How does ADM693 shift the watchdog-timeout to?

Hi Kaos,

This is an old part and the device characterization work done at the time isn't as comprehensive as nowadays.

we don't have enough test data to guarantee this specification over VCC and temperature range.

i would expect the timing specification limits to expand with wider temperature and VCC range, but i don't have data to quantify this.

one thing may help is figure 12, where the reset time is plotted against the temperature, the reset time and watchdog time should be derived from the same internal oscillator, thus radiometric. thus the variation in watchdog time would follow similar curve.

Ren

Hi Ren,

OK, thank you for your kindly answer.

I can explain this to our customer with your info and comment.

Thanks again Kaos

Hi Kaos,

I'm sorry to hear that.

as much as I would like to but we just don't have enough bandwidth to carry out this test.

the typical watchdog timeout vs. Vcc relationship should be easily measurable on the bench by your customer, which should give them an reasonable idea.

for us to be able to guarantee it, we will need more complex measurement setup for a decent volume of samples.

the ADM8696 is an enhanced version of the ADM696, it was released after ADM696 and have better characterization data.

Ren

Hi Ren,

Our customer is producing with AD693 at present.

So on the function test of one production he got the test result one fail that is out of rating 1 to 2.25sec.

The tester swing the supply voltage 4.5V to 5.5V.

Customer needs the min timeout time when each 4.5V and 5.5V are supplied.

How can I explain this to our customer well with use current your datasheet?

Thanks Kaos

Hi Kaos,

To be able to come up with a maximum specification and guarantee it, ADI need to carry out a full yield analysis on this part, which is not viable at this stage.

The best i can do for you is to bench test 3 parts for their watchdog timeout vs. supply and temperature.

This will give you a typical performance characteristic of this specification.

Will this be valued by your customer? or can they do the same on their existing system just as easy?

once the typical number is known, your customer may add some margin to guard band the system.

Ren 

Hi Ren,

Do you have the similar model as AD693?

If you have it, then can you show the chart WD-time out vs supply voltage?

The bench test you suggested is final way to report our customer I think.

Thanks Kaos

Hi Kaos,

I have done some bench testing on ADM693  and here are the results:

this should give the customer some idea on the relationship between VCC, temperature and wachdog timeout period.

Ren

Hi Ren,

I explained your answer to our customer.

But he could not satisfy your answer.

Again The info Customer want to know

1)the value of design WD-timeout

Not only one sample, he needs the design value that was brought from a lot of samples.

The value needs to include the individual variation of ADM693.

Do you have this value?

2)failed ADM693

Customer met one failed ADM693 that was 2.28sec/4.5V at 25C tempereture WD timeout.

So he needs the design data that this failed is real fail or not.

Any comment?

Thanks Kaos