From: Bob Lewandowski (blewandowski@Vixel.com)
Date: Wed Feb 23 2000 - 15:50:12 PST
I would guess that the error you are seeing (~1.8% of CF) could be due to a combination of effects.
1. Is Er measured by vendor at the frequency that you are using, or at some other, typically 10 GHz? Teflon-glass can have some Er variation with frequency. What is tolerance on their Er measurement
2. How thick is the copper layer that the resonators are made of? A couple mils error in length due to over etching or overexposing the photo resist is not unreasonable to expect, particularly if you
are using 2 oz. copper (~2.8 mils thick). To verify dimensions you need some sort of optical measuring system with sufficient accuracy to verify overall length of each resonator. You should also be
able to verify shrinkage with an overall length measurement of the longest dimension of the whole filter. When measuring pcb trace geometries you have to account for the trapezoidal effects of etching.
The effective length at the top surface of a trace (most ectched) will be shorter than it is at the bottom (least etched). This difference can easily be equal to the copper thickness. Even with 1/2 oz.
copper over-etching or over-exposure/over-development of photo resist can change dimensions dramatically. It is a technique that board manufacturers commonly use to trim transmission line widths to
control line impedance.
3. When you press on the top dielectric layer is there a bonding film layer between, or is there possibly air trapped between the dielectric layers? If there is a small amount of air or other material
between the dielectric layers the effective dielectric constant can be changed.
If there was as much shrinkage as your
colleague suggests I would think that the board would warp dramatically. There is always some shrinkage, but a couple % sounds like a lot.
'Solder and see' is typically the order of the day for these types of filters. 8-)
C Deibele wrote:
> I have a question about something that I've heard about recently.
> I am making FIR stripline filters - circa 4 GHz, and the physical
> dimension of the filter is 5"x30". The filter is manufactured using
> Arlon 2.33 LX material, eps_r=2.31 (Arlon measured the dielectric and
> loss tangent), and is put together using an etch/acid wash.
> I designed the circuit using Agilent's Touchstone package.
> My problem: The center frequency of each filter is off by 70 MHz. My
> filters are rather narrow band, so this was a problem for me. I simply
> made individual sections of my filter a bit longer to accomodate this
> frequency offset. Each coupled line in my design was originally 460
> mils, I extended each to 465 mils. This implies that the overall length
> increase is proportional to 3/460 or about .5%.
> Yesterday a colleague told me that it is common with many circuit boards
> to shrink when a layer of copper is removed from a single side. This
> particular colleague even claimed that a board will shrink around 100
> mils or so after etching. Has anyone else observed or heard of an
> effect such as this? This definitely would explain why my center
> frequencies are off.
> Craig Deibele
> **** To unsubscribe from si-list or si-list-digest: send e-mail to firstname.lastname@example.org. In the BODY of message put: UNSUBSCRIBE si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
> si-list archives are accessible at http://www.qsl.net/wb6tpu
**** To unsubscribe from si-list or si-list-digest: send e-mail to email@example.com. In the BODY of message put: UNSUBSCRIBE si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
This archive was generated by hypermail 2b29 : Thu Apr 20 2000 - 11:35:09 PDT