In message snipped-for-privacy@4ax.com, Cursitor Doom snipped-for-privacy@notformail.com writes
It's right all right.
The outer diameter is the same for the 50 and 75 ohms. For the 75, I presume it's not practicable to make the pin diameter smaller and retain its robustness, so the only way to increase the Zo is to remove as much of the dielectric as possible. If you compare the 50 and the 75, you will see what I mean.
I still maintain the principal determinants of the impedance are as I stated previously. The formulas for line impedance are shown on this page and the aforementioned determinants are key.
If they placed the LEDs in the corners of the inner cabinet then they would be quite safe. All the RF energy is concentrated towards the center with very little indeed misdirected.
Very poor web page indeed. No educational value whatever.
They lump together the values of the dieelectric constant and the permeability of free space, the geometry of the configuration and the conversion from natural to base-ten logarithms all together into one magic factor, without any hint of where it all came from.
Shame! That's not 'everythingRF': It's almost nothing! Oh, and there is no such thing as "impedance per unit length".
Jeroen Belleman
I could not get the calculator to come out for me. Getting negative numbers and about 10 times what they should be.
However if you change the Relative Permittivity you will see how much it will effect the impedance of the coax. Go from 1 as air and then 2.1 to
2.6 for teflon and other common insulator/dialectric material and see how much the impedance changes.I doubt that you have ever actually ran the numbers or you would see the impedance change as the material is changed.
The main thing is the ratio of the diameters of the cables, but you still have to account for the material between them just as you will for a capacitor..
You can often find the Relative Permittivity tables where capacitors are.
The critical part of the connector here is a fraction of an inch long, so none of this stuff matters below a few GHz.
There are a lot of silly equations around, that people plug into apps.
Try a wide microstrip in some online calculators. Many use the equation that's in the old Motorola ECL book, and a wide trace reports a negative Z.
In message snipped-for-privacy@4ax.com, john larkin snipped-for-privacy@gct.com writes
Yes. Obviously. I've been retired now for many a year but, IIRC, the 50 ohm was considered 'good' to around 1000 MHz, and the 75 to around 500MHz. [A Google on the individual manufacturers' specs is recommended.] Regardless, both are often used to higher frequencies.
>
Here's a BNC tdr/tdt. It's really not so bad.
In message snipped-for-privacy@4ax.com, Cursitor Doom snipped-for-privacy@notformail.com writes
I'm sure that the principal determinants of the impedance are absolutely correct. As Scotty said, "Ye cannae change the laws of physics". However, we're talking about the practical problems and constraints in the construction of a connector so that it is mechanically viable, and at the same time attempting to maintain the most constant Zo throughout the complete male-female junction.
That calc is nonsense. Reasonable entries generate negative values and a preposterous cutoff frequency.
Negative values for Z0? That would be surprising, because the equation is correct, even though the physics are obfuscated away into a few magic factors.
Example please?
Jeroen Belleman
GIGO. Where on earth did you get those input parameter values from? Try it again with something more realistic.
Indeed! They seem to have botched the inch and cm units. It makes better sense for the other units. It's a weird and wonderful error to make, because it doesn't actually matter in which units the diameters are given, as long as they are the same! The argument of the log is dimensionless!
Shame! Obfuscate the physics and then get it wrong too!
Jeroen Belleman
They are physically reasonable, not garbage. I might build a high voltage coax from a rod in a 1" copper pipe.
No simple equation will predict PCB trace or coax impedance in the general case. Sensible software will warn when the input values are out of the useful range of its equations. This one just displays nonsense.
I guess that a polynomial on D/d might be better. At least it wouldn't go negative.
We use a real e/m simulator to verify capacitances and impedances when we suspect that the dumb programs are being dumb. Or build one and measure it.
this seems to be something related to their units conversion. If you use millimeters it gives sensible-looking answers.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required