Electronics > Beginners
How to find the continuous current and max current requirement for my setup?
(1/1)
redgear:
I am working on building an ebike myself. Right now, I am designing the BMS and I need to know the current requirements of the setup to size my FETs.
I will using a RC motor with a higher power rating, however the power from my battery pack will be limited to 250w. This is the reason I am having trouble finding the max current, continuous current for my setup.
I have to keep the max power under 250w due to legal restrictions. I have planned a 6s3p battery pack. So pack voltage will be 24v, maximum pack current will be 10.41, if I plan to keep the power constant?
I am looking at a reference design from TI TIDA-00449. It states :
--- Quote ---This design aims at ~300 A for 200 μs for SCD and ~200 A for 40 ms for OCD
--- End quote ---
How do they get these numbers?
How can find it for my setup? In addition to the motors, I will be using 2 ESP32's, a micro servo and a Nextion display. I assume the current requirements for these won't matter as much as the motors.
Thanks
Siwastaja:
SOA specification during inrush current to the capacitors, or during an accidental short circuit, is equally important, and a very dangerous catch for the young players, like the summer workers writing appnotes at TI.
A lot depends on how you drive the transistors. Is it an integrated power path controller?
What are the worst-case desaturation detection thresholds - at what voltage max, how long it takes to turn the FET gates off?
During capacitor bank charging, you may have a situation of Vds=10V and Id=1000A for some tens of microseconds. Look at the SOA graphs whether this is acceptable.
All the numbers are available or can be approximated: total capacitor ESR, and wiring R actually limits current, RC time constant approximates the length of the inrush.
Then steady state DC. Look at the worst case (lowest) Vgs your FET driver guarantees. Note that this can be abysmally low for poorly designed IC products with too low UVLO voltage - I have seen products that cannot guarantee better than around 4V Vgs, and these products are supposed to be used with high power (non-logic-level) transistors. A huge catch: such products are completely unusable but they won't tell you that, because they want to sell them.
But if you get far enough to know that your FET driver part can supply, say, Vgs=7V in worst case, then look at the Rds(on) curves for that temperature, calculate power loss I^2*R, look at the thermal specification, and calculate the temperature rise dT = thermal resistance junction-to-ambient times I^2R loss. SMD parts tend to site junction-to-ambient thermal resistance at a pessimistic condition using a small SMD pad on a 1-layer PCB with no extra cooling; you are safe using that number, but if you can provide better cooling, you can do much better.
Don't forget to include a traditional fuse. The failure mode for MOSFETs is they blow short.
You may want to add a TVS diode to add extra robustness againt blowing from short-time overvoltages.
As for how to find the maximum required current - if you have no measured data about how much you actually need, how about summing the maximum specified currents of the motor controllers, etc., and adding a bit of margin.
redgear:
--- Quote from: Siwastaja on August 13, 2019, 12:17:17 pm ---SOA specification during inrush current to the capacitors, or during an accidental short circuit, is equally important, and a very dangerous catch for the young players, like the summer workers writing appnotes at TI.
A lot depends on how you drive the transistors. Is it an integrated power path controller?
What are the worst-case desaturation detection thresholds - at what voltage max, how long it takes to turn the FET gates off?
During capacitor bank charging, you may have a situation of Vds=10V and Id=1000A for some tens of microseconds. Look at the SOA graphs whether this is acceptable.
All the numbers are available or can be approximated: total capacitor ESR, and wiring R actually limits current, RC time constant approximates the length of the inrush.
Then steady state DC. Look at the worst case (lowest) Vgs your FET driver guarantees. Note that this can be abysmally low for poorly designed IC products with too low UVLO voltage - I have seen products that cannot guarantee better than around 4V Vgs, and these products are supposed to be used with high power (non-logic-level) transistors. A huge catch: such products are completely unusable but they won't tell you that, because they want to sell them.
But if you get far enough to know that your FET driver part can supply, say, Vgs=7V in worst case, then look at the Rds(on) curves for that temperature, calculate power loss I^2*R, look at the thermal specification, and calculate the temperature rise dT = thermal resistance junction-to-ambient times I^2R loss. SMD parts tend to site junction-to-ambient thermal resistance at a pessimistic condition using a small SMD pad on a 1-layer PCB with no extra cooling; you are safe using that number, but if you can provide better cooling, you can do much better.
Don't forget to include a traditional fuse. The failure mode for MOSFETs is they blow short.
You may want to add a TVS diode to add extra robustness againt blowing from short-time overvoltages.
As for how to find the maximum required current - if you have no measured data about how much you actually need, how about summing the maximum specified currents of the motor controllers, etc., and adding a bit of margin.
--- End quote ---
Thank You! I will decide the Charge and Discharger FETs after I decide on the other things. I also need to select FETs for balancing my cells. I will be connecting a separate FET for each cell(3 cells in Parallel). I am planning for a 100mA balance current at 4v. How should I select the FETs now?
The reference design uses a CSD13381F4 FET for balancing. Can I use the same? Is there a more reliable and cheaper alternative?
Thanks
Navigation
[0] Message Index
Go to full version