Hello community
I have already spent a lot of time on this issue and I am stuck, so I decided to ask some PROs for ideas.
To give you some background. Forgive me for that wall-of-text.
I want to do some mods on my electric scooter. Stock battery is 52 V / 18 Ah but I want to add another one - external - in parallel. But when you connect batteries in parallel you must take care to equalize voltage first. I wanted to create a device that will allow me to plug in an external battery without equalizing voltages and let my device decide which battery (one with higher voltage) should be used as long as they have equal voltage and both can work together.
First stage was easy: just a pair of op amps, two voltage dividers and wire op-amps with batteries so they gave me something like a logic gate. If voltage @ BAT1 > BAT2, OPA1 = HIGH, OPA2 = LOW. If voltage @ BAT2 > BAT1, OPA1 = LOW, OPA2 = HIGH.
12 V is being supplied from a step down converter that would be attached to “MAIN BATTERY” i.e. internal battery and power all ICs.
Next step was harder. My scooter can draw ~3 kW peak so my device will need to be able to handle up to 60 A current. For prototype stage, without any hardcore driving I believe five IRL540N N-MOSFETs with some tinned fat traces would suffice. At least that's what they do in chinese BLDC controllers. I have discarded P-MOSFETS as they are expensive and inferior in terms of parameters to N-MOSFETS.
But yes - N-MOSFETS. I can not switch batteries to load with them, as Source must be at lower potential than Gate. I firstly thinked about doing something like common VBATT and separate grounds, even done some testing with but… it was tricky. I realized that High-Side N-MOSFET gate drivers exist. After some research I know that there are not a lot of 100% duty cycle high side drivers, I only found two that can be useful and somewhat obtainable - Analog Devices LTC7001 (supports higher voltages) and LTC7004 (lower voltages).
My circuit with LTC7001: (maybe should have 1-2 more components but not critical in this moment)
And that's the moment where troubles begin.
According to the datasheet, LTC7001 pin TS should connect to ground when MOSFET is disabled:
But my setup looks like that:
As you can see, when right LTC7001 has low input and low GATE, left LTC7001 has high input and high GATE (Vin + 12 V).
In this situation, TS pin of the right IC is not connected to ground but to Vload switched by left IC. As a result, the gate of the right mosfet is not discharged to ground but rather is left floating a bit above Vload. I have done some simulation with LTSpice and it shows that my problems are real, in mirror setup of ICs, gate of disabled mosfet is not discharged to 0 V. But maybe I have done something worng, I have zero experience with LTSpice.
My idea to solve this issue was following:
When left IC goes high, transistor to ground B1 is disabled, transistor A1 is enabled, transistor A2 is disabled and transistor B2 is enabled, so right IC that is going low can have its TS pin referenced to ground.
But - A1 and A2 transistor bases should be attached to the same PINs as mosfet GATE, to turn them on ( VG > VE ), but I worry that there might be some undefined states? Or other issues with this. Maybe PNP A1 and A2 transistors could be used to do this switching so then would be pulled to ground instead of pulling to V+ to turn them on. But I would need to use another NPN to pull them to ground, as op-amps could possibly not sink enough current when low.
Does anyone have an idea how this could be solved simpler? Thank you for your support.
Home
Help
Search
Login
Register
