1
\$\begingroup\$

I have a VNA and want to measure the impedance looking into my PCB.

My VNA is pretty old, it's a 8753ES.

My PCB includes a commercial Nordic or Dialog MCU with Bluetooth Low Energy (BLE). The path from the MCU includes bond wires, multiple transmission lines and parasitic components. I've analyzed what I expect the impedance to be using a Smith chart.

The path between the MCU and the antenna has been split and I can plug my VNA into the PCB or the antenna. The total distance from this location to the MCU < lambda/10. It's a single pin antenna.

  1. Can I use a VNA to measure the impedance of the MCU?

  2. If I put the MCU into constant transmit mode (as I have to?) then can I only match at one frequency? (BLE band is 2400MHz to 2480MHz)

  3. Am I going to break my VNA if I plug a constant TX device into it?

  4. I currently don't have the calibration tool and it costs $6K... For this exercise what is the consequence of no calibration? Can I make my own coarse calibrator?

  5. To measure, connect the VNA to my split and simply look at the Smith chart? Only using one connector on the VNA. Can I look at the impedance radar (time domain response) to identify the effect of bond wires etc, or is the radar not straightforward?

    Thanks so much!

\$\endgroup\$
6
  • \$\begingroup\$ Look at Maury Microwave for an alternate cal kit vendor. Last I checked, they had cal kits for ~$2000 or $2500. In the past these guys have manufactured cal kits for HP and Agilent. \$\endgroup\$
    – The Photon
    Commented Mar 14, 2018 at 2:44
  • \$\begingroup\$ I'm not clear what you mean about "connect the VNA to my split" (what's a "split"?) or about connecting to "the PCB" or "the antenna". If you draw a diagram of your system, it might make it more clear. \$\endgroup\$
    – The Photon
    Commented Mar 14, 2018 at 2:53
  • \$\begingroup\$ You only need to do Open and Short cal. not the 10 term. Adjust Smith Chart CAL>Zo> to trace impedance. Tx must be much lower to avoid interference. I would put a directional coupler in the split so you can monitor in situ your Tx s11 into a Schottky diode & DMM. The major issues may be dielectric constant and trace tolerances. \$\endgroup\$
    – D.A.S.
    Commented Mar 14, 2018 at 3:07
  • \$\begingroup\$ By the "split" I was referring to the break in my PCB trace between the antenna and the MCU. On the right is only a 50ohm antenna On the left is a trace (transmission line) then a bond wire then my MCU \$\endgroup\$
    – Tock
    Commented Mar 14, 2018 at 17:28
  • \$\begingroup\$ @TonyStewart.EEsince'75 What result would I be looking for from a directional coupler? Is it 2x 50ohm in parallel? \$\endgroup\$
    – Tock
    Commented Mar 14, 2018 at 17:36

1 Answer 1

Reset to default
2
\$\begingroup\$

I'll answer the parts of your question where I understand what you're asking:

Can I use a VNA to measure the impedance of the MCU?

If the MCU is powered on and transmitting, it could produce spurious readings from the VNA.

If the MCU is shut down, it will likely not have the same equivalent impedance as when it is operating.

If it can be turned on and put in a no-transmission mode, you might be able to get a good measurement.

Am I going to break my VNA if I plug a constant TX device into it?

Not if the output power is below the maximum input level for the VNA.

But again, the signal coming from the TX could cause erroneous readings.

I currently don't have the calibration tool and it costs $6K... For this exercise what is the consequence of no calibration? Can I make my own coarse calibrator?

There are alternate cal kit vendors, but I doubt you'll find a decent kit for under $1000.

You could try measuring some known devices to see how accurate your VNA is without calibration.

It'd be unlikely you could get a good measurement of the \$S_{11}\$ of a well-matched load without calibration (but you could probably verify the load is not totally broken and behaving like an open or short).

\$\endgroup\$
6
  • \$\begingroup\$ Thank you Photon! I think what I can do is put the MCU in receive mode, this is supposed to have the same impedance as transmit mode. For calibration I'll try an open and short calibration as Tony mentioned, that seems to make sense \$\endgroup\$
    – Tock
    Commented Mar 14, 2018 at 17:40
  • \$\begingroup\$ If you're trying to measure how well matched your transmitter is, I think you'll need a load cal (not just short and open). Otherwise there's no way to tell which measured reflections come from the DUT, and which come from the cables, etc. \$\endgroup\$
    – The Photon
    Commented Mar 14, 2018 at 18:06
  • \$\begingroup\$ OK, I got a calibration kit in the mail which includes load. That does make sense because the VNA will be transmitting into a 50 ohm load. \$\endgroup\$
    – Tock
    Commented Mar 14, 2018 at 18:45
  • \$\begingroup\$ I know it's been a while since you answered this question but I am curious about one thing you mentioned: "If it can be turned on and put in a no-transmission mode, you might be able to get a good measurement." So if I have an SoC that has an RF output (eg. WiFi, BT), and the impedance looking into that pin isn't 50 Ohm, what the best way to get the impedance looking into it? If I don't turn on the Tx, I am afraid this output won't biased in the right region. I guess my question is how do you measure the impedance of an active device, if you will. For example, the nRF52 chips don't spec Zo. \$\endgroup\$
    – Big6
    Commented Jul 20, 2020 at 15:21
  • \$\begingroup\$ I thought one could just turn on the Tx, in some low power mode, and measure the impedance with the VNA. But my concern is the power from the Tx skewing the results. \$\endgroup\$
    – Big6
    Commented Jul 20, 2020 at 15:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.