How do I size these components?

[redgear]

I am building a Battery Management System myself from scratch by using TI's reference schematics. I have a few noob questions.

Here is the schematic I am following


I have a question on the following components are sized.

Diodes : D1, D5, D6, D2, D3, D4

Capacitors : C6, C7, C3, C4, C2, C1, C5

Resistors : R7, R6, R16, R14, R17, R18, R15

I posted a question on TI's forum and got the following reply:

Each design orgainzation has its own rules on how to select a component, many use a 30 to  50% derating.  It gives some margin for operation and keeps components closer to a normal operating condition.

D1 and D5 are TVS to avoid transients on the battery and pack.  These can avoid damaging the electronics by limiting the peak voltage of the system.  If they are needed in the system they must be selected according to the needs of the system.  They should not leak during normal operation and clamp before damage occurs to any sensitive equipment.  Some designers do not like this approach for D1 since if it were to short it would be a direct short across the cells and open a fuse to the cells. D6 is a flyback diode, its purpose would be to clamp PACK+ near PACK-.  It would be a pulse current when the discharge FET opens under load, have a fast response, and stand off the normal operation voltage of the battery plus any transient.  D2, D3, D4 are transient limiting zeners which should not conduct with normal voltages and limit for the resistors selected within the voltage of the FET gate or pin limit described.

C3, C4, C6, C7 are ESD suppression diodes, they provide a high frequency path for ESD currents across the FETs or across the pack terminals.  2 are used in series so that a single component short does not short the FETs or PACK. Voltage should sustain the battery voltage.  The size may vary, 0.1 uF is typical. C2 is the charge pump capacitor, it should be selected depending on the load.  It will typically have the charge pump voltage, less than 15V.  See the data sheet for selection.  The BAT and PACK pin capacitors should be 10 nF as indicated, or smaller if needed, see http://www.ti.com/lit/pdf/slua794

R7 is one of the pre-charge resistors.  If precharge is used select values to limit the current to a safe level for the cells in the under voltage condition.  Select a power rating or sufficient resistors to handle the dissipated power.  R6 and R16 are typically 10M resistors, these keep charge off the FETs when not driven by the BQ76200, they are large to minimize load on the charge pump and reduce battery current.  R14 is selected to control the switching speed of the FET and allow the part to operate.  See  see http://www.ti.com/lit/pdf/slua794 .  R17 is not normally used, but could be used as part of the PACK filter.  R18 is the primary PACK filter resistor, it will limit current into the PACK pin but can slow turn off if large.  See the apnote. R15 is a current limit resistor, it would conduct only in transient.  Use the recommended value, see the apnote.
For TVS selection check to see if the series you are considering is rated at the breakdown voltage or the working voltage.  Check the specifications, you want an acceptable leakage at your maximum voltage and temperature.

I would think the 50% derating applies to the nominal voltage to accept the transient, but opinions will differ.  For a 29.4V system a 100V diode seems suitable.

The charge pump capacitor should be selected as either the minimum recommended or scaled from the load capacitance as described in the data sheet http://www.ti.com/document-viewer/BQ76200/datasheet/scalable-vddcp-capacitor-to-support-multiple-fets-in-parallel-slusc161521#SLUSC161521

And going by the reply,
I am planning to a use a TVS rated 33.3v Breakdown and a flyback diode rated at 100v.
All my capacitors, resistors and diodes will have a 50% derating from nominal voltage.

Do you have any suggestions? Does this look fine?  Is there a better way to do it?

[redgear]

BUMP

[T3sl4co1l]

The gate resistors should be split so that there is one resistor per MOSFET.  Then the ferrite beads don't matter.

The BQ- looks to have zeners inside so D2 and D3 aren't necessary.  D4 is superfluous anyway.  D6 is superfluous as, as I mentioned in the other thread, a unidirectional TVS (D5) is already a diode.

If the voltage range is up to 60V, why would you choose a 33V TVS?

Tim

[redgear]

Hi Tim,
Thank you for taking the time to answer. I am a beginner and I am sorry if the questions below are stupid.

The gate resistors should be split so that there is one resistor per MOSFET.  Then the ferrite beads don't matter.

So, I should just add two more gate resistors with the same value of R6 and R16, such that each MOSFET has a resistor connected to its gates and remove the ferrite beads. Correct?

The BQ- looks to have zeners inside so D2 and D3 aren't necessary.  D4 is superfluous anyway.  D6 is superfluous as, as I mentioned in the other thread, a unidirectional TVS (D5) is already a diode.

One of their app note says under headings `Protecting the FET Gate and DSG Pin` and `CHG Circuit` it says:

Code: [Select]

A zener diode may be desired between the gate and source to protect the FET VGS from excessive voltageMay be it's designed for redundancy. Thanks for pointing it out, I will get it clarified with them.
I was also thinking if there would be a better protection than a TVS diode?
Diode D6 is a flyback diode, its purpose would be to clamp PACK+ near PACK-.  It would be a pulse current when the discharge FET opens under load, have a fast response, and stand off the normal operation voltage of the battery plus any transient. (they have mentioned it in their reply they have posted above)

If the voltage range is up to 60V, why would you choose a 33V TVS?

The above schematic I attached is a reference schematic from Ti. My system is a different from theirs. I am a beginner and felt it is best to modify the above schematic to my requirements than to create one from scratch. I will be using a battery that will measure 28.7v when fully charged. I would like to limit the voltage build up to a 30.1v. But I could not find a TVS diode rated at 30.1 so I went with the closest value that is 33v.
I haven't accounted for transients while choosing the diode, Should I?


Edit : Can I apply 50% derating on the capacitors from the nominal voltage - 40v ?

[T3sl4co1l]

So, I should just add two more gate resistors with the same value of R6 and R16, such that each MOSFET has a resistor connected to its gates and remove the ferrite beads. Correct?

Correct.

One of their app note says under headings `Protecting the FET Gate and DSG Pin` and `CHG Circuit` it says:

Code: [Select]

A zener diode may be desired between the gate and source to protect the FET VGS from excessive voltageMay be it's designed for redundancy. Thanks for pointing it out, I will get it clarified with them.

Weird.  Note that, if you do use a zener or TVS to protect gate voltage: put it "outside" of the gate resistors.  The resonant loop created between a zener's capacitance and an active MOSFET typically causes oscillation during switching.  Making sure that the loop between MOSFET and zener includes a resistor, prevents oscillation.

I don't know what they would be protecting against, so I might suggest one of:
1. Nothing
2. A simple zener diode like 1N5244 (or an SMT equivalent)
3. A TVS like P6KE15A or SMAJ15A

Diode D6 is a flyback diode, its purpose would be to clamp PACK+ near PACK-.  It would be a pulse current when the discharge FET opens under load, have a fast response, and stand off the normal operation voltage of the battery plus any transient. (they have mentioned it in their reply they have posted above)

Yeah, I got that.  That doesn't say anything about the diodes being in parallel, or the one diode doing the same thing already.

The one diode will do both already, D6 isn't needed.

The above schematic I attached is a reference schematic from Ti. My system is a different from theirs. I am a beginner and felt it is best to modify the above schematic to my requirements than to create one from scratch. I will be using a battery that will measure 28.7v when fully charged. I would like to limit the voltage build up to a 30.1v. But I could not find a TVS diode rated at 30.1 so I went with the closest value that is 33v.
I haven't accounted for transients while choosing the diode, Should I?


Edit : Can I apply 50% derating on the capacitors from the nominal voltage - 40v ?

Ah.  33V will be fine then.

What is this powering, or being charged by?  What kind of transients are you expecting?

If a charger is connected which is higher voltage or current, and attempts to overcharge the battery, what happens?  Should the diode absorb all that excess energy?  Should the pack voltage be allowed to float up so that no charging current is drawn?  What is that hypothetical charger likely to run at?

What kinds of loads will the pack see?  Inductive, capacitive, motors, controllers, converters?  Regenerative or dumping?  (Most controllers and converters will have capacitance at their inputs, due to local filter capacitors needed for their switching circuitry.)

Capacitors can be rated for:
- Electrolytic or polymer: 35 or 50V
- Ceramic or film: 50-100V
- Don't use tantalum

Most capacitors will be ceramic, because they're small.

C2 I'm guessing can be rated for less voltage.  RTFDS to see what that pin charges up to.  At 1.5uF it will be a larger ceramic (1206 size?) and may be rated 16-50V.  Electrolytic would probably be okay, too.

Ceramic: prefer X7R for general purpose, and C0G under 1nF.  For X7R, check the characteristics.  If you can't find a char. sheet, don't use it.  (Note, some brands keep these data on their website, not linked from the supplier.  So you have to poke through their listings to be sure...)  You're looking for a plot of capacitance versus DC bias voltage.  Make sure it has adequate capacitance at the voltage used.  -30% is a typical threshold.

Tim

[redgear]

Weird.  Note that, if you do use a zener or TVS to protect gate voltage: put it "outside" of the gate resistors.  The resonant loop created between a zener's capacitance and an active MOSFET typically causes oscillation during switching.  Making sure that the loop between MOSFET and zener includes a resistor, prevents oscillation.

Cool. The schematic above has resistors connected between zener and MOSFET -- R13, R14, R15. Are they enough?

I don't know what they would be protecting against, so I might suggest one of:

D4 is to reduce DSG switching loss, D3 is to protect the DSG FET gate, and D2 is for CHG FET gate protection.

The one diode will do both already, D6 isn't needed.

Ok

What is this powering, or being charged by?  What kind of transients are you expecting?

The entire battery pack will be powering a ebike including motor, motor controller, micro controller and a TFT display. It was advised in one Ti's app notes to use 10V transient per cell in series. I will using a 7s design.

If a charger is connected which is higher voltage or current, and attempts to overcharge the battery, what happens?  Should the diode absorb all that excess energy?  Should the pack voltage be allowed to float up so that no charging current is drawn?  What is that hypothetical charger likely to run at?

This is a part of BMS circuit. The above schematic is that of the High Side Gate Driver which will be controlled by a Gauge IC. The gauge IC has programmed limits for charging and discharging currents and volatges based on which it opens or closes the FETs using the BQ76200 High side driver(in schematic above)
Sorry, I should have linked to the entire schematic in my thread. Here is a link to the entire schematic

What kinds of loads will the pack see?  Inductive, capacitive, motors, controllers, converters?  Regenerative or dumping?  (Most controllers and converters will have capacitance at their inputs, due to local filter capacitors needed for their switching circuitry.)

The loads will be low powered BLDC motors(250W) and regen capable controllers. My motors would operate at 10A constant with max of 15A. I guess the current pulled by the controller and display won't matter much as they will be very less.

Capacitors can be rated for:
- Electrolytic or polymer: 35 or 50V
- Ceramic or film: 50-100V
- Don't use tantalum

Most capacitors will be ceramic, because they're small.

C2 I'm guessing can be rated for less voltage.  RTFDS to see what that pin charges up to.  At 1.5uF it will be a larger ceramic (1206 size?) and may be rated 16-50V.  Electrolytic would probably be okay, too.

Ceramic: prefer X7R for general purpose, and C0G under 1nF.  For X7R, check the characteristics.  If you can't find a char. sheet, don't use it.  (Note, some brands keep these data on their website, not linked from the supplier.  So you have to poke through their listings to be sure...)  You're looking for a plot of capacitance versus DC bias voltage.  Make sure it has adequate capacitance at the voltage used.  -30% is a typical threshold.

Thank You!
C2 is a charge pump capacitor for pre-charging purposes. It will typically have the charge pump voltage, less than 15V. I'm still trying to figure out its capacitance as the information about it in the datasheet is vague.