I have to layout this basic power supply circuit in OrCAD layout. Ive got to meet certain requirements such as the ground trace has to be pretty thick, and as many traces as possible have to be on the bottom side of a dual layered board. What is the best way for me to go about placing the parts so a minimum number of traces run on the top? Are there any good strategies for it, or should I keep fooling around with it (like I have been for the past few hours.)
Question for EE's good at board layout
- Thread starter jmcoreymv
- Start date
Well it does have auto place and auto route feature. But Im not sure how to program in certain requirements (such as input and output headers should be on opposite sides of the board), and when it does the auto routing, it doesnt seem to try and place a majority of traces on the bottom.
Ive been trying to lay it out manually for several hours, but theres like an endless number of different possibilities to try and theres got to be some sort of strategy.
Also when I do a DRC in layout, it seems to have a lot of problems with my 100 mil trace. There are like 60 problem circles around it. Any ideas why?
Analog
Lifer
- Jan 7, 2002
- 12,755
- 3
- 0
I hope your not a student in one of my classes
So from what I understand, you have a double sided board, with no internal ground/power planes?
Is this a switching design? That makes it more difficult, as you have to consider transients and high frequency switching noise in the design.
OrCAD is not foolproof, you must place the parts so that the pins give good straight shots to where they need to go. Place parts next to each other that have a lot of interconnects. Orient the parts (rotate etc) to minimize angles on traces, and again, make them as short as possible. Make very sure that high frequency parts have short traces, and are not near other parts that may be affected by EMI, etc (i.e. amplifiers, A/D converters etc).
You may even employ ground line shielding adjacent to thick power lines to minimize radiated noise. Most of my designs employ a minimum of 4 layers, since (in my classes) I teach the concept of ground / power plane shielding, even on simple designs. You also did not mention whether this is a through-hold or surface mount design. Via reduction in OrCAD is a sure fire way to create convoluted routes. HTH.
Its a thru hole design. Can only be double sided (since I will soon be going through the steps to actually etch the board). This is a freshman level class, and we arent expected to know anything about circuit components or interference or anything like that. I dont know if its a switching power supply, how would I be able to tell that?
It has an 2 LM317T's (its a dual variable output supply). It also has a LM7805 in the optional part of the circuit for digital meter readouts. It looks like it has a rectifier bridge of 1n5401 diodes and various other resistors, capacitors and diodes.
I took care of the issues with the 100 mil trace. Problem was that I was doing a manual override of trace width rather than just setting the trace to 100 mil in the nets window. That got my errors down to 10. Now the only errors I have left are on some components, when I run a trace underneath it (on the bottom layer), it generates a yellow circle error in the middle of the component. Any idea why?
Umm... if you're laying out an analog design, do it by hand.
Analog circuitries are extremely picky about how and where you return the feedback signal. Even on a switching power supply, if you route the feedback voltage into the controller using a long trace, you'll get a bunch of noise at the output.
As for placing parts, minimize the loop area between parts. Keep the high current parts close to each other and the high current traces short, and return them to ground at a star point, preferably to the same point as your feedback ground.
Loop area = EMI and distortion/noise at the output.
Bad feedback ground = noise at the output.
Depending on how many amps your supply is, be sure to make your traces wide enough to accomodate the currents. Since it looks like you're doing a two-layer board, I would route power traces on top and ground traces DIRECTLY below.
Analog circuitries are extremely picky about how and where you return the feedback signal. Even on a switching power supply, if you route the feedback voltage into the controller using a long trace, you'll get a bunch of noise at the output.
As for placing parts, minimize the loop area between parts. Keep the high current parts close to each other and the high current traces short, and return them to ground at a star point, preferably to the same point as your feedback ground.
Loop area = EMI and distortion/noise at the output.
Bad feedback ground = noise at the output.
Depending on how many amps your supply is, be sure to make your traces wide enough to accomodate the currents. Since it looks like you're doing a two-layer board, I would route power traces on top and ground traces DIRECTLY below.
The teacher wants as many traces as possible on the bottom side of the board. The odd thing is, functionality isnt really important, we are going to be graded on the final project but not on whether or not it works. Just how well it was made.
The sad thing is I'm almost to my junior year of EE and this thread went STRAIGHT over my head....
Cal Poly SLO. Everything we learn we have a lab for. Thats why our motto is learn by doing i suppose. Kind of a pain in the ass.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter