TPS22810: Limit to Adjustable Slew Rate

Hi Whit,

It is possible for the TPS22810 to have an extended rise time up to 100ms+. That being said, this also depends on the load being driven and output capacitance since these two determine the self heating of the device. 

What is the output current load and capacitance?

Thanks,

Alek Kaknevicius

Thanks for fast response, Alek. My application is to control the power to remote low noise amplifiers located near antennas. These are powered through the coaxial cabling with bias-tees. The coaxial cables can be as long as 30 m. i need to limit the applied voltage risetime so as to not damage the amplifier MMICs.

The normal load current is 230 to 250 mA. The load capacitance is on the order of 10 uF at the input to a remote LDO at the amplifier.

I would like to use this same basic design at 12 V and 5 V. The load parameters would be approximately the same although the load current may be closer to 100 mA for the 12 V design.

Thanks for helping.

Whit

Hi Whit,

At that rise time, the inrush current from the 10uF capacitance would be negligible, so the power dissipation concern would be with the nominal load current. When the switch first tries to turn on, the resistance of the FET decreases until the output begins to rise. At this point, if the load is active, then the switch sees the full voltage drop from VIN to VOUT and the load current is able to pass through the device. As VOUT increases, the voltage difference between VIN and VOUT decreases, so the power dissipated also decreases.

Because power dissipation is the key concern here, I would recommend the DRV package over the DBV package due to the increased thermal performance. Assuming that package, let's take a look at the different cases:

1. 5V/250mA: The worst case power dissipation here is 1.25W. At a DC condition, this would cause a ~94°C rise in temperature, but since the power is temporary, it will not heat up this much. Not too much of a concern here.
2. 12V/100mA: The worst case power dissipation here is 1.2W, so this is very similar to #1.
3. 15V/250mA: The worst case power dissipation here is 3.75W, making this the most strenuous condition. I imagine that thermal shutdown may kick in during turn on, but this can be a way to turn on the load incrementally (VOUT rises, thermal shutdown hits, device is off and cools down, device back on after thermal hysteresis, VOUT rises...).

Two things to consider for this design are the load voltage and ambient temperature. With a high ambient temperature, these cases would become more difficult for the device to handle without hitting thermal shutdown. What I mean by "load voltage" is the voltage on the load that causes the load current to be pulled. The cases above assumed that any voltage on the load would cause a current draw.

Ultimately, the best way to evaluate the device performance here is to take the EVM, modify the CT capacitance, and connect the load you are trying to power. If you monitor the output on an oscilloscope then you can look at the waveform and see if it is acceptable for your application.

Thanks,

Alek Kaknevicius

Thanks again, Alek. This is all excellent info and you've saved me a lot of trouble and helped me avoid making wrong assumptions.

The calculation for 100 ms risetime to 15 V gives a slew rate of 150 V/s or 0.00015 V/us. Using Ct = 46.62/SR from datasheet section 9.3.4 I get Ct = 310 800 pF or 311 nF. This seems to be a fair value (not extremes like microfarads or femtofarads).

The heat dissipation has me a little worried and I will take your advice on the EVM.

Any other advice also welcome here.

Best regards,

Whit