Update & Supplement #2:
I found other information here in the EESE site and in Linsen’s great “Handbook of Batteries” that supports my original answer, but I felt the need to clarify and update some points:
9V NiMH Battery, internal construction - photo
I have an old 9V NiCd battery from GP of 150 mAh, now leaky and defective, that I opened for a comparative view of its internals.
That opened GP’s battery is compared side by side with 2 Panasonic ones of 120 mAh (still so-so) and a couple of newer NiMH 9V from Elgin, marketed as 250 mAh, where the actual capacity seems to be about half of it, so “1XX” mAh.
We can then say that most “9V” Nickel-based rechargeable batteries with 8.4 V nominal voltage are made of a stack of 7 NiMH (NiCd) cells.
Initial & Final discharge voltages of 9V NiMH
An excellent and benchmark source of information is the Handbook of Batteries by Linden & Reddy, which Chapter 29 lectures about NiMH. Its “Figure 29.6.a” reports the discharge profile of the 9V NiMH battery, whose data was provided by GP batteries.
The points to highlight here are:
- Initial discharge battery voltage is 10.0V for lower current drains (i < 0.2C).
- Discharging cut-off voltage 8.0V delivers 100%C at drain current of 0.2C (30~50 mA) and +90%C at 0.5C (75~125 mA).
- Final discharge voltage is set at 7.0V, regardless of discharging current, as an absolute cut-off limit.
Recommended Charging & Discharging for O.P.’s 9V NiMH application:
Adjust “REG2” to deliver 10.0V measured at the cathode of D5 (not at the battery) and with a battery installed, or using a 8.2V Zener diode as test jig (dummy 9V battery load).
This will be the practical final voltage for the battery, assuming a Vdrop = 0.8~0.9 V of voltage drop on D5 (1N4148) for 20~30 mA, lowering to Vdrop ~0.6 V for 1 mA.
D5, R4 and REG2 will provide the charging profile for a kind of “floating” charge for the NiMH, with 10.0 V in real charging situations and up to 10.2~10.3 V if residual charging current is really small (~1 mA).
I suggest R4 = 100R (1W) to limit charging current to less than 30 mA (~0.2C) if V_bat = 7.0 V. At Vbat = 9V, Icharge = 10 mA (0.04C~0.07C).
For a residual charging current of 1 mA, R4_drop = 0.1 V; at this current level, such voltage drop will be more than compensated by the increased cathode voltage in D5.
Upgrades: REG2’s supply & Discharging cut-off:
- you would redesign or modernize the PCB, I would feed the LM317 from the larger capacitor, C1, as I originally said at the end.
- I also would consider a cut-off circuit, even something based on the TL431 and a bipolar or mosfet transistor would be enough to protect the 9V battery from overdischarge below 7.0 V.
Other NiMH related links, from EESE:
Original post:
“And now my big question ...” (and concerns)
Let’s address them by parts:
About the “REG2” as LM317:
- The LM317 requires 3 volt differential to operate properly.
Not exactly, as it is dependent of the load current (and temperature), as seen on Onsemi’s LM317 the dropout will be lower than 2V at charging currents of ~20mA, as this “Figure 10”, extracted and highlighted in Blue marker shows:
.
Because of that, you could maintain the “REG2” connected to the +12V from REG1, as it could provide up to ~10.0V of regulated output to charge the 9V battery, depending also of the output tolerances (typ. 5%) of the LM340T-12.
However, to have more freedom of voltage regulating, I would suggest you to connect the LM317 directly to the unregulated voltage at C1 - see above in Onsemi’s “Figure 8” that the LM317 could provide up to 500mA with a voltage differential of 30V - and in your case, running at 20~30 mA, and Vin-Vout < 20V, dissipating P < 0.5W, it will not need a heatsink for that.
About the 9V NiMH battery:
“ The battery is a 9 volt NiMH ... 8.4 rated voltage... 7 drop off ...10.5 max charging”
It seems the modern NiMH batteries are a stack of 7 element-cells of NiMH, as:
- 7 * 1.2 = 8.4V nominal voltage;
- 7 * 1.0 = 7.0V discharge termination voltage;
- 7 * 1.5 = 10.5V charge termination voltage.
This can be seen for the following manufacturers Energizer, Varta, Ansmann, Tenergy, and some less worldwide-known brands as Elgin could have their customer service providing data with variable degree of detail:
and here this one brings further detail of charging voltage curves, in the highlighted Blue box. The key point here is that the recommended charging voltage should be between 9.1V and 10.0V, otherwise the electrode will not be fully charged - see discussion below.
NiMH - Undercharging, Overcharging and Overdischarging
This Energizer Handbook and Application Manual for NiMH brings further details about the battery chemistry and there are 3 pictures I would like to base my comments:
Nominal voltage = 1.2V is at ~50% State of Discharge and 1.0V is already at knee, voltage termination:
State of NiMH electrodes at fully-discharged, fully-charged and overcharging states:
NiMH voltages and electrode polarization states where I marked the Nominal (1.2V), your application (1.03 ~= 1.0V) and polarization reversal on the first electrode at 0V, not your case in theory, but could happen in 1 cell, in real life situations discharge of a 7 stack of NiMH cells.
Main Conclusion:
Recommended charging voltage and current:
From the above information, we can say that current/modern 9V NiMH should be charged a with a termination voltage of 9.1~9.5V to 10.0V, depending of ambient temperature, and with a floating charging current of 3%C to 5%C to minimize damages due to eventual overcharge of one of the cells, allowing the Oxygen recombination mechanism to happen safely.
The suggested change on the supply voltage of the LM317 would allow better assurance this (essential) higher voltage level is achieved in your charging circuit.
You mentioned charging at 7.24V and this is way too low.
In this case, the battery electrodes might be close to polarity reversal potentials and one of the 7 cells could reverse potential and be irremediably damaged.
’Mind the Gap’ for Board compatibility interchange:
Some changes should be made on the new boards, especially about voltage and current floating-charge values, and they will not affect interchangeability between ‘old’ and ‘updated’ boards.
However, one should keep in mind that further changes are to be evaluated carefully, as they could alter the features in terms of the alarm sensor and LED indications, for instance, eventually impacting the operational experience.
Supplement #1 - Other questions within:
R2 ‘optional’, but always present on all boards:
It seems its main function is to provide a low (very low) current above 7.24V, if the battery stand-by charging current is lower than 0.7 mA (voltage drop on R2 with diode D5 as 0.7V), but in real life 0.7 mA means for a 150mAh battery just 0.4%C, well below usual floating current levels.
D7 and relay:
It is used to avoid inductive voltage spikes when an inductive load (relay coil) is switched-off. Must be maintained.
D6 and Alarm detection:
It seems it is used to establish the supply voltage only to the relay through the High-Low alarm switch loop, but not directly to the relay.
R7 and possibly delayed alarm:
When operation = Normal, alarm switch is not triggered, so its position is Closed, and the relay coil this energized. However when on Alarm status, and depending on how this switch is used inside the sensor, the switch response could be seen as faulty or have a large capacitor still charged, delaying relay response. Possibly R7 is used to avoid false alarms, working as a load, speeding the Alarm response.
