Hi everyone,

I'm seeking some some advice: I'm designing a small underwater datalogging system that's powered by four 24V lead acid batteries connected in parallel. It's main purpose is to be set and left for extended periods of time, hence all the battery power.

I'd like to add the feature that when working with the device (configuration, data retrieval, etc), I can plug in a 48V battery that will take over from the 24V bank.

I'm just wondering if what I have below is suitable. I've been doing some research on connecting batteries in parallel and finding some "don't do this" in some places and "it's fine, just use fuses" in other places, so some advice those who with experience would be greatly appreciated.

 

Putting lead acid batteries in parallel is fine (as in jump starting a car) but I would put a diode in series with each battery to stop a higher charge battery charging a lower charged battery. You can then eliminate the sigle diode you presently have between the 24V battery and the DC-DC converter.

 

I would think about keeping the internal blocking diodes as a dual power connector safety (hot connector with one side plugged in) factor.

 

Ah yes... I had originally planned on using diodes instead of fuses in series with each battery (before I did any research)... would you folks suggest using BOTH? Are the battery fuses necessary?

 

Think of the condition with one battery a lower charge than another, then your charging the lower charge battery with almost infinite current,,,
A diode on each battery output is required,

What sort of current you looking at drawing,
diodes drop a fair bit of power, ( 1 V drop at 1 amp is 1 watt )
there are chips called perfect diodes , and also current share , which use mosfets as very low Vforward drop diodes
https://www.analog.com/en/products/...ion/powerpath-ideal-diodes-load-switches.html

What are you doing for charging ? are you intending to un plug and charge each individually ?

The other thing I'd consider, is what if one battery fails ? If the voltage on one cell gets to low, it should be disconnected, lead acid arn't to bad, but if you move to Lion, then .... discharging a cell to zero leads to permanently dead cell.

Fuse wise, you might want to look at active fuses, either PTC types , or again there are IC's that monitor current and switch off the drive if to high a current. Advantage is these types can auto reset, so if a temp short occures then you don't permanently loose the unit. The IC based ones , also do things like "try 3 times over a period of so many seconds to re connect " and if it still fails then give up, and disconnect permanently,

 

How fast can your DC2DC converter react to a doubling of input? It may be in the datasheet. Worth a try on the bench.

 

Thanks for the input everyone

What sort of current you looking at drawing

It will be drawing no more than 7A

What are you doing for charging ? are you intending to un plug and charge each individually ?

Ya, I'll be unplugging them and charging them individually.

Fuse wise, you might want to look at active fuses, either PTC types , or again there are IC's that monitor current and switch off the drive if to high a current.

This sound interesting, I'll look into it. Again, fuses are necessary for each battery, no?

How fast can your DC2DC converter react to a doubling of input?

I haven't started picking specific components yet, but I'll definitely keep this in mind when selecting a DC2DC.

 

The diodes should work, but as mentioned previously, hot swapping the two different battery banks *might* create inrush current issues and perhaps temporary loss of regulation.

If you find yourself in such a situation, perhaps you should take a look at hot swap controllers. These not only allow an orderly battery bank removal/insertion, some of them have adjustable current limits and/or circuit breaker capabilities.

Additionally, they drive external Mosfets, meaning that the insertion loss will be much lower than using diodes.
For instance:

https://para.maximintegrated.com/en/results.mvp?fam=hot_swap

 

If you find yourself in such a situation, perhaps you should take a look at hot swap controllers.

These seem very useful, I'll definitely look into them.

For anyone interested, here's an updated diagram, ditching the battery fuses, re-adding the diode array, but keeping the additional diode between the array and the 48V battery. I just learned from the person I'm making this with (also not an engineer) that the batteries are actually fully fledged underwater ones, so all the connections need to happen inside the water-tight casing (at least as it stands now - there's chance of adding a separate battery connector case).

Any thoughts are welcome! I'm definitely going to look into the components mentioned above...

 

Think of the condition with one battery a lower charge than another, then your charging the lower charge battery with almost infinite current,,,
A diode on each battery output is required,

Thats only true if the batteries are massively out of balance. If they are maintained and properly charged its normal practice to connect in parallel, thats how it was done in old 48v Stroweger phone exchanges. Indeed, if you put a diode in series with each battery are you going to have 4 separate chargers? Also look at big lithium batteries, its common practice to parallel cells to increase capacity, no diodes there and arguably a worse scenario given their lower internal resistance. And how much power loss/heat are the diodes generating? Ditch the diodes. If you're really worried, by all means put a fuse in each battery link.

 

There is a famous paper with the answer. I'll try to post the PDF, but if it fails here is the citation.
Popular Electronics 1968, page 86, Edwin T Jaynes
"You can parallel storage batteries"

https://bayes.wustl.edu/etj/node1.html
Jaynes, E. T., 1968, `You CAN Parallel Storage Batteries,' in Popular Electronics, 33, 86; as a level 2 postscript file.

 

I'm just wondering if this isn't something an Arduino could do and consume much less power. Also, lead-acid cells have signifcant self-discharge, so might not be the best choice if "extended periods" means months.