Regarding PCB material

[cjrathi]

Hi,
I have designed RF PCB, which is supposed to work at 2 GHz and 5 GHz bands. I have fabricated the PCB locally in Pune. The PCB material is standard FR4 with dielectric constant 4.6 and substrate height of 0.153 mm. This PCB is working quite fine at 2 GHz band (but still losses is more as expected), but I am experiencing very high losses and leakages at 5 GHz band. I think it is due to medium quality PCB used.
So, please suggest me good PCB material with dielectric constant and height of substrate, which will eliminate such dielectric and radiation losses at high frequencies.

Regards,
CR

 

[PlanarMetamaterials]

Any commercial substrate other than FR-4.

The selection depends on what you're using the PCB for. Is it supposed to radiate?

If not, I'd go with something like 60 mil Roger's 3010. (er = 10.2, h = 1.524mm).

 

[thylacine1975]

Aye, I second PlanarMetamaterials - FR4 is a horrible microwave substrate. Aside for the huge variation/unpredictability of permittivity across temperature, manufacturers etc - as you've noticed, it's terribly lossy (in the GHz region) too.
(but it is *cheap*

For an appetite-whetting sample of the range of microwave substrates out there, have a look at the list (+datasheets) offered by Lintek: **broken link removed**
I've recently used Nelco (Park Electrochemical's) MW9350 [https://www.parkelectro.com/parkelectro/product_result.asp?product=Mercurywave 9350] for a cost-sensitive application that needed Rogers-like performance at a few GHz without the pricetag. I must say I'm pretty impressed with it.

Good luck (choosing just one

 

[cjrathi]

Any commercial substrate other than FR-4.

The selection depends on what you're using the PCB for. Is it supposed to radiate?

If not, I'd go with something like 60 mil Roger's 3010. (er = 10.2, h = 1.524mm).

No. PCB is not supposed to radiate. All the power is supposed to get transferred through RF trace ideally. So I guess I will have to use substrate of high er as compared to FR4.

Do you know how much more will Roger's or other material will cost as compared to FR4.?

- - - Updated - - -

Aye, I second PlanarMetamaterials - FR4 is a horrible microwave substrate. Aside for the huge variation/unpredictability of permittivity across temperature, manufacturers etc - as you've noticed, it's terribly lossy (in the GHz region) too.
(but it is *cheap*

For an appetite-whetting sample of the range of microwave substrates out there, have a look at the list (+datasheets) offered by Lintek: **broken link removed**
I've recently used Nelco (Park Electrochemical's) MW9350 [https://www.parkelectro.com/parkelectro/product_result.asp?product=Mercurywave 9350] for a cost-sensitive application that needed Rogers-like performance at a few GHz without the pricetag. I must say I'm pretty impressed with it.

Good luck (choosing just one

Yes.. I observed the performance in 5 GHz band. Its very lossy and giving very unpredictable results. But its giving very nice linear performance in 2 GHz band. So I guess only option to work it in 5GHz band is to use other PCB material. What do you think ?

Do you know how much more this high dielectric constant PCB costs than FR4 ?

 

[sivamani]

For higher speed of operation you should choose material which has Lower er. Rogers may cost you 3 times more than FR4.Try in Isola also..

 

[cjrathi]

For higher speed of operation you should choose material which has Lower er. Rogers may cost you 3 times more than FR4.Try in Isola also..

If I use PCB with lower er than FR4, then I think radiation efficiency of RF trace will increase.

And this will further increase radiation losses. Am I right ?