This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TLC393: What am I doing wrong?

Part Number: TLC393
Other Parts Discussed in Thread: TLV274, , LMC6772, TLV1702, TLV3402, LMC6484, TLV2374

I am trying to design a thermostat and I originally used the op amp TLV274 as a comparator with some success, but I wasn't happy with the response time so I decided to try your comparator,TLC393, in it's place. I installed it as shown in the schematic, powered it on, and it almost immediately burned up.

Am I using the device incorrectly?

Can you explain what is meant by: "Input current, . . ±5 mA" as it appears in the data sheet? From what I understand, the input impedance should be much higher than this.

Perhaps the transitions at the input is too slow? It is based on changing temperature, which can be pretty slow.

The single supply is 12V and the input to the TLC393 max is about 1V to 11.5V.

Should I be concerned with gate to source capacitance of Q1?

Thank you,
Dave

  • Is there anybody home?

  • Hello David,

    What burned up? The 393? The MOSFET? The Peltier?

    A few things I noticed:

    The TLC393 is not rail to rail input device. The input is capable of 0 to 12V.

    The input range of the TLC393 is (V+ - 1.5V), or 0 to 10.5V on a 12V supply. If the voltage is above 10.5V, the TLC393 output could be incorrect (possibly reversed).

    You will need a Rail to Rail input device, such has the LMC6772, TLV1702 or TLV3402, or limit the adjust range to below 10.5V.

    However, the TLV274 op-amp is also not rail-to-rail input, but it may have had different behavior when outside the range.

    Note that your "Voltage set indicator" buffer (U2B) will also be violating the TLV274's input range when the setting is above 10.65V. Again, you need a R-R input op-amp such as the LMC6484 or TLV2374, or limit the setting point to <10.65V.

    You have hysteresis, so you should not have any "chatter" on the slow moving input. If you do still have some chatter, then you may need to increase the hysteresis (decrease R12). Have you monitored the comparator output with a scope to make sure it is not chattering?

    Capacitive loading of a comparator output will not cause oscillations like an op-amp, but it will affect the response time as the output now has to charge that capacitance on every transition, which causes a "rounding" of the edges (particularly the low-to-high transition with a open collector output - as it is a pure RC time constant with the pullup resistor).

    The "±5mA Input Current" is the maximum current that can be drawn through the ESD input devices. There are ESD protection diodes from the input to each supply. If the input exceeds the supplies by >400mV, these diodes will start to conduct. The maximum current through the diodes needs to be limited to 5mA or less (the less the better) to prevent damage. A resistor in series with the input will suffice to limit the current. Since you are not applying an input beyond the supply voltage - then you should not have to worry about the 5ma input current limit.
  • Paul,

    Thank you so much for your reply.

    I guess I forgot to mention that it was the 393 device that burned up. Very odd. I'm pretty sure that i did not exceed the input limits you mentioned. In fact I started the test very conservatively. But then again, anything can go wrong, I suppose. In any case, thank you for pointing out the fact that these are not R-R input devices. I thought R-R output implied R-R input. Silly, huh? I ordered the LMC6484IMX/NOPB and LMC6772AIM/NOPB and hopefully that should take care of most of that issue. However, I will pay closer attention to it and limit voltages if necessary.

    I still think that oscillations may have been responsible for the failure. I did not have time to probe for it and I wanted to get your opinion before I tried again. I've read somewhere that adding a 5pF cap between the output and non-inv input of the comparator would eliminate oscillations in this case. What do you think?

    I didn't stop to think that the output of the comparator is open drain, and driving the mosfet is a matter of determining the capacitance of the gate (a few pF) and using the value of the pull-up resistor in finding the time constant. But I got it now. Thanks.

    5mA max through the input protection diodes. I got it now. Though I don't understand why anyone would intentionally drive the inputs this way. It seems to me that this would be very dangerous territory. A few millivolts in the wrong direction would quickly destroy the device.

    Thanks again,

    Dave