LM3886 Ground Loop Noise - Solutions?- Car Amp Application

[MartyM]

Hi guys. I recently hand-routed the star-ground PCB and built my own LM3886-based PA100 (2 ICs in parallel) mono amps. Work great and sound nice!

Zero noise when used they're used as home home stereo setup. However , my ultimate goal is to use them as the basis for more elaborate car amplifier(s). I already have a prototype SMPS I tested them with. Obviously I am not new to circuit design in general.

Note: I have a 2-amp star grounded car stereo system which normally has ZERO noise. Also the PA100 amp I built has ZERO noise when used as in my home stereo.

Unfortunately, when tested in my car using the front RCA connections from my Alpine DSP, once again ground loop noise (the common "alternator whine") appeared again, as I expected.

Anybody used the chip amps for similiar purposes? Any solutions or practical ideas for eliminating the ground loops? Will I be required to use an op amp circuit on the RCA inputs?

I've tried the usual routes but no relief. I have limited audio amp design experience. Would really appreciate your help! Attached are photos of my hand designed PCB & hand built amp(s).

I would prefer to not have to reverse engineer an existing car amp to find a solution. I know someone on here has gone through the same pains-or has the amp design experience I could learn from...a forum search revealed no directly relevant answer.




Thanks! My schematic is attached (mostly Nat. Semi's example schematic) Please advise if you can!

[Perry Babin]

A ground loop isolator in the RCA lines or floating the secondary ground of the power supply should get rid of the noise.

If you don't want to use either or fhose solutions, you will need to build a balanced input circuit for the amplifier.

[MartyM]

Yes, I realize I could use a transformer based isolator but we all know that's a "patch" and not a real solution. And besides affects frequency response (I've used them years ago and with mediocre audio quality).

The balanced input sounds more reasonable, though I would prefer to keep components out of the audio input "chain."

Hopefully some more ideas will pop up on here also...I'm going to check a few things in the car as well.

[DigitalJunkie]

Do you have a schematic of your SMPS?

I second the notion to isolate the SMPS and amp grounds.

Does your SMPS use feedback? (opto)isolated,or no? -even if not,it's an easy fix.

In the last amp SMPS I "reworked" for a different use,I Had to add a 1K resistor between the output ground(+/-35V),and the input(+12V) ground to eliminate any ground loop issues.
The feedback didn't seem to mind much at all,and still functions,after a minor adjustment (I replaced the feedback R with a pot,to vary the output voltage,it just needed a slight tweak to compensate for the extra 1Kohms.)
You might be able to get away with as little at 10-100ohms.

Make sure the RCA/interconnect cables are in good shape,and not routed near any power leads,etc.

Maybe also check the RCA outputs of your headunit.. I see alot of issues on here where they've 'opened up'..
Get a DMM/VOM,and set it to the lowest resistance scale,and put one probe on the outside (shield) of the RCA,and one on the case/Ground,there should be ~0-ohms (or very low). If not,you need to fix it/have it fixed.

[Perry Babin]

If you mount the amp VERY near the head unit and ground both the amp and the head unit to the same point, the ground loop noise should be minimized.

There are good transformers but they're not cheap. Jensen transformers is supposed to have some of the best.

If you don't want additional components in the audio chain, floating the secondary is just about the only option left. It works. Just about every budget brand car audio amp on the market uses it. About the only time it doesn't work well is when you have multiple signal sources. For example, if you had a 4 channel amp powered by a single supply and 2 channels came directly from the head unit (from the sub out, for example) and the other 2 channels passed through an equalizer before being sent to the amplifier. The two different ground points for the head unit and the EQ would cause another ground loop.

[MartyM]

Hi thanks Perry but my system is not as simple as the average consumer.

Mine goes: head unit -> fiber optic connection & control cable -> digital signal processor -> amplifier(s). There is not a continous analog path between the inputs/outputs to the amp. (It's all in the digital domain). All components are connected to a single junction block leading to the battery via correct copper cabling.

But of course all variables must be checked. The DSP's RCAs do have continuity to system ground.

But at any rate, those are helpful ideas about isolating the power supply's ground. I'll give that and DigitalJunkie's idea a try!

I had a feeling it might be related to that but never found good "do and don't do" info when building the SMPS.

Thanks for the replies I'll post the results A.S.A.P.!

SMPS: +/-35V, 300W (tested max.). My schematic:

[MartyM]

Hey guys here are the results!

I tried your first suggestion (before resorting to a balanced input circuit) and isolating the SMPS worked great!

Also I was able to reconnect my SMPS's ground output to the +12V ground using a 10 ohm resistor as you mentioned. No noise then either!

Against my better judgement I originally removed the optional input ground-to-output ground jumper on my PCB I made and had to separate them physically for testing..but not too bad.

It turned that my ground loop was being created while my Planet Audio amplifiers still had there RCA cables (coming from the signal processor) plugged in when I removed one pair for using as an audio source for my homemade PA100 (LM3886) chip amps.

Isolating the SMPS ground worked great!

Thanks so much guys for your practical ideas. The sound quality is excellent too, by the way.

My dream for some time has been to successfully build my own car amplifier as well as get the info organized for others to do the same.

Now it is possible and I haven't been this happy since the first time I took a girl's bra off!

[Eva]

Instead of a 10 ohm resistor, a better solution is a 10nF to 100nF capacitor (as done in most amplifiers) with an optional damper resistor in parallel (>100 ohms). The purpose of this capacitor in any SMPS is to AC-couple primary and secondary side grounds at RF to prevent common-mode electromagnetic interference due to transformer winding capacitances, but allowing they to float at lower audio frequencies thus breaking the ground loop.

You should also couple the heatsink of the SMPS to the 12V ground (to the same point where input EMI filter capacitors are grounded) with a 100nF capacitor. This will reduce EMI further. A parallel damper resistor may also be added. I don't recommend connecting the main amplifier heatsink to ground directly because this may lead to high currents flowing through unexpected paths...

Have fun

[MartyM]

Thanks! I will be adding those options to the SMPS circuit board when I redo the schematic and reduce the PCB size by about 1/2 (or smaller if possible).

My whole point in this is not only to build my own but also to document this properly (BOM, photos, etc.) since information regarding building your own SMPS is generally pretty poor.

[Eva]

I could probably give you some more tips if you show me your layout. I'm currently laying out the PCBs for a very compact stereo full-range class D car amp