Maker Pro
Maker Pro

Running multiple connections on circuit board ... good or bad?

H

Harry Muscle

Jan 1, 1970
0
Is running multiple connections from one point to another a good or bad
thing to do on a circuit board. I'm working on a layout of a simple
temperature control circuit and I was planning on having the ground run in a
circle, such that almost each point on the circuit would have at least two
different routes back to the ground connection. Is this a good thing to do?
The more routes the better? Or is there a limit? Do the same rules apply
to non ground connections?

Thanks,
Harry
 
J

John Popelish

Jan 1, 1970
0
Harry said:
Is running multiple connections from one point to another a good or bad
thing to do on a circuit board. I'm working on a layout of a simple
temperature control circuit and I was planning on having the ground run in a
circle, such that almost each point on the circuit would have at least two
different routes back to the ground connection. Is this a good thing to do?
The more routes the better? Or is there a limit? Do the same rules apply
to non ground connections?

This does not normally cause problems, and the full extension of the
idea is to have a ground plane. The exception to this involves either
big or fast currents that will cause voltage drop on their ground
lines or two or more points that need very accurate ground reference
voltages. In the first case, you use a separate ground trace to carry
the current, to keep that drop out of the rest of the grounds,
connecting at a point that does not carry current. In the second
case, you connect those nodes together separately from the rest of the
grounds, connecting them at the quietest point in the ground system.
 
H

Harry Muscle

Jan 1, 1970
0
John Popelish said:
This does not normally cause problems, and the full extension of the
idea is to have a ground plane. The exception to this involves either
big or fast currents that will cause vol drop on their ground
lines or two or more points that need very accurate ground reference
vol . In the first case, you use a se te ground trace to carry
the current, to keep that drop out of the rest of the grounds,
connecting at a point that does not carry current. In the second
case, you connect those nodes together se tely from the rest of the
grounds, connecting them at the quietest point in the ground system.

I'm guessing a relay and a fan would fall under the "big current" category.
What about a reservoir capacitor (used to smooth out the DC) of 1000uF?
Should I give it a se te route back to ground?

Thanks,
Harry
 
J

John Popelish

Jan 1, 1970
0
Harry said:
I'm guessing a relay and a fan would fall under the "big current" category.

They are certainly in the category of noise makers rather than noise
sensitive components. I would give their currents as separate a path
as possible, including cancellation of those current's magnetic
fields, by folding the current path into a parallel pair.
What about a reservoir capacitor (used to smooth out the DC) of 1000uF?
Should I give it a se te route back to ground?

The currents between capacitor and steady loads is normally part of
the quiet stuff, while the rectifier currents that charge the
capacitors ar part of the noise producing stuff. Treat the rectifier
currents like you do motor and relay currents. Run them to the
capacitors on separate traces than the grid that connects the
capacitor to its loads so that voltage drops caused by the charging
pulses do not get added to the capacitor voltage as seen by its loads.
 
H

Harry Muscle

Jan 1, 1970
0
John Popelish said:
category.

They are certainly in the category of noise makers rather than noise
sensitive components. I would give their currents as se te a path
as possible, including cancellation of those current's magnetic
fields, by folding the current path into a llel pair.

You mean by running the + and - lines beside each other, correct?
The currents between capacitor and steady loads is normally part of
the quiet stuff, while the rectifier currents that charge the
capacitors ar part of the noise producing stuff. Treat the rectifier
currents like you do motor and relay currents. Run them to the
capacitors on se te traces than the grid that connects the
capacitor to its loads so that vol drops caused by the charging
pulses do not get added to the capacitor vol as seen by its loads.

Got a quick question about running the connection to my caps. Which of the
following two is better:

RECTIFIER------CAP----VOLT REGULATOR-----LOAD

or

CAP----------|------------VOLT REGULATOR-----LOAD
|
RECTIFIER

The bottom option would provide seperate leads to the reservoir caps, but
I'm not sure if that's the proper way of doing it?

Thanks,
Harry
 
J

John Popelish

Jan 1, 1970
0
Harry said:
You mean by running the + and - lines beside each other, correct?

Yes. Run a pair of traces from the switching point with a bypass
capacitor across those rails at the switching device, with the catch
diode close to the switch.
Got a quick question about running the connection to my caps. Which of the
following two is better:

RECTIFIER------CAP----VOLT REGULATOR-----LOAD

or

CAP----------|------------VOLT REGULATOR-----LOAD
|
RECTIFIER

The bottom option would provide seperate leads to the reservoir caps, but
I'm not sure if that's the proper way of doing it?
I like the first one better, because it separates the big charging
pulses from the steady load currents to the regulator. If you show
all the traces, you end up with the rectifier having two lines to the
capacitor, which is right at the input of the regulator, a second
(smaller) cap right at the output of the regulator, and the common
ground node for these two caps as the tie in point of the regulator to
the circuits that need clean, regulated voltage. That way, the big
charging pulses from the rectifiers cause drops on traces that are not
between any devices that have to agree on what zero volts is.
 
A

Addie

Jan 1, 1970
0
John Popelish said:
Yes. Run a pair of traces from the switching point with a bypass
capacitor across those rails at the switching device, with the catch
diode close to the switch.

I was just reading this out of interest, and I am a bit mistified about
this 2nd option. What purpose would a capacitor have before
a rectifier. Surely the capacitor *must* be after the rectifier to make
any sense, and not just a matter of someone's "personal preference" ?
 
J

John Popelish

Jan 1, 1970
0
Addie said:
I was just reading this out of interest, and I am a bit mistified about
this 2nd option. What purpose would a capacitor have before
a rectifier. Surely the capacitor *must* be after the rectifier to make
any sense, and not just a matter of someone's "personal preference" ?

The point is that the rectifiers, capacitor and regulator share a
couple nodes, but there are several ways to run traces to get them
connected together. If traces had no resistance or inductance, any
way would work the same as any other way.
 
C

Chan

Jan 1, 1970
0
John Popelish said:
The point is that the rectifiers, capacitor and regulator share a
couple nodes, but there are several ways to run traces to get them
connected together. If traces had no resistance or inductance, any
way would work the same as any other way.

o.k. seems like I'm out of my depth on this one. Still remember my power
supply circuit ? I wanted to draw the component names as you suggested
to make the discussion easier, but I haven't done it yet. I didn't quite
understand why we need to guard against oscillation. Is this oscillation
picked up from the air or mains, or is it self generating and can be
calculated. The other thing I/you was not sure about was whether the
resistors near the emitter follower transistors where drawn correctly /
or had correct values. Do you think it is worth while to simply measure
some voltage drops across components for varying load voltages/currents ?
 
J

John Popelish

Jan 1, 1970
0
Chan said:
o.k. seems like I'm out of my depth on this one. Still remember my power
supply circuit ? I wanted to draw the component names as you suggested
to make the discussion easier, but I haven't done it yet. I didn't quite
understand why we need to guard against oscillation. Is this oscillation
picked up from the air or mains, or is it self generating and can be
calculated. The other thing I/you was not sure about was whether the
resistors near the emitter follower transistors where drawn correctly /
or had correct values. Do you think it is worth while to simply measure
some voltage drops across components for varying load voltages/currents ?

I don't recognize your name. Are you talking about this post?
http://groups.google.com/[email protected]&output=gplain

Any time you have gain greater than 1 and feedback, there is a
possibility of oscillations (regenerated echoes) being generated.
Before going further than that, I need to know what you are talking
about.
 
J

John Popelish

Jan 1, 1970
0
Chan said:
(snip)http://groups.google.com/[email protected]&output=gplain

sorry, I forgot to mention which posting, you have guest right.
I will post the circuit again. I think the capacitors around the
op amp are simply standard practice to avoid oscillation, and
are not something in particular for a power supply.

Perhaps, but a power supply regulator is a big, power amplifier with
lots of feedback, whose job it is to hold a voltage stable, or switch
to smoothly and accurately regulated current in the event of an
overload, in spite of all variations in load from any value of
resistor between infinity and zero, inductors, to capacitors, (or
resonant loads made of inductors and capacitors) to pulsing circuits.
Their stability is quite a design problem, especially if there is a
fast spec on how quickly the voltage must get back to setpoint after a
big load current spike.
 
C

Chan

Jan 1, 1970
0
John Popelish said:
(snip)http://groups.google.com/[email protected]&outp
ut=gplain

Perhaps, but a power supply regulator is a big, power amplifier with
lots of feedback, whose job it is to hold a voltage stable, or switch
to smoothly and accurately regulated current in the event of an
overload, in spite of all variations in load from any value of
resistor between infinity and zero, inductors, to capacitors, (or
resonant loads made of inductors and capacitors) to pulsing circuits.
Their stability is quite a design problem, especially if there is a
fast spec on how quickly the voltage must get back to setpoint after a
big load current spike.
You may be right. I have a book (how to use op amps) which says
on p.9 that it is very unusual for the 741 to "misbehave", so perhaps
the 10nF from negative op amp input to op amp output is to remove
any ac component from the controlling feedback loop.

Zc = 1/jwc = 1/(6.28*10n*f)

I believe you chose f = 20KHz (why?) which gives us
Zc = 8 ohms
Hence Zc << than transistor feedback loop which is at least
hie of 3 emitter followers.
 
Top