TPS3847: hysteresis

Hi Allan,

Are you able to provide the schematic that the customer is using?

Jesse 

Hi Jesse,

Here is our schematic. Thanks.

Hello Jesse,

Plz refer the attachment, thanks.

Jesse,

Its TSR Jeremy, plz refer the attachment, thanks.

Hi Sean, Jeremy and Alan, 

I also want to confirm that customer is seeing the RESET assert at 11.1V and de-assert at 11.2V. If it does, how long does the RESET stay asserted before de-asserting? Is it he full duration of td=(4.5ms(typ) and 20ms(max)? What is the test condition during failure, also, if possible, a scope shot will help as well.

Jesse 

Hi Jesse,

A quick question: regarding TPS3847108, what is the minimal and maximal hysteresis voltage? I assume the 3.5% of N-going threshold voltage is its typical value, is it?

Actually, the customer did not measure the signal, they were just seeing that the power source for the device would be switched to the internal battery from external DC supply when the external voltage is down to 11.1V, and would be switched back to external once the voltage is back to 11.2V. Please refer to the schematic. 

Do you mean the environmental temperature when you mentioned the test condition? It's at room temperature.

I also tested a couple of devices on my hand, but I didn't see this issue. RESET would assert at around 10.8V and de-assert around 11.2V, or maybe higher, assert at 11V and de-assert at 11.5V. The customer has found that one device out of ten has this issue so far. I'm not sure if there will be more.

Hi Sean,

The 3.5%*Vit is the typical value of the hysteresis with the min and the max factored into the device overall accuracy.  The hysteresis of this device is applied to the P-going threshold voltage (Vit+ = Vit- + Vhyst). 

Yes, ambient temperature would be on factor, but also the signal frequency, signal source and etc would help the debugging process as well. 

Jesse 

Hi Jesse,

I don't quite understand. Since Vhyst = 0.035*Vit-, Vit+ = Vit- + Vhyst = 1.035*Vit-. If Vit- is 11.1V, Vit+ would be about 11.49V, and there is no tolerance, is it correct? 

The external power source is a variable DC power supply, and the internal power is alkaline or lithium AA battery.
Here are the test steps:
1. Both external and internal voltage are about 12V at first, the device is powered the external as designed.
2. Turn down the external voltage to find the N-going threshold voltage. It's 11.1V in this case. The power source switches to the internal battery as expected
3. Turn up the external boltage to find the P-going threshold voltage. It's 11.2V in this case. The power source switches back to the external supply, which was not expected.

Hi Sean,

In the case of the TPS3847108 Vit+(typical) = Vit-+Vhyst = Vit-*(1+.035) = 10.8 * 1.035= 11.178V which is the de-asserting voltage for the device. For reset assertion, it is just the Vit- ± accuracy. For this device the highest voltage that will assert the RESET should be 11.07V.

Jesse 

Hi Jesse,

Understood, but what you mentioned is the general case and is addressed in the datasheet. What information could I provide so that you can help find the reason why the hysteresis voltage is only 0.1V higher than the N-going threshold?

Is there another component similar to TPS3847 but with a higher hysteresis voltage?