DRV8840: Which RDS ON value is valid?

Hi Rick,

Thank you for the reply.

May I understand that DRV8840's FET has completely the same structure and physical characteristics as DRV8842's?

Shuhei

Hi Shuhei,

Yes, the FETs are the same for both the DRV8840 and DRV8842.

The difference is the inputs, PH/EN for the DRV8840 vs IN1/IN2 for the DRV8842, and H-Bridge Logic as described in Table 1 of the datasheets.

Hi,

Let me ask further question instead of Shuhei.

I'm an engineer working with him.

We are on our way to design board that has peltier driver circuit using DRV8840/42 (plus L & C for smoothing).

Our application requires nearly 5A drive for several minutes and 1 to 2A drive for rest of time.

Datasheet of DRV8840 says that its RDSon is 0.65Ohm at first page and DRV8842 datasheet says that 0.2Ohm and web page for DRV8842 also says that it is capable of driving up to 6A.

We did some experiment using DRV8840 and from these information, we have decided to build board using DRV8842 for enough margin.

Please clarify "What is true information" for all of these issue.

At this moment, from your reply, my understanding is that these official information are not true and DRV8840/42 are almost the same except for its control logic.

If so, are there any difference in terms of its spec(performanc) including PWM between DRV8840/8842?

In addition, can we drive our peltier with 5A drive current for several minutes? (I know that it also depends on board design though)

Regards,

Ryo

Hi Ryo,

"Datasheet of DRV8840 says that its RDSon is 0.65Ohm at first page and DRV8842 datasheet says that 0.2Ohm and web page for DRV8842 also says that it is capable of driving up to 6A.

We did some experiment using DRV8840 and from these information, we have decided to build board using DRV8842 for enough margin.

Please clarify "What is true information" for all of these issue."

The RDSon in the electrical characteristics of the DRV8840 and DRV8842 are correct. The RDSon in the front page of the DRV8840 is incorrect.

Also, please note the typical RDSon at 85C is 0.26 Ohm and maximum RDSon at 85C is 0.32Ohm

At this moment, from your reply, my understanding is that these official information are not true and DRV8840/42 are almost the same except for its control logic.

If so, are there any difference in terms of its spec(performanc) including PWM between DRV8840/8842?

As mentioned above, the Electrical characteristics are correct. A review of the datasheet revisions show the DRV8842 front page was corrected in March of 2016. The front page of the DRV8840 was not corrected.

In addition, can we drive our peltier with 5A drive current for several minutes? (I know that it also depends on board design though)

Driving 5A for several minutes may be difficult. The 5A maximum listed at 25C is peak current. RMS current at 25C is listed at 3.5A.

Has a DRV8842EVM been used to evaluate thermal performance? The DRV8842EVM has a large ground plane to dissipate heat.

If you need 5A RMS, the DRV8412 in parallel mode may meet your needs.

Another device is the DRV8701E or DRV8701P with external FETs. With the DRV8701E/P, you can select the RDSon of the FETs.

Hi Rick,

Thank you for your reply.

We were about to order(tape out) this designed board, so it was kind of "shock" but I've decided to change DRV8840 to DRV8412.

Let me ask a few questions about DRV8412.

• L and C for smoothing

In page 22(TEC driver), 2x2 outputs are combined as 2x single channels.

These channels are combined after L and C but will it be OK to just combine 2 outputs and then connect to L & C to reduce number of components?

• Decoupling C value

It seems that C value for PVDD and GVDD/VDD in this datasheet are very large.

In addition, GVDD/VDD does not require that much large current but 330uF is used in datasheet.

Are there any reason to use 330uF for GVDD and 1000uF for PVDD? I'm thinking of using 47uF or smaller(4.7uF might be enough) for GVDD and 100uF(+100uF unmounted) for PVDD .

 

Regards,

Ryo

Hi Ryo,

Let me ask a few questions about DRV8412.

• L and C for smoothing

In page 22(TEC driver), 2x2 outputs are combined as 2x single channels.

These channels are combined after L and C but will it be OK to just combine 2 outputs and then connect to L & C to reduce number of components?

It is not OK. Please refer to section 9.4.5 of the datasheet for an explanation. After the initial 30-nH to 100-nH inductors, the remaining components can be combined.

• Decoupling C value

It seems that C value for PVDD and GVDD/VDD in this datasheet are very large.

In addition, GVDD/VDD does not require that much large current but 330uF is used in datasheet.

Are there any reason to use 330uF for GVDD and 1000uF for PVDD? I'm thinking of using 47uF or smaller(4.7uF might be enough) for GVDD and 100uF(+100uF unmounted) for PVDD .

The capacitors used were for the DRV8432 which can provide 2x the current (14-A Continuous in parallel mode) of the DRV8412. The capacitors can be reduced, but please make sure GVDD/VDD do not violate the absolute maximum voltages.

Please note sections 8.2.1.2.2 and 9.4.2 of the datasheet regarding the GVDD and VDD pins.

Thank you for your comment.

I'm going to use ferrite beads (that has >100nH at <1MHz) before connecting Port A&B, C&D.

Regards,

Ryo