TPS564247: Layout check for tps564247 buck converter + ground layout

Hi,

we will review the layout

Regards,

Manish

hello,

thanks for your interest. i will be waiting, i want to start routing so it is assuring if you can check the layout first.

biggest difference between the 2 layouts is the caps direction.

Hi,

If possible, can you please share the .BRD or .PCB file of the layout.

Regards,

Manish

hello

i still didn't start routing, just components placement. I just need you to verify component placement for now.

as for routing, i will route power input and output using polygons and put many vias to ground as any other pro design. the caps being face to face is not an issue but i thought of asking. please check the images i posted to see the direction of current as well as capacitors + and - terminals direction.

I use kicad too, so i do not think it supports these formats, and not necessary as i mentioned here.

hello

so for layout 1, the big output capacitor is ok with the smaller parallel caps? 

I mean their positive sides are facing together but their grounds are on the opposite sides. 

there will be lots of ground vias for all of them (both sides) so that current can be bypassed to internal 3 ground plains.

i will update you with the layout once i finish it. but the loops will be as small as possible. check the image and see if the loops need anything else.

thanks 

Manish Gautam 

i made the layout for the right side regulator: https://imgur.com/a/z8BS2z3

please check. more refinements will happen but the placement and initial routing with loops should be clear now.

what do you think?

Hi,

The capacitor placement is okay in layout1.

Regards,

Manish

hello

thanks for your feedback

between layout1 and layout2, which is best? 

for layout1, any tip or addition or modification you can mention in order to enhance it?

thanks 

hello,

ok i will follow layout1 and guidelines.

however, for input caps, i use about 4x 10u 0805 MLCC caps with effective capacitance of about 6 uF total at input voltage of 15v. 

now there are huge polymer caps nearby the right-side regulator which can be considered as bulk caps but they are not close to left-side regulator, around 20mm of distance. 

should I add more local bulk caps to both regulators? 

the space is very small and all I can add is either another MLCC which will boost the total effective capacitance to 7.5uF or this tantalum polymer cap from panasonic: https://industrial.panasonic.com/sa/products/pt/POSCAP/models/25TQC22MYFB

it has about 100mOhm ESR at 100Khz (the figure shows it is a bit less), which is not very small but also not high. will adding this very close to the ceramic input caps make a difference? also, after the big EMI input filter, will this even be necessary?

maybe i forgot to mention but i have a simple pi filter at the input of both regulators... it has 220uF -> 10uH -> (220uF || 220uF). inductor is about 10-15mohm dcr and caps are 20-30 mohm esr polymers. I usually have just the 3 polymer caps tho.

I mainly added this to filter out any low frequency ripple generated by the input which is a meanwell power module (15v) to make it easier for the switchers to produce low ripple.

is this ok? will cause any issue?

the local filtering for the switchers is 4x10uF 0805 ceramics with 100nF 0402... will produce effective capacitance of about 6uF. I want to use 22uF instead which will give effective capacitance of around 19-20uF using 6x 22uF 0805 ceramics instead.

please check this

Hello,

I already followed the datasheet suggestion and used approved values which are within the range for output capacitor.

However, for input capacitors the datasheet says a minimum of 10uF with small 100nF. I have more than this and suggested adding more to be of total 19-20uF effective capacitance instead of just 6uF like before.

The input ripple is about 200 mv p-p which I think is too much. I want to minimize it by increasing the input caps to 6x 22u 0805 (effective 19-20uF) as well as having the big C-L-C filter which consists of 220uF + 10uH + 2x220uF polymer caps with inductor.

I want to know if this input filter is ok to ensure very low input ripple in order for the output ripple to be small as well. the type of input ripple i had was low frequency which is why i had the 10uH inductor with the big polymer caps to filter it. 

please advise in details

thanks!

Hi,

Please share your schematic file.

Regards,

Manish

hello

i will send it once i return from work but basically it is this:

source is 15v power module. 

after 15v it gets into pi filter which is 220uF capacitor -> 10uH inductor -> 2x220uF capacitors.

the output of the pi filter is the input for both switchers... which gets into local input caps of each switcher which will be 6x 22uF 0805 ceramics || 100nF 0402 cap.

Pi filter caps are 220uF polymers with 20-30 mOhms ESR. I worry about CLC filter to need damping which i do not want to do, i can replace one of the 6x 22uF caps with the panasonic poscap tantalum mentioned above which has relatively high esr of 80-100 mOhms to aid in damping and would also increase local input capacitance. wanna hear your opinion about this.

the output filters are fine, no need to worry about them.

Hi,

Please share your complete schematic and BOM details, after that I can be able to comment on the damping.

Regards,

Manish

Hello

this is the schematic: https://imgur.com/a/6lByV6l  (sorry, can't post the entire thing here but this is enough for our needs)

BOM:

Inductor 10uH: https://lcsc.com/product-detail/power-inductors_tai-tech-tmpa0605s-100mn-d_C2880077.html

220uF polymer caps: https://lcsc.com/product-detail/Solid-Capacitors_AISHI-Aihua-Group-SPZ1EM221F11O00RAXXX_C160181.html

22uF 0805 ceramic caps (x6 of it in parallel): https://product.samsungsem.com/mlcc/CL21A226MAYNNN.do

one of the 22uF ceramics can be replaced with this: https://industrial.panasonic.com/ww/products/pt/poscap/models/25TQC15MYFB which has 100 mOhm ESR and is solid tantalum polymer.

the rest are ok with me.

for your kind help

Hi,

Currently I am simulation that PI filter and 6*22uf capacitor with 15V input voltage. 

Regards,

Manish

Hello

thank you for your extensive effort.

I noticed you added external resistors of 60m and 100m, are you doing this to mimic the ESR of the caps?

if so, then the 100m is not correct in here since all 22uF are ceramics with very low ESR. By 100m I meant i can replace one 22uF ceramic with that Panasonic POSCAP which is 15uF and 100m ESR.

So, if I kept the 22u or had the POSCAP in its place, will it be nice and stable? Do you recommend I proceed?

The original power module which is the source of 15v I guess its switching frequency is about 130 khz or a bit less (not verified), maybe this can affect the emi filter?

I am interested at the final ripple you had in your simulation, what was it? assuming 200mv p-p of the original.and as for the emi filter, will it negatively affect the final output? i mean stuff like the inductor slowing down the response...etc thank you again!!

Hello,

I tried modifying the circuit to be more realistic by adding the ESR in series with caps + compensated for ceramics derating...etc. I made 2 versions: one with regular 22uF and the other with POSCAP as seen here:

looks like the damping should be ok despite the software does not really show much information so that I cannot verify if it is gonna be like real circuit.

however, i tried modifying the source to mimic 200mv p-p ripple by using pulse or sine but the result wasn't so good. it showed the ripple going to the output.

can you try on your own since you know better to use this software? here is the file after my mods: 

Hi,

I am not able to open this link, due to restriction.

Regards,

Manish

emi filter files.zip

Hello,

I added the link as an attachment here. please check.

thanks 

Hello,

well, I have different measurements but one engineer tried it in his setup and was 200mv while I tested a similar one which was a lot less.

the official datasheet from Meanwell says it is a maximum of 200mV with 0.1u and 47u added capacitors... while their practical test report says it is around 47mV only.

aside from that, I wanted this big filter to eliminate or minimize the low frequency ripple as much as possible so that it does not proceed to the output.

I believe I will build the circuit and use the POSCAP 15uF 100mOhm ESR capacitor which should help adding local bulk capacitance, reliability, and with this ESR can help with damping too.

it is rated for 900mA ripple current but I do not think it would reach that level with all this filtering + paralleling it with 5 ceramics.

In your opinion, will this input inductor negatively affect the transient response? especially if the target device sips power at 50 or 60 Hz rate. 

I mean after the input L there are tons of bulk and parallel caps which should compensate for inductor slowness, but it is nice to see your opinion.

thanks!

Hello,

I know so, but i asked about other things, namely:

1- filtering effect on low frequency ripple.

2- does it have negative effect on total output transient response and will it cause more dipping due to the slowness nature of inductor. We have lots of caps after inductor, thus i guess should be ok.

what is your opinion on this?

thanks