4
\$\begingroup\$

Typical PCBs have a dielectric with a low constant \$\varepsilon_r\$. I'm studying the effect the electrical fields on dielectrics and as an experiment would like to test PCBs with a very high dielectric constant, such as using calcium copper titanate with \$\varepsilon_r \approx 10000\$ (the higher \$ \varepsilon_r \$ the better).

This is not for a "practical application" (I understand why typical PCBs don't have a high \$\varepsilon_r\$ and it makes sense - not slow the signal, low signal loss, etc), but this is for an experiment purpose only.

So I would like to know a manufacturer of PCBs that could make such a PCB with a very high (the higher the better) dielectric substrate. As I'm just studying the electric field properties, I only plan to make the traces on one layer (two at the most) and no chips, resistances, capacitors, holes, etc. Just the PCB would be great!

In summary:

  • PCB with very high dielectric constant and one layer (two at the most)
  • Only traces, no chips/resistances/capacitors/holes/etc
  • minimum trace width ~0.25 mm (lower the better)
  • minimum trace spacing ~0.25 mm (lower the better)
  • Dielectric should support up to 1000 V between traces (I think Calcium copper titanate breaks down at 2 kV/μm so this would work)
\$\endgroup\$
8
  • 2
    \$\begingroup\$ "Dielectric should support up to 1000 V between traces" Unless the tracks are on opposite sides/inner layers, you only have ambient air as isolation between them. \$\endgroup\$
    – winny
    Commented Dec 12, 2023 at 9:02
  • 3
    \$\begingroup\$ "So I would like to know a manufacturer of PCBs ..." That's a shopping question and they're not allowed. See electronics.stackexchange.com/help/on-topic. \$\endgroup\$
    – Transistor
    Commented Dec 12, 2023 at 11:06
  • 3
    \$\begingroup\$ In case you don't find anything on the market, you can make your own "PCB". Just get the substrate you want and use conductive ink to makeshift something. There are semi-professional Printers for conductive ink if you need defined space widths etc. If I had to do that experiment, I'd even try copper foil and just glue it to the PCB. \$\endgroup\$
    – S_G
    Commented Dec 12, 2023 at 11:13
  • 1
    \$\begingroup\$ Kyocera will do custom ceramic substrates for you, if you're prepared to order thousands of them. They're not called PCBs at that point. \$\endgroup\$
    – Hearth
    Commented Dec 12, 2023 at 13:32
  • 1
    \$\begingroup\$ But don't ceramics like those manufactured by Kyocera have really low εr's? Those materials, and Kyocera substrates, are used a lot in our microwave applications, 10 GHz to 50 GHz. \$\endgroup\$
    – SteveSh
    Commented Dec 13, 2023 at 12:35

3 Answers 3

Reset to default
6
\$\begingroup\$

This is a long-shot answer to your question, but since you said this "this is for experiment purposes only" and "... very high (the higher the better) dielectric substrate." I have an idea that you might not have considered.

There are aluminum PCBs and aluminum has an infinite dielectric constant. These PCBs do have a very thin dielectric layer of between 50μm to 200μm. Assuming that this thickness can be specified when you order the PCB, then you could create a customized range of samples which will effectively exhibit very high dielectric constants.

But, if this answer isn't helpful for your particular experiment, just let me know in the comments and I will delete it.

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Answers are supposed to be generally useful to people in the future, so whether it helps the OP in particular is less important than if it is a good answer to the exact question they asked. In my opinion, anyway, so please don't delete. \$\endgroup\$
    – Spehro Pefhany
    Commented Dec 13, 2023 at 16:29
  • \$\begingroup\$ thank you, your answer is "better than my question" since it even solves the problem in a better way \$\endgroup\$
    – mabeco
    Commented Dec 14, 2023 at 22:34
  • 1
    \$\begingroup\$ Very glad to have been of help! Good luck with your experiment! \$\endgroup\$
    – phil1008
    Commented Dec 14, 2023 at 23:57
4
\$\begingroup\$

You can get flexible PCBs manufactured on polyamide backing down to 0.08mm thick. You could consider getting your circuit manufactured on such a substrate, then vacuum bond this PCB, track side down, to your dielectric.

(By "vacuum bond", I mean use a very low viscosity adhesive, with vacuum to remove the air and almost all adhesive, in a similar fashion to laminate build up.)

\$\endgroup\$
2
\$\begingroup\$

For a high dielectric constant, Rogers material is the perfect option (https://www.raypcb.com/rogers-pcb/) (Rogers 4350b, Rogers 4003c, Rogers 5880, etc.) I am associated with the company RayPCB.

Here is dielectric constant vs frequency for your reference:

enter image description here

\$\endgroup\$
10
  • 1
    \$\begingroup\$ If you represent the firm in question you should declare your association. \$\endgroup\$
    – colintd
    Commented Dec 13, 2023 at 9:00
  • 1
    \$\begingroup\$ I don't think this is a spam answer - it's quite relevant to the question. In fact, I was trying to remember the name of these PCB materials when I wrote my answer but could recall it. \$\endgroup\$
    – phil1008
    Commented Dec 13, 2023 at 12:35
  • 1
    \$\begingroup\$ @phil1008 - Hi, Re: "I don't think this is a spam answer" FYI on this site, spam also means "undisclosed promotion", where "undisclosed" means that it is missing a disclosure statement from its author within the text of the post, if it mentions, links to or even just infers anything affiliated to an entity (e.g. company) with which they have a connection. By that definition, this is still a spam post - no disclosure statement from its author, yet it mentions (links to) their company. || The only reason it has not been deleted is I'm giving its author a short time to edit it & follow the rules. \$\endgroup\$
    – SamGibson ♦
    Commented Dec 13, 2023 at 13:01
  • 1
    \$\begingroup\$ Fair enough. His/her user name is "RAYMING PCB' though, and they look like a new user. Looks like a disclosure statement was added. I worry that sometimes people are too quick to cry "spam" and then people pile on, but don't reverse their votes later. \$\endgroup\$
    – phil1008
    Commented Dec 13, 2023 at 23:01
  • 1
    \$\begingroup\$ @phil1008 - I'll come back to this properly, but to respond to an important misconception that you mentioned: "His/her user name is "RAYMING PCB' though" That is not enough for disclosure of affiliation. Usernames can be changed with no visibility on the posts they made. This was thrashed out ages ago on one of the metas (either Meta.SE or Meta.SO, I forget which one) leading to the current rule on promotions. To be clear - the disclosure must be text within the post. Nothing else is acceptable, for good reasons. || As I said, I'll get back to this but there's more to do in the background. \$\endgroup\$
    – SamGibson ♦
    Commented Dec 13, 2023 at 23:26

Not the answer you're looking for? Browse other questions tagged or ask your own question.