question

itllgrowback avatar image
itllgrowback asked

SmartShunt miscalculating Consumed Ah

Hi folks - I recently built a cabinet for a LiFePO4 house battery and Victron Orion-Tr Charger in the back of my truck, primarily to supply a Dometic 12v fridge. The setup works great, but the SmartShunt is miscalculating the amount of current used.

For example, yesterday, the SmartShunt went from 66% SOC to 100% SOC in three minutes. I anticipated it would, because the voltage was increasingly nearing the charger's specified Absorption voltage (14.4v) as I drove. The charger had already been on Absorption charge for nearly an hour of driving; and as I watched, it steadily increased voltage and decreased current, as it should. It was still reporting 66% SOC and -19.7Ah right before hitting the absorption voltage, and then it jumped to 100% SOC once the conditions for a full charge had been met (charge voltage, tail current, 3m wait).


victron-issue.jpg


I believe the discrepancy has more to do with discharge calculations than charge calculations: in another experiment, the SmartShunt showed a 0% SOC at 12.58v which should be somewhere in the neighborhood of 5% SOC for the LiFePO4 (and where I don't intend to take it regularly); but more to the point, it was showing that I had used 72.8Ah, from my 54Ah battery. So somehow the calculations are off in a way that reports more current loss than has actually occurred. I observe that the reported SOC is always lagging on each and every recharge cycle.

I'll post my settings below to see if there's something I can adjust to get accurate measurements.

screenshot-20230126-163004.jpg

- I don't believe the issue is with any of the full charge settings like Charged voltage, Tail current, or Charged detection time, because based on indicated voltage, full charge is being reported at the correct time.

- I don't know much about the Peukert exponent but perhaps it's something to look at. I've read here that the PE doesn't account for temperature, but I don't know how significant a factor that could be.

- Charge efficiency factor is set to 99% as a default for LiFePO4 batteries - we can adjust that if it could help.

- I don't know how the Current threshold is used, but this is the default, and we can adjust it.

- The Time-to-go averaging period is 3m by default; I changed that to 12m trying to even out the cycles of the fridge's compressor - when it's running versus when it's off - which report very different Time-to-go values.

- My battery is a Dakota Lithium 54Ah LiFePO4


Any help appreciated - I've read here on the Victron support forum that the defaults are only starting points and that massaging the variables is necessary in some cases - but this is off by such a significant amount.



SmartShunt
victron-issue.jpg (284.5 KiB)
screenshot-20230126-163004.jpg (334.1 KiB)
8 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.

klim8skeptic avatar image klim8skeptic ♦ commented ·
Share a screenshot of your SmartShunt settings plz.
0 Likes 0 ·
itllgrowback avatar image itllgrowback klim8skeptic ♦ commented ·

I don't know why the embedded links didn't work - I used the interface to add them. Let me add a post with that info.

EDIT: I was getting a message that the post was awaiting moderator approval, and since being approved, the links all work now.

0 Likes 0 ·
klim8skeptic avatar image klim8skeptic ♦ commented ·
More likely you have the charger negative connected to the battery, rather than the system side of the shunt.
0 Likes 0 ·
itllgrowback avatar image itllgrowback klim8skeptic ♦ commented ·

The negative lead from the charger goes to the system side of the shunt, and then the battery side of the shunt goes to the battery - and the positive from the charger goes to an ANL fuse, and then to the battery.

Here's a logical diagram of the way it's set up:

battery-charger-schematic.jpg

0 Likes 0 ·
battery-charger-schematic.jpg (1.3 MiB)
pwfarnell avatar image pwfarnell itllgrowback commented ·
Have you got a ground connection anywhere that you have not shown.
0 Likes 0 ·
itllgrowback avatar image itllgrowback pwfarnell commented ·

No, the cabinet is self-contained past the charger. The charger's input negative is grounded to chassis, and the charger's input positive comes directly from the starting battery.

But apart from the charger inputs, everything is as you see in the diagram - with the exception of the low-voltage input to power the shunt, which comes from the ANL fuse. I need to add that, now that I think of it.

0 Likes 0 ·
falsevagabond avatar image falsevagabond itllgrowback commented ·

What does your installer say about this or are you the installer?

tunnel rush

0 Likes 0 ·
itllgrowback avatar image itllgrowback klim8skeptic ♦ commented ·

Here's a photo of the actual install - I know it lacks context but with the labels, you will see that the 8ga conductors from the charger output connect where the diagram indicates. The charger is mounted to the opposite face of this surface and its inputs/outputs come through from the other side.

Anyway, the only things connected to the battery come from the Fuse (+) and the Battery Minus side of the shunt (-).


dscf8267.jpg


0 Likes 0 ·
dscf8267.jpg (2.9 MiB)
7 Answers
klim8skeptic avatar image
klim8skeptic answered ·

The charger settings look right.

The SmartShunt settings look ok.

The SmartShunt has automatically synced the battery properly. Battery voltage > 13.2v + current < 2.16a for 3 minutes.

There is just the oddness of the -72.8 Ah drawn from a 54Ah battery. It is also odd that an 18a charger can charge the battery 53.1Ah in 32 minutes??????

The anderson plugs are for the Orion's power input?

Puzzling.

1 comment
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.

itllgrowback avatar image itllgrowback commented ·

Yeah the big Andersons are where the 6awg charger inputs connect.


I'm not sure what you mean about the 53.1Ah in 32 minutes - but the screenshot regarding the -72.8Ah was from a different day. Perhaps that's where the confusion entered in. That episode happened previously, after I started to notice there was always a jump in SOC as it synchronized.


0 Likes 0 ·
JohnC avatar image
JohnC answered ·

Hi @itllgrowback

Make sure your Smartshunt is never powered down. This may go against the grain in a 'truck', but it needs continuous power. Not much though. If it's been powered down it will need to see a realistic sync to be of any use.

The Consumed Ah figure is actual, as measured from the last sync to 100%. This isn't adjustable via calibration. If it's wrong the unit is faulty. This doesn't apply to the Ah 'counted back' under charge, as it must be adjusted for Charge Efficiency to give a realistic figure.

How you get 72.8 Ah from a 54 Ah battery I can't explain. Maybe the batt is capable of more, down to a lower V where it's not supposed to be?

I don't think your settings have caused this. Not to say they're all suitable. I'd raise the Charged V up to say 14.2V (0.2 below Abs), as 13.2V risks false syncs. The 1.05 Peukert is probably ok, but if you want Ah to match SOC then set it to 1.0.

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.

itllgrowback avatar image itllgrowback commented ·

Thanks @JohnC - the SmartShunt is powered by the house battery itself (directly from the downstream end of the 50A ANL fuse - you can see that in the photo of the installation above if you can find it). So it's never disconnected from power, assuming the LiFePO4 has life.


On the Charged Voltage, I read on the forum to set that 0.2-0.3v below the charger's float voltage (13.5v). The manual also ties charged voltage to float voltage. That's how I arrived at 13.2v but it does make sense that it should be referencing the absorption voltage instead. As a side quest, should I raise the float voltage as well (13.5v) for this battery? In any case, I can raise the charged voltage to 14.2v on your recommendation; it makes more sense than 13.2v.

It always seems to sync properly though - according to voltage and tail current (as I observe them in real time), it syncs properly every time. It's just the Ah consumed in the meantime (and therefore the SOC) that's wrong.

Unless there's a way to test it, perhaps it will need to be returned. Is there a channel I should use for initiating that? I'm eligible for a return through Amazon for another 4 days (!) so I might be able to do it that way rather than through official Victron channels.

Thank you for the help!

0 Likes 0 ·
JohnC avatar image JohnC ♦ itllgrowback commented ·
@itllgrowback

The 13.2V is more suitable for Pb batteries in Float mode. 14.2V for Lithium, which is normally considered fully charged when it reaches Absorb.

I wouldn't rush into sending it back, you have a 5yr warranty. These things are reliable as hell, and I don't actually recall anyone else complaining here that the A (and subsequent Ah) is wrong. Try testing it's output with known loads or another meter. If you found something I'd be really surprised.

The relationship between Ah and SOC is adjustable. But the Ah drawn from a genuine sync should be truth.

0 Likes 0 ·
itllgrowback avatar image itllgrowback JohnC ♦ commented ·

Thanks @JohnC - is it worth disconnecting the System side and doing a zero calibration?

To test with a known load, I have a pair of spare 12v lamps that I've used before which draw 8.8A together. I could put my Fluke inline to confirm current and plug them into the output in place of the fridge. I can do a synchronization and then run those lights for an hour or two and compare that to what the Shunt reports.

I'll do that this morning and see what it shows.


0 Likes 0 ·
Nev avatar image
Nev answered ·

Your charged voltage setting of 13.2V is far too low and as a result will always report 100% early.

My 4 cell 300Ah LiFePO4 battery is set at 14.1V.


1 comment
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.

itllgrowback avatar image itllgrowback commented ·

Thank you - I address this in a comment above, and I think you're right. I'll raise that up to 14.2v.


That being said, it doesn't have any problems syncing correctly. I've never had a false or early synchronization. Even if the charged voltage value is too low, it still waits for the tail current (2.1A) for three minutes, and thus it always syncs correctly. The issue I'm reporting is about Ah Consumed being inaccurately calculated/reported.

There's no issue with synchronizations at all.

0 Likes 0 ·
snoobler avatar image
snoobler answered ·

Might also be worth selecting battery current and voltage under trends tab. This will log 45 days worth of the selected settings. Might give you additional clues.

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.

Alexandra avatar image
Alexandra answered ·

@itllgrowback

10:50 to 10:53 is three minutes. Your charge detection time is 3 minutes. So not random. It synchronised then because of your other settings. If your settings are close to perfect for your batteries you will see amost no synchronisation to get to 100%.

As is mentioned charged voltage is set too low in your settings. It must be higher than float but not as high as the absorption voltage.

Also, i know lifepo4 say they are 99% but in reality i have not seen that. So you may need to adjust that.

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.

itllgrowback avatar image itllgrowback commented ·

I should perhaps have said "once the battery synchronized, it jumped from 66% to 100% SOC." The three minutes I used to illustrate the issue simply represents the last two screenshots I took - I could have posted one from five, or ten, or 20 minutes prior, showing 65%, 65%, or 64% SOC, for example (I have those captured as well). I just thought the two closest screenshots would illustrate the point best. The issue is that the SOC being reported is impossibly low, until the moment of a new synchronization, when it finally jumps up to 100% - and it happens that way on each and every recharge cycle.

The voltage and tail current settings all behave perfectly, the charger works perfectly, and if I didn't have a Shunt at all, there would be no issue to report. But something in the way this Shunt is reporting AH Consumed is off.


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

@itllgrowback

Peukerts affects discharge as well as charge capacity. That is why a battery has a c rating as well as a capacity given at the c rating.

You can get a higher AH drawn from a lithium as you have seen on yours. Although i have not seen it 20AH higher. (barring a fake amp reading. So check your zero current as you mentioned above.)

Charge efficiency on lithium is not 99% especially if the battery is passive balanced which most are. So you do need to change that as well.

1 Like 1 ·
itllgrowback avatar image
itllgrowback answered ·

I have some results to report - @JohnC , @Alexandra , @klim8skeptic - if you're still interested, this was interesting to me:

This morning I ran a test with a known load - two spotlights from an old install. Together, the Shunt is reporting they use roughly 9.1A (unfortunately, my Fluke's 10A fuse has blown at some point so I was unable to verify that current externally - I can test that tomorrow when a new fuse arrives, and it'll be interesting to see if the Fluke reports the same current as the shunt does).

But, what I observed over a two-hour steady discharge and a two-hour steady recharge, is a predictable discharge/recharge in this test, and accurate SOC numbers. The charger I plugged into the unit to recharge it is from Dakota, and was charging at roughly 9.2A, almost the same rate as the lights discharged the battery - that's entirely coincidental, but it resulted in the experiment taking almost exactly the same amount of time to recharge as it did to discharge. More to the point, though, the SOC behaved exactly as you'd expect:

discharge-cycle.jpg

charge-cycle.jpg



So then the question becomes, why am I getting such odd results in my actual use case? Again, that's a Dometic CFX55IM Fridge, powered by the battery full-time, but whose compressor runs only intermittently. Then the charger also runs intermittently when the truck is running and the Orion-Tr senses the alternator is pushing out amps. So it's much less direct than my experiment - things are discharging and recharging at different rates and periods intermittently throughout the day.

I'm currently running a new test, same as the one I did this morning, but with the Dometic plugged in instead of the spotlights. That'll take more time, because the fridge only uses some 3A, and only when the compressor is running - so it's not discharging as quickly. And when I recharge it for this test, it'll be with the Dakota charger again versus the Orion-Tr I that would be in play in normal use. I'll report back those results when I have them, but I might just let it run overnight to discharge further, and then charge it up tomorrow while I can monitor it.

Thank you for your help - I hope we can find out what's going on, whether something in my install or other.






discharge-cycle.jpg (467.3 KiB)
charge-cycle.jpg (447.8 KiB)
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.

klim8skeptic avatar image klim8skeptic ♦ commented ·
I reckon at some point in time, you have topped the battery up with the Dakota charger, but have bypassed the SmartShunt. Battery Soc has increased, but the shunt cant measure current going in. Hence being able to draw 72Ah out of your battery.


Just a theory tho.

2 Likes 2 ·
itllgrowback avatar image itllgrowback klim8skeptic ♦ commented ·

The way I plug in the Dakota charger is through the same place as I'd plug in a device such as the fridge, in what's labeled as "12V Output" in the schematic I posted earlier - so that branch of the circuit still connects to the System end of the Shunt.


Here's a photo - I love using Anderson Powerpoles, and in this case I use the same port whether running a device or plugging in the charger.

img-20230129-083842674.jpg


So nothing connects to the battery without running through the shunt.

I don't know if that could still be related to the issue somehow.

(I use the Dakota at times if for example I don't drive the truck for a few days and want to keep the fridge from hitting the low-voltage cutoff, which is at 11.8v and lower than I like to take the battery if I can help it.)


0 Likes 0 ·
img-20230129-083842674.jpg (3.9 MiB)
klim8skeptic avatar image klim8skeptic ♦ itllgrowback commented ·
@itllgrowback Well that blows that theory out of the water.

Your last screenshots show the SmartShunt is correctly wired up when using fridge or charger.

One last thought regarding the battery going from 66% (-19Ah) to 100%, and the -72Ah discharge figure.

If a manual Soc was set to 60% when the battery was near full, that would produce the odd results observed.

Just another theory though.

1 Like 1 ·
itllgrowback avatar image itllgrowback klim8skeptic ♦ commented ·
I really appreciate you hanging in and trying to help! I'm trying to figure it out and leave it all at the door so I'm open to any thoughts. I'm about to make another update which shows a normal discharge/recharge cycle so I'm still unsure what's going on.
0 Likes 0 ·
itllgrowback avatar image
itllgrowback answered ·

Well, after running the fridge overnight, on its normal periodic on/off cycles, I discharged it down to 20%, which is about the lowest I prefer to let it discharge. Then, I plugged in the Dakota 10A charger again, and let it charge back up.

It took about 5 hours as you can see below, and though I didn't observe the synchronization in real-time (it happened before the hour mark, and I wasn't watching), it looks like everything behaved as you'd expect:

charge-after-dometic-overnight.jpg

And so the question remains: what's going on in my normal use scenario? Everything in these tests seems normal, but in normal use things are wacky. So I decided to reset the history on the SmartShunt, and turned on the fridge again, with normal settings, and I'll log what happens in the week to come.

As a sidenote, I got a new fuse and tested the current with the Fluke compared to the Shunt, while the accessory spotlights were connected, and they were within a margin I'd consider negligible: 8.82A on the Fluke compared to 8.76A on the shunt in the same moment. So take that for what it's worth - I think it settles any discrepancy between the shunt and another device (but let me know if you disagree - I'm a hobbyist).

So I'll report back soon to let you know what numbers I see in the days to come, with the normal intermittent discharge/recharge cycles.


charge-after-dometic-overnight.jpg (837.1 KiB)
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.

JohnC avatar image JohnC ♦ commented ·
@itllgrowback

Impressed with your work, and your reporting of it. Well done. You tried using the Trends graphs too?

Look forward to seeing your 'wacky' normal use. I've a theory that frequent batt 'mini' cycling might exaggerate any small mistuning of Peukert and Charge Efficiency settings, so consider that as a possibility if you see strangeness.

The Ah Discharged is gospel though, but nowhere does Victron report Ah Charged.

0 Likes 0 ·
itllgrowback avatar image itllgrowback JohnC ♦ commented ·

Thanks John - I've learned that if you're asking for help, it pays to be precise. :D


I haven't looked into the Trends graph much - any pointers or suggested parameters to compare?

You might be right about the mini cycling - it's definitely a lot of start/stops in both discharge and charge cycles in my normal use, so there could be something in that.

0 Likes 0 ·