DRV8801: Undervoltage Lockout and nFault indication

Hi Gregor,

The datasheet specification when the voltage on the VBB pin is increasing (ramping up) the UVLO threshold could be a maximum of 7.5 V. The DRV8801 was designed and guaranteed to operate from 8 V to 38 V. When an UV is detected the outputs of the driver will be disabled. When VBB > VUV the outputs should be operational although nFAULT may be indicated until VBB > VBBNFR. The VBBNFR is for reporting only and does not affect the device functionality as described in the datasheet. See below. Does your device function normally even when the nFAULT was asserted due to UV reporting?

Other options could be;

– DRV8870: 6.5-V to 45-V, 565-mΩ, shunt

– DRV8251: 4.5-V to 48-V, 450-mΩ, shunt

– DRV8251A: 4.5-V to 48-V, 450-mΩ, mirror

– DRV8231: 4.5-V to 33-V, 600-mΩ, shunt

– DRV8231A: 4.5-V to 33-V, 600-mΩ, mirror

There is also a DRV8801A-Q! option in WQFN package, https://www.ti.com/lit/ds/symlink/drv8801a-q1.pdf with modified UVLO thresholds.

Regards, Murugavel

Hi Murugavel,

thanks for the quick reply.

The IC is fully functioning in the state that i described when nFault is low. So from that point of view, everything is ok.

But my question is: What's the voltage threshold (for VBB) at which the nFault will go low when VBB is decreasing? Let's say VBB is at 13V and starts decreasing, at what voltage level will nFault trigger?

This is the info that i cannot find in the datasheet.

And this is important too, because we have a microcontroller evaluating nFault, so i don't want to trigger if it's just VBB that was briefly below 11.5V or something like that.

Plus: could you recommend a driver IC which is as similar as possible to DRV8801 but has a lower level at which VBB triggers the nFault?

Thanks and best regards,

Gregor

Hi,

Thak you for your questions. Monday Feb 19th is national holiday in US. Please expect some delays.

Hi,

ok, that's fine. I would still be interested in the answers to the questions in my last message.

In the mean time i looked at alternative ICs. DRV8876 seems to be what i'm looking for. But for this IC I would need the same information:

Is there a voltage level at which nFault triggers a low voltage condition but at which the Outputs of the IC are still working? Similarly to what we see on DRV8801?

Thanks and best regards,

Gregor

Hi Gregor,

The UVLO detection threshold to disable the outputs and nFAULT reporting UV threshold are two separate entities in the DRV8801. Like mentioned in the datasheet "Diagnostic Output" description nFAULT will be asserted until VBB > VBBNFR, which according the specifications could be a max of 13.8 V for the DRV8801. Because you are using this reporting input to your MCU looks like it may not be desirable the way it behaves although the driver output functions normally. 

For the DRV8876, the behavior is different. This may be suitable for what you were looking for. 

Regards, Murugavel

Hi Murugavel,

thanks. But to be honest, you still didn't answer my question:

With VBB starting high (for example 15V) and decreasing, at which level will the nFault pin report a UV situation?

And more important right now for me: how is this behavior in the DRV8876? Does it also have a level at which nFault indicates a UV situation but the outputs are still working?

Thanks and best regards,

Gregor

Hi Gregor,

For the DRV8801. I highlighted a sentence in section "8.3.8 Diagnostic Output" paragraph image. And also mentioned in my reply "Like mentioned in the datasheet "Diagnostic Output" description nFAULT will be asserted until VBB > VBBNFR, which according the specifications could be a max of 13.8 V for the DRV8801.".

Based on this statement, the VBB voltage threshold for nFAULT pin output reporting fault with logic LOW would be VBBNFR which is specified typically 12 V and a max. of 13.8 V. With 12 V power supply this may mean the device be constantly reporting a fault on the nFAULT pin. That's why I mentioned "it may not be desirable the way it behaves although the driver output functions normally.". The UVLO internal fault detection and disabling output happens at 6.5 V typical and 7.5 V max. So UVLO internal and nFAULT reporting are decoupled in this device. We do not have an EVM for the DRV8801 with us at the moment to verify this. We may be able to get hold of one in a few dates and verify. If any changes/updates with the observed behavior I'll let you know.

With the DRV8876, I mentioned previously the behavior is different and highlighted some information from its datasheet. In summary when the device is operating at 12 V for example, if the VVM (VBB) voltage falls below UVLO VVM falling threshold nFAULT pin will be asserted to logic LOW reporting a fault as well as disables the bridge both at the same time. Then for recovery from fault, after VVM increases above UVLO VVM rising threshold nFAULT pin will be logic HIGH (with the pullup resistor) and the bridge will be re-enabled automatically, again these two happen at the same time.

We had an DRV8876EVM with us and able to verify UVLO behavior described in the datasheet is true. With our 12V setup with a motor when the voltage was decreased nFAULT was reported at 4.4 V and the bridge was disabled and when the voltage was increased, at 4.48 V nFAULT was cleared and the bridge was enabled and normal operation was restored.

I hope this information addresses what you were looking for. Thanks.          

Regards, Murugavel

Hi Murugavel,

thanks a lot for clarification. Sorry that i was confused. I thought VBBNFR would only mark the level of resetting the nFault, not also the triggering (also because we're operating many many DRV8801 at 12V and none has the nFault triggered...).

But in this case i'll switch to the DRV8876.

Thanks again and best regards,

Gregor

Hi Gregor,

No worries. I interpreted the datasheet description and made that statement.

You said, "also because we're operating many many DRV8801 at 12V and none has the nFault triggered...". Thanks for sharing this information.

This was why I wanted to test the behavior of the DRV8801 with an DRV8801EVM. I'll let you know as soon as I get an EVM and check out the exact behavior. Based on your experience it appears that the behavior might be different.  

For the DRV8876 the datasheet description and the device behavior were aligned as expected. Thanks.

Regards, Murugavel

Hi Murugavel,

ok, I would be very interested in your findings.

On my sample, i found the nFault pin to be triggering somewhere around 11.0 ... 11.2V.

Best regards,

Gregor

Hi Gregor,

Thanks for the additional info. I'll keep you posted.

Regards, Murugavel

Hi Gregor,

I expect the EVM by Wednesday - just letting you know.

Regards, Murugavel

Hi Gregor,

The EVM delivery got delayed the updated ETA is Thu evening. We should be able to test the threshold operation by Friday and give you an update. Thanks.

Regards, Murugavel

Hi Gregor,

I received the EVM today. See below my observations with a small DC motor connected with PWM input in PH/EN mode 25% duty cycle.

VBB = 6V, Motor runs normally, nFAULT = LOW reports a UV fault. Increased VBB in 0.01V increments until nFAULT = HIGH (no UV fault).

VBB = 11.82V, Motor runs normally, nFAULT = HIGH no UV fault. Decreased VBB in 0.01V decrements until nFAULT = LOW (UV fault).

VBB = 11.02V, Motor runs normally, nFAULT = LOW reports a UV fault. Decreased VBB in 0.01V decrements until motor stopped, VUV detected.

VBB = 4.46V, Motor stopped, nFAULT = LOW unchanged. Increased VBB in 0.01V increments until just above VUV threshold by observing motor running.

VBB = 5.47V, Motor runs normally, nFAULT = LOW continues to report a UV fault.

Please note this test was with one sample of the device. These thresholds are expected to vary device to device. However the expectation on nFAULT behavior does not change. It is tied to VBBNFR specification. In this sample the threshold measured were < the typical value specification in the datasheet. The datasheet does not have a minimum specification. So lower than typical values may be expected with other samples as well. Some samples may report above typical value because the specification allows it. So no guarantees that all devices produced will have VBBNFR threshold < 12V.

Regards, Murugavel   

Hi Murugavel,

thanks for your investigation. My conclusion would be as follows:

I you wanna operate the circuit at 12V nominal (which we wanna do), you should not evaluate nFault pin. Because your voltage might drop to a level where nFault triggers (goes low) and it might never recover.

I think i'll go with DRV8876.

Thanks again and best regards,

Gregor