LM5066I: Function and impact of PWR-Resistor

By the way, is this an error in the datasheet?

Hi Christian,

Thanks for picking up this device. What is the max. load current for your system?
Rsns should be modified to support higher currents. Can you please fill the design calculator tool www.ti.com/.../toolssoftware and send me.

Best Regards,
Rakesh

Hi Christian,

It is an error in the datasheet. The numerator and denominator are swapped.
We will make it right in the next revision of the datasheet. Thanks for bringing to our notice.

Best Regards,
Rakesh

Hi Rakesh,

Please see my filled design calculator tool: LM5066I_Design_Calculator_REV_C.xlsx

The maximum current in this system will be 10 A continuously. The R_CL1 and R_CL2 resistors nearby to the R_SNS are assembled to increase the minimum current limit to more than 10 A.

I double checked some of the values in the design calculator and came up with more things, that are unclear. First of the SOA of the used MOSFET SIR870ADP. According to its datasheet diagram it is safe to drain 5 A at a V_DS of ~33 V for 10 ms.

When using the initial dv/dt values shown in the circuit diagram in the first post and R271 = 100k, a load of 3.3 Ohm high power resistors and 3x 220 uF capacitors can be switched on within ~15 ms. The average current during that startup ramp would be 4.5 A, which fits into the SOA. Also it was possible to start into short circuit and connect a short circuit or the load while running (multiple hot plugging) without much temperature rise. But the design calculator tool shows a critical value in the "SOA margin during start-up" field in this case. The FET Power diagram starts with 150 W and falls linear to 0 in ~16.5 ms (=1.24 Ws).

Ch1: Voltage on load
Ch2: Enable signal of LM5066
Ch4: Current through load

So I tried the recomended values of the design calculator tool, mainly 200 nF for the dv/dt capacitor. This of course results in a much longer rampup time of ~150 ms, but with the same average current of 4.5 A, which then is not longer within the SOA. The FET Power diagram in this case starts with 7 W and falls linear to 0 in ~330 ms (=1.16 Ws).

Please tell me if I made an error in reasoning. Maybe I'm simplifying the current ramp too much or neglecting the thermal impedance?

Best regards,
Christian

Hi Christian,

On the first test waveform, the output came up in 15ms.. so for 33V input it is roughly 2V/ms ramp rate at Vout.
2v/ms rate charges 3x 220 uF capacitors with 1.32A charging current and at the same time the 3.3Ohm load discharges... As an example at 6V vout, the 3.3Ohm discharging current dominates leaving no-current to charge 3x 220 uF capacitors. I am wondering how Vout came up? May I am missing something. Please cross check at your end, if needed we will discuss over call.

Best Regards,
Rakesh

Hi Christian,

Any update to close this?

Best Regards,
Rakesh

Hello Rakesh,

I did some investigations with the circuit and detected some more problems. My current assumption is that the serialized coils cause a voltage peak when connecting a short circuit on the output, destroying the switching FET. This problem is postponed for the moment.

Instead there is an even stranger issue with a nearly identical LM5066i circuit. When powering this with 49V, the output voltage can be switched on, but seems to turn off randomly after some time (~60 seconds). After that, the voltage can be turned on again but turns off immediately again. I will do further tests with this and may open an seperate post, as it is a different problem.