I have a Cosel DC-DC converter. It's pretty straightforward, 2 inputs and 2 outputs. It is an isolated converter, and I was thinking it would probably be best if I put a ground plane beneath it on the PCB, but which side should I ground connect the ground plane to? The input ground or the output ground?

Or would there be two ground planes half on the output side connected to the output ground and half on the input side connected to the input ground? Could something happen if the ground plane gets too close to the ground that it is not connected to?

EDIT: JUst noticed there is a case pin and the manual connects it to the input ground for shielding. I assume I would connect a ground plane to the input ground then to complement the case shield? I am still not sure how to pick the ground to connect it to though.

 

I would connect the PCB ground plane to the output ground, but it would probably make little difference, if any.

__________________
see my website: www.geocities.com/russlk

 

It was just I was told that isolated converter means that the output and input grounds don't have to be at the same potential so I was wondering which to have the ground plane at. Or does it not really matter as long as it's not floating?

 

Can you please post a schematic.

__________________
I also post at the following sites:
http://www.stop-microsoft.org http://www.heated-debates.com
Screen name: Aloone_Jonez
And http://www.silicontronics.com, same screen name as here.

 

really depends on what you are doing and why you need an isolated DC:DC converter

IF it is because the output floats to otehr potential's then DO NOT gnd secondary side GND plane!!!

Think of a DC:DC converter make two GND points
GND1 and GND2. GND1 can obviously be tied/reference to GND (there is a high chance the PSU you are driving the DC:DC with is earthy and thus earthing the primary side is neither here nor there.

However.. earthing the secondary side (even if it is just a ground plane) will produce an earth-loop. At best the earth-loop with be for HF stuff via the capacitance between secondary and the plane, at worse you remove the isolation (if you really did it BAD)


Personally What I do is have a plane on the primary side of the DC:DC and tie that to the 0V of the DC:DC (and thus normally earth), the secondary side I also make a plane but tie it to the 0V of the secondary side of the DC:DC with lots of capacitance across the SEC V-0V rails to firm it up

its very good

__________________
Nothing is impossible.
Once a problem is realised, the rest is just details

 

You can disable smilies by going into advanced editor and clicking on the "disable smilies in text" box.

So DC:DC converter appears as DC:DC converter and not DCC converter.

EDIT:
I've notice you've figured it out for yourself.

__________________
I also post at the following sites:
http://www.stop-microsoft.org http://www.heated-debates.com
Screen name: Aloone_Jonez
And http://www.silicontronics.com, same screen name as here.

 

I was considering tying the two ground planes together with a capacitor. Here's the DC-DC converter...not much to look at:

http://www.trcelectronics.com/Cosel/...ztw_manual.pdf

http://www.trcelectronics.com/Cosel/pdf/zus25.pdf

I actually don't need it to be isolated, it's just that the converter I have just happens to be isolated. I wonder if I can connect the two grounds together to force it to be unisolated...but I am not sure what is inside the converter and connecting grounds can cause...funny things to happen.

 

The whole point of being isolated is so you can do whatever you want, I'm somewhat bemused why you are using groundplanes though? - but there should be no problem joining either ONE of the output rails to either ONE of the input rails (I don't see how 'ground' comes into it?).

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

What should I be saying instead of ground? Input/output neutral? Input/Output 0V? Input/output reference? Because the one that makes sense is the one I mean to say.

I just wanted to use a ground plane to try and reduce some noise.

 

It's really dependent on exactly what it's refering to?, but 'ground' really refers to something which is 'grounded' - chassis would perhaps be better?, or 0V which doesn't imply a metal chassis.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

I'm still having a bit of trouble with ground vs neutral. I got the impression that when people say grounded they are actually referring to neutral, but I am not sure.

Neutral is like 0V, but grounded is connecting to an infinite potential (or infinite ground potential, but I am particularily certain about what this means exactly, the closeset thing I can remember was an effectively infinite supply of electrons) like the earth or a large metal frame? I do have a big metal frame and an enclosure sitting on the metal frame with the PCB inside. But I wasn't planning on connecting the ground planes to those. Should I do that rather than connect the ground plane to the input or output reference? And connect the converter case to the metal frame instead of the input reference (like the manual suggests?). I'm having a bit of trouble wrapping my head around this shielding stuff.

Do you know of any articles about ground planes? The closest thing I learned about a ground plane so far in class is "virtual charge" method for solving problems in my electromagnetics class, but nothing practical so far about using it. I haven't had much luck getting specifics with google.

 

Neutral != Earth/Ground

Likewise having an earth-plane be it a common-wording (and can be taken the wrong way) has alot of noise-imunity benefits

On hte PRIMARY side of the DC: DC converter having a plane for 0V is great and somewhere along the line that 0V is going to be tied to earth (or it should)

on the SECONDARY side of the DC: DC having a plane that is connected to the isolated 0V provides benefits as well JUST DO NOT TIE DOWNTO EARTH!!!


I recently design and got layed out a gateboard with a TRACO DC: DC converter on it - 6layer board

PRIMARY SIDE:
0Vprim
tracking
+12V
+12V
tracking
0V prim

with 0Vprim tied back to a DC: DC converter 270V->12V with earth connection

an Inch gap bridged by opto's and TRACO

SECONDARY SIDE
0Vsec
tracking
+15Vsec
-15Vsec
tracking
0Vsec

and around the edge of the board vias shorting the top-plane to the bottom plane in multiple locations to form a kind of faraday cage
this gateboard is the most robust and stable I have made

__________________
Nothing is impossible.
Once a problem is realised, the rest is just details

 

Why are you wanting to use ground planes anyway? - what is the complete project?.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

It's an aluminum framed robot, and the converter is for the logic power supply and will sit on a PCB with the other electronics inside an aluminum enclosure bolted to the frame. I was going to attach a the negative terminals of the main battery to the frame (as a dead end, not so current would flow through it in normal cases).

I need to shield the electronics from the noise produced from the motors which are very very large by comparison (and the DC converter as well which is also considerably larger than the rest of the electronics).

 

so you really don't need the isolating aspect of the DCDC converter?

__________________
Nothing is impossible.
Once a problem is realised, the rest is just details