SLA battery charging monitor question

[Mark_R]

Hi,

I'm working on the design of a piece of equipment which has a backup 7AH 12V sla battery. I'm using a switching power supply which has built in battery charging.

==>> **broken link removed**

The only problem is that the charger does not indicate battery failure when on AC, of even if the battery is not connected. I was thinking of running the battery circuit through a DPDT relay and periodically (like 15 min) disconnecting the charging circuit and connecting a load resistor across the battery and sampling the voltage. Probably would be controlled by the systems µC such that the test would not happen during AC fail when on battery.

Wondering if anyone has a better idea? Perhaps monitoring the charge current through a shunt, no float current = no battery?

Thanks.

 

[ericgibbs]

Hi,

I'm working on the design of a piece of equipment which has a backup 7AH 12V sla battery. I'm using a switching power supply which has built in battery charging.

The only problem is that the charger does not indicate battery failure when on AC, of even if the battery is not connected. I was thinking of running the battery circuit through a DPDT relay and periodically (like 15 min) disconnecting the charging circuit and connecting a load resistor across the battery and sampling the voltage. Probably would be controlled by the systems µC such that the test would not happen during AC fail when on battery.

Wondering if anyone has a better idea? Perhaps monitoring the charge current through a shunt, no float current = no battery?

Thanks.

hi Mark,
What type of battery failures are you expecting.?

 

[Mark_R]

Main concern is disconnected battery, but knowing if a cell has failed would be nice.

I have had the battery on charge for a few hours and it's drawing 135ma of charge current. I'm going to let it float for a day or two and see what kind of current it's pulling. I don't know if the current will drop off to almost nothing at full float charge, if it does then the current sensing idea won't work.

Perhaps I could look for something like .50ma = battery OK? Then I could come up with an optoisolator circuit that output a digital signal for the controller when the threshold was exceeded.

 

[MrAl]

Hi,


I am not sure if i understand what you want to do 100 percent, but in my experience i have found that LA batteries fail mostly in the mode where they will no longer hold a charge very well. This means the output voltage drops faster with load then when the battery was new. This makes me think a load test is the best test to do and watch the dv/dt while the battery is loaded. A test when the battery is new as well as a compilation of tests as the battery ages should tell you what to look for as the battery ages. That way you could probably tell faster how the battery is performing as the load test is being run.

If you just want to know if the battery is disconnected, then yes a current measurement should tell you right away if that has happened.

 

[Mark_R]

This is a piece of remote telemetry equipment monitoring wells, etc. It's run by AC mains normally, but has a backup 7AH 12V SLA battery, so it can continue to transmit for a few hours during a power outage. One of the main purposes is to transmit A/C fail so someone can get a pump truck / generator to the site to maintain operation, thus the need to be able to transmit on battery after a power fail.

I'm planning on using this ==> **broken link removed** power supply (with integral charger) as I doubt I can design an on-board supply for what this costs.

The power supply provides an output for AC fail, but not for loss of battery. I need to detect that so I can send an alarm back to the mothership to let someone know that there is no backup supply. Without that, no one would know the unit is off line until it comes up missing on it's daily check-in.

Detecting a disconnected battery is vital. It would be nice to know if the battery has failed and is only being held up by the charge voltage, but that's not that important if complicated to achieve. The PM schedule for the equipment calls for an annual check of the battery by a technician who can do a detailed battery test. I'm just looking for a major battery failure indication with the circuitry, not weather weve lost 20% of capacity or whatever. The battery is over-sized to begin with.

I know i could periodically swap in a load resistor and do a voltage check from time to time, but I hate the idea of a mechanical relay operating every few minutes forever. I'm thinking some form of voltage / current monitoring would be continuous and not subject to wear, plus that creates a brief period when the battery is not available should the power fail.

 

[RCinFLA]

SLA will have a recognizable characteristic under float charging.

With a float charge of 13.5v to 13.8vdc there should be some charge current flow that can be detected. For a 7 A-H battery it may be only 3-25 mA's depending on condition of battery.

Professional UPS systems do an impedance test periodically on batteries to ensure their Rs is within a range that ensures the battery is in acceptable condition. This amount to putting an AC coupled signal onto the battery and measure current associated with the AC loading by the battery. For a 7 AH battery an AC signal amplitude to produce about 1 amp AC current should suffice. Assuming an Rs of 10-50 milliohms this would only be a 50 mV AC reading across battery. A low voltage transformer capable of 1 amp can be used. A series coupling cap value is selected to provide a 1 amp AC current sourcing to the battery. For a 6.3 vac 60 Hz transformer, series cap would be about 400 uF. A higher voltage transformer would lower cap value.

 

[Mark_R]

The AC is probably way more complicated than I need here.

OK, So let say I opt for detecting 3ma as a battery present signal. If I go with a .05Ω shunt, that would provide a 150 µV signal at 3ma.

What would be the best approach for detecting this? I'm assuming an op-amp? i would need to drive an optoisolator. I'm a bit rusty (OK, VERY rusty) in op-amp circuits.

Thanks for the help.

 

[RCinFLA]

150 uV is pretty low but if your sure you can amplify it without too much noise added would be okay. Maybe figure 0.1 to 0.2 vdc drop with max battery load current for resistor value.

 

[Boncuk]

Hi Mark,

using an UC3906N (TI) the battery will be float charged as long as mains supply is present.
A relay cuts off the load while the battery is being charged in order not to get erroneous readings for the IC. (suggested by TI)

You might connect the load supply voltage to relay (K1) pins 14 and 24 for mains power supply.

If mains fails the relay is deenergized and completes the circuit between battery and load via the NO contacts.

Attached are the schematic and a PCB design.

Remark: The PCB is designed for tropical area use and therefor the fan runs continously when the charger is active.

Regards

Boncuk

 

[arunb]

using a shunt to monitor the charge and discharge of the battery would be ideal, but I think you may have to make one. I am also working on a similar project for monitoring several batteries in my lab. I plan on using a shunt and a PIC micro-controller to monitor the charge and discharge cycles. A watch crystal is used to keep an accurate timing.

 

[Mark_R]

Hi Mark,

using an UC3906N (TI) the battery will be float charged as long as mains supply is present.
A relay cuts off the load while the battery is being charged in order not to get erroneous readings for the IC. (suggested by TI)

You might connect the load supply voltage to relay (K1) pins 14 and 24 for mains power supply.

If mains fails the relay is deenergized and completes the circuit between battery and load via the NO contacts.

Attached are the schematic and a PCB design.

Remark: The PCB is designed for tropical area use and therefor the fan runs continously when the charger is active.

Regards

Boncuk

Thanks, the circuit looks good, but I'm planning on using a pre-made power supply charger.
See attached drawing, for clarification....
I need to sense the battery power passing thru the PCB to detect a missing battery (the blue writing).

 

[joemcder]

Check the charging voltage with the battery 1)charging, 2)floating, 3)disconnected.

In some (cheap) charging circuits the voltage will rise with no load. If you are lucky a simple voltage monitor will do the job.

Lets hope that the charging circuit designer was a lazy, cheap SOB

Joe

 

[k7elp60]

Monitoring the charge current to the battery may be the wrong way to go as the when the battery is being charged with a float charger the charge current may drop to a level difficult to monitor, as all ready been mentioned. I think the thing to do would be to use a power mosfet as a switch to switch off one lead to the charger. Always monitor the battery voltage. If battery voltage is below a predetermined level the voltage sense circuitry sends an alarm signal. I'm thinking somewhere around 11.6V as the battery still has some charge left. I have designed a low voltage disconnect circuit to diconnect the load from a discharged SLA battery that uses a IRF4905 p-channel mosfet, could be easly modified to switch the charger from the battery.

 

[Mark_R]

Monitoring the charge current to the battery may be the wrong way to go as the when the battery is being charged with a float charger the charge current may drop to a level difficult to monitor, as all ready been mentioned. I think the thing to do would be to use a power mosfet as a switch to switch off one lead to the charger. Always monitor the battery voltage. If battery voltage is below a predetermined level the voltage sense circuitry sends an alarm signal. I'm thinking somewhere around 11.6V as the battery still has some charge left. I have designed a low voltage disconnect circuit to diconnect the load from a discharged SLA battery that uses a IRF4905 p-channel mosfet, could be easly modified to switch the charger from the battery.

My original thought was to periodically disconnect the battery and switch in a small load resistor, then check the battery voltage. Not only would this detect a disconnected battery, but also one that is being "held up" by the charger.

I don't understand why the designer of the charger didn't include a battery disconnect indication.

Would you mind sharing your IRF4905 circuit?

 

[mneary]

I haven't tried this, and I know better than to suggest something I never tried, but....

A failing battery would tend to have a high impedance. Wouldn't it be simple enough to apply a heavy load (say 1C or even 10C) to the battery for a few milliseconds and judge the battery's condition from the voltage depression? The heavy load would have to be enough so the charger can't supply it, only with a good battery. This saves the extra circuitry to disconnect the charger, more to reduce failures than to cut cost.

This could have a pitfall, if it accidentally convinces the charger to come out of float mode.

 

[Mark_R]

Check the charging voltage with the battery 1)charging, 2)floating, 3)disconnected.

In some (cheap) charging circuits the voltage will rise with no load. If you are lucky a simple voltage monitor will do the job.

Lets hope that the charging circuit designer was a lazy, cheap SOB

Joe

I checked that, the voltage is well controlled. It only climbs 15-20 mV when disconnected from trickle.

 

[k7elp60]

The attached diagram is for the battery disconnect circuit. The TP is for adjustment of the disconnect voltage, and it is very close to 1/3 of the applied voltage. The 47uF capacitor gives the circuit some stability in that if a load is connected to the battery the voltage my dip for an instant. This capacitor prevents the circuit from disconnecting the load in that condition.

 

[Mark_R]

The attached diagram is for the battery disconnect circuit. The TP is for adjustment of the disconnect voltage, and it is very close to 1/3 of the applied voltage. The 47uF capacitor gives the circuit some stability in that if a load is connected to the battery the voltage my dip for an instant. This capacitor prevents the circuit from disconnecting the load in that condition.

Thanks.

I'm also considering using one of thees intelligent battery charge controllers ==> **broken link removed** , they seem pretty slick. About $11 in small quantities, provides a 3-level charge and a µC interface for different types of failures. I have a couple sitting on my desk. Data sheets attached...

See page 3 of the application note for another version of a low voltage switchover circuit usnig a MOSFET.

 

[k7elp60]

The AG102 charger looks like a neat charger. One thing I find interesting is that if the input is 9V it will still charge a 12V battery.
Having quite a bit of experience with SLA batteries and chargers I agree the the concept of this charger and the output voltages for the bulk and float charge voltages. The one draw back that I see is that the charger module is monitoring the charge current in the - lead to the battery. This means that the - lead of the battery can not be connected to a common ground. This may cause a problem is some circumstances.