Troubleshooting noise problem on a bluetooth audio receiver

[victor]

http://tinypic.com/view.php?pic=a4rmrl&s=8

I picked a couple of these cheap bluetooth audio receivers on eBay. I could anticipate that these would not be state of the art by its price 14 bucks shipped. But I thought I would give a try. The device identify itself as PT-810.

My intention was to use it on my car, so hacking one of these inside the pioneer that has aux in and USB would be very easy.



So here comes the problem, they work but with a awful noise. But surprising it was very quite when I powered it from a 3.7v li-ion battery (I soldered a battery directly on a USB female). Powering from my MacBook USB port also proved that these can sound good.


I tried bypassing the hell out of it, on a breadboard without much success, I played around with various caps values from pF to thousands of uF. Without good results.



I took some macro shots of the dongle itself, sorry I couldn't get the hole board in focus on a single picture (I made a composition of various shots)


Anyone could me help how to make these usable?

[HalfSpace]

Hi victor,

What does the noise sound like?

Is it a hiss or a buzz and does the sound change when you touch the receiver with your hand?

Try connecting it to your pioneer using the 3.7v battery but also connect just the negative side of the USB. It could be an earth loop problem.

HalfSpace 

[victor]

This is how it sounds like.


Sounds like data, when it goes to standby, it does bip-bop every second or so.
This is what I got so far,
- dongle powered by the stereo usb and the audio cable going to the aux port: NOISE.
- dongle powered by battery or macbook usb and the audio cable going to the aux port: quite, near full volume you can hear very little noise.
- dongle plugged on the stereo usb without the audio cable connected: nothing.

this makes me think that the noise is coming from the dongle audio signal, because if I disconnect the audio cable the noise stops.
The stereo adc is not picking rf noise, because powering it from another power source would make no difference (unless the rf is being picked by the ground, but it would pick noise without the audio cable).

my first guess was that the stereo usb was poorly filtered, and the noise is going through the dongle DAC power rails and superimposing the noise in the audio signal, so powering it from a battery should stop the noise, as it did. And the noise could come from the Pioneer uC multiplexing the lcd or the usb host controller, but the timing of the noise matches more with the bluetooth status than with the stereo.

touching the circuit doesn't change the noise, but when I was playing around, I connected a usb cable power and ground to the power rails on a bread board and then plugged the dongle to the breadboard, I connected like some capacitors of 1000, 440, 220, 10, 1 0.1uf like then of various values in parallel, and made very little difference, tried to add some inductors and a big ferrite bead, not much improvement. But a weird thing is that if I moved everything around with the breadboard and long cable about 1meter long, the noise changed in amplitude, not in timing, or frequency.

the best result I got so far was to connect a 1000uF cap directly to the usb port, and reduced the noise significantly, not by the amount I want but it is something.
But the problem is that I can't like add like 5x1000uF caps to a dongle to make it work, it would be such a ridiculous solution, and of course it wouldn't fit anywhere.



I know that bypassing and decoupling are different things, I'm not a EE, and I know that these monopole rf radios is highly dependent on it's ground plane. so dealing with decoupling bypassing, rf propagation, ground plane, ground loops, local grounds, when mixing all that stuff together you can get in all sort of weird behavior that is very hard to predict.

[peter.mitchell]

I think this may be the problem;
The bluetooth receiver is physically very close to the input of the stereo when it is plugged it to the stereo for power, but is distant when plugged into the macbook so it appears to be radiated EMI, with a small amount of conducted EMI.
To me this looks like a combination of lack of shielding and high impedance inputs on the stereo.

How I would try and fix;
Add a 4.7k resistor between gnd and L, and a 4.7k between gnd and R somewhere on the cable, preferably closest to the stereo, but anywhere should help (possibly even on the jack on the bluetooth board). With the bluetooth thingo outside of its case, wrap the board in tape to insulate it, then, tack a tiny bit of uninsulated wire to the GND pad on the usb "fingers", wrap that around the board a few times just to hold it in place, then, put the usb connector housing back over the end and tack solder the ground to the shield, then wrap the entire thing in aluminium foil, and stuff it back into its case. It should be pretty well shielded now, and the resistors should make it a bit less susceptible to noise.

[HalfSpace]

Hi victor,

Try running the audio receiver using the 3.7v battery but also connect just the GND side of the USB to the pioneer USB port. If the noise disappears then the noise is from DC current pulses being drawn from the USB port when audio receiver transmits back to the audio transmitter.

HalfSpace 

[victor]

This time I used a portable speaker, it is a lot easier than testing this inside the car, the speaker have its own battery.

Powering each from separated batteries and just connecting audio, no interference what so ever. But when I connected the negative terminal to a ground point on the speaker, awful NOISE.

At least it doesn't seems to be irradiated noise, because moving it closer to the speaker or covering the whole thing with my hand doesn't seems to affect it.

I'm giving my best to troubleshoot this, I will try any suggestion.



Is that a ground loop issue, how can I fix this?

Edit
I quite doesn't understand this phenomenon,  the audio gnd is the same ground of the USB inputs, so jus connecting audio cable already make the speaker on the same ground potential, so the extra wire to ground shouldn't make any difference at dc perspective, again I don't fully understand what is a ground loop.

I also found out that the circuit have 2 voltage regulators, one from 5v->3.3v and other 3.3v->2.5v they are far apart. The main circuit seems to run at 2.5v, I tried bypassing the 2.5v rail with larger caps from 1uF to 1000uF made no difference, again the same thing on the 3.3v made little to no difference, and the when I connected a 1000uF in parallel to the 5v input cap it reduced the noise a lot.

[bgsteiner]

That noise you are hearing is high frequency noise from the IC either from EMI or bad isolation on the UART that generates the audio signal. You would not be able to correct this with capacitors. As for how it sounds when plugged into he macbook it probably has to do with Radiative EMI vs conductive which is lower. and with your last test the noise is going to be prominent on the power side due to a lack of isolation so having the separate power source helped with the mac book.

[HalfSpace]

Hi victor,

I have done some LTSpice simulations to try to explain the noise problem with your Bluetooth module. M1, V2 and R2 represents the modules transmitter being turn on and off drawing current from the power supply and the displayed waveform is audio output measured across GndSig and AudioSig.

Bluetooth Noise1.jpg is a LTSPICE model of your last video with the USB ground wire connected to the battery connection. A noise waveform voltage at AudioSig of 9mV peak to peak is generated by the voltage drop across R3 (power lead resistance) and R7 (PCB resistance) when the Bluetooth transmitter turns on and off.

Bluetooth Noise2.jpg is a LTSPICE model with the USB ground wire connected to the PCB. The noise waveform voltage has now dropped to 1.2mV peak to peak because you have removed the voltage drop across R4 (power lead resistance) and now only have the voltage drop across R7 (PCB resistance). In your video the noise level was less at that connection.

Bluetooth Noise3.jpg is a LTSPICE model of your pioneer stereo connection to the Bluetooth module. Note the audio waveform amplitude is about 11mV peak to peak. The problem is R5 which represents the internal ground connection between the audio connector ground and the USB ground in the pioneer stereo but it cannot be removed.

Bluetooth Noise4.jpg is the only way I can think of that will solve the problem. Using an isolated 5v to 5v convertor will remove the R5 ground connection issue. The link below is just one many DC to DC convertors that are available. You will need to measure the peak current draw of the Bluetooth module to select the correct convertor required.

http://au.element14.com/recom-power/r1s8-0505/dc-dc-unreg-1w-1kv-05v-05v-smd/dp/1793128

HalfSpace 

[victor]

I still don't quite understand why this is happening, but I experimented with it recently and could get it working reasonable well. The noise still there but at some acceptable level. My first intention was to hack it inside my car front panel stereo.

Again I'm no EE engineer, but I did my best to do a clean job, like using the shortest path possible, I added a ferrite beads to both vcc and gnd, and added a low-pass decoupling filter. I have only one ground point going from the module to the amp, I just didn't had any space for shielding or adding a dc-dc isolated converter.

I would like to receive any critique and judge what I did.

I will write a guide on instructables later.

My low pass filter is a 10ohm resistor and a 220uF electrolytic cap, the cap was the largest value I could fit, and the resistor was the only value I had that was low impedance. I think I could go higher impedance, but I'm not sure

The ferrite beads was from an old diskette drive, I tried a single pass and 1 turn.

My antenna I'm very proud of it, I think it's pretty neat, I used a savaged laptop wifi antenna, the antenna was too big so I just used the tiny cute coax, I soldered it and cut the end of the coax about a 1/4 wave length of 2.4ghz, I could fit it on the top most part far away from the rest of the board. The range was a good 5 meters, from outside the car with the doors closed.

[Radek77]

Hi Victor, I also have used the same Bluetooth PT-810 bluetooth receivers in some audio projects and intially experienced all the problems you describe in here as well. The problem with this particular dongle is that audio-output-gnd is not at the same level as the power-supply-usb-ground. The only good solution to the problem is to galvanically isolate those grounds. One solution is to use an isolated DC/DC converter like "SIM2-0505s" to power the bluetooth receiver or to use an "Audio Ground Loop Isolator" on the audio-output from the device, but these guys are typically much larger. There could possibly be another option to use some differential op-amp solution to convert the the L/R signals relative to the noisy A-GND from the receiver and bring them back to be relative to the power ground (but it's also more work). So in my case I have chosen the isolated DC/DC converter option for powering this dongle. Btw. did you find any other simplier final solution (as I can see your post is a bit dated already)?

[victor]

An isolated DC/DC converter would probably work, because powering it from a battery solves the problem. But I had no space to fit one, nor I had one laying around. A low pass filter did not eliminated the noise completely, but to a point that it was acceptable, and only audible at max volume without any song playing. Next time I would probably make it to a heavier filtering like with 2-order LCR filter, I would choose the largest cap I could possibly fit, the increase the impedance while vcc still stable and had enough current to operate. It doesn't draw much btw.

I forgot to post my tutorial here http://www.instructables.com/id/Add-bluetooth-to-your-car-stereo/?ALLSTEPS

[advancednewbie]

I have a XS3868 module off eBay.  I had the exact same noise problem.  Completely solved the problem using an op-amp in differential configuration between the audio source and audio ground prior to amplifying further  (must be done with each left and right signal separately).  No more noise   Thanks all.

[timotet]

@advancednewbie, care to show us the circuit you used?

thanks

[Brumby]

I'd be interested too.  I have a PT-810 that I tried for a BT upgrade on an old juke box - but had the same problem.  I, too, suspected some isolation issues but I couldn't justify any more effort on it at the time.

Might revisit it...

[advancednewbie]

Hello again,

So here's a quick sketch of my audio circuit made up of LM386 op amps.  There are a few things I would have done differently - and that is to replace the first op amp on the left to have two separate difference amplifiers for the Right and the Left channels and change the input configuration a little on the second op amp to make it more like a weighted summing op amplifier configuration.

Hope this helps.

Cheers,

Warren

[advancednewbie]

I would really like to hear your opinion / reply on the how the circuit I posted is going.  Right now the range in which I can connect is just terrible.   But I am trying to make a shower radio - so of course, I have a highly sealed project box that may be limiting the range.  Anybody have a perfect solution?

[leereyzhe]

Tried your circuit, it didn't work so well. Although, I probably connected something wrong since I didn't do a through job.

Anyway, this module really sucks, has been giving me so many headaches. 
but I finally figured out how to eliminate 99% of the noise, at least on my breadboard setup. I am using a 12V power source, and a LM836 op-amp, with a switching step down module set to 5V for the PT-810.

Do not connect the audio - input of the LM386 to ground. connect it directly to the audio - output of the PT-810 module. also, I put a 1000uf cap between +5V and GND right before the power inputs of the PT-810.
this will kill 95% of the noise, and it is only noticeable when nothing is playing.
Then I connected another wire directly from the audio - output of the PT-810 to near the step down module. pretty much eliminated all the ground loop noise. 

I can post a schematic and pictures if needed.

Now I just have to deal with the annoying static white noise. 

[dwi2rzq]

Tried your circuit, it didn't work so well. Although, I probably connected something wrong since I didn't do a through job.

Anyway, this module really sucks, has been giving me so many headaches. 
but I finally figured out how to eliminate 99% of the noise, at least on my breadboard setup. I am using a 12V power source, and a LM836 op-amp, with a switching step down module set to 5V for the PT-810.

Do not connect the audio - input of the LM386 to ground. connect it directly to the audio - output of the PT-810 module. also, I put a 1000uf cap between +5V and GND right before the power inputs of the PT-810.
this will kill 95% of the noise, and it is only noticeable when nothing is playing.
Then I connected another wire directly from the audio - output of the PT-810 to near the step down module. pretty much eliminated all the ground loop noise. 

I can post a schematic and pictures if needed.

Now I just have to deal with the annoying static white noise. 

Could you post your schematic and pictures for us? Thanks

[ruiseixas]

Hi Victor, I also have used the same Bluetooth PT-810 bluetooth receivers in some audio projects and intially experienced all the problems you describe in here as well. The problem with this particular dongle is that audio-output-gnd is not at the same level as the power-supply-usb-ground. The only good solution to the problem is to galvanically isolate those grounds.

So, why you don't just connect them together? Why audio-output-gnd and power-supply-usb-ground aren't connected?

[bgdan]

Thank you Radek77 ... your solution worked perfect. I tried yesterday to integrate the same Bluetooth to my desk speakers and got the exact problem with the noise. Read your reply and used the DC-DC convertor and problem solved. Thanks!

[victor]

The low pass solution worked for me quite well, I didn't touched it anymore and I still use on a daily basis. But I really don't understand the basis of the problem, solved it more or less by trial and error, and settled when it was good enough.

What is the problem of the module itself, what should they have did to make it "just work"?
I know superficially about the ground-loop issues on long path, and you can use the star configuration to avoid it and so on... but where is loop in this case? and how using a isolated supply does solve the problem?

*sorry for bringing up the thread, but I have been REALLY busy for the past months, I'm just catching up the unread replies.

[adras]

I'm wondering what would happen if you connect the USB ground to the audio output ground. Could that also solve the issue?

[dr_p]

Quote from: victor on 11-10-2017, 01:04:50>[...] But I really don't understand the basis of the problem


the BT module consumes high current in bursts (what you hear as noise), this current obviously flows through the (power) ground.
Wires have resistance, even if very small, so there will be a voltage drop on the ground wire proportional to the module current usage.
The audio signal is generated at the module side, referenced to module ground, but keep in mind this module ground is moving up/down proportional to current flow.
The Player receives some audio data that is referenced to a variable ground, so it reads this a variation in audio signal.


For example, an exaggeration:
Player supplies 5.00V, module consumes bursts of 1A, in those 1A moments, module receives 4.90V (100mV lower), 50mV drop on the +5V wire, 50mV on the GND wire (current goes in series through all components). So module GND is now 50mV higher than Player GND. Module generates a 200mV audio signal ( referenced to module GND), but Player receives a 250mV audio (referenced to Player GND). So you just inserted a 50mV "noise"


Quote from: victor on 11-10-2017, 01:04:50>[...] how using a isolated supply does solve the problem?


when using an isolated supply, these GND voltage drops will still occur, but before the isolator. After it, the module power GND is isolated from Player power GND, the only point where they are connected is Audio output ground. But this will not have massive amounts of current going through it, so it will not generate voltage drops, so no noise will be added.


Adding large capacitors near the module also helps (as you found out), because those current burst are drawn from the capacitor, not through the Player-module GND wire, so module GND stays more stable, so noise goes away


Also, you can use an audio isolator (a transformer basically) that takes that 200mV generated by the module and galvanically isolates it , so that the audio output GND is not connected to power GND anymore, so it will not be moving up/down with current flow, and noise goes away