diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Everything Else (http://www.diyaudio.com/forums/everything-else/)
-   -   Single Device - Ground Loop hum (http://www.diyaudio.com/forums/everything-else/211417-single-device-ground-loop-hum.html)

redjr 23rd April 2012 09:41 PM

Single Device - Ground Loop hum
 
Hi all - I recently put together a simple 3 source selector and remote volume control unit (kit from China). After firing it up for the first time I learned I have a ground loop, or induction hum. I seem to have it narrowed down to the shielded output cable (twin lead cable w/shield) to the right and left channel RCA jacks. During some troubleshooting I can physically move the cable and increase the level of the hum. I've checked and doubled checked my grounding topology and believe it to be correct, but obviously I've got a problem somewhere.

My best practice question is... when wiring up RCA input or output jacks, do you only connect the signal lead going to/from the PCB device to the jack, and then connect the ground side of the jack to the grounding star? I've read this is the best practice and that's how I hooked it up in this device and now I've got hum. :eek: I've also read that you can keep the cable shielded (minimize stray interference) by connecting only one end to ground.

Enzo 24th April 2012 02:09 AM

Have you tried this setup with complete cables? That is cables that do not omit ground connection at an end?

redjr 24th April 2012 02:53 AM

Quote:

Originally Posted by Enzo (Post 2998376)
Have you tried this setup with complete cables? That is cables that do not omit ground connection at an end?

I haven't yet, but that was my next test. Based on what I have in place that scenario will most likely create a loop. ;) I also believe I need to use better coaxial cable too.

gootee 24th April 2012 03:43 AM

I might not be the cable but while you're discussing that I will try to tell you what I think I know. Also, some photos of your wiring etc would probably help a lot.

There are a few ways that you could get hum. The two main ones that I can think of are 1) conductor pairs that form enclosed loop areas and 2) not using star grounding, such that large or dynamic currnts in a shared ground return induce hum voltages at a signal amplifier input's ground reference point.

Without seeing how your unit is laid out, it's difficult to say. So I'll just try to give some generalities about "enclosed loop area", first, and maybe you'll see some egregious example of it and solve the problem immediately.

Firstly, any conductive loop will have a time-varying current induced in it by any time-varying electromagnetic field in the air. (And any loop with a time-varying current will induce a corresponding time-varying EM field in the air.) How well a field can induce a current (and vice versa) depends in part on the how much geometric area is enclosed by the loop.

If a current does get induced in a loop, then any impedance in the loop, and even the impedance of the conductors themselves, will have a voltage induced across it.

This loop-area effect is significant. So you really do want to minimize the enclosed loop area, for every loop where it might matter.

The main loops to worry about are the ones that could be formed by pairs of conductors, like the signal input and its ground, and the AC mains pair, and the DC power rail and ground pairs, and output signal and ground pairs, etc.

For example, you would not want the input signal wire to go one way and its ground wire to take a different route. In fact, you want them to be as close together as possible, ALL the way to wherever they need to go. The best simple way to accomplish that would be to twist them tightly together, with lots of twists. (The twisting is even better than just having them close together.) You would also want to tightly twist together each of the other pairs of conductors that were mentioned above.

Even better would be to use a twisted pair, as above, but then also have a shield around the twisted pair. But NOTE that the shield would NOT be connected to the signal ground, since one of the wires in the twisted pair already carries the signal ground. The shield would only be connected to chassis ground, and usually works better when only one end is connected to chassis ground, usually on the end that is nearer to the input jack (since that's usually near the chassis). The input jack would need to be electrically isolated from the chassis, so that the signal input ground wouldn't be connected to the chassis at the jack.

I'm not sure how your inputs are connected, or what they connect to, but in a single channel amplifier I would typically take the twisted-together input signal and signal ground wires all the way to the two ends of the resistor that is usually connected from the amplifier input to ground. Then the connection to star ground would usually go from the ground end of that resistor to the star ground. Sometimes, people take the connection to star ground from the input jack's signal ground, insteaof from the amp's input's ground reference point (the ground end of the mentioned input resistor).

Star grounding is about not letting the wrong two (or more) things share a conductor for their ground connections. For example, an amplifier input's signal-ground reference point (the ground end of the input resistor) should not share ANY length of conductor with the ground ends of decoupling/bypass capacitors, or with any local power supply smoothing caps. So the amp input's ground reference point needs it own separate conductor, back to the star ground point.

If the input's ground reference DID share a conductor to ground with the decoupling or bypass or power supply filtering capacitors for any power amplification device, the time-varying currents through those capacitors and their connection back to star ground would cause voltages to be induced, back at the non-star-ground end of the conductors, including on all conductors that share any common segment. In the case of an amplifier input, the bouncing ground-reference voltage would effectively be arithmetically summed with the input signal voltage, which would be "a bad thing". Any 120 Hz that was "swallowed" by the capacitors between the power rails and ground would be likely to show up as 120 Hz hum.

Sorry. I have to just cut this off and get to bed. Later.

gootee 24th April 2012 04:05 AM

RCA jacks have only a center conductor and an outer "shield ground". They are not able to use shielded cable with two conductors inside. RCA cables use the shield as the signal ground conductor. Thus, they are not actually shielded cables, when used that way, since the signal ground conductor, being the shield, is not, itself, shielded by anything.

If you are using regular RCA jacks and RCA cables, with one center conductor and one outer "shield" conductor, then you usually would want to keep the shield connected, and make sure that the bulkhead-mounted jacks are isolated electrically from chassis ground.

the only time you might want to disconnect one end of the shield conductor would be if there is already another connection between the two units' grounds, forming a ground loop. But it would usually be better to solve that in a different way, because the "other" ground connection would typically be farther away, creating "enclosed loop area" with the signal conductor if the RCA cable's signal ground (aka the "shield" conductor) were disconnected.

redjr 24th April 2012 04:19 AM

I put much better shielded, coaxial cables on tonight and connected both signal and ground up at both ends. I also removed my star ground wire going to the PCB since it is grounded through the RCAs lugs. I reduced the hum significantly, but you can still hear it if you put your ear up to the speaker. You'd never hear it at normally listening levels through speakers, but I'd know it's there. :D So something is still amiss. It's late. I'll try more tomorrow.

Enzo 24th April 2012 08:00 AM

Hum will alwaqys be 60Hz or other multiple of the mains frequency, but hum can come from many many sources. It will sound the same because it all relates to the mains. But hum is not generic. Each source of hum has its own cure.

In other words, you could have some hum from ground loops, and some from poor shielding on cables, and some from poor ground connections and some from who knows what. All the cable shielding in the world won;t decrease hum coming from poor power supply filtration for example. SO you may have eliminated a lot of hum with a cable change, but that does not mean the remaining hum is related to cables.

SO what you need to do is determine the various sources of hum that are active in your unit, then address each cause appropriately.

redjr 24th April 2012 12:44 PM

I agree there could be multiple sources of hum. Perhaps the cable hum was louder then a secondary source of hum and was therefore masking it initially - until I fixed that hum, or partially fixed it. I need to investigate further.

redjr 24th April 2012 02:34 PM

Now that I grounded the output cables ground/shield to the PCB, I was able to disconnect one of the star ground wires and further reduced the hum. I also better twisted some AC mains wires so that should have helped too. I may beef up the filter caps to see if that improves things as well. It's at a manageable level now though - at least for this device. :D

redjr 24th April 2012 08:55 PM

Now I'm chasing ghost! :eek: Let me explain...

1) I just finished building my first, class A, 100wpc amp using pre-assembled modules. Okay, I cheated a little bit not building it from scratch. ;) I was very meticulous in planning, layout and wiring of the amp. In particular, adopting the star grounding strategy. Of course I tested voltages along the way and everything was checking out beautifully. After it was completed I measured the DC-Offset voltages on each channel and they were both less than 4.0mv. Ok good. Complete thread on that build can be found here.

2) To test the amp I used an old school Heathkit pre-amp I had built years ago and by all accounts is functioning properly. As the source I used a very good Sony portable CD player (D-303) using the line out. I also used just simple cheap interconnects between the devices. The amp was hooked up to some Optimus Pro LX5 speakers serving duty on my bench.

3) Using a favorite CD I proceeded to test the amp. It performed flawlessly, with no noticeable or at least audible hum. I listened through phones, paused the player, and could not hear any hum. OK. Even I was astonished that there was no hum present. For my first major DIY amp I was pleased. I let it burn-in for a couple of days at low volume levels. Nothing blew-up! :D

Enter my most recent DIY source selector/volume control/pre-amp project thingy (SS) that has been causing trouble on the ranch in the hum department...

4) To test this device I used my new amp and the same Sony CD player, along with the same interconnects. I was purposely trying to keep some of the factors the same to ease in potential trouble-shooting. I even used the same mains outlet.

5) Using the same favorite CD I played it through the SS device, adjusted volume and listened. At normal levels everything appeared to sound fine. However, when I paused the CD that's when I noticed the hum. OK I thought. My SS device is introducing the hum. After all the amp had tested clean prior to introducing the new SS device. So I focus my energy on the new device and start to double-check my wiring, grounding (star approach), and cabling. The fruits of that effort are described earlier in this thread. A few changes were made and I was able to reduce the hum to very low levels. Good I thought, but not totally gone.

6) So, in an effort to do more testing, I decided to pair the new SS device up to a little store bought Class T amp, and remove my new amp from the mix.

7) Using all the same devices (except for the amp) and interconnects, I proceed to hook everything up to the Class T amp. Powered everything up and had a listen. At the very least I was expecting to hear the same amount/level of hum I had experienced with my other amp. But to my amazement there was absolutely no evidence of any hum. Nadda! I paused the player, listened closely to the speaker and through the same phones and - no hum. Super quiet. Things are getting spooky around here. :eek:

8) So for a final test I decided to bring the SS device into my office and hook it up to my Yamaha amp I use at my computer. The same CD player, CD, interconnects, the whole shebang. The speakers for this test were different. Some little AR speakers I use on my desk rig. The CD played beautifully, plenty of volume, the SS device was functioning properly, etc. Nice. Paused it, listened with the speakers up close to my ear and through phones and not the slightest hint of any hum.

So... What does all of this mean? Is my spanky new amp to blame, even though it sounded just fine though my Heathkit pre-amp? Is there an issue with mains outlet and my new amp. Something not wired correctly or backwards - neutral and hot leads reversed on my IEC connector? It shouldn't matter since they are both going to the primary side of the tranny.

Since I had no hum on my last two tests I'm about to consider this case solved and closed. I think I'll re-test my new amp with a couple different pre-amps and see if any hum exhibits itself. Sorry for such a long post, but felt I needed to lay it all out there. Thanks to gootee and Enzo for their observations and help.


All times are GMT. The time now is 03:41 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2