LM5122: huge (not desired )current

Hi Vern,

Can you share the parameters of your design? VIN range, VOUT, IOUT, switching frequency. Values can be checked by following the equations found in the datasheet.

-Garrett

glad you answered Garrett. I was asking that designer how he derived 5.6uF.
I wrote I am designing LM5122 24v in 45v out 4A interleaved, 2A each.
I can change my frequency. webench will not allow my freq choice, or keeps changing it, result is 33uH to 83uH. depends on my Iout.
using LM5122 eq 26 pg 37, and actual switching freq = 500khz, my inductor = 24.3 uH.
so his 5.6 uH is too low.
to get actual switching freq of 500khz, I must set LM5122 master to 1000 khz, right?
webench won't allow 500khz.
while writing this, just tried lm5122 again, now it allows 500 khz! previously I used 2.2A and 2.5khz, and wbench changed to a lower freq. this time I started with 2A. perplexing.
maybe there was an update since Sunday?
problem still: webench selected 39 uH, and inductor range is 33.35 to 100.1. should be 24.3 using LM5122 eq 26 pg 37.
webench does not display inductor value, nor inductor peak amps, in Operating Values. can see it in BOM. please add inductor value and inductor peak amps in Operating Values.
the tps43061 webench uses freq better, results come closer to actual equations. so I also use it for my LM5122.

Hi Vern,

I'm sorry to hear to you are having issues with WEBENCH.

Following Equation 26 of the datasheet, the calculated inductance is going to depend on the output current, the switching frequency, the duty cycle and the desired ripple ratio. Typically the nominal input voltage can be used to come up with a reasonable inductance value.
Assuming [VIN = 24V, IIN = 48V*2A/24V = 4A, RR = .30, Fsw = 500kHz] is going to yield 20uH. If RR is 20% then the inductance is going to be ~30uH. What is the minimum input voltage? What is the maximum input voltage? WEBENCH is most likely trying to optimize efficiency the design based on the available inductors in the library. This might be why the numbers are different.

Regarding the switching frequency please refer to figure 31 of the datasheet. Typically the master phase should be configured as MASTER 1 mode (OPT = GND, FB = Resistor divider) and the slave phase should be in SLAVE 1 mode (OPT = GND, FB = VCC)

Thanks,

Garrett

hi Garrett and Gerold, having a blessed day?
ok, nice extra info. my input is fixed at 24v.
yes, I have MASTER 1 mode (OPT = GND, FB = Resistor divider) and the slave phase should be in SLAVE 1 mode (OPT = GND, FB = VCC)
to clarify, to get actual switching freq of 500khz, I must set LM5122 master to 1000 khz, right?
with Gerold's reply, the webench high inductor values are explained further--webench adds 25%. please add a note for this. you cannot imagine the grief and wasted time I spent trying to get webench to agree to your equations. sigh.
pretty much, all eval boards and pdf datasheets do not add 25% to inductor value. sigh again.
I don't want to add 25%, I'll deal with tolerances with my parts. example, hand picking inductors.
by webench adding 25%, nothing works, no solution.
creating a custom inductor has software errors, the allowed range is pre-set, say 47-83 uH, precludes putting in 22uF, for example.
also, a custom inductor can not be deleted. I finally found a way to modify by changing the limits of search. that is odd.
I did start a fresh new LM5122 webench, but somehow it retained my other webench custom inductor. so that was a dead end too.
all this is valuable webench feedback, sure need webench improvements asap.

thanks, nice details. see my last reply.
that designer used 5.6uH, must have not used webench due to its inductor issues.
doesn't 5.6uH seem way too low to you?

Hi Vern,

To find inductors that are not in the recommended range, you can try clicking on the pencil symbol in the select alternate parts window-> changing the lowerbound value and then clicking search. This should give the inductors in the lower range that is expected. Again, when you are outside the recommended range, the stability and other performance parameters of the design has to be thoroughly verified by you using simulations to to have confidence that the design works as expected. But the tool will give you flexibility to pick other components in our database if you decide to do so instead of creating custom components for each design.

I am not sure about the 5.6uH in the other customer design - if it was an intentional change by the customer or not. 

Regards,

Gerold

hi, if you don't mind continuing with these inductor values. please.
I do get a warning about instability.
once I pick my custom 22uH inductor, no other parts ever show.
but after seeing many software issues with webench, I have no faith.
I will have to look into instability, but if not webench, what to use?
so far, we have these inductor values for very similar designs ( I went thru every TI eval kit):
1. 1 customer uses 5.6uH, does not report any instability.
2. Coilcraft has at least 4 solutions, and much less power loss than webench, same conditions. therefore I consider this a MAJOR webench software error. these coils never show under any input values close to mine. my Coilcraft coils can handle the 4A without satuation.
3. ANY inductor < 33 shows instability warning.

YET, YET, YET....
4. LM5122 pdf uses a 10.6uH at less frequency than mine. how can that be? is it unstable, too?
5. LM5122 -2PH eval, I bought, very close to mine, uses 10 uH how can that be? is it unstable, too?

look, I'm at wits end. pulled my hair out 4 months ago, has not grown back.
please focus on this. I was top top Senior FAE at Atmel, I know what this situation is.
you need to escalate webench fixing to top marketing manager. and now.

Hi Vern,

    Would you be willing to have a call to discuss these open issues? It might be quicker to view the results that you are seeing. Please send me a note at

g-dhanabalan@ti.com

Regards,

Gerold