Fluke 732A Battery Life Extension

[Evans2688]

The Fluke 732A voltage standard uses sealed lead acid batteries that I have found typically last only 3 years or so, based upon my experience with 12 units (between three of us Voltnuts) over many years of operation.  This is in contrast with some people on the forum stating that they have gotten 10 years of operation with the batteries.  Note that the float voltage of the batteries is set at 27.0 volts, which is the recommended value for either four 6 volt or two 12 volt batteries in series.  In my case, I have two 732As with two 12 volt sealed lead acid batteries that have never lasted more than three years in operation.

It should be noted that I have experience with the operation of many 12V sealed lead acid batteries lasting at least 10 years but these are not batteries in series with one another.  These batteries are in 12 volt alarm systems and lighting emergency backup systems.

The 732A batteries I have experience with fail silently with no warning until a unit shuts down from a power failure after only a few minutes.  The batteries are typically very hot to the touch at this point and obviously not working.  A post mortem on the batteries typically finds one battery fried and the other working somewhat but not fully.  A good test is to measure the current through the battery with 13.5 volts applied. A new battery will typically have less than 2 mA of current flowing thru it whereas a marginal battery may have some 10s of mAs flowing through it (or none).  Note: a new battery may take several days to settle down to these low values.  Presumably, the cells take time to “form”.

My speculation is that the batteries are well matched when new and the current thru the series connection of cells results in a near equal splitting of the 27.0 float voltage so that each cell sees approximately 13.5 volts across it.  Over time, the cells age differently and the voltage across the two cells diverge from one another.  The one with the higher voltage dissipates more power (equal currents thru both batteries) and heats up more, which leads to more degradation in the hotter cell, creating faster degradation in that cell.  If this is, in fact, what is happening then a potential cure would be to use a circuit to maintain the 13.5 volts across each cell for a given 27.0 float voltage.

I have developed and built a external Battery Management  System (BMS) for my Fluke 732A battery assembly that plugs into the rear connector (modified) of the Battery module. It does not impact the operation of the 732A battery charger circuit in any way, yet maintains an equal split of the voltage across the two batteries regardless of whether the circuit is in bulk charge (28.8 VDC) or float mode (27.0 VDC), or how different the battery characteristics have diverged (up to a point).  So far the circuit has worked exactly as expected but only time will tell if it extends the life of the batteries (but I have relatively high confidence that the circuit will extend the life of the batteries).

The circuit of the BMS is shown in Figure 1 and consists of an LT1010T Fast ±150mA Power Buffer, an LM317T, and an LM337T.  The LT1010 is used as a voltage splitter to maintain an equal voltage across the two batteries.  The LM317T and the LM337T are used as current limiters to prevent damage to the LT1010 in the event of a short or heavy current load >100mA through the LT1010 output. The LT1010 will either source or sink current as required to keep the voltage across the batteries equally split from the 732A battery charger circuit.  The LM317T and the LM337T do not impact the circuit at all unless the LT1010 needs to either sink or source more than 100 mA to balance the voltage across the two batteries.  Obviously, if more than a few 10s of mAs is required to keep the voltages balanced then something is wrong with one or both of the batteries.
So far the circuit has only seen <1 mA through the LT1010 since the batteries in the 732A are new and well balanced.  I also included a rotary switch so I can monitor the balance current from the LT1010 as well as the charge/discharge current through the batteries and the voltage across the individual batteries and the voltage from the 732A battery charger circuit. The rotary switch goes to two terminals that I connect a DMM to monitor the voltages and currents, with the currents represented as 0.1mV per mA on the meter.

I may have done overkill on using TO-220 packages (I had them available in my inventory) for the ICs based on what current levels I’m seeing but I wanted to make sure I had plenty of thermal margin in the event that a large unbalance occurs in the future.  If it turns out that TO-220 packages are not needed, then an 8 pin dip package could be used for the LT1010 and TO-92 packages could be used for the LM317 and LM337 ICs, resulting in a very small BMS package that could possibly be fit inside the 732 battery module.  My goal is to see if the battery life can be extended by the use of the BMS circuit and optimization can be done in the future.  Also, active current monitoring may be added to the circuit for the balance current along with an alarm circuit to alert me if the batteries become excessively unbalanced so I don’t have to do manual monitoring.  All future enhancements if warranted.

[bastl_r]

based upon my experience with 12 units

I wish I had one of your problems. 
Can you change anything about the tension? If so, you could try to build a replacement battery with 4 LiFePO4 cells connected in series and charged at 3.4V/cell.

regards

[picburner]

I have two Fluke 732A and I opted to insert the BMS inside the battery compartment, replacing the original PCB.
Naturally, with just two 12V batteries only, the electronic circuit is much simpler.
I noticed that the BMS intervenes both in the charging phase and in the discharging phase for a short period, this happens
even though the batteries are new and from the same batch.

One thing I noticed while dismantling the old PCBs is that the old protection PTCs were completely out of tolerance (resistance much, much greater than the rated resistance at room temperature)
this in both devices.

[MegaVolt]

I made a simple internal board.
The operational amplifier in the model is not the same as that installed on the board. In my opinion there is TL061 there.

I'm not sure if this board changed anything

[MegaVolt]

I have two Fluke 732A and I opted to insert the BMS inside the battery compartment, replacing the original PCB.

Can you share the schematic and files for making PCB?

[picburner]

Here is the schematic and the required gerber files.
Gerbers and ncdrill are made with the Jlcpcb requirements.
The jumpers P2 and P3 are used to deactivate the circuit if required.

[MegaVolt]

Here is the schematic and the required gerber files.
Gerbers and ncdrill are made with the Jlcpcb requirements.
The jumpers P2 and P3 are used to deactivate the circuit if required.

Thank you very much! Can you tell me what options you chose for the gold plating of the edge connector?

[floobydust]

A suggestion for any revisions is to add a comparator or op-amp that keeps an eye on the midpoint voltage. If it's out of spec, then light a fault LED or more. Otherwise I think you won't know things are amiss.
A very common problem with SLA batteries is they corrode and eventually a cell shorts. So it's now a 10V battery, not 12V and a charger will try to bring that up and basically roast. The battery also overheats and even melts in UPS applications where charging current can be high.

[picburner]

@ MegaVolt:

ENIG and Gold Fingers. On such a slim PCB the factory can't make the 30° finger chamfered , you have to think about it yourself.
For Gold Thickness I choose 1 U", ​I think that's enough.

[MegaVolt]

On such a slim PCB the factory can't make the 30° finger chamfered , you have to think about it yourself.

what thickness did you specify?

[picburner]

I meant that the PCB is not big enough for that type of work (30° finger chamfered).
If instead you mean the thickness of the PCB logically it must be identical to the original, i.e. 1,6mm.

[alligatorblues]

I've found there are only a few good battery brands. Allstate, Enerays Genesis, and a few others. Batteries also need to be cycled every year or so. The 732A has 15 hour run time on good batteries. It's worth it to change them every 3 years.