5v plane and 20h-rule
joe_ls , 05-03-2018, 01:09 PM
Hi
During the ISL6235A power supply project I created the 5V plane as it was mentioned in the course.
This +5V plane is on L4, two ground layers are on L2 and L5. The whole stack-up is made with 6 layer.
After finishing the project I still have some questions:
1) The +5V plane on L4 is way smaller than the outline of the board. Is the reason for this design decision based on the "20H Rule"?
2) In my layout I created a via fence on the edge of the board, connecting both GND-planes together. I don't know how much noise there is in general in this design, but do you agree the via fence could possibly improve the shielding and reduce noise? Or is this useless?
3) The green boundary is +Vin on L1 (TOP). Quite in the middle the copper is broken via vias several times. Robert made quite a fuss ;-) to measure the current rating at other locations, but not this one. I changed the shape and made it wider around the vias to enhance the current rating. Good, bad or ugly? What do you think?
All comments are most welcome.
Thank you!
Joe
During the ISL6235A power supply project I created the 5V plane as it was mentioned in the course.
This +5V plane is on L4, two ground layers are on L2 and L5. The whole stack-up is made with 6 layer.
After finishing the project I still have some questions:
1) The +5V plane on L4 is way smaller than the outline of the board. Is the reason for this design decision based on the "20H Rule"?
2) In my layout I created a via fence on the edge of the board, connecting both GND-planes together. I don't know how much noise there is in general in this design, but do you agree the via fence could possibly improve the shielding and reduce noise? Or is this useless?
3) The green boundary is +Vin on L1 (TOP). Quite in the middle the copper is broken via vias several times. Robert made quite a fuss ;-) to measure the current rating at other locations, but not this one. I changed the shape and made it wider around the vias to enhance the current rating. Good, bad or ugly? What do you think?
All comments are most welcome.
Thank you!
Joe
robertferanec , 05-04-2018, 10:07 AM
1) No real reason why it is so small. The currents on this plane are small, so I only made it big enough to cover all the +5V VIAs and track. I also kept in mind, that GND can be noisy, so maybe smaller power plane may be less sensitive to pick up the noise. But, that is only my feeling, if you like, you can make it bigger, but I would not do it as big as the GND planes because of the 20H rule.
2) I read about VIA fence. I have done this fence on some of my boards, but have not see any difference comparing to boards without fence (both kind of boards pass EMC/EMI oki). I need to say, I normally design digital boards, the fence can make differences in different kind of boards e.g. RF.
3) If this is the one in the middle, there will be only small currents flowing on that part of plane. Feel free to adjust the shape and use any shape you like
2) I read about VIA fence. I have done this fence on some of my boards, but have not see any difference comparing to boards without fence (both kind of boards pass EMC/EMI oki). I need to say, I normally design digital boards, the fence can make differences in different kind of boards e.g. RF.
3) If this is the one in the middle, there will be only small currents flowing on that part of plane. Feel free to adjust the shape and use any shape you like
joe_ls , 05-04-2018, 11:11 AM
Hi Robert
Thank you for clarification.
Are there some documents or books you recommend about VIA fencing?
Cheers
Joe
Thank you for clarification.
Are there some documents or books you recommend about VIA fencing?
Cheers
Joe
robertferanec , 05-06-2018, 06:24 PM
I have not read anything special. Just some articles on Internet.
I have not found anything special about fencing in design guides, so I have not really searched much about the topic.
I have not found anything special about fencing in design guides, so I have not really searched much about the topic.
joe_ls , 05-07-2018, 06:26 AM
Found a paper called "Effectiveness of PCB Perimeter Via Fencing":
[...] The rule of thumb that a stub starts to become reactive as it
approached 1/8th of a wavelength has been known and followed
since the early 70’s [5]. This same principle shows that the
space between vias also starts to look reactive as the spacing
approaches 1/8th of a wavelength. [...]
But It doesn't look like there is an easy answer when to use VIA fencing. It is effective for some frequencies, but not all. It does help but not always. There is no easy rule of thumb to follow.
[...] The rule of thumb that a stub starts to become reactive as it
approached 1/8th of a wavelength has been known and followed
since the early 70’s [5]. This same principle shows that the
space between vias also starts to look reactive as the spacing
approaches 1/8th of a wavelength. [...]
But It doesn't look like there is an easy answer when to use VIA fencing. It is effective for some frequencies, but not all. It does help but not always. There is no easy rule of thumb to follow.
Use our interactive Discord forum to reply or ask new questions.