What's this element (component) to separate Digital and Analog GND?

[user 37518]

Fair enough - if 0.5mm conductor length makes a significant difference then it does. But that isn't down to a Zero Ohm resistor.

I didn't mean to upset you, a zero ohm resistor is even worse than 0.5mm of copper, it is much more complex, it has non linearities, it is a distributed system, it is definitely down to it, there is no such thing as a zero ohm resistor at high frequencies. Thinking of its parasitic capacitance being shunted by the near 0 ohm resistor only works at low frequencies, I was trying to explain why this only works at low frequencies. In RF and digital circuits, specially modern ones, it is highly advisable to think of components and tracks as transmission lines.

The reason why the 0.5mm copper affects the signal is not so much because of its resistance, its because of its characteristic impedance and its electrical length, which depends on the L and C of the transmission line; that C that you say its shunted by the resistor, is in fact, not shunted by the resistor when the resistor works as a transmission line, the resistor can create a discontinuity in the transmission line due to the difference between the copper line characteristic Z and the resistor characteristic Z, a part of the signal might reflect back, etc... it is a very complex subject. So just saying that it just boils down to a near zero ohm resistor in series with some nH is simply not true at high frequencies. You can't say that its not down to a Zero Ohm resistor, because the Zero Ohm resistor can't be divorced from the parasitics that accompany it.

I don't want to create an argument, this is simply one fact one must grasp when dealing with high-speed circuitry, a cable is no longer just a cable, resistors are no longer just "R", capacitors are no longer just "C", etc...

Make a search for SOLT (short open load through) calibrator, companies sell you a short circuit, an open circuit, a load (tipically 50 ohms) and a through barrel to calibrate an RF system, yes a short, why? because there is no such thing as a short at all frequencies, you can't just take a piece of cable and short the ends of a coax cable and expect it to be actually a short. Also take a look at the price tag, that tells you that Zero Ohm is low frequency illusion..

Here is one, for only a mere $4352 dollars, I'll tell you that is the most expensive "necklace" you'll ever wear:

https://www.pasternack.com/50ohm-3.5mm-female-open-short-load-26.5-mhz-pe5ck1012-p.aspx

 

[Analog_Fan]

Here is one, for only a mere $4352 dollars, I'll tell you that is the most expensive "necklace" you'll ever wear:

https://www.pasternack.com/50ohm-3.5mm-female-open-short-load-26.5-mhz-pe5ck1012-p.aspx

that's almost enough for 2 x Tata nano.



lol

is it made from some extra terrestrial noble metal?

 

[Newmarket]

I didn't mean to upset you,

I'm not upset by electrical characteristics !

, it has non linearities, it is a distributed system, it is definitely down to it, there is no such thing as a zero ohm resistor at high frequencies.

Well in absolute terms there is no such thing as a "Zero Ohm" resistor full stop. But we know what we mean by it.
Non-linearities / distributed networks etc. These are all truisms.

The reason why the 0.5mm copper affects the signal is not so much because of its resistance, its because of its characteristic impedance and its electrical length, which depends on the L and C of the transmission line; that C that you say its shunted by the resistor, is in fact, not shunted by the resistor when the resistor works as a transmission line, the resistor can create a discontinuity in the transmission line due to the difference between the copper line characteristic Z and the resistor characteristic Z, a part of the signal might reflect back, etc...

Yes. I know.

So just saying that it just boils down to a near zero ohm resistor in series with some nH is simply not true at high frequencies. You can't say that its not down to a Zero Ohm resistor, because the Zero Ohm resistor can't be divorced from the parasitics that accompany it.

It's a model and in many cases a rather useful one. If a circuit / frequency is such that a more complex model gives a benefit in design/analysis then switch up to that and so on.
Every understanding is a model at the end of the day and can be refined. eg wrt Wave Particle Duality. Wave theory alone doesn't completely describe light / em behaviour. But it remains a very valid and useful model.

I don't want to create an argument, this is simply one fact one must grasp when dealing with high-speed circuitry, a cable is no longer just a cable, resistors are no longer just "R", capacitors are no longer just "C", etc...

It's not really an argument. It's all true. Just a difference of emphasis.
Indeed, I have often tried (with varying degrees of success) to convey this type of understanding to other engineers / designers, mainly of the mechanical breed. eg Why we need a braid at this point and not the wire they have put in despite me specifying a braid,. Or why we need to take a co-axial cable right upto a pcb rather than transition to the usual thin twisted pair used for general internal wiring that is easier to fit (but upsets the eye-pattern etc etc.

 

[user 37518]

It's not really an argument. It's all true. Just a difference of emphasis.
Indeed, I have often tried (with varying degrees of success) to convey this type of understanding to other engineers / designers, mainly of the mechanical breed. eg Why we need a braid at this point and not the wire they have put in despite me specifying a braid,. Or why we need to take a co-axial cable right upto a pcb rather than transition to the usual thin twisted pair used for general internal wiring that is easier to fit (but upsets the eye-pattern etc etc.

My basic rule of thumb to discriminate between "regular" and "freaky" electronics is that if you want to buy a substrate to make a PCB from a company that won't sell it to you unless you give them a good reason why you need it, then you know you are messing around with some serious "black magic" electronics...

 

[neil.johnson]

Sadly after he died, his web-page disappeared. I used the Internet Archive to store most of the interesting pages on my hard drive.

Oooofff that's unfortunate. I know Henry passed away a while back (there was a nice write-up in the IEEE EMC magazine) but a real shame his internet legacy has been taken down