Menu
▾
▴
1 Attachments
#183 Errors in CPW line calculations
Hi!
I have been recently playing with QUCS trying to do some simple excercise in CPW lines (CPWG -- grounded and CPW -- suspended in air). I found some scenario that gives suspiction results as compared to other apps that calculate the same. I would like to report this -- maybe will help to improve QUCS in the future.
For my play with CPW I wanted to calculate params of a CPW line with dimensions like shown in the PDF attached (CPW with W = 25um, S = 20um, t = 1um, substate is Si or 11.8 and height is 275um).
The discrepancy between Z0 results (53.8 Ohm vs. as much as 56.6 Ohm) in both cases made me run the same calculationn using the TXLine application (from AWR) and LineCalc (from Agilent). The TXLine is telling me 53.9 Ohm vs. 53.5 Ohm, while LineCalc says 52.85 Ohm vs. 52.63 Ohm. Tu sum up:
CPW CPWG
QUCS: 53.8 Ohm 56.6 Ohm (sic!)
TxLine: 53.9 Ohm 53.5 Ohm
LineCalc: 52.9 Ohm 52.6 Ohm
TXLine differs slightly from the LineCalc. However, the difference is nothing compared to what the QUCS calculated for the same line. I wonder what might be the reason for such huge differences.
Regards
Pawel
Thanks for the detailed report on the CPW(G) lines impedances. As you may know there are no exact formulas for computing the parameters of CPW(G) for the most general case, so the different results might be due to different approximations used.
But it seems the results are different even for the case of zero-thickness lines, which has a closed-form solution. I will take a closer look at the formulas used.
There was actually a bug in the CPW code (details at https://github.com/Qucs/qucs/pull/306 ). I have corrected it now and the results for the CPWG with zero thickness are in good agreement with LineCalc. For practical non-zero thickness values the results are also similar.
Note that TXLINE appears to use a different set of formulas, as the impedance computed is strongly frequency (and conductivity) dependent for your case.
This correction will (very likely) be in the next 0.0.19 release.
Fix is now merged https://github.com/Qucs/qucs/commit/d7169ef6ed9e29ebfd58326d6f0d9ca576d04918