LMP7300: Measurements/Simulations do not match Design Equations

Erik Anderson41

Part Number: LMP7300
Other Parts Discussed in Thread: TLV6700

Hi,

The design requirements are simple: Detect when the battery pack reaches 57.4V to shutoff the charger (positive going). Then, detect when the battery pack discharges to 54.5V to turn on the charger (negative going).

See attached schematic/notes.

I've followed the design equations correctly as far as I can tell.

But unfortunately my simulation and measured results are way off. What am I missing? This should be such an easy design... argh..

Thanks for any help,

Erik

over 7 years ago

0 Erik Anderson41 over 7 years ago

Intellectual 925 points

My problem is actually quite similar to:
e2e.ti.com/.../1939385

But still I don't understand, I've swapped HYSTP and HYSTN, as well as swapped IN+ and IN-. My hysteresis is ~100mV, less than the max 130mV in datasheet.

0 Jonathan Nguyen over 7 years ago in reply to Erik Anderson41

TI__Expert 4555 points

Eric,

What is the full voltage range coming from battery pack B1?

Regards,
Jonny

0 Erik Anderson41 over 7 years ago in reply to Jonathan Nguyen

Intellectual 925 points

Thanks for the reply, about 50V to 60V. 14*S 4.1V lithium ion.

0 Jonathan Nguyen over 7 years ago in reply to Erik Anderson41

TI__Expert 4555 points

Erik,

I believe that there is an issue with the part when asymmetrical hysteresis is applied.

Could you try applying symmetrical hysteresis (Like Figure 16 in the Datasheet) and see if the simulated/measured values are the same (or at least much closer) to the calculated values?

Regards,
Jonny

0 Erik Anderson41 over 7 years ago in reply to Jonathan Nguyen

Intellectual 925 points

Ah okay. I will try that and report back, thanks so much for the help.

0 Erik Anderson41 over 7 years ago in reply to Jonathan Nguyen

Intellectual 925 points

In reference to figure 16 (symmetric hysteresis), the datasheet says:

"For precise level detection applications, TI recommends symmetric hysteresis values less than 5 mV to 10 mV"

But in my application I need about +-53mV of hysteresis. So, I'm not sure what to do.

But I did symmetrical hysteresis simulation anyways to see what happens. Now it trips at my low threshold 54.5 all the time, as if there is no hysteresis. If I flip the IN+ and IN- I get my high threshold 57.4 all the time, as if there is no hysteresis. I don't understand. Attaching my excel spreadsheet and LTSpice sim (you will need to recreate the LMP7300 symbol from the SPICE model).LTspice.zip newcomparator.xlsx

0 Erik Anderson41 over 7 years ago in reply to Erik Anderson41

Intellectual 925 points

I guess this part won't work for me. Can anyone recommend another comparator that will work for me? I need:
1. Handles 12V supply
2. Includes reference (~0.2%, doesn't matter what the reference voltage is)
Thanks!
Erik

0 Jonathan Nguyen over 7 years ago in reply to Erik Anderson41

TI__Expert 4555 points

Hi Erik,

Sorry for the late response. I did some digging up of the old design files and I found that the way the part implements hysteresis is that not only does it set the amount of hysteresis at the trip point, it also offsets it as well which is why the design seems to fail.

For the 53mV of hysteresis that you need, you are correct that the datasheet only recommends between 5-10mV. That was a detail that I missed and I apologize for that.

We also have the TLV6700 that I think would work better in your application. It is a 1.8V-18V Supply with 0.4V integrated reference and open drain output. It is designed as a window comparator so you can more easily set your thresholds. It has roughly half as much Iq as the LMP7300 however the trade off is the speed. Is there concern for the speed of the comparator detecting low/high thresholds for the battery pack?

The datasheet for the TLV6700 can be found here: http://www.ti.com/lit/gpn/tlv6700

A brief design guide can be found here: http://www.ti.com/lit/an/snva832/snva832.pdf

Regards,

Jonny

0 Erik Anderson41 over 7 years ago in reply to Jonathan Nguyen

Intellectual 925 points

No concern for speed. I will check out the TLV6700. Thanks a lot for digging into this for me.

Amplifiers

Amplifiers forum

LMP7300: Measurements/Simulations do not match Design Equations