LM5010: 15V converter problem

softstart pin goes silmuntanously with the Vcc pin shortly down to GND even, bellow; blue curve in following plot; pink SWT, yell Vout(15V), green inductor current

Hi 

I didn't see any input or out capacitor at U8, I check TPS26623, it typical schematic, it have I/O cap. cause mosfet is a transient, so without enough input and output cap, the LM5010 may not dealing with the transient current with 10uF cap. so you may need adding some I/o cap and increase the output cap of LM5010, see whether the issue is still there

for the SW spike and VCC dip, it may cause some coupling or resonant when transient happened, you can check whether they still keep same waveform after increase the capacitors 

1. Thanks Daniel for advise, I will try to add some caps. It is still not clear for me what boost my input voltage.

HI Christopher

 When driver circuit draw a current from dcdc, from figure 3, I can see the Vin have a drop first, then when the load current go to 0, the dcdc come into the DCM, but the energy store in the input line can't be consume by the load , so it will charge the input cap to raise up the Vin.

Thanks

Daniel,

I am not sure about that, voltage before my input series diode D8 is constant, all the time at 48V, checked on Scope, thus I suspect it comes from the other side, am I not right?

" I can see the Vin have a drop first" - yellow in Fig3 is 15V output, that's true it goes down firstly, due to small cap and limited current output from the LM5010. But what boost input voltage? It is nice visible on fig 5 - blue is Vin, behind D8 diode, same voltage boost is on SWT pin.

It is in continuous mode when voltage goes up, actually, it starts to be higher when inductor current (green) starts to rise up, due to demanded higher output current, LM5010 gives then "full power", seems like the duty is limited to 50-60%

In the following Fig I have changed C30 to 47u cap and added directly to the output 22u, and I did short circuit between Vin and Out of the TPS, to be sure it is not limiting.

How do I actually know if LM was switched off due to UVLO or no?

I do not know if it is not clocking( after Vcc blue gos down) due to UVLO or due to steady state, small load.

From my observation increasing C30 to 2x47u or 2x47u +10u does not help to reduce voltage drop on the output in first shoot. It is always around 8V deep which is a lot for 15V DCDC. I have also tried to increase capacitor direct connected between 15V out and GND to 270u, it does not help also. I reduced Mosfet driver bootstrap caps to 1u, this can not be that 1uF is so high  load for the 270u cap. It should transfer energy immediately without noticeable voltage drop.

I have problem with identifying where does this drop comes from ( yellow)? Layout ?

------

Probably here is the problem, but why; blue is FB yellow Vout(15v)

Hi

1 can you share the layout?

2  can you remove the U8, and use a wire connect U8's VIN and VO and measure the current from dirver when Vo drops?

3 for identify the ULVO, you can check SS, below is the datasheet descripted of Soft Start function:

An internal switch grounds the soft-start pin if VCC is below the undervoltage lockout threshold, if a thermal shutdown occurs, or if the circuit is shutdown using the RON/SD pin.

Thanks

Hello Daniel,

this is proper measured Feedback signal and Vout, no crocs on probe, short ground connection. If we assume 6V drop on yellow with divider ratio of 6 I should get 1V drop on feedback blue, but I got 2V drop there. Still a little confusing.

1. problematic right now

2. TPS is bypassed, LM output is wire connected with TPS output. Plus I added 47u from output to ground. I will measure the load current and will update this post.

3.thx

--update----

Green is now the load current, to be honest..that i did not expected, so high. This would clarify so high voltage drop. I must check if all the Low side mosfets are not switched simultaneously, this starts to charge bootstrap caps. I have one 1u ceramic cap for each IR2181 driver, in total 3pcs.

--update 2

yes all the 3 Low side H bridge mosfets in the first initial phase they start at the same time which surge the current. workaround would be adding some delay, or  add dedicated cap for each driver next to it, currently I have only typical 100n for each and one common the 15V output.

Hi

I think the large load transient is the cause.,. and  because  you have two  2.9ohm resisitor series with C30, so no matter your increase the C30, you will always find the voltage drop is still so large. I suggest you can try type2 with smaller resistor or even type 3. for type 3, if you increase C30, it will have help on reduce the VO drop. 

For your SW or Vin increase, cause there is some parastici inductance between Vin input and Vin of IC, so when you "short" the output, the input current will become larger, after the load current decrease, the energy stored in the parastic inductance will charge the input cap of IC which will make the input of IC increase as well as SW.

Thanks

This might be the key 6A x1.45 Ohm gives  huge drop. However I had additional 270u direct from from +15 to gnd. When shorted, firstly the 270u 'responses' and it's internal dcr is the key.

Thanks for the help, now I LL try to improve gates sequence and add some hardware solution. Regards