Analog_Fan
Well-known member
What's this element (component) to separate Digital and ANALOG GND?
market as 600Z, on the lower part.
What's this element (component) to separate Digital and Analog GND?
- Thread starter Analog_Fan
- Start date
Help Support GroupDIY Audio Forum:
What's this element (component) to separate Digital and ANALOG GND?
market as 600Z, on the lower part.
feed-through-capacitor?
5v333
Well-known member
Ferrite bead?
wire jumper with a ferrite bead (inductor).
It is pretty common practice to name analog and digital ground networks differently so the PCB layout artist knows to keep them separate, but they ultimately need to get connected to work.
JR
It is pretty common practice to name analog and digital ground networks differently so the PCB layout artist knows to keep them separate, but they ultimately need to get connected to work.
JR
Analog_Fan
Well-known member
some answers from stack exchange:
http://www.analog.com/en/content/raq_groundingadcs/fca.html
than i found what
https://ez.analog.com/data_converters/high-speed_adcs/w/documents/2665/ad9243-grounding
Texas Instruments also has a nice one:
Thank you:
http://www.analog.com/en/content/raq_groundingadcs/fca.html
than i found what
https://ez.analog.com/data_converters/high-speed_adcs/w/documents/2665/ad9243-grounding
Texas Instruments also has a nice one:
Thank you:
JohnRoberts
Once i used star grounding and a 1ohm resistor near the power connector.iampoor1
Well-known member
Sometimes they will also be labelled as a resistor and then a 0 ohm resistor is installed 
Back in the early days of machine insertion for PCBs, they actually used zero ohm resistors in place of jumpers.Sometimes they will also be labelled as a resistor and then a 0 ohm resistor is installed
At peavey we had a machine that inserted real wire jumpers.... in different lengths.
JR
user 37518
Well-known member
The problem with zero ohm resistors is that there is no such thing as a zero ohm resistor, specially at high frequencies. Or better put, a zero ohm resistor is not a zero ohm impedance.
ommmmm
Analog_Fan
Well-known member
i never used a bead, although i have seen them in circuits.
any recommended brand? for SMD.
any recommended brand? for SMD.
you can buy smd inductors. When I put a class D amp IC in my drum tuner I put inductors in series with the output in combination with Cs to gnd.
JR
They (analog and digital gnd pins on the converter IC) generally need to be connected with a low impedance. Putting a ferrite bead in between is probably a bad idea in general. The vendor datasheet or app note is usually a good starting point, but don't assume the app note schematic is the optimum that can be achieved. You would be much better off keeping the return impedance as low as possible, meaning full board ground plane with direct connections from pins to plane, and segregate analog and digital return currents by spacing on the PCB rather than separating the copper. High frequency return currents will follow the path of least inductance, so you can control where the currents flow in the ground plane by how you route the signal and power.
user 37518
Well-known member
The ferrite bead produces a high impedance at high frequencies, so for audio frequencies its basically a short but keeps those dirty digital edges out of the analog ground. I reckon its not the most elegant solution.... But in this case, (looks like that chip is an ADC), analog and digital grounds might be best left connected like you mention, I am thinking more of a micro controller or something like that.
True but so what ? No conductor eg pcb track is 'Zero Ohm'
Well - I've (obviously not) tried using a 1206 superconductor. But found it hard to solder
There are different schools of thought on the preferred way to handle this type of 0V connection.
Henry Ott (RIP - he died last year) used to have some very good stuff on this on the web but I don't think it's available now ?
The standard manufacturer application note guidance on joining AGND and DGND is commonly to do this at one point at the chip. But in reality this often falls down in practice. What is the best when you have several converters etc on the same pcb etc...?
It's also worth a check if the AGND and DGND pins of a device are actually joined inside the device itself.
Similar considerations for any other 'GND' pins eg you might also get pins designated as 'PGND' relating to "Power" etc.
user 37518
Well-known member
It goes beyond having zero resistance, they have parasitic inductance and capacitance, at high frequencies they present a high impedance, that was my point, not the fact that you need a super conductor.True but so what ? No conductor eg pcb track is 'Zero Ohm'
Well - I've (obviously not) tried using a 1206 superconductor. But found it hard to solder
Obviously I was joking about the superconductor. The parasitic capacitance from end to end is essentially bypassed by the (very very near) zero ohm resistance. There is series inductance but with a metal film type it is very small - down in the nH range. I don't think anyone is using wirewounds for zero ohm links.
user 37518
Well-known member
I know one or two professional PCB designers that would disagree with you on the severity of this issue, and Im not talking about stuff in the GHz range, these resistors can also wreck havoc when it comes to EMI, zero ohm resistors should be avoided in digital circuits with fast rise/fall times, period, clock speed is not the biggest issue, but fast transition times.
You are considering the bypassing of the capacitance by considering a lumped parameter model, again, with very fast rise and decay times this is not necessarily true. When I designed RF amps a mere 0.5 mm of extra conductor length would alter the entire circuit dramatically, accurate models of passive components need to be used for simulation, you can't assume a linear and lumped parameter model for the passive components, the model you propose is an extreme oversimplification. Digital circuits with fast rise and fall times can and should, in many cases, be treated this same way.
Last edited:
Speedskater
Well-known member
Sadly after he died, his web-page disappeared. I used the Internet Archive to store most of the interesting pages on my hard drive.
Fair enough - if 0.5mm conductor length makes a significant difference then it does. But that isn't down to a Zero Ohm resistor.
