SuperCap balancing

Hi @anishkgt,

Rp1 will control how much current bypasses the cap and, therefore, how responsive it is to rapid charge events to prevent overcharging. The lower the resistance value, the more current it will bypass and the harder it will work to prevent the cap from overcharging. The trade-off is more power to dissipate in the P-channel MOSFET and the resistor. Based on the info in their documentation, it would seem that 100mA should be plenty for an application like yours. Since the Rds-on of the FET should be in the milliohm range, nearly all of the voltage will drop across Rp1 and you can calculate the value for Rp1 as:

Rp1 = Vcap / I

where “I” is the current through Rp1, or 100mA in this case.

The voltage across Rx2 determines the gate voltage of the P-channel FET. You want to scale it such that the P-FET is just beginning to turn on when the threshold voltage of M1 is reached. At that point, 1uA is passing through M1 and Rx2. The value of Rx2 will depend on the gate threshold value of the P-FET, so you may have to experiment a bit. In the reference design, they scaled it so that about 1mA passed through the P-FET when Vth for M1 was reached. Just a few millivolts above this, the current increased rapidly to 100mA, which clamps the capacitor voltage.

For this design, it is not as critical that the power resistors be right next to the FETs, as it is not a high-speed signal path. I would not go with the ALD810028, as it’s threshold voltage is too high for your caps. The ALD810027 or the ALD810026 would be better choices to protect your caps, with the 0026 giving you a little bit of a safety margin.

Notes on your design:

You are not leaving any margin for your current sense resistor. You will be passing 1W across it at the full 10A charge current, and the resistor you chose is rated for 1W. I would go with at least a 2W resistor in this case.

I don’t know where “CS” goes, but it looks odd to have a 10K resistor in series with “CS” as shown. Unless you have a pull-down resistor elsewhere, the voltage at “CS” will likely float when the phototransistor is off. If you place the resistor between the emitter (pin 4) of the phototransistor and ground and tie “CS” to the emitter also, then “CS” would be low when the phototransistor is off and high when it is on. If you want reverse logic, connect the resistor between 5V and the collector (pin 6), tie the emitter (pin 4) to ground, and connect “CS” to the collector (pin 6).

Also, a minor note: you have the connections for the P-channel MOSFETs, Q10-Q12, correct, but the internal drawings are wrong. You have them drawn as N-channel MOSFETs instead of P-channel MOSFETs. They should look like this:


Was wondering why a P-Channel MOSFET is used control the switching ? Why not an N-Channel ?

Because of the way FETs have to be turned on. To turn on an N-Channel FET (N-FET), one must raise the Gate voltage above the Source voltage (Vgs) by some amount (anywhere from 1V to several volts, depending on the particular FET). That works fine if the N-FET is on the low side of the load, but in the circuit referenced, the FET is on the high side of the load (Rp1).

Using an N-FET on the high side of a load, as soon as it starts to turn on, the voltage drop across the Drain-Source region (Vds) becomes very small because its resistance (Rds-on) drops to a low value. This raises the Source voltage up to nearly the Drain voltage, and since the Gate voltage can go no higher than the Drain voltage in this circuit, Vgs drops below the level necessary to fully turn on the FET. The only way to fully turn on an N-FET on the high side of a load is to raise its voltage above the Drain voltage, which can only be done by providing another higher voltage source or by using a booster circuit just for that purpose.

Alternatively, a P-channel FET (P-FET) is turned on by dropping the Gate voltage sufficiently below the Source voltage (-Vgs). Note that with a P-FET, the arrangement of the Drain and Source are reversed in the circuit, with the Source placed at the higher voltage point. So, when placed on the high side of the load, the P-FET will be off when the Gate and the Source are at the same voltage, and it will begin to turn on as Vgs becomes more negative. Since the Source is tied to the upper supply, it is easy to accomplish this without an extra supply or boost circuit.


I trust this finds you well.

I am reaching out to inquire about Supercapacitor charge balancing modules (SABMB6XX) circuits.

My requirement is to balance 6x(48V 165F Supercapacitors) connected in series making a combined voltage of 288V and 27.5F. In total we are going to have 3 sets ( 6 x (48V 165F Maxwell Supercapacitors) in series) making a total of 18 Supercapacitors.

Does this board works for 48V Supercapacitors? If yes then can you please suggest which SABMB6XX board to go for as per our requirement. I’ll be grateful.

Note from datasheet: SABMB6XX is optional with specific ALD9100XXSALI units installed. XX = 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30 and how to get the values for R1, R2, R3: USER SPECIFIED VALUES FROM OPEN CIRCUIT TO ZERO (0.0) OHMS as per our requirement.

Looking forward to hearing from you soon.
Best Regards,

Hi umer,

The Advanced Linear Devices (ALD) products, including the SABMBxxx products, are designed for balancing individual super capacitor cells (typically 2.5V to 2.7V per cell) connected in series. They are not designed to balance higher voltage super capacitor modules connected in series. Those enclosed modules are comprised of a bunch of individual cells packaged together into a higher voltage module, as your 48V Maxwell module would be. The ALD products would be the type of parts that might be used inside those 48V modules.

Looking online, I found this document, which appears to be a user’s manual for the BMOD0165-E048 B01 Maxwell module, a 48V 165F module. On page 6 of that document it states the following:

I have found similar statements in documents from other vendors we carry, such as for the very similar Eaton XLR-48R6167-R. It would seem that, as long as the modules have similar active internal cell balancing, it is not necessary to employ external balancing measures. This is not to say that one should ignore the signals from the diagnostic cables that these units provide, which can be used to monitor the status of each module and can be used to warn of over-voltage or over-temperature conditions of each module.

1 Like

Hi David,
Thanks a lot for the clarification. Means a lot.

1 Like