3
\$\begingroup\$

Circuit structure:

  1. An AC to DC power converter MeanWell 220V to 5V. The output is 5V. The power grid runs at 50Hz.

  2. We build DC-DC converters(Power Module) with multiple orders of Low pass filters to provide different DC voltages needed for our main circuit: +5V, +15V, -15V

  3. Main circuit uses the voltages generated above. It provides control signals to its target and reads feedback signals. It connects with target through 2.54 mm dupont wires, which are around 0.3 meters long. It will adjust control signals according to the feedback signals. The target device has purely differential receiver and shall not be influenced by common-mode noises.

enter image description here

Problem:

  1. The whole circuit runs normally at our companies lab. The circuit is not connected with earth.

  2. When used in some Lab with many devices(scopes, signal generators and etc.), it does not function well any more. The feedback signal looks like following, which is 50Hz noise. Noises

  3. When the circuit ground is connected with earth, the 50Hz noise will be reduced greatly. It starts to work.

  4. There are still 50Hz noises which do not exist when in my Lab and it harms the device performance.

Questions

  1. Where are the 50Hz noises from?

  2. Why the noises are reduced when connecting the circuit with earth?

  3. How to remove the 50Hz noise totally?

\$\endgroup\$
13
  • 4
    \$\begingroup\$ You haven't filled in your country in your profile, but I guess that the 50 Hz comes from the background radiation of the power grid. It's quite common issue. \$\endgroup\$ – AndrejaKo Feb 1 '16 at 14:08
  • 3
    \$\begingroup\$ Your question begins by mentioning a power supply, suggesting that it might be the cause (?). But your circuit might have other sensitive points, long wires to pick up the 50 Hz hum. \$\endgroup\$ – Bimpelrekkie Feb 1 '16 at 14:11
  • 1
    \$\begingroup\$ "Does connecting the circuit with earth reduce the 50Hz noise?" Sometimes it does, sometimes it does not, sometimes it makes no difference ! It all depends on your system setup/connections/ circuits. \$\endgroup\$ – Bimpelrekkie Feb 1 '16 at 14:13
  • 3
    \$\begingroup\$ All I see is a picture of some signal and no schematic and a statement about it not functioning well anymore. Nothing to go on. \$\endgroup\$ – Andy aka Feb 1 '16 at 14:28
  • 1
    \$\begingroup\$ Also, you say something about "The Circuit is tested in China": 1. Which circuit? The supply or yours? 2. Do you know they tested it, or do they say they did? \$\endgroup\$ – Asmyldof Feb 1 '16 at 14:57
2
\$\begingroup\$
  1. The 50Hz noise is mains hum. Any time you pick up noise at the same frequency as your mains hum, assume it's from the mains (once, maybe twice in a lifetime it'll be a wrong assumption, but every other time it's right)

  2. Pretty much the only way to get rid of picked-up noise (mains hum, emi/rfi, whatever) is to ground it out.

  3. From your updated question post, it looks like the noise is appearing on a (possibly long) wire carrying a non-differential signal & coming as data feedback from a (mains powered?) "slave" device.

Either the hum is originating from within the "slave" device due to some power supply noise to it (likely due to poor wiring & other noisy devices on same circuit), or it is being picked up by the wire as EMI/RFI.

If it is being picked up as EMI/RFI (some part of it almost surely is), then switching to differential signalling could "fix it."

If it is coming from the "slave" device somehow, then there may be a need to add some 50Hz filtering &/or grounding in that device, to eliminate the noise before it gets onto the data line.

\$\endgroup\$
11
  • \$\begingroup\$ Could you please explain why ground it would reduce the picked-up noise? For the filtering in power supply module, we use: LC filters(filter high frequency noises) and LDO to filter low frequency noises. For main circuit, every OpAmp has two caps(10uF and 0.1uF) at its power supply rails. \$\endgroup\$ – richieqianle Feb 2 '16 at 6:25
  • \$\begingroup\$ Re 1., if there's ever a question of whether that 50Hz (or 60Hz) hum is coming from the mains, trigger your scope on the mains (usually a built-in option on the trigger source select) and look at the noise. If it stays in phase with the trigger for a minute or more, you can go from 98% sure to 99.99% sure that's the source. \$\endgroup\$ – Entropivore Feb 2 '16 at 9:10
  • \$\begingroup\$ @richieqianle shorting to ground helps with power mains noise in 2 ways.#1 Ground is where all power mains current "wants" to go, it completes the circuit. Mains power generators, outdoor utility poles, power utility transformers, etc. are all connected to the ground, sometimes using it for lightning protection, for a reference voltage, and even as a "fourth phase"/neutral. Any time you use your power at 220V, you're connecting one pin to a live phase, and the other(s) to ground to complete the circuit (different in the USA). Connecting mains noise straight to ground gives it a "short circuit" \$\endgroup\$ – Robherc KV5ROB Feb 2 '16 at 16:33
  • 1
    \$\begingroup\$ #2 The earth works effectively like an enormous capacitor. Any time lightning hits the ground, it lowers the charge of this ground-air-clouds capacitor. Due to the enormous storage capacity of the planet's capacitance, any noise you introduce to ground is of too low energy to cause a noticeable change in potential, so effectively the noise gets stored in a capacitor so large that it forms "the ultimate" low-pass filter. \$\endgroup\$ – Robherc KV5ROB Feb 2 '16 at 16:38
  • \$\begingroup\$ Can you test your device where it is picking up this 50hz hum flr us, and determime if you can find the hum in the power supply output, or only in the main ckrcuit, please? \$\endgroup\$ – Robherc KV5ROB Feb 2 '16 at 16:41
3
\$\begingroup\$

As others have said, the 50Hz is almost certainly mains-related. You say the target has a differential receiver, but you don't say if the "main circuit", which is receiving the feedback signal from the target, uses a differential receiver, or what the signalling levels of the feedback signal are. Even assuming a differential circuit, any real differential receiver has a limited common-mode range, which you could easily be exceeding. Connecting your system ground to earth ground likely reduces the common-mode 50Hz noise to a level that your receiver can tolerate.

Switching power supplies such as the one you show generally will have a capacitor bridging across the gap between the primary (mains) and secondary (output) side (commonly referred to as a "Y" cap). Its function is to reduce interference generated by the power supply switching circuits. Because of this capacitor, and the parasitic capacitance between the primary and secondary of the transformer itself, the output of the power supply can have a very large (>100V) AC voltage with respect to earth ground superimposed on it. This AC signal has a relatively high source impedance, because the capacitor is relatively small, and thus has a high reactance at the mains frequency. Because of this high impedance, it is not a safety hazard (unless the capacitor shorts, which is why specially rated capacitors are used for this function), but if you have a high impedance signalling circuit, this can overwhelm it.

When you test in your lab, if both the main circuit and the target are floating with respect to earth ground, then everything is riding this AC component together, and there is no problem. When you take it to the real world, if the target device has a connection to earth ground (even a small capacitive connection), now you have a large potential difference between the main circuit and the target, and the problem occurs. By also connecting your main circuit to earth ground, you greatly reduce the potential difference between the main circuit and the target, so your receiver is able to function.

You also don't give much information about the distance between the main circuit and the target device, the type of cabling used, and so on. The more information you provide, the better the answer you're likely to get.

\$\endgroup\$
6
  • \$\begingroup\$ Thanks for providing such a great answer! 1. The receiver is single-ended. 2. The connection between main circuit and target is through dubont wires, which are shown in the diagram in edited post. 3. By connecting main circuit with earth, do you mean a path is provided for the large ac signal? But connecting with earth is still through 10MOhm resistors, how would it help? \$\endgroup\$ – richieqianle Feb 2 '16 at 6:46
  • \$\begingroup\$ Can I provide some sort of earth on the power module? I really want to make the usage easier. \$\endgroup\$ – richieqianle Feb 2 '16 at 6:47
  • \$\begingroup\$ And do you mean the Y capacitor makes the output of wall wart PSU not truly floating any more? \$\endgroup\$ – richieqianle Feb 2 '16 at 7:46
  • \$\begingroup\$ Aha! So the single-ended feedback receiver makes this scenario even more likely. I'm not familiar with what you refer to as 2.54mm DuPont wires. I assume 2.54mm is the connector pitch, but I don't know if the wires are twisted, shielded, or whatever. If you could provide a link to a product datasheet that would help, but I don't think that's the most important point. Connecting the main circuit to earth puts it at a roughly equal potential with the target (i.e., both have their circuit ground reference at earth potential). 10MOhm may be too high a resistance. You might have to use less. \$\endgroup\$ – Entropivore Feb 2 '16 at 8:57
  • \$\begingroup\$ Yes, I think connecting the DC return ground to earth in the power module would work. You could probably tolerate a few kilohms in that path, but I think 10Meg is too high. Also look out for the problem that if someone substitutes a different wall wart with the same nominal specifications (voltage & current), it might have a different Y cap and things could stop working again. You have to design for the worst case. And yes, the Y cap makes the wall wart output not truly floating. You may be able to get a custom wall wart with a smaller Y cap or none at all, but that's not a good solution. \$\endgroup\$ – Entropivore Feb 2 '16 at 9:06
1
\$\begingroup\$

According to travel websites and Wikipedia, China uses 50Hz power.Any alternating current gives off electro-magnetic radiation with the same frequency.This is a likely source of the noise. If you test any circuit in the same way, you are likely to find the same noise. Try this before looking for other solutions.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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