TLV3402

Team,

Here is more info

Do you need the comparator to be Open-drain or can it also be Push-pull?  

Push- pull is OK as long as it can go rail to rail on the output.

 Also you mentioned that when the input transition from high to low is slow you see the propagation delay to be over 2msec.

Can you tell me the following?

1)     What are you powering the device Vcc with?

Currently I am using 15V from some of them and 12V for others.  This is due to wanting to decrease power dissipation in my regulators and also to guarantee that input voltages are less than or equal to VCC.  I could rearrange for lower voltage if I really had to, but it would increase heat and I would have to change some other resistors values and it would increase risk spikes exceeding VCC.

2)    How slow is the input transition from High to low?

The problem input is where the negative input is fixed at 2V and the positive transitions down from 2.5V to below 2V at a rate of around 0.5V / ms.

3)    Can you send me the schematic circuit so I can see you’re actual set-up with the reference voltage you are using?

Schematic won’t help much.  Very simple reference on one pin and resistor divider on the other.  The resistor divider voltage can vary from fast changing to very slow changing as described above.

So I need input range from 0 to VCC and output high to either be VCC or open-drain.  Basically, low input leakage current when output is high.

Oooohh...the experts should answer soon - I can't suggest a comparator without looking through the look up tables (a factory apps can do that).  But the overdrive is what is required for a speedy response of a comparator.  The input voltage needs to overcome the reference voltage + input offset voltage + overdrive.  You can see the tables in the datasheet that illustrate the propagation delay.  The more overdrive then the faster the response. This can be a bigger problem when adding hysteresis because a designer needs to take into account the offsets and overdrives necessary to achieve the expected response time. This is often due to the fact that most people use mV levels of hysteresis so a low offset helps here.

Off the top of my head without crunching numbers, there is no mention of hysteresis which could be the problem as I mentioned above (required overdrive beyond offset) and the transit time of his ramp may play a role.  I will have to think about that.

I will be back later if a TI engineer has not answered.  It is movie time son!

-Ken

Hi Johnny,

Keep in mind that the TLV3402 is a nano-power comparator and it isn't exceedingly fast, but a 2ms propagation delay time is long. The propagation delay is highly influenced by the input overdrive level. Overdrive is simply the voltage beyond and over the comparator trip voltage at which it changes state. Thus, a 20mV overdrive represents an applied input level that is 20mV beyond the voltage where the comparator trips.

It is normal behavior for the propagation time to become extended for analog comparators as the overdrive becomes less and less. Also, the low-to-high and high-to-low response times become longer with lower overdrive contributing to a longer propagation delay time. Figures 12 through 17 in the data-sheet provide examples of the slowed output response. So low overdrive voltage should be one thing to check. Also, take a look at the loading on the output of the comparator. A high-resistance, high-capacitance load will have long time constant.

Different comparators will have different propagation and rise and fall times, but a slowing of speed characteristics with low overdrive is common between them. Therefore, selecting a faster comparator should result in a shorter propagation time. It may be slowed relative to a high overdrive input condition, but it should be faster than a slower comparator such as the TLV3402. A faster open-drain comparator is the TLC393. Keep in mind that with increased speed comes at the expense of higher supply current.

Regards, Thomas