VANILLA POWERED

For weeks I’ve been watching numbers flicker across VictronConnect. Watts. Amps. State of charge. Solar yield. Bulk. Absorption. Float. Interesting individually, certainly, but usually just little snapshots of what Vanilla happened to be doing at that particular moment.

Then I downloaded the data.

And suddenly Vanilla stopped being a collection of screenshots.

She became an experiment.

Buried inside two innocuous CSV files was an electrical diary stretching across late summer and into autumn: more than a thousand hourly battery readings, thirty days of detailed solar history and enough evidence to reconstruct exactly how my campervan has been living, charging, draining and recovering.

The results genuinely surprised me.

Because the biggest discovery isn't simply that my new solar installation works.

It's how profoundly the entire electrical personality of Vanilla appears to have changed.

The first file came from my Victron SmartShunt and contained 1,107 valid hourly battery readings covering 17 August to 2 October 2026.

Then came the really juicy file.

My Victron SmartSolar history.

This told me what the 200 W PV Logic panel on Vanilla's roof had actually harvested.

Not theoretical solar output.

Not what a 200 W panel might produce in laboratory conditions.

Not somebody on a campervan forum announcing that their 200 W panel once produced enough electricity to illuminate Wolverhampton.

Actual energy harvested by my system.

Across the thirty complete days from 2 September to 1 October:

☀️ TOTAL SOLAR HARVEST — 8,470 Wh ⚡ AVERAGE PER DAY — 282 Wh 📊 MEDIAN DAY — 300 Wh 🔥 BEST DAY — 570 Wh 🌧️ LOWEST COMPLETE DAY — 140 Wh

Add the beginning of 2 October and the recorded total nudges up to around 8.49 kWh.

Eight and a half kilowatt-hours.

From a 200 W panel bolted flat onto the roof of a campervan.

That puts the whole thing into perspective.

At typical 12 V-system charging voltages, that's broadly equivalent to something in the region of 630–660 Ah of energy processed through the solar system during the period.

Obviously battery charging isn't quite as simple as dividing watt-hours by twelve, because voltage changes and losses exist, but as a sense of scale it is remarkable.

That roof isn't decoration.

It is working.

The SmartShunt data reveals something even more dramatic.

Before the new 200 W PV Logic panel and Victron MPPT installation at the end of August, Vanilla's leisure battery lived a very different life.

Looking at the clean pre-installation period from 17–27 August:

Average SOC: about 60% Median SOC: about 62% Lowest SOC: 34.5% Time below 50% SOC: nearly 19%

Then look at 30 August onward:

Average SOC: 91.5% Median SOC: 95% Lowest SOC: 55% Time below 50%: ZERO

That is not a little improvement.

That is an entirely different operating environment.

Before the change, Vanilla frequently lived somewhere in the middle of her battery.

Afterwards, she spent much of her life bouncing around the top.

Across the entire 47-day dataset, the battery spent an extraordinary 580 recorded hours at 90% SOC or higher.

That's 52.4% of every recorded hour.

There were 194 readings showing a full 100%.

And from 29 August onward, nearly 70% of every recorded hour was at 90% or above.

Before 29 August, Vanilla reached 100% SOC on only one day out of twelve.

Afterwards?

Twenty-nine days out of thirty-five.

Now, I need to be careful here.

That doesn't mean solar alone caused every percentage point of improvement.

Vanilla was being driven. The Victron Orion DC-DC charger was working. EHU was occasionally connected. Usage changed. Weather changed. I wasn't running some pristine laboratory experiment in a white coat while Vanilla sat on rollers.

But the timing of the change is difficult to ignore.

Something changed dramatically at exactly the time I dramatically upgraded the solar system.

The roof started pulling its weight.

Possibly my favourite discovery is what happens when I average the battery movement by hour.

You can actually see Vanilla waking up with the sun.

Around midnight the battery is normally discharging at roughly:

00:00 — −0.74 Ah 03:00 — −0.71 Ah 06:00 — −0.78 Ah

Nothing unusual there.

The fridge is alive. Monitoring electronics are awake. Little background loads are quietly nibbling away.

Then daylight arrives.

08:00 — +0.61 Ah 09:00 — +1.61 Ah 10:00 — +2.55 Ah 11:00 — +2.49 Ah 12:00 — +1.66 Ah

There it is.

A beautiful little electrical sunrise.

Production climbs through the morning, peaks around 10:00–11:00 and then gradually declines:

13:00 — +0.54 Ah 14:00 — +0.42 Ah 15:00 — +0.46 Ah 16:00 — +0.62 Ah 17:00 — +0.20 Ah

Then evening arrives.

18:00 — −0.22 Ah 19:00 — −0.85 Ah 21:00 — −0.96 Ah 23:00 — −0.78 Ah

Sun down.

Battery back on duty.

What makes these numbers even more interesting is that they're NET battery movement.

If solar is producing 5 A while the fridge and other loads are consuming 2 A, the SmartShunt doesn't report the solar panel's five.

It sees the three arriving at the battery.

So the solar system is often doing more work than the battery graph alone suggests.

This is why having both sets of data suddenly makes everything much more revealing.

The SmartSolar tells me what came from above.

The SmartShunt tells me what happened below.

Now we're doing proper campervan forensics.

🌙 WHAT DOES A NORMAL NIGHT COST?

I've often wondered what Vanilla genuinely consumes overnight.

Now I have an answer.

Looking at 46 nights where I can compare the battery at 20:00 and again at 07:00 the following morning, the median SOC reduction is:

6.25 percentage points.

The average is almost exactly the same at 6.26%.

That's incredibly useful.

It means I now have a genuine AGM baseline for a normal Vanilla night.

Not a guess.

Not a calculation from the fridge specification.

Not “I think it dropped about seven percent.”

Actual historical behaviour.

Some nights, of course, are considerably heavier.

One night dropped 25.5 percentage points.

Another dropped 20%.

Another 17.6%.

Something was clearly happening on those nights beyond the normal background load.

And that's exactly what makes this data exciting.

With enough logs, unusual behaviour stops looking random.

It leaves fingerprints.

And then we arrive at 28 September.

This day is glorious.

Electrically speaking, somebody kicked the front door in.

Vanilla started around 96% SOC.

By 11:00 she'd reached 100%.

Everything looked lovely.

Then something started eating electricity.

13:00 — −5.07 Ah 14:00 — −7.65 Ah 15:00 — −6.72 Ah 16:00 — −6.30 Ah 17:00 — −4.86 Ah 18:00 — −3.18 Ah 19:00 — −2.02 Ah

SOC responded accordingly:

11:00 — 100% 14:00 — 87% 15:00 — 79.5% 16:00 — 72.5% 17:00 — 66.5% 19:00 — 60.5% 23:00 — 57.5%

During the whole day the SmartShunt recorded approximately:

−46.19 Ah of discharge

against:

+6.57 Ah of battery charging

giving:

−39.62 Ah NET.

It was the largest net-loss day in the entire dataset.

But here is where the new SmartSolar data makes this fascinating.

The solar panel produced:

300 Wh.

Maximum PV output reached:

102 W.

So 28 September was NOT a failed solar day.

The panel was generating perfectly respectable energy.

Something else simply consumed more.

A lot more.

And the shape of that load catches my eye.

Several consecutive hours of roughly 6–8 A of battery discharge equates to somewhere in the ballpark of 75–100 W at typical 12 V battery voltages.

Anyone familiar with my recent electrical tinkering may already be raising an eyebrow.

I regularly use the Anderson connection from Vanilla's leisure battery to feed the DC input of my Jackery power stations at around 90 W.

Can I prove from this CSV that I was charging a Jackery on the afternoon of 28 September?

No.

And I won't pretend I can.

But electrically?

Let's just say the suspect fits the description. 😏⚡

Vanilla begins the next morning wounded.

00:00 — 57% 04:00 — 55% 07:00 — 55%

Then somebody opens the taps.

07:00 — +3.83 Ah 08:00 — +15.30 Ah 09:00 — +11.20 Ah 10:00 — +7.05 Ah 11:00 — +3.05 Ah

SOC rockets upward:

07:00 — 55% 08:00 — 65% 09:00 — 78% 10:00 — 86.5% 11:00 — 98.5% 12:00 — 100%

Net battery movement for the day:

+36.09 Ah.

At first glance that might look like an astonishing solar recovery.

Then we check the SmartSolar file.

Solar harvest on 29 September:

180 Wh.

Maximum PV power:

58 W.

And suddenly we know something important.

The sun did NOT perform that recovery alone.

There simply isn't enough solar energy in the day's MPPT record to explain a net 36 Ah increase in the battery while simultaneously running Vanilla's normal loads.

Therefore another charging source was involved.

Alternator/DC-DC charging.

EHU.

Or a combination of charging sources.

That's the moment these two datasets stop being merely interesting and become genuinely powerful.

The SmartShunt says:

“The battery charged.”

The SmartSolar replies:

“Don't look at me, mate. I only supplied 180 Wh.”

Case cracked.

The best solar harvest of the thirty complete days happened on 26 September.

570 Wh.

That's nearly three times the panel's nominal 200 W rating expressed as energy over time, which is exactly why watts and watt-hours should never be confused.

The most interesting part?

Maximum PV power was only:

124 W.

The panel didn't briefly explode into some glorious 200 W supernova.

It simply worked for hours.

The MPPT spent a staggering:

722 minutes in Bulk.

No Absorption.

No Float.

And despite harvesting 570 Wh, the battery still ended the day with a NET movement of:

−16.61 Ah.

That's brilliant.

It means Vanilla had somewhere to put the solar all day.

Loads were consuming energy. The battery required charging. The MPPT kept harvesting.

The system behaved less like a drag racer and more like a truck hauling freight all day.

And frankly, as someone who drives trucks, I approve.

Then we get the opposite phenomenon.

8 September.

Solar yield:

140 Wh.

Maximum PV power:

66 W.

Looks disappointing.

Except the Victron spent:

27 minutes in Absorption

and:

293 minutes in Float.

That's nearly five hours in Float.

So was the panel rubbish that day?

Not necessarily.

The battery simply wasn't asking for much.

Once the battery reaches the appropriate charging stage and Vanilla's loads aren't demanding everything the panel can provide, the controller throttles production.

Solar panels don't manufacture electricity simply because sunshine exists.

There needs to be somewhere for that electricity to go.

This completely changes the way I look at the daily yield figure.

A low number doesn't automatically mean:

“Poor solar.”

Sometimes it means:

“Battery full. Job finished.”

And that is a much nicer problem to have.

🔋 BULK. ABSORPTION. FLOAT. REPEAT.

Across the thirty complete days, my Victron SmartSolar spent roughly:

More importantly, it reached Float on:

20 DAYS OUT OF 30.

That's one of the most reassuring findings in the data.

Two-thirds of the time the system was completing enough of the AGM charging process to enter Float.

So when I ask myself:

“Is 200 W of roof solar genuinely useful on Vanilla?”

The answer from the data isn't merely yes.

It's demonstrably yes.

There will always be days when the loads exceed production.

There will always be British weather.

There will always be a Jackery somewhere nearby looking hungry.

But the system is regularly getting the battery into the upper charging stages.

That is exactly what I wanted.

⚡ 147 W FROM A 200 W PANEL. SHOULD I BE DISAPPOINTED?

No.

My highest instantaneous recorded output was:

147 W.

That's 73.5% of the panel's nominal 200 W rating.

And I'm perfectly happy with it.

The panel rating is based on controlled test conditions.

Mine lives horizontally on top of a VW campervan in Britain.

Its relationship with the sun involves roof angle, vehicle direction, cloud, temperature, shadows, season and the occasional tree deciding to become an electrical engineer.

More importantly, the battery has to WANT the power.

If Vanilla is sitting at 100% and the fridge isn't asking for much, the Victron isn't going to force 200 W into a battery for the entertainment value.

That's why daily energy harvest matters far more to me.

16 days produced at least 300 Wh.

4 produced at least 400 Wh.

The record produced:

570 Wh.

In September.

From a campervan roof.

I'll take that.

⚠️ AND THEN THERE ARE THOSE VOLTAGES...

The MPPT file contains a few rather alarming-looking minimum battery voltages.

27 September — 11.00 V 28 September — 10.63 V 29 September — 10.58 V 1 October — 11.13 V

At first glance 10.58 V looks hideous.

But the SmartShunt adds crucial context.

On 29 September, when the MPPT saw that low voltage, battery SOC was still around 55% at its lowest.

So I don't believe Vanilla had a rested AGM battery genuinely sitting at 10.58 V.

Much more likely we're seeing short-term voltage sag under load, wiring voltage drop at the MPPT, or a combination of the two.

And this gives me yet another benchmark for the next stage of the experiment.

Because very soon the AGM is going.

And the Fogstar 105 Ah LiFePO₄ is coming.

This may be the thing I'm now most excited about.

Until I downloaded these CSV files, I thought I had a reasonable understanding of Vanilla's electrical system.

Now I have a baseline.

A proper one.

When the Fogstar goes in, I can measure the new system against exactly the same questions.

How much does SOC fall overnight?

How firmly does voltage hold under load?

What happens while feeding a Jackery at around 90 W?

How quickly does the battery absorb morning solar?

How rapidly does the Orion replenish it while driving?

How much longer can I operate the Webasto, fridge and lighting overnight?

How much harvested solar can the lithium battery actually accept?

How often does the MPPT end up throttling because the battery is already full?

And perhaps the most important question of all:

How many days can Vanilla realistically operate without needing an orange EHU cable?

This won't be an internet comparison.

It won't be AGM versus lithium specifications copied from manufacturers' websites.

It will be:

versus

VANILLA LIFEPO₄ DATA.

Same van.

Same fridge.

Same solar panel.

Same MPPT.

Same SmartShunt.

Same idiot connecting Jackerys to Anderson plugs in the name of science.

Now THAT is an experiment.

Another lovely little event appears on 1 October.

The battery begins at:

85.5%.

Then gradually falls:

03:00 — 80.5% 06:00 — 75.5% 08:00 — 70.3% 10:00 — 66.1%

Then charging begins.

15:00 — 68.4% 16:00 — 74.6% 17:00 — 80.7% 18:00 — 84.8% 19:00 — 92.8% 20:00 — 100%

I know I connected EHU.

And the data captures it beautifully.

After reaching 100%, charging doesn't abruptly disappear.

It progressively tapers:

21:00 — +1.64 Ah 22:00 — +1.19 Ah 23:00 — +0.91 Ah 00:00 — +0.78 Ah 01:00 — +0.66 Ah 02:00 — +0.55 Ah 03:00 — +0.49 Ah 04:00 — +0.43 Ah 05:00 — +0.38 Ah 06:00 — +0.32 Ah 08:00 — +0.24 Ah

You can literally see the charging system settling into maintenance.

No guessing.

The fingerprint is right there.

The biggest lesson from all this is that campervan power is much more interesting than a battery percentage.

Watching SOC alone is like trying to understand a motorway by looking at how many vehicles are parked at the services.

The SmartShunt tells me what's happening to the battery.

The SmartSolar tells me what the roof is contributing.

The Orion contributes energy while driving.

EHU occasionally arrives with a sledgehammer.

Meanwhile the fridge, heater, lights, USB devices and occasional Jackery experiment are pulling energy in the other direction.

All of those systems overlap.

That's why the data matters.

Over this period I recorded:

🔋 1,107 valid hourly battery readings ☀️ 8.47 kWh of complete-day solar harvest ⚡ 570 Wh best solar day 🚀 147 W highest solar peak 📈 91.5% average SOC after the late-August change 💯 29 days reaching 100% after 29 August 🌙 Approximately 6.25% typical overnight SOC drop ⬇️ 34.5% lowest SOC in the whole record 🔥 −39.62 Ah worst net battery day ⚡ +36.30 Ah strongest net recovery day ✅ Zero recorded SmartSolar errors

But perhaps my favourite statistic is still this one:

After 29 August, Vanilla never recorded an SOC below 55%.

Not once.

Compare that with the beginning of the dataset, where she remained below 50% for more than thirty consecutive hourly readings.

That's transformation you can actually see.

🚐 VANILLA ISN'T JUST OFF-GRID. SHE'S BECOMING MEASURABLY OFF-GRID.

That's where all this has changed my thinking.

When I fitted the 200 W panel, I wanted more independence.

Less worrying about EHU.

More freedom to stop somewhere beautiful without immediately wondering where the nearest blue socket was.

What I didn't expect was to end up building such a detailed picture of how Vanilla actually uses energy.

The solar installation has already demonstrated that it can harvest meaningful power day after day.

The SmartShunt has shown me exactly where that energy goes.

The strange days have become explainable.

The heavy loads are visible.

The overnight consumption has a baseline.

EHU leaves a signature.

Solar has a signature.

Driving has a signature.

Even charging a Jackery probably has a signature.

And now we're about to change the biggest component in the system.

The battery itself.

The AGM chapter is almost finished.

Fortunately, before it goes, it has left me something extraordinarily valuable.

Data.

Thousands of little electrical breadcrumbs showing exactly how Vanilla behaved.

Soon I'll fit the Fogstar 105 Ah LiFePO₄.

Then we'll start recording Chapter Two.

And somewhere a few months from now I'll open another collection of CSV files.

AGM on one side.

Lithium on the other.

Same Vanilla.

Same roof.

Same lifestyle.

Then we'll finally discover what the upgrade really bought me.

Not in advertising claims.

Not in specifications.

In nights.

In harvested sunshine.

In heater hours.

In Jackery charges.

In days away from EHU.

And ultimately, in freedom.

That, after all, is what the whole system was built for.

THE DATA DOESN'T GUESS.

I’ve now gone through every valid hourly battery record in Vanilla’s SmartShunt history.

Not just the dramatic days. Not just the solar peaks. Not just the moments when the battery hit 100%.

All of it.

There are 1,107 valid hourly readings, and when the whole lot is added together the numbers become surprisingly revealing.

Across the dataset, the SmartShunt recorded approximately:

870.67 Ah flowing into the battery.

779.62 Ah flowing out of the battery.

That gives an arithmetic net of +91.05 Ah across the period.

More importantly, the battery handled around 1,650 Ah of total two-way recorded movement.

That is a serious amount of electrical traffic through one campervan leisure battery.

And that figure is actually conservative, because the readings are hourly net values. If the battery was simultaneously receiving solar while also powering the fridge, heater or another load, those flows can partially cancel each other before appearing in the exported figure.

So the real electrical activity was almost certainly higher.

Vanilla spent 375 recorded hours charging and 731 hours discharging.

There was only one hour that registered exactly zero.

So roughly two-thirds of the time, the battery was supplying power.

Yet the battery still finishes the dataset ahead because the charging periods are generally much stronger than the discharge periods.

That gives Vanilla a very distinct electrical personality.

Most of the time she quietly sips power.

Then every so often a charging source arrives and absolutely floods the battery.

A typical charging hour was around +1.00 Ah.

The average positive hour was approximately +2.32 Ah.

A typical discharge hour was around −0.84 Ah.

Average discharge was approximately −1.07 Ah.

The biggest charging hour was +18.35 Ah.

The biggest discharge hour was −7.65 Ah.

So Vanilla normally spends hours gently using energy, followed by relatively short periods of aggressive replenishment.

That is exactly what I would expect from a campervan supported by solar, DC-DC charging and occasional EHU.

The most reassuring thing in the entire dataset is how predictable the normal background load appears.

Of the 779.62 Ah recorded leaving the battery, most of it happens at fairly modest currents.

Around 42 Ah was consumed during hours below 0.5 A discharge.

Approximately 235 Ah occurred between 0.5 and 1 A.

Around 241 Ah occurred between 1 and 2 A.

Approximately 160 Ah occurred between 2 and 4 A.

Only about 101 Ah occurred during hours exceeding 4 A discharge.

That means more than 60% of every recorded amp-hour leaving the battery happened in the 0.5–2 A range.

That looks completely believable for Vanilla.

Fridge cycling.

Monitoring systems.

USB devices.

Lighting.

Control electronics.

The quiet background activity of a campervan that is never completely asleep.

At around 12.5 V:

0.5 A is roughly 6 W.

0.8 A is about 10 W.

1 A is approximately 12.5 W.

2 A is roughly 25 W.

So the ordinary baseline appears completely sensible.

I do not see evidence of some mysterious permanent 4 or 5 A parasitic drain hiding in the electrical system.

That would have jumped out immediately.

Instead, large loads appear as distinct blocks.

They arrive.

They operate.

Then they disappear.

That looks much more like equipment being deliberately used than an electrical fault.

This is one of the most useful discoveries.

Looking at nights where there was no meaningful positive charging between approximately 20:00 and 07:00, the median overnight draw was around:

9.23 Ah.

That works out at approximately:

0.84 A average.

At normal 12 V battery voltage that is somewhere around:

10–11 W continuous average load.

That figure lines up beautifully with the median negative hourly reading.

So I now have a genuine real-world Vanilla overnight baseline.

A relatively quiet night appears to cost around 9 Ah from the leisure battery.

That is going to be hugely useful when the Fogstar lithium battery goes in.

I’ll be able to compare exactly the same overnight behaviour on AGM and LiFePO₄ rather than relying on memory or rough battery percentages.

Separating the data into daytime and night-time makes the solar influence obvious.

During roughly 08:00–17:00, the battery received about:

+721 Ah of charging

while supplying around:

−206 Ah.

That leaves a daytime net gain of roughly:

+515 Ah.

During evening and night, around 18:00–07:00, Vanilla received approximately:

+149 Ah

while supplying about:

−574 Ah.

That leaves a night-time net of around:

−424 Ah.

So roughly 83% of all recorded charging happened during the daytime.

And roughly 74% of all discharge happened during evening and night.

That is exactly the pattern I wanted from a solar-supported campervan.

The roof fills the bucket during the day.

Vanilla drinks from it overnight.

The really interesting stuff lives in the relatively small number of hours where discharge exceeded 4 A.

Only 17 hours out of the entire 1,107-hour dataset recorded more than 4 Ah leaving the battery.

Yet those 17 hours consumed around:

100.92 Ah.

That is nearly 13% of every recorded amp-hour discharged during the whole dataset.

So the big loads matter enormously.

They are rare.

But when they happen, they dominate consumption.

This is one of the clearest heavy-load events.

07:00: −6.85 Ah

08:00: −6.78 Ah

09:00: −7.01 Ah

10:00: −2.49 Ah

11:00: −5.66 Ah

That is approximately:

−28.79 Ah in five hours.

The first three hours alone average almost:

−6.9 A.

At around 12.5 V, that represents roughly:

86 W net.

That is far too much to be normal fridge operation.

Something meaningful was running.

A sustained 80–100 W electrical load fits this pattern extremely well.

One possibility is charging one of my Jackery power stations through the Anderson feed at around 90 W.

The CSV cannot prove that.

But electrically, it fits remarkably well.

Another very clear event appears late in the evening.

21:00: −7.36 Ah

22:00: −5.02 Ah

That is approximately:

−12.38 Ah in two hours.

Average draw:

6.19 A.

At typical battery voltage, that is somewhere around:

75–80 W average.

Again, far beyond the usual background load.

Because it happened at night, solar cannot have been masking or contributing anything meaningful.

Something substantial was switched on.

I would not expect a diesel heater alone to average that sort of power continuously for two whole hourly blocks once it had finished its initial startup phase.

So again this looks more like a significant electrical device or power-station charging event.

This one looks slightly different.

21:00: −4.50 Ah

22:00: −4.33 Ah

23:00: −3.19 Ah

Total:

−12.02 Ah over three hours.

That is around:

4 A average.

At typical voltage, roughly 50 W.

That feels more like several ordinary campervan systems operating together.

Fridge.

Lighting.

Heating.

Entertainment.

Phone or device charging.

Possibly multiple loads stacked together.

At 07:00 the battery recorded:

−4.17 Ah.

At 08:00:

−4.84 Ah.

So around:

−9.01 Ah disappeared in two hours.

What makes this particularly interesting is that 26 September was also the best solar day in the SmartSolar history.

The panel harvested:

570 Wh.

Yet the SmartShunt still finished the entire day at approximately:

−16.61 Ah net.

That is a fantastic demonstration of why solar yield cannot be judged in isolation.

The panel had its best day.

The system simply consumed even more.

The solar installation was not struggling.

It was working hard while Vanilla was working harder.

This remains the most dramatic electrical day in the whole dataset.

At 11:00 the battery was at:

100%.

Then the drain begins.

12:00: −1.69 Ah

13:00: −5.07 Ah

14:00: −7.65 Ah

15:00: −6.72 Ah

16:00: −6.30 Ah

17:00: −4.86 Ah

18:00: −3.18 Ah

19:00: −2.02 Ah

The five-hour block from 13:00 to 17:00 removed approximately:

30.60 Ah.

That is an average net discharge of around:

6.12 A.

At approximately 12.5 V, that is somewhere around:

75–80 W net.

But there is an important twist.

The SmartSolar data shows the solar panel harvested:

300 Wh that day.

Maximum PV output reached:

102 W.

So during at least part of that heavy battery drain, the roof was simultaneously producing electricity.

That means the actual electrical load must have been higher than the SmartShunt negative value suggests.

If solar was adding 2 or 3 A at a particular moment while the SmartShunt still showed −6 A, then the loads might really have been consuming 8 or 9 A.

That could easily place total system demand around or above 100 W.

That again fits very neatly with a roughly 90 W Jackery DC charging load plus Vanilla’s normal background consumption.

I cannot prove that from the logs.

But the electrical fingerprint is extremely convincing.

Another interesting morning sequence appears here.

08:00: −3.12 Ah

09:00: −6.87 Ah

10:00: −2.48 Ah

That is:

−12.47 Ah in three hours.

Then the direction completely changes.

11:00: +4.96 Ah

12:00: +4.85 Ah

13:00: +2.08 Ah

14:00: +1.35 Ah

15:00: +5.02 Ah

16:00: +1.77 Ah

17:00: +1.51 Ah

The SmartSolar harvested:

390 Wh that day.

This looks like a heavy morning load being switched off, followed by the solar and charging systems taking control and rebuilding the battery.

It is almost a textbook example of how the SmartShunt records NET battery current rather than individual sources.

The incoming current tells just as interesting a story.

Approximately 34 Ah came in during very gentle charging hours below 0.5 A.

Around 46 Ah arrived between 0.5 and 1 A.

About 90 Ah arrived between 1 and 2 A.

Approximately 197 Ah arrived between 2 and 4 A.

Around 189 Ah arrived during 4–8 A charging hours.

But the really revealing category is above 8 A.

There were only:

28 hourly readings

where charging exceeded 8 Ah.

Yet those 28 hours delivered approximately:

314 Ah.

That means only a tiny number of very strong charging hours account for around:

36% of every recorded amp-hour entering the battery.

That is the difference between solar and a strong external charger.

Solar is the marathon runner.

DC-DC and EHU arrive carrying buckets.

This is probably the clearest example of substantial external charging.

10:00: +14.89 Ah

11:00: +18.35 Ah

12:00: +6.37 Ah

That is approximately:

39.61 Ah in three hours.

The first two hours alone added:

33.24 Ah.

A genuine +18.35 A average into the battery during an hour represents well over 200 W of net battery charging power.

And the source has to provide even more because Vanilla is consuming energy at the same time.

This is not ordinary campervan solar.

This is serious charging.

It strongly suggests DC-DC alternator charging, mains charging, or another major charging source.

From 2 September onward, the SmartSolar file gives me the maximum solar output for each day.

That lets me calculate whether a SmartShunt charging event could physically have come from solar alone.

29 September is a perfect example.

The maximum solar output recorded that entire day was only:

58 W.

Even at a very low 12 V battery voltage, 58 W could provide less than:

5 A.

Yet the SmartShunt recorded:

08:00: +15.30 Ah

09:00: +11.20 Ah

10:00: +7.05 Ah

Solar cannot possibly account for that.

Another charging source was unquestionably operating.

That is the beauty of having both datasets.

The SmartShunt says:

“The battery gained energy.”

The SmartSolar says:

“Here is exactly how much of that could possibly have come from me.”

The rest must have arrived through another route.

This is the cleanest example of what looks like a major external recharge event.

The battery begins at around:

55% SOC.

Then:

07:00: +3.83 Ah

08:00: +15.30 Ah

09:00: +11.20 Ah

10:00: +7.05 Ah

11:00: +3.05 Ah

SOC climbs:

07:00: 55%

08:00: 65%

09:00: 78%

10:00: 86.5%

11:00: 98.5%

12:00: 100%

The whole day’s solar harvest was only:

180 Wh.

Maximum PV power was only:

58 W.

So most of this recovery absolutely did not come from the roof.

This has exactly the sort of shape I would expect from Vanilla being driven and the Victron Orion DC-DC charger aggressively replenishing the leisure battery.

EHU could also produce a similar effect, but the charging pattern is classic strong external charging.

1 October is particularly useful because I know EHU was connected.

That gives me a confirmed charging signature.

The battery had fallen to around:

by 10:00.

Then charging begins.

15:00: +3.60 Ah

16:00: +8.51 Ah

17:00: +5.73 Ah

18:00: +3.63 Ah

19:00: +2.79 Ah

20:00: +1.99 Ah

SOC reaches:

100%.

But then the really revealing part happens.

Charging keeps going.

21:00: +1.64 Ah

22:00: +1.19 Ah

23:00: +0.91 Ah

00:00: +0.78 Ah

01:00: +0.66 Ah

02:00: +0.55 Ah

03:00: +0.49 Ah

04:00: +0.43 Ah

05:00: +0.38 Ah

06:00: +0.32 Ah

08:00: +0.24 Ah

That is a beautifully clear charge taper.

Strong initial replenishment.

Gradually reducing current.

Then maintenance.

The positive charging run lasted approximately:

19 consecutive hours.

Total recorded charge during that continuous period was around:

34.46 Ah.

Because much of it happens overnight, it obviously cannot be solar.

That is the mains charger leaving its signature in the data.