2

enter image description here

I've already asked a similar question in the past, but I want to make sure I understood things correctly.

Attached is the current lab setup for my chip testing. The green square represents the main PCB board, and the little blue square represents the chip. The chip needs separate VDD voltage for its analog and digital circuits, but their '-' terminals are tied to the same layer (so-called 'GND layer') of the green PCB. In addition, there are many RF connectors are attached to the board. The body of these RF connectors is also connected to the 'GND layer' of the green PCB. This is also highlighted as a gray wire in the figure. For the debugging purpose, the RF test equipment (signal generator and oscilloscope) will always be in use while USB communicates to the chip.

From here, I have two questions regarding the ground loop.

Q1: To minimize the ground loop effect, do you think I should use a '2 prong cord' for the 2017 Macbook Pro shown on the right side of the figure? I think I should use 2-prong cord because using a 3-prong cord will create an extra earth ground connection to the mains, creating an extra low-impedance path for the ground loop. There are already multiple ground loops due to the connection to the RF test equipment; hence I need to keep the # of GND loops as minimum as possible.

Q2: Possible concerns about the ground loop if '-' terminal of PSU is connected to the earth ground in the given situation. In this situation, the GND is already connected to the earth GND due to the connection between PCB and RF test equipment. Thus, I believe the power supply should not provide extra earth ground to minimize ground loop effects. Please correct me if I am wrong on this.

JYelton
  • 32,302
  • 33
  • 134
  • 249
Emm386
  • 577
  • 2
  • 14
  • 3
    For Q1, have you even measured if the laptop supply has any output referenced to earth? If you have not, then you should. If the power supply has 3 prongs, it must be connected to matching outlet. Connecting a power supply with 3 prongs to an outlet by using only 2 prongs will create more problems than solve as you are misusing the supply. – Justme Aug 13 '23 at 19:44
  • For note 1 of your diagram I recommend you read my message here https://electronics.stackexchange.com/a/677017/274961. You risk creating loops with your oscilloscope probe. – Vincent Aug 13 '23 at 20:41
  • @Vincent thanks for the heads up. In the figure, the frequency of the target signal is around 5 GHz, and I will be using an RF coax cable with scope's input impedance of 50 Ω. I guess in this situation, the GND loop inductance will be minimized. – Emm386 Aug 13 '23 at 21:23
  • @Justme thanks for the heads up as well. this is good suggestion. I will double-check this later once I get back to the lab. – Emm386 Aug 13 '23 at 21:25
  • @Emm386, As all GNDs are connected to the same point, this depends on the distance between the "same point" and the inner conductor of your coaxial cable. It's this surface area that needs to be kept as small as possible. – Vincent Aug 13 '23 at 22:32
  • @Vincent hm... but the distance b/w GND and inner conductor is determined by the target characteristic impedance (50Ω in this case) and signal integrity (as smooth transition as possible from on-board stripline/microstrip to coax). In this sense, how would you address GND loop inductance in the RF connector layout? – Emm386 Aug 13 '23 at 23:06
  • @Emm386, Sorry, I misunderstood your diagram. I hadn't seen note 2, there's a connector. I thought you had a probe whose ground would be connected far from the measurement point. – Vincent Aug 14 '23 at 06:41

1 Answers1

0

I'm answering my own question in case other students/people encounter the same problem later.

  1. Macbook Pro's 3-prong charger's earth GND is indeed connected to the shield of the USB-C via 1MΩ resistance. Because of 1MΩ resistance, I shall not worry too much about the extra GND loop coming from the Macbook pro, but I think I should place a USB isolator to make sure about this.

  2. RF test equipment's connector body is connected to the earth GND. This can cause an extra GND loop. Thus, I should plug these RF test equipment into the same multi-tap to minimize the GND loop issue.

  3. The '-' terminal of the power supply shall not be connected to its earth GND. This may cause an extra GND loop.

  4. All in all, the GND loop may not cause too many issues in this configuration. but following the above suggestions can avoid GND loop and plus minimum CM noise coupling b/w earth GND and floating GND. The latter is because the DUT's GND is indeed connected to the earth GND via the RF test equipment's body connection.

Emm386
  • 577
  • 2
  • 14