0
\$\begingroup\$

I want to build my own 120A Electronic Speed Controller and decided to try it with the DRV8305 IC. I am by no means a hardware expert, so go easy on me.

On page 45 of the DRV8305 datasheet: https://www.ti.com/lit/ds/symlink/drv8305.pdf?ts=1603013837741&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FDRV8305 You will see SN1, SP1, SN2, SP2, SN3, SP3 as the connection points for 3 sense resistors. In many other datasheets, I see that all low side sources are connected together in a single shunt resistor. Page 47 describes just using 1 shunt, but I am not sure what the trade-offs are. What information do I get by using 3 instead of 1 and can I just use 1 to drive a sensor-less BLDC motor?

Another question I have is how to choose a shunt resistance value. In many datasheets examples of the rated current are like 10A, but I want a rating of something like 120A (common ESC rating). In the examples, the shunt resistance is something like 5 mOhm, but for 120A, I would need something like 500 uOhm. At that resistance, isn't the resistance so low that the PCB trace resistance dwarfs the sense resistor? If my sense PCB traces are even 1mm different from each other, that could be ~3 mOhm, about 6x the shunt. What is the solution here?

Thank you :P

\$\endgroup\$
2
  • 1
    \$\begingroup\$ This will not be an easy project without previous practical experience. You'll likely want to spend some time studying actual solutions typically used for your application, and not try to do a clean-sheet original design using some promising looking IC. If you really want to use the TI part, start with their reference design. If this is for an RC vehicle, start with the designs of actual hobby ESC's as documented by those who reflash them with open firmware. Whatever it is for, look at the implementation detail of incumbent solutions, otherwise you're just likely to burn FETs. \$\endgroup\$ Commented Oct 20, 2020 at 14:43
  • \$\begingroup\$ I agree it will not be easy, but I am still going to try. I researched MOSFETS with a low Rds_on, high continuous current, high voltage, high avalanche ruggedness, and a high power dissipation (PSMN1R0-30YLDX). I also intend to use 24 of them all cooled with the PCB and a heatsink and small fan. I have been referencing my blown up 120A ESC the best I can and if I do this right, it should be much more capable. I also am doing as you say and keeping pretty strictly to the reference design. \$\endgroup\$
    – James
    Commented Oct 21, 2020 at 2:45

1 Answer 1

1
\$\begingroup\$

If you'd take time to read a bit more through the datasheet, you'll find that two pages below, in the Current Sense Amplifiers section, this text appears:

The DRV8305 provides three bidirectional low-side current shunt amplifiers. These can be used to sense the current flowing through each half-bridge. If individual half-bridge sensing is not required, a single current shunt amplifier can be used to measure the sum of the half-bridge current. Use this simple procedure to correctly configure the current shunt amplifiers. [...]

So it's up to you to determine if you want to use three separate sensing resistors, or not.

As for the values for 120 A currents, you may want to look into current transformers for that, or Hall sensors, since the last sentence in that section states that with 5 mOhms resistors:

These values allow the current shunt amplifiers to measure ±33 A across the sense resistor.

Which might suggest that 0.5 mOhm would suffice for your application, if it weren't for the dissipated power: 0.0005*120^2=7.2 W.

In short, if you have reasons to monitor the current through each leg, do so at the expense at three current sensors, oterwise it's just like using one single sensor on the supply. Don't just guide yourself on "what others are showing", because each "other" datasheet has its own applications in mind. It's all about the current requirements.

\$\endgroup\$
3
  • \$\begingroup\$ Thank you for the reply. I did read that on page 47, but was wondering what advantage there is to using 3 different sensors. Luckily, I did manage to find this neat article from Texas Instruments about the different methods and their advantages: ti.com/lit/ug/tiducy7/tiducy7.pdf?ts=1603234989423 And 7.2W power dissipation seems very reasonable considering 500uOhm 8W shunts exist and if the motor is drawing 120A at a nominal 11.1V, 7W is <1% of the watts being drawn from the battery. \$\endgroup\$
    – James
    Commented Oct 21, 2020 at 2:34
  • \$\begingroup\$ @James You'd better be careful with that thinking: the maximum allowed total dissipated power (8 W) is not the same as the maximum dissipated power in your application (7.2 W). Usually, that 8 W means that the resistor is used for no more than 4-5 W (around half). Best listen to Chris Stratton's advice. I'm only saying this because I get the impression the eagerness is not tempered enough with caution. \$\endgroup\$ Commented Oct 21, 2020 at 6:48
  • \$\begingroup\$ That is a fair point, which is why I will be switching to a 10W (there isn't anything much higher unless I want to run multiple resistors in parallel). However, 120A is not likely to be reached consistently (my guess is about half that), and there should be quite a bit of fast moving air over it. But, if I notice it gets crazy hot, I will add little heatsinks to them...or redesign the board in a rev2. This is a personal project and I want to learn by doing and as well as by making mistakes...just...as few of them as possible. I will be be taking Chris' advice as well to some extent. \$\endgroup\$
    – James
    Commented Oct 22, 2020 at 1:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.