AGM Battery Voltage - What's the deal?

[RockJGC]

Does anyone else notice that the battery voltage on their AGM battery drops quickly and stays pretty low? My battery is less than a year old and I see 13.8V while charging (engine running) and then I can watch the voltage drop all the way down to around 11.4V within just a few minutes after the engine is turned off. With conventional lead acid batteries, the voltage typically stays at around 12.4V to 12.6V, which indicates a full/nearly full charge. But on my AGM, the voltage drops much lower very quickly. However, the battery still tests just fine (other than the low voltage) and still starts the car without any issue at all.

Is this normal behavior of AGM batteries? One of the reasons I ask is because I see similar behavior on our Jeep and our Acura, both of which also use factory installed AGM batteries.

[k.sigota]

I have noticed very weird behavior when the battery is not coded to the car. Other than that I've never seen this on any of my cars.

Out of curiosity did you code the battery to the car when you installed?

[DrGER]

The previous owner of our '11 Golf Variant wagon put in an H6 AGM in 2018. The behavior you describe is consistent with this AGM, while the Golf does not have battery energy management and was designed for a conventional flooded H6 battery with PR-J1D. I have yet to see any "low voltage" DTCs since we bought the wagon 2+ years ago. --g

[Smac770]

Dropping down into the 11's on engine off is indicative of a car problem, not the presence of an AGM battery.

Is the car reporting the battery voltage incorrectly in block 18 compared to what you measure? Or are you also measuring ~11v. What is the car reporting as the battery SoC? Charging voltage of only 13's and not 14's when the SoC is under 80% would seem an error on the J533's part. What voltage is being commanded by the J533, block 14, when the engine is running compared to what you are measuring, say across the jump start posts?

Of course, if the wrong battery type is in the car, compared to what is defined in the adaptation, that'll be a problem too.

[RockJGC]

Okay, so forget what I said yesterday. What I was doing was using my recently updated Carista app to watch the battery voltage on my phone. I'm not sure how they are pulling battery voltage data, but it's way off. I'm talking more than a full volt off.

When I got home yesterday, I parked in my garage and a few hours later I popped the hood and checked battery voltage at the jump posts. I got 12.51V, which is good, but still a little lower than I was expecting considering that the vehicle was driven 3 hours earlier. So I got out my battery tester and checked the battery. What I got was not too good: According to the tester, my battery only had 289 cold cranking amps. State of Health is 46%. Internal resistance is 8.8 milliohms, which seems pretty high for a battery that's about a year old.

Yes, the battery was coded when I installed it. I think I may have just gotten a bad battery. I can't find any reason why the CCA are so low.

[Smac770]

Scan tools are going to be reading a measuring value or measuring block entry. It should be reporting value 1 from block 18 from the J533 (19-CAN Gateway). That's the voltage reported by the J367 on the negative battery post to the J533 via LIN. But it could be reading block 1 from the J533, or even grabbing the value from the J623 ECM.

12.5 is exactly where it should be. The J533 is going to target 85% SoC, not 100% SoC. Some kind of overcharging safety buffer.

Make sure the battery tester is aware of the battery type. If it's AGM, you need AGM aware tester/charger. Using wet flood stats against an AGM will give incorrect results, and vice versa. Can that tool go to the CAN Gateway and read blocks directly? Battery voltage is one of the values in block 18, computed SoC is one of the values in block 19, computed internal resistance and battery capacity are in block 20.

But those values were done at the jump start posts under the hood? What values are reported if performed at the battery posts themselves?

But back to Carista, it's reading that from a car control module. It's not making up the value. So something in the car is seeing a full V less than what you're measuring?

[RockJGC]

Scan tools are going to be reading a measuring value or measuring block entry. It should be reporting value 1 from block 18 from the J533 (19-CAN Gateway). That's the voltage reported by the J367 on the negative battery post to the J533 via LIN. But it could be reading block 1 from the J533, or even grabbing the value from the J623 ECM.

12.5 is exactly where it should be. The J533 is going to target 85% SoC, not 100% SoC. Some kind of overcharging safety buffer.

Make sure the battery tester is aware of the battery type. If it's AGM, you need AGM aware tester/charger. Using wet flood stats against an AGM will give incorrect results, and vice versa. Can that tool go to the CAN Gateway and read blocks directly? Battery voltage is one of the values in block 18, computed SoC is one of the values in block 19, computed internal resistance and battery capacity are in block 20.

But those values were done at the jump start posts under the hood? What values are reported if performed at the battery posts themselves?

But back to Carista, it's reading that from a car control module. It's not making up the value. So something in the car is seeing a full V less than what you're measuring?

I'm not sure if you looked at the attached pictures, but it shows the battery type chosen on my tester, which is a flat AGM as opposed to a spiral AGM. I didn't check at the battery....I was too lazy to pull the spare tire out. Either way, my tester only has the ability to connect to the battery/jump posts.....it can't read any blocks directly. It's just basically an electronic battery tester.

As for Carista, I sent them a message through the app and they responded by telling me that "the adapter gets the voltage reading directly from the vehicle when it's connected." I got a kick out of that response, but it's about as vague as possible. I'll see if I can get more info from them. It's definitely far off from the actually battery voltage.

My only concern right now is that I have a one-year-old battery that's giving me less than 300 CCA already, which is less than half of it's rating.

[DrGER]

...
As for Carista, I sent them a message through the app and they responded by telling me that "the adapter gets the voltage reading directly from the vehicle when it's connected." I got a kick out of that response, but it's about as vague as possible. I'll see if I can get more info from them. It's definitely far off from the actually battery voltage.

Yeah, Carista is from Prizmos Ltd out of Sofia, Bulgaria, and they don't offer much (any) info regarding how the app gets its "live data" "directly from the vehicle when it's connected". I'll be interested to see if you can get any technical information about this from them (funny, I have my Carista bluetooth adapter with me in the office today to run a quick scan on our VW wagon this afternoon). --g

[RockJGC]

Yeah, Carista is from Prizmos Ltd out of Sofia, Bulgaria, and they don't offer much (any) info regarding how the app gets its "live data" "directly from the vehicle when it's connected". I'll be interested to see if you can get any technical information about this from them (funny, I have my Carista bluetooth adapter with me in the office today to run a quick scan on our VW wagon this afternoon). --g

Yeah, Carista isn't exactly the best out there, but I have free lifetime with them since I helped with Beta testing back a few years ago. They recently added battery registration and throttle body calibration to their tools for the B8 platform. I think something else too, but I'd have to connect again to see exactly what they have available now.

[DrGER]

Yeah, Carista isn't exactly the best out there, but I have free lifetime with them since I helped with Beta testing back a few years ago. They recently added battery registration and throttle body calibration to their tools for the B8 platform. I think something else too, but I'd have to connect again to see exactly what they have available now.

Same here -- got the adapter 5 or 6 years ago and did the Beta testing thing, too. I usually keep the adapter in the glovebox on longer drives out of town in either of our VAG cars, just in case... --g

Update: Just tried to run a scan with Version 7.2 beta-2 and it wanted me to purchase the PRO version to do anything useful in our Golf wagon. This happened at least once in the past. Time to e-mail their "customer support"... --g

[Smac770]

It's possible the Carista has its own voltmeter internally. All scan tools are supplied ground, term 30, and term 15 on pins on the OBD-II port. The plug-in device could easily just measure the presented voltage.

As for the battery being at half cranking amps, that presumes the tool is correct and using the same metrics as whatever numbers you're comparing against. If you do a counter test at an AutoZone or such, does that test to the same results. Also, do the results differ with the battery disconnected from the vehicle.

Yeah, saw the AGMF when looking at the pics a second time, but had already posted the text.

The EA888 does TBA on each shutdown. There's no actual TBA basic setting. I believe there's a TBA test basic setting though.

[RockJGC]

So here's Carista's reply to my second question, asking for more information on where the voltage data comes from:

"The OBD2 adapter itself has a voltmeter and it's value gets shown on the app's main screen next to where the connected car's VIN gets shown. We also have for some models a live data tool that's able to read what the ECU themselves see as voltage coming from the car's electrical system. The ECUs have integrated voltmeters for the purpose of over/undercurrent protection that we're able to read via the OBD2 in the live data tools mentioned."

[Smac770]

You might test the pins at the OBD-II port to see if the voltage there is low relative to the rest of the car, or if the internal meter on the Carista is just off.

Test 1 to 4 and 5, and then test 16 to 4 and 5. 4 and 5 are redundant ground connections, they splice together before going to the center tunnel left side ground post. D refers to fuse panel D, the one on the passenger end of the dash.

B8 OBD-II port.jpg