Rec: end-current cutoff at low amps

[john61ct]

Originally Posted by Maine Sail

The other problem you have, other than the OFF trigger, point will be the ON trigger point.

A well-calibrated coulomb-counter may well be the best centralized control point, well avoiding the shoulders on both ends.

I believe a bottom limit of 10-15% SoC would be safe, leave the BMS-type failsafe to deal with voltage drop separately, maybe a relatively high bank-level cutoff, lower per-cell setpoint disables just that 12V block.

Voltage disconnect at the whole bank can be set very low at the high side, since no charge source will ever be set (intentionally, times of normal usage) at 14V or above.

Is there any need in this scenario for OVD at the per-cell level?

[john61ct]

Originally Posted by Ding Duck

Have a look at the Allegro ACS758 / 759 Hall effect Current sensor. They do up to 200A in a combined unit, but options to go to 1000A. The 200A version is about $10 on ebay from China. Plenty of You Tube stuff on integrating these into Arduino.
All you have to do then is to decide on how your going to control the output.

thanks very much!

However it seems the versions that tolerate high amps don't have the resolution needed at the low end.

Coulomb-counting units like Bogart Pentametric will work with 100A, 500A and 1000A shunts, you do lose some resolution going higher, but OK for this use case.

They are designed to work with computer inputs like USB and ethernet, even old-school RS232, so perhaps a translation layer can be written to feed the Arduino inputs.

[john61ct]

Too bad Gibbo hasn't come up with an updated Smartgauge to work with LFP banks, I've got one coming in to work with my eventual Firefly Reserve/starter bank, along with some (not sure working) Smartbank units I found cheap.

Merlin has come out with a few SoC monitoring units since their SG purchase, haven't seen any hands-on reviews of those yet.

I guess I should check into what control / data output options they might offer?

[john61ct]

Originally Posted by a64pilot

I know my Magnum inverter charger is able to be set to drop out of Absorption to float at a set number of amps.
However I believe it then drops into float, I know you want off, but if you set float voltage low enough, would that be acceptable?

Yes, I would rather avoid float, since the (likely negative) effect of even low-voltage float on LFP longevity is such an unknown.

In fact that's one reason I'm looking for (to create) this functionality

*** independent of any particular charge source ***

since so few even top-end good one$$ let you disable their float stage. Everyone designs for lead chemistries in the general-use non-proprietary space, while the LFP-specific vendors are milking the relatively wealthy early adopters with be$poke integrated proprietary "solution" packages.

Part of my thinking is that separating the monitoring and control intelligence from the charge current source, can free the user up from having to pay the premium for that intelligence over and over again, in the Alt's voltage regulators, wind/solar controllers, fuel cell charger AND the shore charger.

Maybe an Arduino-based unit can be made reliable and robust enough to act as such a central "charging control hub" in a marine, or at least mobile context. Something like what this guy was looking for: http://www.cruisersforum.com/forums/...ad.php?t=23858

And again, this use case will actually see mains power only rarely.

[a64pilot]

Surely there is a low voltage disconnect available?
Set float voltage to a point under the disconnect voltage and charger gets disconnected when it is in float.

Excuse my ignorance in most things Lithium, but you can't use voltage to determine SOC very well I assume, you use an amp counter?
If so then you have to also have a way that an amp counter will kick the charger back into absorption, which would be above the low volt cut off and it would then be reconnected to the bank. I think maybe a Magnum BMK will turn charger back on based on number of amps out, but do not know that

[john61ct]

Yes, low voltage cutoffs are readily available, including ACR/VSR combiners, even ones with custom setpoints, but they get expensive at high current capacities.

But they're off topic here, not relevant to the use case.

Yes, above I was talking about using shunt-based coulomb counters as a possible alternative, but I would prefer the more reliably accurate basis of realtime measurement of current on the charge buss.

Don't assume any charging algorithm intelligence from the current source, for the sake of clarifying my need here, let's just say it is a DC power supply set to 13.8V, and the bank is as usual dictating the CC-CV transition to Absorb, and what I'm looking for is that

THIS device will switch off the current flow to the bank, when that flow drops to an X amps setpoint. Either by

opening / isolating the two circuits

and/or

turning off the charge source.

I'm not also not concerned here about restarting the charge cycle, but thanks for pointing out the Magnum BMK.

[Ding Duck]

Hi John, unit looks like it's got everything in it.
Still think you will run into the same problems. Accuracy is +/- 1.5% +/- Least significant digit. To me that means it could be anywhere from -1.5 to +1.5% plus the LSD factor. For 100A, you could be out by 3 Amps. You will need to determine what the LSD will be based on an 500 or 1000A Shunt etc.
If your using this for SOC, then you need a calibration setting initially and then periodic re-calibration.
The issue is the low amps on such as wide band. Generally, once you below say 20% the error starts to increase. Obviously you need a certain amount of power to drive whatever you are using to sense the current and if your at 2A, then the output will be 0.0005V (500A/50mV Shunt). Having along length of cable could change this value as well.
Cheers

[john61ct]

Sorry I don't follow.

Originally Posted by Ding Duck

Hi John, unit looks like it's got everything in it.

What unit?

> Still think you will run into the same problems. Accuracy is +/- 1.5% +/- Least significant digit. To me that means it could be anywhere from -1.5 to +1.5% plus the LSD factor.

No need for great accuracy, not using this to measure SoC.

As stated in the OP, goal is to isolate when current amps drop. If shooting for a cutoff of .025C, e.g. 5A on a 200AH bank, then stopping at 7A or even 10A is fine. Even pushing current until 2AH isn't going to do "damage", but with LFP sooner is better than later.

[DeepFrz]

Is this what you want?
Arduino based Alternator Regulator: Welcome,

[john61ct]

That's one of the more active projects I found, referenced above, might use as a starting point.

The guy's very helpful, hangs out in the trawler forums, boat's name's "Viking Star". His wife's got her own blog as well.

[thomasow]

Originally Posted by john61ct

That's one of the more active projects I found, referenced above, might use as a starting point.

The guy's very helpful, hangs out in the trawler forums, boat's name's "Viking Star". His wife's got her own blog as well.

And that would be me! Do not monitor the Web based forums much, but noted a spike in traffic and traced it partly to here.

John, not fully sure what you are looking to do - but in reading you are seeking an ability to monitor current over a wide range. On my regulator I am using the INA226 which features a 16 bit ADC on the current shunt, giving very fine resolution and wide dynamic range. Accuracy is another matter, largely dependent on the shunt used. You could take one of the blank PCBs, solder on only the components needed - then use the Feature-out port to drive an external relay... Just an idea,

On using acceptance current for transitions: This is, to be honest, the reasons I created this alternator regulator. Being a full time cruiser (and away from the docks 8mo or so a year) it became very apparent to me the limitations of time-only based transitions in the real world. (Having to literally stop the boat, turn off the engine and restart, to 'trick' the regulator back into recharging my battery as it entered float before full charge no matter how I configured it was the last straw).

Just picking the right 'timer value' is a guess, largely mitigated by FLAs ability to accept overcharging to a limited manner - but add in any dynamic house load (Maybe even running the washer/dryer while underway), initial SOC, temperature, batteries less accepting of slight overcharing - and in the real-world it is very problematic to truly recharge batteries w/o proper measurement - let alone keep from overcharging a battery when say leaving the dock with an already high SOC. There are techniques for mitigating these (time-to-achieve-setpoint-voltage timers, Field drive % thresholds, etc) - all help but are still have limitations.

Almost without except battery manufactures specify measuring acceptance current as the prefer way of deciding transaction into float. So that is what I did - measure, not guess.

MainSail - I see your point for liability (which I suspect is why most regulator tend to undercharge, issues do not show up right away and is easy to discount). It is not only install issues, but also failures which need to be consider and addressed - protect the battery at all points. Of course many boats already have a current shunt at the battery - is simple to reuse that, and that is what most of the folks using my regulator are doing. Then comes a matter of following the battery manufactures recomended charge profile + safeguards.

Couple of other thoughts I picked up in this thread: External signaling to turn off charging (ala, Li based batteries getting a signal from a BMS) - have that feature enabled. Other topic seems to touch on the idea of a battery monitor -- FWI: For a more capable deployment / simplified install a few of us are well along in developing a CAN based comms architecture to allow a simple monitoring device at the battery to instrument voltage, current, temp, SOC, etc; as well as coordinate charging sources to meet the batteries needs. All build upon proven industrial technologies and open standards and again with fall-back/fail-over techniques designed in. Simplifies install as well, reducing the number of sensing wires needed back to the battery. Perhaps reducing potential for errors
There is more to this story, but have gone on enough already

John: Let me know if you are interested in a blank PCB to build up. Assuming that meets what you are looking for. I might even have a 'scrap' board around that might work for you.

[john61ct]

Wow, thanks for showing up Thomas, like tcp/ip magic!

I'm planning on buying on of your ready-built SAR units, soon as some customers pay their invoices faster than I have to pay bills 8-)

Do you think for this specific control functionality I can piggyback on that? I assume to trigger a remote relay

or should I build a separate single-function unit?

In looking at your docs, I've noted some questions, unrelated to this thread topic, but figure while I've got you 8-)

_____
Your SAR requires P or N type alternator only? Or can work with B/A as well?

______
Suggestion for rev4: allow for two shunts, to allow both "belt saver" and "small engine" amps downrating AND accurate current-based stage transitions.

In the meantime, if I rig a three-position switch [A to Alt output shunt, *B to Batt input shunt,*C disable current-sensing (go to voltage/time-based algorithms only)] how will I tell the SAR to change modes?


I need "STOP charging competely when ABS current drops to .025C", ideally also would be able to send a signal to a kill switch or relay/solenoid to, for example, cut DC gennie's fuel supply.

_____
I assume you allow for "no float" (kill all field current?) when alt is on a propulsion engine?

[john61ct]

And should really say Thanks for all your work on this project, Viva FOSS!

[thomasow]

Thanks John, will look forward to hearing from you later!

On your questions P/N vs A/B: These are different names for the same thing. High Drive or Low Drive, and yes, the regulator is configurable for either via a jumper wire.

2x shunts for 'belt saver' AND monitoring battery current: You can accomplish this today with a single shunt - place it on the battery for accurate translation determination. One of the 'fall back' modes for the regulator is to apply the pullback % to the field drive for the belt-saver / small alt-mode as well as adaptive idle pullbacks (Useful for small engines to prevent stalling near idle) when the regulator notes there is no shunt on the alternator its self (or the shunt has failed). Not as accurate, but that really mostly comes into play when using this regulators for a small DC generator.

-OR- using the CAN (Control Area Network) one could instead place the shunt at the alternator, get not only finer resolution on the pull-backs, but also be able to monitor the alternator current via the status outputs. Then a device is installed at the battery to monitor battery current and communicate that to the regulators via the CAN (device coming this summer/fall)

I would not try to switch the shunt between uses, that seems a recipe for failure. Put the shunt on the battery and let the pull-backs work via % modifiers. That way you will be able to see your trigger point. The field-drive % modifiers work rather well in most cases.

BTW: This regulator will function just fine w/o a shunt, it falls back to time-based decisions. Allows for simpler installs if one wants, but is mostly there to cover failures with the shunt.

Signal engine stop: There is a compile option in the source-code for the feature-out port to go active when the regulators enters FLOAT mode. Perhaps this could be used for your kill-engine signal?

Not sure I understand your question about kill-field when on propulsion engine. One of the uses of Feature-in is to force the regulator into Float-mode, it is mostly used with LiFeP04 installs for a BMS to signal the regulator to stop charging. Some folks configure FLOAT to fully turn off the regulators (no field drive), others to a low 'carry' voltage, and one promising approach is to configure the regulator actively regulate to 0A acceptance current when in 'float'. Prevents any additional energy being sent into the battery while allowing the alternator to carry house loads.

=================================

At to your other project, (or is it covered by the above???) - perhaps you can elaborate more exactly what functions you are looking to do.

[john61ct]

Originally Posted by thomasow

Signal engine stop: There is a compile option in the source-code for the feature-out port to go active when the regulators enters FLOAT mode. Perhaps this could be used for your kill-engine signal?

Sounds good!

Originally Posted by thomasow

Not sure I understand your question about kill-field when on propulsion engine. One of the uses of Feature-in is to force the regulator into Float-mode, it is mostly used with LiFeP04 installs for a BMS to signal the regulator to stop charging. Some folks configure FLOAT to fully turn off the regulators (no field drive), others to a low 'carry' voltage, and one promising approach is to configure the regulator actively regulate to 0A acceptance current when in 'float'. Prevents any additional energy being sent into the battery while allowing the alternator to carry house loads.

Yes, that last is a great concept!

My use case, when maneuvering in a dicey situation, you sometimes need to ensure all available engine horsepower is driving the prop. Of course a manual kill switch at the helm is the usual, and KISS reliable solution.

Same on land with pulling a max load in the mountains, transmission getting overheated.

Originally Posted by thomasow

apply the pullback % to the field drive for the belt-saver / small alt-mode as well as adaptive idle pullbacks

field-drive % modifiers work rather well in most cases

I really need to get deeper into your project!


"Other projects" include: looking to create a small DC generator/charger, as well as getting the most out of an HO alt on the propulsion engine.

One day, a central control that would integrate solar/wind shore charging, the two fossil-fuel sources and SoC/coulomb-counting monitors like SmartGauge, Trimetric, etc would be ideal, think your code could be the foundation for something like that?

If so seems to me the CAN functionality would part of the solution.