Charging LiFoPO4 without DC-DC charger

[tpbrady]

DDW,

It was the Wakespeed Pro with Bluetooth without a shunt connection. The sense line was connected at the battery bus so it was essentially measuring the battery. The measurement was done with a VOM at the alternator. The app returned a reading higher than the actual measurement. I did not do any logging. That difference went away when the ArgoFET was removed.

If I had to guess, I would say the ArgoFET was putting out some RF that caused some problems with the internal Wakespeed measurements. It would easily be coupled onto the DC conductors.

I like the concept of using a battery isolator as it decouples the LFP from the lead acid when being discharged. That eliminates the parasitic loads of lead acid banks when connected to ACRs that stay connected above 13v. The real problem I see in refitting old boats to LFP while retaining lead acid deals with DC+ cabling which uses a common charging and discharging path. That’s the downside of inverter chargers. If you separate DC loads from charging it gets a lot easier but the amount of cables goes way up.

One thing I learned yesterday was a DC-DC charger with a lithium profile can still cause the BMS to shut down charging. I had all 3 LFP batteries do that yesterday. It was a LiTime charger that floats too high,13.8v. I was configured that way until I sort out regulators. The battery isolator effectively prevented the alternator from caring as it had two lead acid banks connected.

Tom

 

[twistedtree]

The shunt inputs to the Wakespeed, or any device measuring current through a shunt, are very sensitive. They measure voltage across the shut with a typical range of 0 to 50mV. If those inputs are not connected, there is a lot of opportunity for phantom readings just from nearby electrical noise. I think the best practice with the Wakespeed is to disable any unused shunts in their config software so it ignores any spurious input. An alternative would be to short together the sense leads so they are always sensing 0V between them.

 

[twistedtree]

But I worry that the Balmar, which can't determine when to switch to float, would keep LFP too high most of the time.

Can you say more about this? I figured it could be made to work, but it sounds like I’m missing something.

 

[DDW]

On the Wakespeed I think you'd have to disable the shunt, shorting them would result in an implied zero charge and could cause problems if it were paying attention.

The Argofet should be a more or less linear device operating in switch mode, not creating RF. But I could see a plausible interaction with the PWM used to regulate the field. It would be interesting to put an oscilloscope on the field, and compare it to the current waveform to the Argofet. Without the Argofet, there is a more direct connection to the battery, which is a very effective low pass filter. Perhaps a largish filter cap across the alternator output near the Argofet would solve the problem. Or even from the Argofet input to output (to pass HF across to the batteries). Ordinarily, the alternator output is connected directly to a lot of capacitance and inductance, the whole DC network. If the Argofet is not passing the high frequency components of the current waveform on, then the circuit between it and the alternator is very small and isolated. However I don't understand why the Wakespeed would be reporting this voltage, and not the voltage sense voltage from the DC bus. This is effectively a servo system, and too little inductance or filtering on the servo drive output (the alternator, in this case) can cause it to be unstable.

The problem I have had with the Balmar is it has at best a rough guess of alternator output (based on field PWM duty cycle) and no idea at all what is going into the batteries (vs. other loads). Its algorithm attempts to switch to float after either the field duty cycle goes below a certain threshold, or a certain amount of time has passed at absorb voltage. To switch back to bulk, it looks for voltage sag for some period, then you are back to guessing when to go to float.

I could not make it work well trying every setting possible, the only workaround was simply to abandon the decision making and have a fixed absorb time. Switch back to re-bulk still seemed to happen when you didn't want them. The Wakespeed was much better behaved (at least after 2.50 firmware). Because it can read the shunt and know what the charge current is to the battery rather than to the inverter or refrigerator. The switch to re-bulk had many more components and is more accurate.

 

[tpbrady]

DDW,

My assumption is that the Wakespeed reads the alternator voltage from the power lead connected to the alternator output. It was surprising that it could be that far off with the ArgoFET in the line.

I did observe that the Wakespeed dropped to float as soon as the voltage reached the target bulk voltage with no shunt configured in the app. The voltage would build as the LFP bank was charged and when it reached 14.2v, my set point, it went to float. The batteries were at 99% SOC as reported by their BMS. While it was in float I could watch the Victron Smart Shunt show charging and discharging cycling back and forth. I don’t why it does that, but it seemed like it was trying to figure out the house load by measuring voltage as it went on and off to result in 0 net amps going into batteries. It eventually seems to stabilize there. The Wakespeed needs a shunt to work reasonably.

Tom

 

[DDW]

I am logging the voltage from the serial port (wrote an app to do that). The voltage is reported (and logged) every second, and it is definitely the sense line voltage, not the power line to the regulator. I do not have the Pro version with the app so can't compare it.

The hunting a bit around zero in float I would not consider unusual for any charger. How many amps to do see going or coming (with nothing else on)? The alternator has to be producing some amperage to maintain float voltage, as there are always at least some loads (the regulator itself at the very least).

It sounds to me like the Wakespeed charge profile is doing what it is supposed to do, even without the shunt. With AGM, you want 2 hours minimum absorb time (can be set up in the regulator) for the charge tail. With LFP maybe 1/2 hour, to allow balancing. In a mixed system it seems like a mixed bag, I guess I would go with the 1/2 hour, on the theory that AGMs being used for engine start or thruster are never lower than say 95% SOC and don't need a long absorb soak. That is what I intend to do anyway.

The only question is, why is the Wakespeed app reporting the odd voltages, where is it coming from, and is it real?

I intend to separate the DC bus to isolate the engine and nav instrument electronics to the AGM engine start battery, so that a complete disconnect of the LFP house doesn't make the engine and nav go down. Did you do anything like that?

 

[tpbrady]

DDW,

I am pretty sure the alternator voltage reading in the Wakespeed app has to be from the power lead which is connected to the positive post. When I removed the ArgoFET, the readings with the VOM matched what the app was reporting from the alternator.

I really had no way to power the electronics from a separate source without a complete rewire which would have taken a few weeks of removing and replacing wire. Since my primary nav is using Coastal Explorer on a laptop and I intend to keep the port half of my old house bank for the remainder of the year, I can maintain nav continuity until I switch over to the old house bank.

Tom

 

[Marco Flamingo]

This got too complicate to tackle with all my other projects that need to happen before our summer cruise that starts at the end of June. So it's on pause for now. To be able to use the full potential of my 160 amp alternator the best path looks like Wakespeed or something similar. This will most likely cost less than 2 DC-DC chargers.

Pause is good. One needs time to digest the info. I was two years between the idea and the installation of LFP. And then I went way off the tracks with all LFP, only LFP. I know that a lead acid start battery is traditional, like having a land line telephone. But can anybody figure out a way to eliminate a land line telephone or a lead acid start battery? Call me crazy, but I think both are possible. Sure would make things simpler.

And that was a selling point for me. If a gadget or system is complex, it better do a lot of things well. There is usually an ongoing thread on TF about the best cappuccino machine to have aboard. For me, unless it can also wash dishes, bait a hook, and stores in a ditty bag, I'm not interested. Lead acid does one thing. Provide a bunch of amps. Turns out, LFP can also do that. Better.

Given my biases, here's what I would be thinking about in your situation. First, as you have identified, your alternator's external regulator may be a problem. It sounds like it isn't programable. My old first generation Balmar (ARS-512? or something like that), doesn't have a "lithium" profile, but that is a good thing. It, and later models, can be modified to fit my particular needs. And by particular needs, I mean changing parameters to take the best advantage of LFP. That is the way to get rid of lead.

Well, that and buying enough BMSs. I know, we think we are buying LFP batteries, but the secret is we get a "free" BMS with each battery. You don't get that with any of the lead battery variations. How is cell #3 doing in that old AGM? Who the hell knows! With LFP you get a Bluetooth BMS. And you get one free with each battery!! I know, most think they are getting more amp hours, lots more amp hours, with each additional LFP, but the additional BMSs might be more important. And, as you've noted, those BMSs are only $170 per 150Ah on Ebay. Step right up, folks!

I think you said that two 150Ah LFP would be better than your 6 lead acid gold cart batteries. Better in more amp hours? But you have room for 6, right? Would 6 150Ah be too much? Has anybody ever said, I have too many available amp hours on my boat? No, but that's not the real issue. You would have also purchased 6 BMSs. That eliminates any fear of the often alleged bugaboo of a "BMS shutdown causing alternator damage" when one single LFP battery is connected to the alternator. With a reasonable sized bank, if battery #1 shuts down because of a cell overvoltage, batteries #2, 3, 4, 5, and 6 keep on keeping on and you won't even know (unless you are presently monitoring the Bluetooth app). That battery balances itself and will soon reset whether you've seen the incident or not. No reason for that battery to notify the alternator of an impending shutdown. Let the BMSs you've purchased do their job.

But wait, there's more! You also get 6 LFP batteries with a capability of discharging 120 amps each continuously. 660 amps in your pocket, so to speak. Let's see what can we do with that! A Ford Lehman needs less than that to start. A thruster much less. Same with a windlass. And if you check the BMS parameters for the EcoWorthy 150Ah (with an app like Overkill) you will find out that each battery can actually put out 300 amps for several seconds. You could pull the anchor, use the thruster, and start your engine all at the same time (not recommended). But clearly you won't need a lead start battery. In fact, since you have two group Group 24 lead start batteries, you have even more room for LFP. 8 150Ah LFP batteries and no need for gadgets and doodads feeding this over to that when such and such triggers the thingamajig.

There is, of course, the issue of cables and fusing. But you will avoid the issues of DC to DC strangling your charging ability. Your 160A alternator, detuned by something like a Balmar external regulator to output 130A and a bank of 6-8 can take all of that without a problem. No gadgets necessary. For a bank of 6 150A LFP, that's about a .2 charge rate. Which actually isn't that important because of the amount of amp hours you would have aboard. 900 amp hours. You could leave the marina fully charged and do a long weekend even with no charging from the alternator. Another possibility caused by way too much LFP.

I'm up past my bedtime.

 

[SteveK]

Marco, no one said an LFP bank cannot start an engine, it can, that is why I have a cross over in case. But like LFP evolution for house banks, so is the LFP start battery now taking shape. When I need new start batteries they will be on the shelf to buy.