Author Topic: Does impedance value of a voltage rail change at different points ?  (Read 1139 times)

0 Members and 1 Guest are viewing this topic.

Offline muthukural001Topic starter

  • Regular Contributor
  • *
  • Posts: 211
  • Country: in
Hello All,

Today, my colleague asked me to fill impeance value of the voltage rails.He told me to take impedance value of the same rail at different point.For example, a 5V usb input is going to three DC-DC converters for 3.3, 1.2, 1.8. He told like, take impedance value of 5V input rail at USB connector, 3.3V DC-DC converter's input side,1.8V DC-DC converter's input side and 1.2V DC-DC converter's input side...
Is he right?...My assumption is all the values are same...Please correct me if I am wrong...


Thanks,
Muthu
 

Offline AndyC_772

  • Super Contributor
  • ***
  • Posts: 4315
  • Country: gb
  • Professional design engineer
    • Cawte Engineering | Reliable Electronics
Yes, it does vary, because the power plane has inductance and resistance. If it were a perfect conductor, there would be no need for ICs to have decoupling capacitors physically located next to power pins.

It's incredibly difficult to measure accurately, though, especially at the kinds of frequencies where this inductance has a significant effect. High end CAD software (eg. Cadence Allegro) can simulate the PCB layout and derive the impedance of a supply at different places on the PCB.

To do it in a meaningful way, you'd need a network analyser and some specialist training in how to probe the board correctly. I'm guessing you don't have these.

What equipment do you have?

Offline muthukural001Topic starter

  • Regular Contributor
  • *
  • Posts: 211
  • Country: in
Hi Andy,
Yes, we do not have network analyzer. We will measure with Multimeter...

Ok, Generally, how much ohms difference they will have? milli Ohms (OR) Ohms ?.


Thanks,
Muthu
 

Offline AndyC_772

  • Super Contributor
  • ***
  • Posts: 4315
  • Country: gb
  • Professional design engineer
    • Cawte Engineering | Reliable Electronics
It's highly unlikely that a multimeter will tell you anything useful whatsoever.

If you need to know the impedance of a supply, this implies that you need to know for a good reason.

If the board were being used to supply very large currents to a resistive load like a heater, then it might make sense to measure the dc resistance (not the impedance) of the tracks on the board.

However, I'm guessing that this isn't the case, since you're using dc/dc converters to produce low voltage rails which are typical of low power digital logic. It's much more likely that, if you need to know the impedance at all, you need to know it across a range of frequencies that a multimeter simply cannot measure.

Ask what this information is needed for, and how it will be used. It may well be that you don't have the ability, or the need, to make a measurement which is correct and meaningful for the purpose.

Offline radiolistener

  • Super Contributor
  • ***
  • Posts: 4135
  • Country: 00
Ok, Generally, how much ohms difference they will have? milli Ohms (OR) Ohms ?.

it may vary in range from milliohms to gigaohms, it depends on frequency, power supply and wires.
 

Offline muthukural001Topic starter

  • Regular Contributor
  • *
  • Posts: 211
  • Country: in
Hi,

By "frequency", what do you mean ? Is this the switching frequency of the DC-DC converter (Or) others?. Please let me know...
 

Offline radiolistener

  • Super Contributor
  • ***
  • Posts: 4135
  • Country: 00
For AC current, resistance depends on frequency, because it consists of active (energy loss) and reactive (inductance or capacitance phase shift) component. This complex value, which includes active and reactive components is named "impedance".

if you're need to measure internal resistance of DC source, you can measure it by adding some known load resistance and measure voltage with load and without load. When you know source voltage and voltage drop on some known resistor, you can calculate internal resistance of power source.

Internal resistance also depends on wires, because wire resistance will be added to source resistance.
« Last Edit: May 21, 2019, 11:23:22 am by radiolistener »
 

Offline brabus

  • Frequent Contributor
  • **
  • Posts: 330
  • Country: it
Hi,

By "frequency", what do you mean ? Is this the switching frequency of the DC-DC converter (Or) others?. Please let me know...

muthukural001, just one question: do you have a clear picture in mind of the difference between "resistance" and "impedance"? The problem may be here.
 

Offline schratterulrich

  • Regular Contributor
  • *
  • Posts: 50
  • Country: at
    • Elektronik & Layout
If you want to fully determine the impedance of your power supply it is not a single value. Instead it is a function of frequency as it is the case with capacitors and inductors.

To get a better understanding I have written a tool to simulate the impedance of power planes in combination with capacitors in the past.

I want to give an example:
This is a powerplane-pair with a distance of 0,7 mm between them.

You can see the locations of caps and measurement points.

Here you can see the simulated impedance at the different measurement points.


some insights that can be derived from it
At low frequencies (up to 1 MHz) the 100 µF Elko dominates the impedance. The higher impedance at Zin(i002) is from the resisitivity of the copper planes. Distance to the capacitor is not important.

At higher frequencies (~ 10 MHz) the local decoupling caps (100nF) are effective. Here the distance plays a big role. Zin(i002) has no local 100 nF cap so its impedance is higher.

At even higher frequencies (100 MHz) you can see the effect of the plane inductance. So Zin(i002) shows higher impedance due to the thin connection to the rest of the plane.

At frequencies above 500 MHz (depends on size of planes) board resonances are dominant. You cannot improve the impedance in this region very much with decaps.

You can find the tool at https://leiterplatte.jimdo.com/pdn-sim/
 

Offline pwlps

  • Frequent Contributor
  • **
  • Posts: 372
  • Country: fr
Today, my colleague asked me to fill impeance value of the voltage rails.He told me to take impedance value of the same rail at different point.For example, a 5V usb input is going to three DC-DC converters for 3.3, 1.2, 1.8. He told like, take impedance value of 5V input rail at USB connector, 3.3V DC-DC converter's input side,1.8V DC-DC converter's input side and 1.2V DC-DC converter's input side...
Is he right?...My assumption is all the values are same...Please correct me if I am wrong...

The impedance has a resistive part and an reactive part (probably mostly inductive in the case of PS rails), which is important depends on your application. For example if you draw little current but it is a pulsed current with fast edges then the inductive part will dominate.
Without resorting to network analyser measurements, if you only need to check if inductance might be a limiting factor for your PS rail you can connect a fast switching load (e.g. a pwm controlled mosfet) and look at transients induced on PS lines. It is not obvious to translate such measurements in terms of impedance (there will be reflected transients etc.) but it may provide a qualitative test.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf