it all varies (a bit). B.t.w. great you are educating us about this as it's I think the fine line between chemics and the combination of Voltage, Current, charging variances and Temperature which will get a difference in measuring. Keep up the good work and thanks again for sharing.

[@OffGridGarageAustralia]

Thanks Edward! That is correct, it said there are 7.4Ah missing. But were they really missing or was the shunt just thinking they were missing and the cells were full? I think it's not consistent either as you said temperature plays a role as well as other factors. Regardless what has caused the drift, it will be very hard to reproduce that. The first couple of full charges after the calibration were consistent until I went really deep down and up again.

[@philbrooke-little7082]

I think the difference in internal resistance causes the drift in the balance while the drift in overall capacity is more likely a combination of rate of charge and overall charge efficiency. I find with my own BMS that a charge efficiency of 96% brings it back to pretty much exactly the same point. I do it by it simply deducting 4% of the current being charged when it accumulates negative amp hours. On my lead acids the efficiency using the same method is 80% calculated by Ah in to Ah out calculation. I suppose I could automate that so it effectively learns the efficiency over time and applies it to the soc calculation this should give a system that doesn’t require regular synchronising to keep it accurate.

[@willtaylor8904]

Surely B and C are one of the same thing. In that the higher the charge current the more internal losses you get. Predominantly due to heat effects on the chemistry...

[@i.do.not.exist.9992]

I can see the SOC from my Smartshunt and my JBD BMS. They are close, but never 100% in sync. The moment the cell voltages affect the BMS SOC, the difference becomes even greater between the two, as the Smartshunt does not take into account the volts per cell, only the total voltage of the bank. Using the BMS SOC is "more accurate" on a Victron system in my experience so far, I'm still learning. And it does help a LOT if the JBD BMS, with an RS485 port added, is also connected to the system via a VenusGX/Cerbo or Venus OS on a Raspberry Pi, using an RS485 to USB cable. It controls the charging better.

[@stoomheier]

Nice video. I have 4 x pylontech 2000C connected via a Cerbo to a Victron Multiplus2. Charge with solar energy up to 89% (maximum charge voltage 50.8) (pylontech has 15 cells therefore 3.338 V / cell. SOC 17 April = 89% at 50.66 V, SOC 18 April = 88%, SOC 19 Apr = 87%, Apr 20 SOC = 86%, Apr 25 SOC = 79% every day with a 50.66V. Where is it going wrong?

[@rickard1802]

Imagine a world where everything that happens is expected and nothing changes. Staring at numbers all day would be quite boring ;)

[@OffGridGarageAustralia]

That is so true. Thankfully there is something happening all the time.

[@michaelwilkes2519]

I have the same issue with my victron smart shunt. I'd been wondering if it was due to charging while underload. I noticed some paper work on your desk for the DALY SMART BMS, which I also use. The daly is even worse at tracking SOC than the victron smart shunt. Pro Tip, if the DALY starts showing nonsensical information, unplug the BLUETOOTH dongle, then reattach. I suspect it might not like being close to other BLUETOOTH gear. (Only guessing)

[@am1rb]

all of the above ! how ever for (a)-smart shunt , you can refine it and make it drift less by setting correctly the "charge efficiency factor" in the smart shunt setting by making a controlled experiment of full cycle charge-discharge and calculate the ratio between charged Ah and discharged Ah by this test you will also know the real capacity of your battery . and btw even this parameter is depended on on the charging/discharging current (speed of charge / discharge) as you clearly showed in your experiments - so when you set this parameter do it based on a charge/discharge rate that has minimal effect on this parameter . in my installation i even don't know if it drifts since i always charge to 80% and never discharge bellow 20% so the smart shunt does not have a chance to self calibrate ! however i monitor on the smart shunt the accumulated charged/discharged Ah since last calibration (many months ago ) and I see that the ratio between charged/discharged Ah is kept about constant and converges to the parameter that I set (~98% on my system )

[@bboyda4399]

Andy I think it would be a great test to see how things progress over an entire season, just to see if the drift is predictable or chaotic. It might also end up insignificant.

[@OffGridGarageAustralia]

Yes, that could be a great test. Maybe it does not drift any further and even goes back down again. I'll think about it... I'll try over the next 6 months and see how this goes.

[@edwintorres5850]

I have a question the negative cable from the mppt solar charger where is connected at the battery or at the system minus of the shunt? Thanks in advance

[@monsieurd.6890]

Drifting also depends on battery balacing taht can change a little bit, time it took for bulk charge (depends on the clouds), etc

[@johnbee1107]

Shunt drift depends on your charge/discharge currents. At small currents the shunt is less accurate. Also charging efficiency CEF is depending on SOC. Charging in the flat part of the voltage curve is almost 100% efficient. In the steep part of the curve CEF is more around 90%. SmartShunt cannot not compensate for that. So regular synchronisation is necessary in most applications.

[@OffGridGarageAustralia]

Thanks John. I thought the shunt will be more accurate at lower currents as we have less heat loss in the battery (which the shunt does not account for). The charging efficiency at different SOC is a very good point! I didn't think about that.

[@johnbee1107]

@@OffGridGarageAustralia A problem is the resolution of the analog/digital converter inside the smart shunt. Victron is +- 10mA according to the manual, which is very good. This means at 200mA discharge current you are 5% off. At 20A you are only 0,005% off. In my experience permanent loads under 1A (e.g. alarm systems, wifi router or inverter standby) cause most smartshunt drift. This drift is even bigger at the BMS build in shunt as china BMS builders usually use cheap a/d converters. So I have matching BMS and Victron readings at higher currents but at low loads they drift apart.

[@AveRage_Joe]

I smell moar testssssss coming!!!😀

[@OffGridGarageAustralia]

Hahaha, what, me more testing? No way!

[@aureltanasan1413]

I can read on his face at the end, a plausible answer: it will be on the next testing vid.

[@harrymuurling2742]

Personaly i think C because thats the most logic, I am charging at 3,46 and then half hour CV, it works for me the best, maybe also because i have second 2020 cels from Ups systems but they are very close at 22 mamps

[@leondavibe]

balance happens by adding a resistor to the cell, turning some energy into heat so every balance stint will help the shunt drift imo by tiny amount there is a saying if you pick up enough pennies you will have dollars

[@KevIsOffGrid]

If you wait even longer - like 3 months, you'll find a similar 97-99% at "full" it wont drift further. Of course the shunt has charge co-efficiency you can change and puekert - so it also depends on this. Not been as clinical as you with my playing, but I've fond the victron immensely consistent. I rarely hit "full" either by MPPT settings or using power before it hits full.

[@OffGridGarageAustralia]

I thought so too, that the shunt is fairly consistent it its measurement. I rarely have very small currents which the shunt may not be able to measure, there is always something running somewhere.

[@heinvandenbroeck3184]

I don't know about the Victron smart shunt, but in my Victron multiplus GX I can setup a "charge efficiency" of the battery. This charge efficiency will depend on the internal resistance of you battery. So the internal shunt measurement will be compensated with the "charge efficiency" of your battery and and this will result in a more accurate S.O.C. I think.

[@OffGridGarageAustralia]

I can set this up as well. What number have you put in for LiFePO4?

[@heinvandenbroeck3184]

@@OffGridGarageAustralia at this moment I have set it to 0,98. I am not sure of this efficiency. It is still try and error...it will never be perfect I think.

[@OffGridGarageAustralia]

@@heinvandenbroeck3184 Thank you. Mine sits on 0.99. Google says for LiFePO4 0.99 or 1 is OK. Maybe that's part of the drifting and it accumulates the error over time.

[@ivanio2008]

Good morning A tip for lifepo4 24v After reaching the maximum charge in my case 27.52v. What is the voltage to set to resume charging the battery pack? Thanks

[@SenatorPerry]

C. And F. F, of course, being that as the voltage rises it represents a smaller amp hour capacity. When you have a small inbalance between cells it represents a diminished amp hour as one cell rises .04 to compensate for a battery that no longer has a need to increase .04 at a lower overall voltage. In other words, you are replacing a .04v at 2 AH with .04v at 1.8 AH. Then your balancer with a resistor lowers the .04 inbalance over time losing capacity as the shunt is between the BMS and the Inverter. Over time this small difference takes place hidden from the shunt as it doesn't know what the BMS is doing. My Juntech shunt is absolutely the worst garbage as it also eats up power running the wifi chip on the shunt that over days/weeks it causes a miscalibration as it doesn't know what its own power usages are....

[@fdsf2307]

Hi from germany, will ich need the shunt for the diy 16s lifepo4 battery?

[@hugodeandres1497]

About the charging speeds… y hey just doña matter if you are absorbing. The reason why you get more capacity when you charge slower is due to charging voltage rise. Thus, different SOC at the same voltage cutoff. But if you absorb and ley the currency fall, the voltage rise caused from charging slowly fades and doesnt affect it

[@firpofutbol]

Just bought a Smart Shunt thinking it was never going to have these issues. Why? Because its VICTRON! That bias we all seem to have for Victron products. Even when they perform like average product, we still make excuses for why it is still greater than other products. Frequent recalibration not something I want to have to worry about. Think I'm going to return it now.

[@ab_ab_c]

That is a good question--what is the cause of the smart shunt SoC drift?. I'd have to know a lot more about exactly how the shunt keeps track of the Amp hours & how accurate that process is; but, my guess is that the shunt cannot accurately account for the energy lost as heat during charging & discharging at different rates.

[@uhjyuff2095]

Makes sense.

[@teardowndan5364]

If you count Ah instead of Wh, then internal resistance has no net effect since 1Ah is still the same amount of ions going through the battery's electrolyte regardless of internal resistance. You will still get differences/drift from mismatched internal leakage current, which can be a few %/month.

[@uhjyuff2095]

Nice! I 100 percent agree. These batteries cannot hold a charge very well especially at a high state of charge.

[@martinhansen9517]

My smartshunt drift around 3% in 30 days. Do you have the settings for the auto calibration?

[@OffGridGarageAustralia]

Is it a Victron smart shunt? It's in the settings-battery of the shunt. Three numbers are to be set to reset the shunt to 100%: - charge voltage - tail current - time

[@rogerkerkmann]

I know charge efficiency factor affects the reading but what about the current bleed from the BMS? does that go through the shunt??🤔 Let acid was easy you reset it every time 😁

[@IndependentNewsMedia]

All 3.

[@uhjyuff2095]

Quiz time!

[@daveduncan2748]

Probably all, to some degree. But I think the charging speed probably accounts for the cast majority of it. Otherwise, the smart shunts would be nearly useless. And the batteries would lose capacity like crazy. I think there's a reason that the big brand tell you to charge to 3.6V or 3.65V.

[@daskasspatzle2396]

20/55/25 abc. Just a guess. Thanks for sharing :-)

[@shanesummerhayes7101]

Did you calculate how much unmeasured current the BMS is using??

[@OffGridGarageAustralia]

I showed this in a past video a while back. It is less than 5mA the BMS takes including the relay, so ~0.1Ah per day.

[@jadu79]

I'm guessing B and C on your question but a question is if you intend to test a few other things like putting the balancing should start close to what you average the total volts? - sets the battery charge to 55V and no limit when it should stop and then sets the balancing to start at 3.45V? (realize that your BMS will have to work a little more when you take in 50A than when it's possible, but then you still have energy left over and it should be relatively full at 3.4375V, but without a time limit, you avoid the problem of cloud edge? (just one thought or set the package to 55.1v and it starts at 3.45v?) Another question when you want to streamline most things if you intend to redo a little when you then move the battery and build the new one is my idea that you solder or screw a separate cable and not where the screws to the connectors are (as DM has on her battery with a small screw similar to M3 but if possible solder is preferred) and from it you pull a 1.5mm cable to a distribution connector so 90-98% of the distance to BMS is a "thick" cable and then it is easier to connect fer cables there as in your case have the connector there for your active balancing as well so it is easier to push it there if you want without having to screw anything and think it should be much smoother reading value the shorter the distance are the thin cables which connector is on? (I think the type used for wiring and the thin cables can be folded a few times if they are too thin or a little solder may be a better option)

[@OffGridGarageAustralia]

Thank you for your suggestion. I will connect the balance cables directly to the busbars with an M4 screw with the next setup. It is just easier when I have to change something. I have 1mm cable already here for that and will also fuse all connections for the balancer.

[@jadu79]

@@OffGridGarageAustralia you also think that the connectors used for splicing and branching are easy to use, I have seen everything from 1-1 to 1-3 that are spring-loaded so you no longer even need to screw them as before. Suppose you may have other dimensions but here two sizes are used the most and it is 1.5mm2 and 2.5mm2 in cabling and sometimes where there is only one lamp to have, we use 0.75mm2 and these cables are light and cheap to find. think the contact will also be better when the nut is to cut in that it does it in the right place and not on a ring contact and I do not know where to put the balancing now but I think it is best to have it close to the battery voltage so it starts at 3.45v and packs to 55-55.1v or reached that direction when 3.45 is well when it starts rise faster (and as long as everyone is even it does not need to balance at all and it gives a little longer time as it still can not take down so much so they should make them stay smoother and where do you have it breaks the relay? under 3v and 3.65v? Then I saw your video about your 4S and they must have had good sellers or something else as they are still expensive but another BMS that the ones I have seen people are happy with is DIYBMS which is on an ESP but it is also not as easy but seems good for those who like to do things themselves

[@timburk6097]

If only the victron smart shunt could be easily and cheaply made to communicate longer distances. My power shed is about 150 feet from my house.

[@charlesrichards5389]

I learned elsewhere that if you want more accurate battery monitoring, you need to get one with an improved algorithm. Instead of suffering from drift, your monitor becomes more accurate as it accumulates more data points. In other words, it's not so much the shunt as it's the smarts of the electronics monitoring the shunt. When you cycled the cells at different amperages for efficiency testing (awesome experiment, by the way), did you note the amount of drift each time? I'd be interested to learn if the amount of amperage itself influenced the amount of drift.

[@OffGridGarageAustralia]

That's a good question, I did the test for efficiency testing only once for each current setting. I could repeat that test and several times to see if it drifts. Good thinking.

[@charlesrichards5389]

@@OffGridGarageAustralia Thanks! You could cycle the cells several times at 5A, measure drift, reset the shunt, cycle them several times at 10A, measure the drift, reset the shunt, yada yada yada... or save a LOT of time by just cycling the two extremes (5A & 40A) and comparing the two drifts. But if drift due to different amperages is confirmed, you would be handcuffed to always charging and discharging at the same exact amperage every time just to keep your shunt more accurate. Not practical. If you could tell your shunt to "plug in and use Andy's data points" that would solve the problem (and be totally awesome), but that's not possible (as far as I know). So I think the only solution is to either find an shunt with an improved algorithm, or just live with occasionally having to reset your current shunt.

[@OffGridGarageAustralia]

@@charlesrichards5389 Yes, that is all very time consuming. The shunt can calibrate automatically every time I fully charge the battery. It was set to this until I started the test and I turned this function off. So every time I hit 3.45V and absorbed for 30min and the current went under 5A, it reset to 100A. That works fine and is close enough to show the SOC.

[@charlesrichards5389]

@@OffGridGarageAustralia Yeah, I get it. Automatic mode will hide drift, regardless of what causes it. That's why I don't entirely trust automatic mode to correctly provide available power as a percentage. When the battery is charged, the shunt will calibrate to display 100% even if the cells have gradually degraded over time from, let's say, 280Ah to 250Ah. It's just presenting the "new normal" as 100% and can give someone who isn't paying attention a false sense of security. I also suspect that some shunts give too much weight to voltage as SOC because LiFePO4 voltage curves are so darn flat. But like you said, it's close enough (at least until you decide to go down the rabbit hole🤓).

[@OffGridGarageAustralia]

@@charlesrichards5389 The percentages will always be correct, 100% is always 100% when fully charged no matter how far the battery has degraded 😉 But the shunt also shows the missing Ah until full which is smart too. They won't change either over the years. Just the full capacity will, so the 280Ah will never change in the smart shunt and a full capacity test need to be done every year/two years if someone wants to keep this actual. I'm not too fussed about it, just want to see what happens when I change parameters and how the system reacts. Thanks Charles.

[@petermerle]

You will get less drift if you are able to set charge efficiency exactly to long term average to better than 0.1 % . Victron only allows 1% resolution. You then only have to deal with variations ( which you cannot compensate for via charge efficiency ) . I have been finding a drift uncertainty ( or variation of 2-3 Ah/ 1000AH discharged ) I calculate charge efficiency to be about 98.6%

[@briansousa6521]

I have 2 48v sets of the Lishen 272AH batteries running for the last 6 months. I changed The battery efficiency setting in the Victron Smart Shunt to 98%. It has been two months since I calibrated it last and it is still dead on. It was drifting 3% with the original victron recommended settings.

[@kevinmills5293]

Who is to blame? I agree with your conclusion “all of the above” but if you look at the tail currents between the then and now full charges, they are different so you’re not comparing at the same state of charge. Having said that, the amps, volts and watts reading on the “before” full charge state, don’t add up.

[@YouTubeviolatesmy1stamendment]

On mine I set the charge efficiency to 100% even though it's not really 100%, so mine drifts so that when I reach 100% I really have a little bit more to go

[@OffGridGarageAustralia]

Got mine on 99%. It's really a matter of fiddling around with the numbers until it's the best outcome for one.

[@paulg1802]

Seen Daly BMS.. got one been working well for 2 months.. but it too slow to balance...

[@pau1phi11ips]

It's easy to connect to Venus OS tho.

[@briansousa6521]

@@pau1phi11ips how do you connect the daly smart bms to victron venus. I have them and would love to do that. Can you share your process?

[@paulg1802]

Unfortunately to those using phocos anygrid inverter there appears there is no way to communicate with Daly BMS..

[@neilmarshall977]

What’s the user manual under your phone?

[@OffGridGarageAustralia]

Daly BMS. Video comes in a few hours. Won't be pretty 😉

[@mariog5569]

Maybe you set the peukert Exponent wrong. You can calculate it (see manual smart shunt) by yourself if you have measured the different capacities for different currents. In your example it is: (Log10(109/5)-Log10(102/40))/(Log10(40)-Log10(5)) =1.03

[@OffGridGarageAustralia]

I changed it a few times but found 1.01 works the best for me. All other settings cause even more drift.

[@keyem4504]

I believe the drift is caused by 🐸 jumping around.

[@geroldgoerke5729]

Does LifePo have a Peukert effect?

[@aaronwalker1507]

No

[@rickard1802]

@@aaronwalker1507 puekert effect is voltage drop while under load. Not energy lost the higher the load. Its in wikipedia =) lead acid voltage drop under load will trigger low voltage limits earlier than same capacity lifepo4 but. With this said, lifepo4 do have peukert effect but it is around 5% where as lead acid is around 25% Quoting wiki below: It is a common misunderstanding [2] that the energy not delivered by the battery due to Peukert's law is "lost" (as heat for example). In fact, once the load is removed, the battery voltage will recover,[3] and more energy can again be drawn out of the battery. This is because the law applies specifically to batteries discharged at constant current down to the cutoff voltage. The battery will no longer be able to deliver that current without falling below the cutoff voltage, so it is considered discharged at that point, despite significant energy still remaining in the battery.

[@Ibian666]

Personally I don't see the need to have an exact percentage displayed. Lithium is not good enough to justify the price, with normal batteries you get a good idea of how full they are from the voltage and you can easily calculate necessary charging and capacity.

[@Zorlig]

That's hardly any drift at all, nice!

[@graemebushell7531]

This internal resistance thing, I don't think is right. Charge is conserved, so the ampere hours added and removed from each cell in a battery is always identical in a charge or discharge process (assuming no self discharge). Otherwise the cells would gain a net positive or negative charge. The internal resistance makes no difference to this. As far as I can tell the only thing which can cause cells to drift apart is differing rates of self discharge in each cell. This happens whether you are charging / discharging or not. What is causing your smart shunt to "drift" is, IMO, integration error. It's an unavoidable numerical feature. The more you integrate up and down the more the error will accumulate.

[@dyslectische]

Reason? I see on my own battery its not always the same. So what do i do ? Just not care about it

[@Nic7320]

How do you characterize the AH capacity of your cells? Using the manufacturer's stated AH rating can't be all that accurate, and the Smart Shunt relies on that AH value to compute %SOC. Remember, a SOC meter does not directly measure SOC. It can't. It estimates SOC based on integrating Amps (both positive and negative) over time, and estimates SOC from the 100% starting point. If the AH value is wrong, the SOC estimation will always be off. Did you adjust your AH capacity down when you derated your SOC usage from 10 to 90%

[@OffGridGarageAustralia]

I have used the 280Ah of the specs, so fully charged is 280Ah. This is not accurate of course and may even vary with temperature. But this is as close it can get with the effort I'm putting in. So discharging 1Ah and re-charging 1Ah and we should be back at 280Ah at full charge. This alone depends if I discharge with low or high current and how long it takes until we're back at 280Ah (self discharge). The shunt does coulomb count and also has a parameter for charging efficiency which I have set to 0.99. This again may be correct under certain situations and is not with other circumstances. My capacity setting in the shunt never changed. It was always at 280Ah even I charged to 3.45V only. But you may remember that from another video: I fully charged and balanced the battery to 3.65V and set the shunt to 100%. I then let it discharge slowly a few Ah and recharged it to the old setting of 3.45V including absorption. It showed us 99.5%SOC by the time. Still the capacity was set to 280Ah in the shunt, so it knew there are 0.5% (~1.4Ah) missing to be at real 100%SOC.

[@Nic7320]

@@OffGridGarageAustralia Look at these two videos by Texas Instruments: Battery Chemistry Foundations, and Battery Monitoring Basics. It's tech heavy, and their example covers Li-Ion chemistry, but it gives you a broad overview of all the issues involved with SOC estimation.

[@edwardbyrd7667]

So, did Diego find his own way out of the water jug, or did he need rescue. Drift is caused by D all of the above

[@OffGridGarageAustralia]

Every day he comes and goes for a swim before he goes hunting.

[@bigbeef2654]

I hate shunt drift. It has ruined my life.

[@lanealucy]

All

[@garys-half-baked-offgrid-dream]

@Off-Grid Garage, Andy I hope the Daly BMS manual is idiot proof! Not you, me lol. After much indecision I orderd a 250Ah bad boy a week ago. 😲

[@honumoorea873]

At 1C you got near 35w of power lost as Heat.... It's not 1.5W....for a 16S pack, it's nearly 500W of Heat... (at 1C)

[@BrazzaB1]

Typical Off-Grid Garage - Daly BMS manual littering the bench!

[@OffGridGarageAustralia]

In tomorrow's video! It won't be pretty... 😉

[@BischesseHunting]

70 out of 100 can not calculate percentages, that's more than 90 percent :-))))

[@edwinkania5286]

Angry pixies!

[@stephsoltesz6731]

Observation on your own statement. You previously calibrated the Shunt to when you were charging to 3.650 and from your comments, it seems you never recalibrated the Smart Shunt with you charging to 3.45x as your "Self Determined 100%".

[@OffGridGarageAustralia]

Yes, that is correct, but should the shunt not be at 100% again when charging to 3.65V even without calibration again? In theory it should, but...

[@offgridtruckerdude7133]

The shunt does not account for the power your bms uses, that big azz relay it uses also

[@LupoDIY]

especially as it kicks in quite early at a lower voltage already with Andy's settings, right? Still it is just a some mA, but over 16 days it might be one A or something like this. Could have a small impact on the result.

[@dmeyer8937]

DALY BMS???

[@honumoorea873]

Tweak pukert factor...?

[@SVAdAstra]

What causes displayed SOC "drift"? In declining order of importance: 1) Charge voltage and current, which can't remain constant with PV charging sources. 2) Charge efficiency of LiFePO4 cells is about 99%. If you set the "Current effiency factor" inaccurately you will never hit the 100% target, or hit it too soon. I've set mine to 99%, but I've calculated the actual value to be 99.4%. Unfortunately, Victron doesn't let you configure to tenths of a percent resolution - and errors accumulate over cycles. 3) Capacity declines with ambient temperature about 0.4% per degree C below 20 C. So, if the ambient temperature is cooler, it will appear that you haven't put back what you took out (discharge/charge delta), or conversely, you have less than the displayed SOC during discharge. 4) Incorrect Puerkert exponent setting. This setting is only an approximation and doesn't exactly correlate to LiFePO4 behavior (it's a lead-acid battery setting). I've set mine to 1.04 with good results, but it depends on several independent varibles -- mostly cell impedance. It's all so complex that, everything taken together: the Smart Shunt just requires periodic recalibartion to yield a reasonably accurate SOC result.

[@OffGridGarageAustralia]

Thank you, that is a great comment. I'm running 99% as well for charging efficiency and 1.00 for the Peukert exponent. Ss you said, this is not 100% accurate all the time though, so there is the error to start with.

[@ronny1734]

Could it be wire resistance ohms temperature ect who knows 🤣🤣🤣😁🤫

[@uhjyuff2095]

My answer is "none of the above." Your battery is now a 274ah battery!

[@pau1phi11ips]

I don't think that's the answer. Battery is too young to lose 6Ah

[@uhjyuff2095]

@@pau1phi11ips Just a guess. Here is another guess! The battery is losing charge do to a self discharge day after day.

[@keyem4504]

A Daly BMS die the new battery? You will probably hate the app. It's crappy....

[@oldschoolfoil2365]

Why are we in 2022 using such primative power supplies like batteries? damn these oil barons. im making my own hydrogen with shop bought stainless steel drinking bottles and steel wool filaments a handfull of AA torch batteries inflating home made balloons with a tea candle under them creating quite an impressive explosion at 1000m + in the air.

[@hommerdalor6301]

a+b+c+x

[@mfgxl]

I am not going to be wasting my time and money on a stupid shunt!

[@harryharry8384]

just charge the dam things to 3.65 and set static ballance down to 3.5v...... what are you messing around at float charging lithiums..... and i d have thought it obvious its down to charging voltage which is basically variable dependant on bright or dull day and i experience an exaggerated version of this when using a 14.8v lead acid charger when in tandem with a 14.6 lithium charger. ps. you can set the cell voltage the smart bms flicks to 100% indication in settings as its bog standard is 13.6v which equates to around 75% if a lithium specific charger is used. (mine is set to 3.6v or 14.4 if multiplied by 4) you seem to be spinning this out for media success rather than sorting the problem which i find a bit missleading for folk ...

[@OffGridGarageAustralia]

Oh, Harry, it seems you don't understand what this is all about, right??? Poor guy... 🤦‍♂️

[@offgridwanabe]

% this is why I only monitor the cell voltage and the spread nothing else matters.