Grounding and noise advice requested.
Jul 19, 2013 at 4:00 PM Thread Starter Post #1 of 12


100+ Head-Fier
Aug 10, 2012
Hi all,

I'm just in the process of finishing to put together an original hybrid design (MOSFET AB output stage). Unfortunately I've run into a bit of an issue.

I first built the sand output stage and wired it up directly to the RCAs to test for background noise. It was dead silent in both channels. Next I built the valve PS, the input stage and added the pot. This is when the trouble began. The first thing that I noticed was a high pitched grounding noise whenever I got close to the to the case, even though everything is grounded (pot is a blue velvet grounded through one of the screws). This noise increases as the volume increases, up to about 4/5 of the turn, then subsides. This sounds like simple grounding and I'm going to go back in and create 3 star grounds (1 for the valve PS, 1 for the valve input, and one for the Sand bits).
Any advice for actually creating these star grounds? The issue I always run into when trying to implement a star ground is all the wires springing off whenever I try to solder one more to the node.

The other noise that I heard was a low frequency (<200Hz) 'whopper whopper' noise that doesn't vary with volume, I also tried bypassing the input stage and this noise didn't diminish meaning that it must be coupling from the valve power supplies to the SS stage. A bit of experimentation soon revealed that the noise was coming from the valve heaters of all places. The heaters are running from DC, (originally an SMPS, which I replaced when I heard this noise with a power resistor which drops the 18V rail to 12V). How can this be putting serious noise into the other stages? It's the same DC as used for the other rails just passed through a resistor! The two heater wires are tightly twisted. Any ideas? Anything? I'm at a real loss here.

One more question that shouldn't be to hard to answer is about wiring the pot. I'm using shielded wiring from the RCAs to the pot. Do I take the ground from the RCA to the pot and then to a star ground, or the RCA to a star ground and a star ground to the ground of the pot? And at which end do I connect the shield?

Any responses to any of these questions would be greatly appreciated.
Jul 19, 2013 at 4:59 PM Post #2 of 12
I'm having trouble picturing what you are doing and how the voltages are obtained.  mains transformer?  Try taking the heater directly from the AC through another regulator.
Star grounding is going to depend on what you have in the case and whether or not the case is earth.  FWIW, I don't really use star ground exactly.  The scheme (unless with PCB grounded to chassis as some designers use) I always use is this.  Mains earth to chassis.  Ground loop breaker via rectifier or diode bridge, power resistor and capacitor in parallel to ground bus.  AC side of rectifier to earth.  DC side to ground bus.  All voltage regulators grounded on the DC side.
RCA wiring, shield on the input jack side.  Floating on the other end (pot side).  I normally like to take ground at the pot input to the amp, input side of the amp, post pot.  I have no theories there, just convenience, and what I take as a ground reference should match what the amp sees.
Jul 19, 2013 at 5:23 PM Post #3 of 12
Thanks for the response, just for clarification, I have a dual 18V AC toroid, which outputs into a pair of diode bridges, followed by 10mF caps, then LT1084s that regulate +/-18V DC. For the heaters, (2 12AU7s, so 300mA at 12.6V) I've wired them in parallel with a 22R resistor between +18V and one side and the other side of the heater to ground. For some reason this nicely regulated DC is creating a buttload of noise which ate up 3 hours of my afternoon :D
If I'm still not clear, I'll post a schematic tomorrow.
Jul 20, 2013 at 1:07 PM Post #6 of 12
Well, I spent a whole day on this, and around lunchtime I had it to a point where I was pretty happy with it, but I decided that some of the noise was coming from the heater cables, so I replaced them with shielded ones and moved a few things around, now it's much worse than it was. I tried moving things back but everything I do makes it worse. I think I'll probably have to rip everything out and start again... :frowning2:
I can't even ground the pot/case properly as it whistle unless I touch it, even though there is a good connection from mains earth to the case and then from the screw in the pot to the same location.
Anyone got any suggestions for implementing ground busses before I take such drastic action?
My current guess is that I'm getting ground loops (fairly sure it's 50/100Hz) because I have a point to point stage and then a stripboard stage, which have different paths to ground...
Thanks for your help,
Jul 21, 2013 at 5:20 PM Post #8 of 12
I installed a ground loop breaker as soon as I heard how noisy the amp was, however I didn't, and still haven't installed the diode bridge. From what I've read the resistor and capacitor make the loop breaker and the bridge is there for safety in high power amps. This isn't very high power and I don't have any spare diodes capable of more than 3A to hand. Should I add this anyway? (I noticed AMB's amps don't have the diodes either).

I redid the grounding today using a solid core 2.5mm2 bus, which snakes it's way through most of the case with all the power supply grounding at one. End where it's connected to the loop breaker and the sensitive stuff at the other end. According to RMAA the worst peak in noise is at 50Hz and is just below -70dBA. This is barely audible in my DT880s but still a bit louder than I'd like in my shure SRH440s (very sensitive and closed). RMAA gives the Amman overall noise rating of -81dBA, which isn't too bad for a lot of valve amp (although my previous one was about 10dB better).

Another idea I have is to put the supposedly recommended 120R series resistance in series with the headphone jack which should significantly reduce the noise level in both set of headphones (more so in the 40R shures). I'll try this tomorrow but wondered if anyone had any opinions on the subject.

I also sometimes get a high pitched whistling interference when the pot is between 1/3 and 4/5 of the way on. (About 5-6kHz) this goes away if I touch the case. I've tried grounding the shell of the pot to different points using the screw on the back of it (blue velvet). Then I realised that I've had this pot in three different amps now and every time it has exhibited the same problem not matter how it's been earthed. Is this a known problem? I'm going to try replacing it with a cheap alpha I've had lying around and used with success before, though it seems a shame to do so.

Thanks for your help, I'm pleased to have this amp to the stage where I can actually listen to music on it, but I would just like to get a few more percent :)

Jul 22, 2013 at 11:21 AM Post #10 of 12
Sorry, missed that you have a breaker in place.  Yeah, the bridge is safety and won't help much or any with regards to the noise.  Try building a quick 12.6V regulator and see if you get noise by pulling the heaters onto it's own circuit instead of off the amp rails.  I haven't seen many circuits do it your way, and I've never done it except for the "starving student" type amps.  Another thing to try, maybe first, is to put the heaters in series between the 2 18V rails, and center it to ground instead of loading only one rail.  Hope for some cancellation/balancing of some sort.
Jul 22, 2013 at 11:43 AM Post #11 of 12
I'm just about to try stringing them between the two rails, I've returned to the smps (cheap lm2596 module) to regulate their supply and it was actually quieter than a resistor, leading me to believe that it is related - at least in part - to loading. Before I poked around too much I had to build an e12 type circuit to be protect my cans, as I'm not used to amps without output coupling caps.
WRT the bridge, if it's there for safety I'm surprised Ti Kan doesn't recommend it to go with the ground loop breaker he says to put in in single box amps, you can't exactly accuse the Beta22 of being 'little'!
Jul 22, 2013 at 12:01 PM Post #12 of 12
That did it!!!
I put the SMPS module between the 2 rails (still set to 12.6V) with the heaters in parallel. (Screw the dissipation for this test!), with my most sensitive headphones there's only a slight pink like noise in the background. 
Thanks so much for all you help chaps! I'll post some information on this project soon, as I'm really quite pleased with it. 


Users who are viewing this thread