Author Topic: LM25085: Vout dropping out at higher Vin  (Read 1830 times)

0 Members and 1 Guest are viewing this topic.

Offline KristoffTopic starter

  • Newbie
  • Posts: 6
  • Country: us
LM25085: Vout dropping out at higher Vin
« on: August 21, 2020, 04:28:49 pm »
Hey all,

I posted this on the TI forums here, but haven't made any progress, hoping someone here may be able to help.

I used WEBENCH® Power Designer to design the DC part of a  24 VAC to 5V 3A and 3.3V 3A supply using the LM25085SD/NOPB.  The 5V and 3.3V circuits are identical except for the R17 and R23 which are used to set the output voltage.

When the circuit is powered via 24VAC transformer VS is typically around 34VDC and both the 5V and 3.3V hover around 2V.

Lowering the input voltage such that VS is ~12VDC results in the expected 5V and 3V output.  If I slowly increase the input voltage, eventually the rails will drop to again.  I've primarily been experimenting with the 5V rail and the drop out point seems to be around VS=19VDC, but also seems to vary with the output load.

I'd appreciate any suggestions / troubleshooting steps.  Please let me know what other info I can provide to help.







In both scope captures below

CH1 (yellow): Vout

CH2 (blue): switch output FET, Diode and inductor (T1, D6 and L1)

There is no load on Vout.

When VS = 19VDC      Vout = 5V

When VS = 20VDC   Vout = 1.38VDC



I attached a 12 Ohm load the the 5V output and repeated the measurements.  With the 12 Ohm load Vout drops when VS is greater than ~12V:

CH1 (yellow): Vout

CH2 (blue): switch output FET, Diode and inductor (T1, D6 and L1)


Here is VS=12V with 12 Ohm load on Vout:


Here is VS = 13V with 12 ohm load on Vout:


Same setup with time scale adjusted to see multiple switch events:



The positive side of the input cap and the voltage at Vin on the LM25085SD are steady at VS in both the good and bad scenarios.  No dips, very little ripple.

The voltage on the source side of the FET is also a very steady VS.

Here are some captures of the PGATE output with a 12 Ohm load (~400mA draw on Vout)

CH1 (Yellow) is Vout

CH2 (Blue) is the PGATE output of the LM25085SD

VS=11V   Vout = ~4.96V


VS = 12V   Vout = ~4.96V


Vs = 13V   Vout = ~2.05 V


VS = 15V   Vout = ~0.07V







 

Offline jkostb

  • Regular Contributor
  • *
  • Posts: 69
  • Country: np
Re: LM25085: Vout dropping out at higher Vin
« Reply #1 on: August 21, 2020, 05:35:06 pm »
Have you checked that your design meets minimum on time and off time time conform data sheet? If you increase the input supply voltage then the buck regulator needs to decrease duty cycle (=VOUT/VIN). But if you decrease duty cycle then the on time will decrease. All buck converters can only regulate if on time is higher than parameter specified in datasheet.

Further have you verified that buck converter is still working in continuous mode when the voltage drops under loaded condition?
 

Offline KristoffTopic starter

  • Newbie
  • Posts: 6
  • Country: us
Re: LM25085: Vout dropping out at higher Vin
« Reply #2 on: August 21, 2020, 07:38:07 pm »
Hey jkostb,

Thanks for the suggestions.

From the data sheet: The minimum on-time, which occurs at maximum VIN, should not be set less than 150ns
[ Attachment Invalid Or Does Not Exist ]

At Vin = 40V I calculate Ton as 192.9nS which is within the specified range.  I'm not sure about Toff.  I calculated it to be 1.350 nS as Vin=4V, but I'm not sure if that is correct or what it should be.

How can I verify that the converter is in continuous mode?

Thanks!






 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15802
  • Country: fr
Re: LM25085: Vout dropping out at higher Vin
« Reply #3 on: August 21, 2020, 09:32:07 pm »
Also, you said that the outputs vary depending on the load.
Is there a minimum load above which it starts regulating properly?
 

Offline KristoffTopic starter

  • Newbie
  • Posts: 6
  • Country: us
Re: LM25085: Vout dropping out at higher Vin
« Reply #4 on: August 21, 2020, 10:19:55 pm »
A higher load appears to lower in Vin value at which the output becomes unstable.

With no load a Vout is a stable 5V up to Vin of 19V.
With a 12Ohm load Vout is stable at 5V up to Vin of 13V.
 

Offline Benta

  • Super Contributor
  • ***
  • Posts: 6420
  • Country: de
Re: LM25085: Vout dropping out at higher Vin
« Reply #5 on: August 21, 2020, 11:33:18 pm »
My guess is that your inductors are saturating, that would fit with the symptoms. As you give no specifications to those (although they are the crucial point in a converter design), I can't say much more.

 

Offline KristoffTopic starter

  • Newbie
  • Posts: 6
  • Country: us
Re: LM25085: Vout dropping out at higher Vin
« Reply #6 on: August 22, 2020, 02:40:14 pm »
Hey Benta,

The inductor is: IHLP3232DZER100M01

It was recommended by TI's power designer tool.

 

Offline Benta

  • Super Contributor
  • ***
  • Posts: 6420
  • Country: de
Re: LM25085: Vout dropping out at higher Vin
« Reply #7 on: August 22, 2020, 02:51:28 pm »
Hey Benta,

The inductor is: IHLP3232DZER100M01

It was recommended by TI's power designer tool.

Hmm. They look OK to me. Did you buy them from a reputable source?

 

Offline KristoffTopic starter

  • Newbie
  • Posts: 6
  • Country: us
Re: LM25085: Vout dropping out at higher Vin
« Reply #8 on: August 22, 2020, 08:07:49 pm »
Yea, all the components were ordered from Digi-Key.
 

Offline Benta

  • Super Contributor
  • ***
  • Posts: 6420
  • Country: de
Re: LM25085: Vout dropping out at higher Vin
« Reply #9 on: August 22, 2020, 08:22:45 pm »
Then I'm out of ideas.
 

Online magic

  • Super Contributor
  • ***
  • Posts: 7458
  • Country: pl
Re: LM25085: Vout dropping out at higher Vin
« Reply #10 on: August 23, 2020, 06:42:11 am »
Disconnect all sensitive loads and halve the value of R14 or R17 to see if the problem is with FB or ISEN?
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15802
  • Country: fr
Re: LM25085: Vout dropping out at higher Vin
« Reply #11 on: August 24, 2020, 12:49:16 am »
Also, one obvious thing to check would be the soldering. This IC comes in packages with a thermal pad. Have you properly soldered it?
 

Offline Phoenix

  • Frequent Contributor
  • **
  • Posts: 442
  • Country: au
Re: LM25085: Vout dropping out at higher Vin
« Reply #12 on: August 24, 2020, 04:52:37 am »
Disconnect all sensitive loads and halve the value of R14 or R17 to see if the problem is with FB or ISEN?

This looks like a good path of investigation. The chip has OVP and OCP comparators - if noise gets in either of them it can trip early like you're seeing. You can also try simplifying the feedback network down to two resistors to see if that's a path for noise too.
 

Offline KristoffTopic starter

  • Newbie
  • Posts: 6
  • Country: us
Re: LM25085: Vout dropping out at higher Vin
« Reply #13 on: August 29, 2020, 07:31:59 pm »
I built another board and only populated the 5V supply circuit.  I also simplified the feedback circuit (did not load R15, C28, and C29).  I removed the fuse and bridge rectifier and I'm providing VS from a benchtop supply but still seeing the same behavior.

I reduced R14 to 0.05 Ohms but didn't see any change in behavior.

One new thing I noticed is that as I increase VS a volt at a time I can hear a slight buzz/whine when VS gets close the the point where the output cuts out.  I pressed on caps and think the noise is coming from C15.  I replaced the 1nF cap with a 10nF but still no change in behavior.

I then replaced Cout (C30) with 3x 10uF 10V caps in parallel.  The previous Cout was 22uF 6.3V.  Still no change in behavior.

Finally I halved the value of R17 to 1.66K.  This resulted in the expected output voltage around ~8.8V.  I was also able to increase VS beyond the previous point of failure. 
With Vout =5V and no load the output was stable to Vin ~19V. 
With Vout = ~8.8V and no load the output as stable at least to 25V (this is where my benchtop supply maxed out).  However when generating VS from a 24 VAC supply with the bridge rectifier to get a VS of ~36VDC the output voltage dropped to ~2V.

Any suggestions on how to interpret these findings and what to try next?

Thanks!


 

Offline Ton

  • Regular Contributor
  • *
  • Posts: 53
  • Country: dk
Re: LM25085: Vout dropping out at higher Vin
« Reply #14 on: September 02, 2020, 06:21:06 pm »
Hi all

I once had a nice fight with the same SMPS controller, in our case it was the switching noise coupling to the current measurement.

and as your scope captures show then you have a quite a lot of ringing in your circuit (depending on how you have used your oscilloscope probes.)

For us it was solved by increasing the Rsen (R14/R20) from around 20mOhm to 82mOhm (and adjusting Radj (R13/R19)) for a better current value signal, and adding a RC snubber on the junction between the Mosfet drain(T1/T2), coil (L1/L2) and diode Cathode (D6/D9) to GND, the snubber values has to be calculated based on the ringing frequency.

you could also try to lover the dV/dt of the switching node by inserting a smal resistor (around 2-4 ohm) in the gate signal. in my case there was no room for a gate resistor, so we had to fix it with a snubber.

I think in my case we ended up using a 30 ohm resister to gnd, in series with a 470pF to the switching point, this was with a ringing frequency of 124Mhz, the ringing frequency wil depend on componnts and layout.

it ended up running pretty well in the end.
« Last Edit: September 02, 2020, 06:22:58 pm by Ton »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf