Trace Impedance Calculator that Doesn't Assume Air???......

[Smokey]

I need to make a board with 120ohm differential traces (Edge Coupled Microstrip), but it also needs to be potted in epoxy.  It looks like all the impedance calculators assume that the dielectric above the board is air.
Are there any calculators that have the material above the board as a variable so I can set it to the value for the epoxy I plan on using and see what that does to the trace impedance?

Note: Yes, I know I can make myself a spreadsheet from the full equations, but I figured if that exists already it would be good to also make that known to other people where to find that easily.
... I also know I could probably make it an edge coupled stripline on internal layers and not worry about it, but I would have to add extra layers to the board to do that.

[rsjsouza]

Perhaps calculate as if it were a stripline? I have a vague recollection there is at least one calculator (from Saturn PCB, perhaps?) that allows using a split medium.

Otherwise, you might have some luck running a simulation on ADS, HFSS or even the freeware Qucs Studio.

[twospoons]

Use Saturn PCB toolkit, but treat your stripline as embedded. Your potting epoxy Er will be pretty close to your board Er, so the result should be pretty close to correct.

The other option is to use finite element software like FEM4.2, which should give you a more exact solution, but is much harder to iterate to get to 120 ohms.

[Benta]

The dielectric constant of FR4 is a bit over 4, and pure epoxy is in the same range (3...5).
I'd just go with that.

[Smokey]

Use Saturn PCB toolkit, but treat your stripline as embedded. Your potting epoxy Er will be pretty close to your board Er, so the result should be pretty close to correct.
...

I like it!  Duh, FR4 core is essentially epoxy.     I should have realized that.  I shouldn't need to get any closer than that for this application.  Good call everyone. 

[Smokey]

Hmmm... Saturn Toolkit doesn't have embedded Edge Coupled Microstrip??  What am I missing? 

These guys mention it directly, but otherwise there isn't much out there I can see https://www.polarinstruments.com/products/si/help/edge.coupledembeddedmicrostrip.htm

[Terry Bites]

https://www.eeweb.com/tools/edge-coupled-microstrip-impedance/

[selcuk]

It is not under "Conductor Impedance" tab. It is on "Differential Layer" panel under "Differential Pairs" tab in Saturn PCB.

If you have time I recommend you to simulate the transmission line with OpenEMS or a similar tool. The text books and calculators are okay for theoretical formulas but they sometimes differ a lot from practical results. You may need to consider PCB structure, epoxy, air, plastic enclosure, track geometries etc. Think about a 4-layer PCB, having core, prepregs and solder mask layers having different dielectric constant. But it may be too much an effort if this is just a track connecting a connector with an IC in short length.

[nctnico]

Where it comes to potting, I'd recommend looking at polyurethane. AFAIK this shrinks less and thus causes less issues with components being broken and so on.

[Smokey]

It is not under "Conductor Impedance" tab. It is on "Differential Layer" panel under "Differential Pairs" tab in Saturn PCB.

If you have time I recommend you to simulate the transmission line with OpenEMS or a similar tool. The text books and calculators are okay for theoretical formulas but they sometimes differ a lot from practical results. You may need to consider PCB structure, epoxy, air, plastic enclosure, track geometries etc. Think about a 4-layer PCB, having core, prepregs and solder mask layers having different dielectric constant. But it may be too much an effort if this is just a track connecting a connector with an IC in short length.

Thanks.  Found it.

Wow.  Ok.  Recap: I'm targeting a differential impedance of 120ohm, on a 2 layer 62mil board with 0.5oz copper.

If I run the numbers for non-embedded edge coupled microstrip (air on top), I get 24mil conductor width and 10mil spacing (with a conductor height of 61mil).

......For embedded with a total H2 of 400mil (all else the same), I get 8 mil conductor width for 120ohm. 

Glad I ran those numbers.  If I had left it at 24mil width and potted it, it would have only been 87 ohms.

[Smokey]

Where it comes to potting, I'd recommend looking at polyurethane. AFAIK this shrinks less and thus causes less issues with components being broken and so on.

As it turns out, for this application the potting is mostly to fill a void, and serves very little other purpose.  So I'm looking to use the material that would reduce the cost the most.  Would polyurethane be less expensive than epoxy?