Change Pot in Mini Buck Converter

Status
Not open for further replies.
ThomsCircuit said:
I thank you for the formula and the info. It is evident that ceramic is what is needed here. ive updated the project to 0805 style SMDs
Click to expand...
All components have problems. Different components have different problems.

High capacitance ceramic capacitors have a problem that their capacitance reduces with the voltage. It is worth checking the data sheet to see how much they reduce. I suspect that 22 uF 35 V 0805 capacitors will have much less than 22 uF if 35 V is applied.
 
I went ahead and changed the size of the vias. Based on a few videos i watched they are not to touch. I have a few with spoke errors. LMK if i need to have spokes on these vias. These are a first for me.
 
VIAs, that will have no component soldered in, should not have spoke or thermal clearance, they should be in solid copper! That PCB program is weird...

I'd be worried about changing to tiny tracks on high current circuits, such as the main power capacitors!

You need to think of every PCB trace as adding unwanted resistance and inductance, and always keep them short and wide, where any power at all is involved.

Also, all cap grounds need via arrays to actually link to the ground plane, as there are now no through holes..
 
rjenkinsgb said:
VIAs, that will have no component soldered in, should not have spoke or thermal clearance, they should be in solid copper! That PCB program is weird...
Click to expand...
DS does not appear to have an auto via stitching feature. There are a few options to control the behavior though. I set the vias plane connection to non isolated and that seems to have achieved what we are after.

I have beefed up all the tracks. I have done what I can to shorten their lengths while following guidelines.
Just so I understand correctly. The four tabs I highlighted in white also need vias. If you could use this photo to show where i should place them.

p.s. Updated the caps to 1210 style as per design specs.
I thought that the caps size was a way to show its voltage rating. While that may be the case for tiny un-marked caps these 1210 style come in a wide range of voltage ratings. While Im using the caps as defined in the schematic, I cannot locate 35v versions of 22u in a 1210 case. So is it OK to use a smaller 1206 case 22u 35v or do these need to be 1210?
 
Vias around Cap ground pads. This would be my best guess. LMK if they are in the correct place. There are 4 spokes from the pour.

With pour. Are they too close to the pad?
 
Last edited:
rjenkinsgb said:
Also, all cap grounds need via arrays to actually link to the ground plane, as there are now no through holes..
Click to expand...
I take it that I have done this correctly.
I have looked. DS does not do via stitching, shielding, or any other automated process where vias are added in groups.
I have done my best to understand the guidelines but im probably more puzzled now. Meaning I know that they connect the top pour to the bottom pour and that is good. But not a whole lot on their size, hole size, spacing from each other or the pad. Mine are touching, 30 width, 15 hole. In some cases the pad is very close to the via. Just want to be sure this is ok. Thank you so much for your help. I will be able to make 10 of these for $22.00
 
ThomsCircuit said:
Vias around Cap ground pads. This would be my best guess. LMK if they are in the correct place. There are 4 spokes from the pour.

With pour. Are they too close to the pad?
Click to expand...

If your using all SMD devices, why are you using vias at all?
Use multiple planes with wide spokes to make device connections.

Similar to this (LM2673 Pos to Negative 3A regulator:



The bottom side is a single copper ground plane
 
Last edited:
eTech said:

If your using all SMD devices, why are you using vias at all?
Click to expand...
I believe these are Stitching vias. technique used to tie together larger copper areas on different layers, in effect creating a strong vertical connection through the board structure, helping maintain a low impedance and short return loops.
This is the authors PCB. while the vias that are centered on the pads of caps of c2 and c5 are to connect the SMD caps c1, c7 that are on the opposite side of the board there are some vias randomly placed on the board. I believe those are to help reduce noise. Im concerned with the number needed and their distance from each other and the pad they should be. I cant seem to find a clear answer. They are all on the ground plane.
 
eTech said:
Use multiple planes with wide spokes to make device connections.
Click to expand...
The reason it was suggested ,
"...Also, all cap grounds need via arrays to actually link to the ground plane, as there are now no through holes.."

Since my ground plane is on top i dont really need them?
Do I need a bottom ground pour? If so how to i connect them? (top n bottom pours)
 


Every time a via is used, its the same as creating a tiny resistor between layers.
I don't see any benefit, for this design, from the vias if thru holes aren't used in the first place.
A single ground plane on bottom for shielding, and limited sized ground plane on top for connections would have sufficed, in my opinion, and planes and wide spokes on top would surely be enough for the current capacity requirements.

But its up to you....
 
ThomsCircuit said:
Since my ground plane is on top i dont really need them?
Click to expand...
It's only a functional "ground plane" (or power plane) if it is near enough continuous, so it provides good connectivity everywhere.

Having one split up by many tracks is not much use, unless it's bonded to another continuous plane.
 
You need a ground plane on bottom for help reduce noise.
Use the external wiring pads to make the connection between ground planes.
 
ThomsCircuit said:
Are you saying the pads should touch so there are no tracks?
Click to expand...
No; just that a "plane" on the same layer as all the other tracks will have gaps all over the place that dratically reduce its effect - it's only joined in odd places.

The other side can be continuous, so that's the more important one to give a low impedance ground path. The partial one on the track layer can be "boosted up" by plenty of through VIAs to the full ground on the other side.
 
rjenkinsgb said:
No; just that a "plane" on the same layer as all the other tracks will have gaps all over the place that dratically reduce its effect - it's only joined in odd places.
Click to expand...
Ohhhhhhh. I get it. (dont happen often)
Isolated grounds from the pour. I see. you know i do spend time to make sure there are as few of them as possible. I nudge components and tracks to allow the pour to be continuous. I just thought it was a good thing to do. OK. I get it.
 
Status
Not open for further replies.

Similar threads

A
  • Solved
Buck/Boost Converter
Replies
7
Views
2K
Tony Stewart
J
  • Question
Removing gel pot for lamp body repair
Replies
10
Views
1K
rjenkinsgb
M
  • Locked
  • Question
Attempting To Design A Lower Power Version Of The Arduino Pro Mini
Replies
1
Views
409
JimB
  • Question
Buck converter with transformer output
Replies
14
Views
3K
throbscottle
A
  • Question
12V LM317 variable power supply, pot got burned.
Replies
19
Views
4K
Nigel Goodwin
Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…