DRV8837: What is the efficiency of DRV8837?

Hi Tzion,

We will investigate and reply by Wednesday.

Tzion,

Most of our devices do not specify efficiency because it usually doesn't make sense to quantify without a loaded motor in the system. The DRV8832 is an older device, so I'm not sure why we put that information in the datasheet at that time. I believe that number comes from dividing electrical power into the chip by electrical power out of the chip into some load.

How are you calculating efficiency? Did you identify if the losses are mainly mechanical or electrical?

Part Number: DRV8837

Hi,

I ran the same scenario for both TI drivers with 3xAA batteries connected to TI DC booster TPS63050

In both cases I ran the eval boards with the same VM of 5V.

On both I choose a 50% duty cycle so motor will run in 2.5V.

Running both till total battery depletion.

DRV8832 ran for 23.5 hours

DRV8837 ran for 17 hours

Am I missing anything here? After all DRV8832 has a higher RDSon + Rsens.

One more issue is about the internal shoot through protection, does it have it?

Thanks,

Tzion

Hi James,
We ran the same setup for DRV8832 and DRV8837 with AA batteries and a DC 5V Booster.
Both were driving the same motor (Maxon brush DC motor) under 100mA
Both ran in 50% duty cycle that represent 2.5V.
Tests were done till batteries were depleted.
On DRV8832 it took about 23.5 hours, on DRV8837 it took 17.5 Hours.
That's why I want TI to test the efficiency so we will know the right driver for our battery operated pump.
Thanks,
Tzion

Tzion,

For battery-powered applications, the main device specs that I could see affecting efficiency are operating current and Rds(on).

The total operating currents of the two devices are similar, according to the datasheets (1.4 mA for DRV8832, and 0.7+0.8 = 1.5 mA for the DRV8837). However, the DRV8832 may draw additional current if you have a resistor on the VREF pin. Overall, I would not expect this difference to cause a 6-hour reduction in battery life.

In terms of power dissipation, the Rds(on) and sense resistors will be the biggest points to dissipate power before it even gets to the load. The DRV8832 has almost twice the Rds(on) (LS+HS = 450 mOhm) compared to the DRV8837 (LS+HS = 280 mOhm). Considering this data, I would expect the efficiency of the DRV8832 to be worse if we calculate efficiency as electrical power out divided by electrical power in. However, the power dissipated by the FETs in either device is < 5 mW compared to the 250 mW power dissipation by the motor. Also, the difference in the power dissipation between the two devices should not cause a 6-hour difference in operating time over a 24-hour testing period.

I think that the reason you are seeing this discrepancy in your battery lifetimes is because AA batteries can vary in their capacity. You may also want review your prototype boards for other possible current leakage paths that may contribute to this discrepancy.

Edit: Fixed typos

The DRV8837EVB operates the motor in 15Khz PWM (we set it to 50%).
The DRV8832EVB operates the motor in 40Khz PWM (we set it also to 50%) - PWM is being generated internally.
Can this reason cause this difference of 6 hours operating time?
Tzion

Hi Tzion,

This post was merged with your previous post to keep the responses in one place.

All of the DRV8x devices with integrated FETs have internal shoot through protection.

Hi James,

I think there is a small confusion.

DRV8832 has a P-Channel and N-Channel and has HS+LS=450mΩ + Sense resistor (0.2Ω)

DRV8837 has dual N-Channel and so it has HS+LS=280mΩ

So DRV8837 should be better and more efficient although the test results show different?!

Tzion

Hi James,

So from what I read between the lines you don't expect almost any differences between these 2 motor drivers?

I ask a lot on this because the DRV8837 is much cheaper and my project is very tight on cost but w/o compromising on quality (healthcare product).

We saw that the DRV8837EVM consumes about 5mA with no load (MCU + LED), so we removed the 5mA from the results of the DRV8837 and the results become much closer to each other.

I will send you the Excel file and graphs.

Tzion