Brace yourselves... It's ANOTHER 12V/230V fridge question

[Alan de Enfield]

5 minutes ago, captain flint said:

'Ere, Alan. I'm not doubting you may be right about this and that I may have the wrong end of the stick in some way, but that's not what it seems to say here

 

Easy to be confused - to add to those dimensions you have to add the thickness of the insulation (which will be between 0.8mm and 1mm wall thickness) 

So the thickness of the 'wire' at 3mm OD will only have approximately 1.4mm OD of conductor which is about a 1.5mm2 conductor.

 

That table would look to apply to a single strand of wire - boat cables are manufactured for multistrand conductors and there are many intertices which add to the diameter but not to the CSA.

Edited November 1, 2023 by Alan de Enfield

[captain flint]

2 minutes ago, magnetman said:

It is all about the knees. 

 

Keep away from these areas and the batteries will last for ages. 

 

Someone who knows did explain it to me once. Its to do with the ions. 

 

They don't like being too excited or too dormant. 

 

Something like that anyway. 

 

I like LTO batteries as they are abuse tolerant. 

 

I have no idea where to find a battery's knees. I guess I'll have to do some googling...

1 minute ago, Alan de Enfield said:

 

Easy to be confused - to add to those dimensions you have to add the thickness of the insulation (which will be between 0.8mm and 1mm wall thickness) 

So the thickness of the 'wire' at 3mm OD will only have approximately 1.4mm OD of conductor which is about a 1.5mm2 conductor.

Ah, right, well in my case the diameter of the copper is 3mm. c. 5mm plus the sheathing

Turned it off at the fuse box and got my steel rule onto the little bit of exposed copper I can see at the little junction box thingy

 

[Alan de Enfield]

3 minutes ago, captain flint said:

I have no idea where to find a battery's knees. I guess I'll have to do some googling...

 

About half way between its ankles and its hips.

[ditchcrawler]

17 minutes ago, captain flint said:

Thanks, Alan. Kind of what I was expecting to hear but with some extra info, I was ignorant about CSA vs diameter, so that's another tiny (and very basic) chip off the enormous rock of my ignorance. 

Ah well. It was nice to dream of a new fridge. But I guess now I need to focus on dreaming of new cables, first. I wish I could be advised online about the best way to replace my fridge cables, but I think it's the kind of thing you really need to be on board in person to be able to say anything much sensible, right.

If you don't know what your cable runs are you cant know how long they are so you need to establish that before you can work out the cable size. BTW I have a 240 volt fridge

[magnetman]

Knees for a LiFePO4 battery:

 

 

Multiply the left column by 4 for a 12v equivalent battery. 

 

You want to keep away from the areas where there is a rapid change in voltage. SOC is state of charge. There is very little to be gained by charging to anything over about 3.4*4 which is 13.6v. 

 

 

Edited November 1, 2023 by magnetman

[ditchcrawler]

12 minutes ago, captain flint said:

'Ere, Alan. I'm not doubting you may be right about this and that I may have the wrong end of the stick in some way, but that's not what it seems to say here

That is the diameter V Cross section of the copper conductor, not the total wire diameter including insulation 

[magnetman]

So basically in the case of an LFP battery keep it between 3.2 (12.8) and 3.4 (13.6). 

 

(data from here https://www.researchgate.net/figure/SOC-versus-OCV-Curve-for-an-18650-Lithium-Iron-Phosphate-LFP-Battery-Cell_fig1_304655337 

Edited November 1, 2023 by magnetman

[captain flint]

1 minute ago, ditchcrawler said:

If you don't know what your cable runs are you cant know how long they are so you need to establish that before you can work out the cable size. BTW I have a 240 volt fridge

weeeelll.. I mean. I know how much cable sticks out at each end. I can measure that. Then I can measure the travel round the cabin sides from one hole where they go into the cabin lining to the other hole where they come out. Of course, they might be meandering all over the place in there. I can't see. But I bet they aren't, at least not much.  And really, whilst acknowledging my profound ignorance about the technical stuff we're discussing here, I'm hazarding a guess that this sort of measurement-assisted estimate is good enough for our purposes here. I mean, I don't want to do down the ingenuity and knowhow that goes into canal boats, but it's not NASA, right?

According to that rough measurement it's 6m but I called it 6-7 to account for the fact I can't actually see all the cable

2 minutes ago, ditchcrawler said:

hat is the diameter V Cross section of the copper conductor, not the total wire diameter including insulation 

yeah as per my previous message to Alan the diameter including insulation is more 5mm, the diameter of the copper itself is 3mm. Well, obviously, measuring diameter with a steel rule is not exactly an exact science, but I'm saying 3mm near as hot dammit. 

5 minutes ago, magnetman said:

So basically in the case of an LFP battery keep it between 3.2 (12.8) and 3.4 (13.6). 

Thanks!

[Tracy D'arth]

14 minutes ago, captain flint said:

I have no idea where to find a battery's knees. I guess I'll have to do some googling...

Ah, right, well in my case the diameter of the copper is 3mm. c. 5mm plus the sheathing

Turned it off at the fuse box and got my steel rule onto the little bit of exposed copper I can see at the little junction box thingy

 

Its not the diameter of the copper but the Cross Sectional Area, the area of the cut end, a very different thing.

[captain flint]

8 minutes ago, Tracy D'arth said:

Its not the diameter of the copper but the Cross Sectional Area, the area of the cut end, a very different thing.

Thanks, Tracy! Yeah, I found this table useful. I do hope it's correct, as that would mean it looks like my cables are about right
 

So after a little learning, I'm thinking my cables should in fact be OK, and I am leaning towards a 12V fridge.

 

Problem I have is every time I read which people like, and think, right then, the consensus seems to be Inlander/Shoreline/Dometic/Whatever is the way forward, I find a post saying, "Well, our Inlander/Shoreline/Dometic/Whatever was utter rubbish and broke down all the time."

 

Ah well. You shouldn't ask for people's opinions if you don't want a range of views.

 

At least it's not as bad as wood finishes. I swear, I have 3 friends who are all really experienced professional carpenters, and any time I ask them about finishes, they will say, "Well, for what you need here, loads of people would say X or Y but they all all 100% wrong for these [very impressive sounding] reasons. What you want is Z". But of course they never agree with each other!

Edited November 1, 2023 by captain flint

[Tony Brooks]

1 hour ago, Tracy D'arth said:

Beware that battery capacity monitors tell lies unless you regularly  re-calibrate them.

and have set them up properly in the first place.

[Tracy D'arth]

1 minute ago, captain flint said:

Thanks, Tracy! Yeah, I found this table useful. I do hope it's correct, as that would mean it looks like my cables are about right

That table may be for solid single core cable wire. Your boat must be wired with multi strand flexible cable, not domestic copper cable with one solid core or the 7 strand type which is not flexible cable.

As a rule of thumb the cable should be one millimetre in CSA for ever metre of cable between the fridge and the batteries. 

[captain flint]

1 minute ago, Tracy D'arth said:

That table may be for solid single core cable wire.

 Ah. yes. Good point. Ta.

[magnetman]

 

 

If you are marginal about having a fridge you could consider a compressor coolbox. 

 

They are not as good as fridges and don't have a condensation drain but something like the Dellonda (uk company buying chinese things probably made by slaves) 15 litre cool box does work. You need to clean them quite often due to the lack of condensation drain. Compact, efficient and top loading so you don't have to lick the floor to get the beer out. 

 

I quite like mine but have no need for a fridge because the shops have them and I shop every day. 

 

Great for cold beer in summer though. 

[captain flint]

8 minutes ago, magnetman said:

If you are marginal about having a fridge you could consider a compressor coolbox. 

I'll probably stick with a fridge (if nothing else I will just keep the old one I am currently running) as I don't shop every day. But thanks for the tip.

[IanD]

24 minutes ago, Tony Brooks said:

and have set them up properly in the first place.

Set up SoC monitor in the first place to suit the batteries : absolutely. This should be done for any commercial system with an internal BMS or one which is matched to the cells.

 

Recalibrate (regularly or not) : definitely not needed, AFAIK all LFP BMS self-calibrate by measuring changes in voltage/current slope, they track any shift in the "knee" positions,

 

Depending on the BMS, they may also redefine 0% and 100% SoC *as seen by the user* to be these slope change points -- for example given the curve posted above by @magnetman they may set 0% (user) to be 10% (3.2V/cell) and 100% (user) to be 95% (3.4V/cell). Which would lose about 15% off the theoretical cell capacity (which is invisible to the end user) but means they can be treated as "idiot-proof" anywhere from 0% to 100% SoC.

 

I suspect this is what manufacturers like BYD do, they just specify lifetime energy use in MWh with no restriction at all about SoC range (e.g. 46MWh for a 15.4kWh battery).

Edited November 1, 2023 by IanD

[nicknorman]

2 minutes ago, IanD said:

Set up SoC monitor in the first place to suit the batteries : absolutely. This should be done for any commercial system with an internal BMS or one which is matched to the cells.

 

Recalibrate (regularly or not) : definitely not needed, AFAIK all LFP BMS self-calibrate by measuring changes in voltage/current slope, they track any shift in the "knee" positions,

 

Depending on the BMS, they may also redefine 0% and 100% SoC *as seen by the user* to be these slope change points -- for example given the curve posted above by @magnetman they may set 0% (user) to be 10% (3.2V/cell) and 100% (user) to be 95% (3.4V/cell). Which would lose about 15% off the theoretical cell capacity (which is invisible to the end user) but means they can be treated as "idiot-proof" anywhere from 0% to 100% SoC.

 

I suspect this is what manufacturers like BYD do, they just specify lifetime energy use in MWh with no restriction at all about SoC range (e.g. 46MWh for a 15.4kWh battery).

Bearing in mind the very flat voltage vs SoC curve over much of the SoC range, I would have thought that even a fairly good BMS would struggle to keep accurate track of SoC if it is kept around the middle and doesn’t approach the knees from time to time.

[IanD]

21 minutes ago, nicknorman said:

Bearing in mind the very flat voltage vs SoC curve over much of the SoC range, I would have thought that even a fairly good BMS would struggle to keep accurate track of SoC if it is kept around the middle and doesn’t approach the knees from time to time.

That's true to a certain extent, the amp-hr counting is pretty accurate but there will be some accumulated error, also in voltage measurements -- though these are consistent down to the millivolt level. The slope in the flat region is about 1mV/1% SoC (for a single cell) which is 16mV/1% for a 16S (48V) battery, this is perfectly measurable.

 

As soon as it ever gets near either knee the endpoints get recalibrated, and once both have been done the battery health/capacity can be updated.

 

Here's an example of what mine's done over the last couple of weeks -- moored on shorepower initially, then solar (not enough to keep up) with occasional shore charging bursts.

 

 

Edited November 1, 2023 by IanD

[captain flint]

38 minutes ago, captain flint said:

 Ah. yes. Good point. Ta.

 

My multi strand cable has 81 strands. I'll properly measure one of the strands with a micrometer when I get the chance (and before I take any decision), but as a rough estimate, 3 lined up very carefully are just about 1mm, 4 are way over, and 0.3mm is a fairly standard diameter for these things according to a couple of cable manufacturers. If each strand is 0.3mm then my cable would be 5.725 mm2. It should ideally be more like 6 or maybe 7mm2, you but I guess I'm wondering if such a small shortfall will make a big difference to things. Not sure anyone can answer that but you never know

30 minutes ago, IanD said:

A bunch of clever stuff I didn't really follow

The installer told me no need for me to recalibrate, but that I should charge them up fully every now and then to keep readings accurate

[IanD]

7 minutes ago, captain flint said:

 

My multi strand cable has 81 strands. I'll properly measure one of the strands with a micrometer when I get the chance (and before I take any decision), but as a rough estimate, 3 lined up very carefully are just about 1mm, 4 are way over, and 0.3mm is a fairly standard diameter for these things according to a couple of cable manufacturers. If each strand is 0.3mm then my cable would be 5.725 mm2. It should ideally be more like 6 or maybe 7mm2, you but I guess I'm wondering if such a small shortfall will make a big difference to things. Not sure anyone can answer that but you never know

The installer told me no need for me to recalibrate, but that I should charge them up fully every now and then to keep readings accurate

That's making sure that 100% SoC is accurately known -- if you look at the picture I attached you can see the very rapid rise in battery voltage at this point.

[LadyG]

1 hour ago, captain flint said:

Thanks, Tracy! Yeah, I found this table useful. I do hope it's correct, as that would mean it looks like my cables are about right
 

So after a little learning, I'm thinking my cables should in fact be OK, and I am leaning towards a 12V fridge.

 

Problem I have is every time I read which people like, and think, right then, the consensus seems to be Inlander/Shoreline/Dometic/Whatever is the way forward, I find a post saying, "Well, our Inlander/Shoreline/Dometic/Whatever was utter rubbish and broke down all the time."

 

Ah well. You shouldn't ask for people's opinions if you don't want a range of views.

 

At least it's not as bad as wood finishes. I swear, I have 3 friends who are all really experienced professional carpenters, and any time I ask them about finishes, they will say, "Well, for what you need here, loads of people would say X or Y but they all all 100% wrong for these [very impressive sounding] reasons. What you want is Z". But of course they never agree with each other!

I have just ordered a Coolmatic , CRX80, to replace the original one which is 23 years old.

The new one will be quieter, and have a good door seal, be more efficient and has an optional freezer / larder fridge option.

I live in hope it will work as advertised.

It's expensive [~£699.95], and you have to shop around to get best price.

Edited November 1, 2023 by LadyG

[captain flint]

2 minutes ago, LadyG said:

I have just ordered a Coolmatic to replace the original one which is 23 years old.

...
I live in hope it will work as advertised.

 

Good luck, if it lasts as long as your last then it will be doing pretty well I guess!

I was wrong about mine, it's not a Coolmatic. It's a Coolzone (?!). It has a danfoss but the fridge itself is pretty cheap and cheerful, not very well insulated

 

[nicknorman]

38 minutes ago, IanD said:

That's true to a certain extent, the amp-hr counting is pretty accurate but there will be some accumulated error, also in voltage measurements -- thought these are consistent down to the millivolt level. The slope in the flat region is about 1mV/1% SoC (for a single cell) which is 16mV/1% for a 16S (48V) battery, this is perfectly measurable.

 

As soon as it ever gets near either knee the endpoints get recalibrated, and once both have been done the battery health/capacity can be updated.

Thread drift alert.

 

Yes I guess that measurement to 1mV or so is not that difficult - the chip I use has that sort of accuracy if temperature compensation is applied. One would need to model the cell internal resistance though, which I don’t think is fixed even at fixed temperature, unless the calc is only done when current drain is near zero. My setup uses the BMV712 to track SoC based on current integral, and it’s not bad at doing that. I did envisage some algorithm to tweak the BMV SoC (my BMS can write a new SoC to the BMV) but I never got around to it. Instead I just fully charge every month or so and the BMV resets to 100%. Much easier!

[Alan de Enfield]

48 minutes ago, captain flint said:

 

My multi strand cable has 81 strands. I'll properly measure one of the strands with a micrometer when I get the chance (and before I take any decision), but as a rough estimate, 3 lined up very carefully are just about 1mm, 4 are way over, and 0.3mm is a fairly standard diameter for these things according to a couple of cable manufacturers. If each strand is 0.3mm then my cable would be 5.725 mm2. It should ideally be more like 6 or maybe 7mm2, you but I guess I'm wondering if such a small shortfall will make a big difference to things. Not sure anyone can answer that but you never know

The installer told me no need for me to recalibrate, but that I should charge them up fully every now and then to keep readings accurate

 

 

84 strands of 0.3mm wires (noted as 84/0.30mm) is a standard size and is 6mm2 (56/0.30 is 4mm2) the next standard size above 6mm is 7mm which is 97/0.30 which is the absolute minimum you need if the run is 7 metres. Ideally you should go up to at least 8.5mm which is 120/0.30mm.

 

These are automotive cables manufactured to BS6862 and are the most common used in boat wiring.

 

I had over 10 years workimg for a cable manufacturer, inventing new cables and writing specifications for the BS, the NCB and the Aerospace industry.

 

There is a noticable difference in volt drop between 6mm2 and 7mm2 wires. 

 

This is the installation instructions from the manual for the 12v fridge I bought - you'll notice that they recommend a cable size of 10mm2 for a 7 metre run.

 

 

Edited November 1, 2023 by Alan de Enfield

[magnetman]

Or put a small battery by the fridge and use smaller wires to feed the battery.

 

 

Compressor fridges don't run on a continuous tour of duty. They have breaks.