Ground noise?

[dteck]

Hi everyone.
I'm having a little trouble with a small project I'm working on. I'm trying to make a portable audio module. I'm trying to use an rn52 bluetooth module or one of the cheaper Chinese alternatives.  And I thought I might have some problems when I tried to breadboard the circuit and I got a horrendous popping and clicking noise on the speakers.
A little searching told me that breadboards are not good for audio testing since they let in a ton of noise. So I started hooking parts up directly and things quieted right down.
But now my problem is that I want to do battery powering. So I picked the tps61202 boosting switch mode power chip. It just needs a couple caps and a tiny inductor and you are good to go. And everything works great with fresh batteries. But once they start to run down a bit I start to get clicking on the speakers again.
The audio amplifier I'm using is a class d pam8403.
The speakers are some 2" 3watt 4 ohm Dayton audio speakers I got from parts express.
if I power things from my bench supply, usb port, or a cell phone charger everything is fine. It's just the batteries. So I'm wondering if maybe it's switching noise that's being picked up as the batteries run down since the amp isn't differential.  Or if I'm missing something.

As a side note I don't think it's a fall off of current output as Vin drops since the clicking stops when you start playing audio.

Sorry for the sparse detail I'm on mobile. I can post some pictures of my setup when I get back home.

[T3sl4co1l]

Bad grounding / routing / layout / filtering?  Absolutely possible...

You didn't indicate if you made any measurements or observations (or what "clicking" is), but I'd be willing to guess the controller is in a low power (bang-bang) regulation mode, and this is causing modulation which is picked up by the amplifier.  Normally, a controller has a buzzing or whining sound when in this mode, but perhaps it's doing it slow enough that it sounds like a "click".  The amplifier, in turn, is disturbed by RF energy coupled in by poor layout or something like that.  The amplifier itself should be very well filtered and grounded, because it is producing RF as well, which needs to be contained as well as possible to keep offset and distortion low, prevent oscillation, and radiated emissions.

Tim

[dteck]

Hi Tim,
The amplifier is one of those cheap Ebay finds. eBay auction: #281430483882 so it is possible its got a bad layout. but when I use a power source like a benchtop supply or a USB port or cell phone charger I don't get any of the clicking. So that's why I suspected it could be switching noise from my boosting supplies.
Now that I'm home I took some measurements using two different boosting supplies that give me the popping sound. One is the TPS61202 that I laid out and the other is a no name Ebay booster. These are from the output pins on the power supply boards.  It looks like both are are dipping from 5v down to 2v at 9-9.5 Hz (the no name brand the time div was 25ms, and the TPS time div was 50ms).
Do you think adding a cap on the output would help with the clicking?

Also I'll upload a video here in a minute so people can hear the problem.

[codeboy2k]

Just a thought.. can you try to measure the bluetooth current at the same time ?

It's one possibility that the radio is doing something at a 10hz rate, i.e. transmitting data, then it will drain more current from the switchmode converter and this will drop the voltage if it cannot supply the needed current.

There is only a 10uF capacitor on the output of that boost converter, so I'd experiment and put a 100 uF there instead.
And also you can experiment with putting a small 10uF cap directly at the power terminals on the bluetooth radio too.

But it would be good to know if the 10hz brownouts are correlated to 10hz current draw from the bluetooth device. So if you can measure the current there on the scope on a second channel that would be good.

Probably you don't see it with the bench supply simply because the bench can provide the needed current during these 10hz pulsed loads.

[dteck]

Hi codeboy2k

I thought I would try and see if you were right about the rn52 drawing too much current. So at first I removed it from the circuit. Then I found that if you touch the PAM8403 board the clicking would come back. I added a few wires to the L and R inputs (used just USB breakout I had on the desk. The wires should be isolated they are used for power and ground) and found that if I touched those the clicking would increase.
I also found that If I touched the TPS61202 board while the speakers were clicking in this setup the noise would stop.
Ill link in a few videos I took of the different tests.
The clicking on the no name booster also shows the output indicator led flicker in time with the audio clicks.

Clicking with TPS61202
https://youtu.be/cutqze_G17s

Clicking with no name booster
https://youtu.be/cqOK17c8mG4

No Clicking USB/Wall wart
https://youtu.be/6wWpVodM18U

Clicking with no audio input - finger press
https://youtu.be/IU6ZggMllv4

Clicking and silencing TPS61202
https://youtu.be/vkF07dqoGYc

Clicking no name - touching USB
https://youtu.be/G-OAEeXEcow

[codeboy2k]

It only shows me that you've clearly got power related problems.  Those long wires add inductance to the power leads, and your touching the boards is adding parasitic capacitance, which either stabilizes it and ends the clicking or destabilizes it and starts the clicking. 

You need to monitor both the current and voltage going into each component, and also coming out from the batteries, to see where the power problem is originating. 

You seem to only get clicking when the batteries are used, never the USB port or a wall wart or a bench power supply is used. This tells me that the current requirements are not met by the batteries, they can't supply the instantaneous current needed and the voltage drops.  Maybe you need a larger input capacitor at the DC-DC converter board. There is only 10uF there now, make it 100uF, so it has more bulk capacitance to handle the transient currents.

If you don't have the equipment to make proper measuremements, then you're just guessing and grasping at straws.

As a final test, try more batteries in parallel.