• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Balanced to Unbalanced audio problem

Status
Not open for further replies.

jjmalove

New Member
Hello,

I need to make a circuit to do the following:

A modem on a dev board is spitting out a balanced speaker output, and receiving a balanced microphone input.
I'm plugging that into a USB Sound Card, this guy:
https://www.amazon.com/gp/product/B001MSS6CS/ref=oh_aui_detailpage_o03_s00?ie=UTF8&psc=1

It has a unbalanced stereo output for speaker, and an unbalanced mono input for microphone.

We get bad echo when trying to communicate. I opened up the USB sound card and the GNDs of the speaker and microphone jacks are tied to common GND, which I believe is unbalanced and the source of our problem. Our modem is expecting to be plugging into a balanced microphone and speaker.

FYI when a headset is plugged into the dev board and communicating these same lines it sounds great. What we are trying to do is instead convert it to a USB output to act as a sound card. This is a mandatory part of the process that due to the nature of the project cannot change.

So my understanding is somehow I can fix this using a transformer, but I'm not exactly sure how. Given this (and any more information I can provide if I missed something critical), does anyone know the step(s) I'm missing here to fix it?

Thanks!
 

audioguru

Well-Known Member
Most Helpful Member
I do not think balanced or unbalanced can cause an echo. The problem is probably caused by the simple fact that the mic can hear the speaker which produces an echo or can even produce acoustical feedback howling, but the mic cannot hear the headset that sounds great.
 

jjmalove

New Member
Let me take a step back. Maybe I'm overthinking the problem and there is a simple answer. If you could also just confirm my questions/statements I would appreciate this because audio is really not a strong point for me.

Our modem dev board has an RJ9 port on it. If we plug a phone into the RJ9 port it sounds great. No issues. The datasheet specifies this to be an analog signal for mic and speaker. This would be an analog output from the phone microphone to an analog input of the modem speaker. And an analog output from the modem microphone to an analog input of the phone speaker, correct?

The datasheet doesn't specifically say whether they are "balanced" or "unbalanced", however doing some measuring with a multimeter, none of the four lines, MIC+ MIC- EAR+ EAR- are shorted together, none are shorted to ground, and all have 100+K ohm or into megaohm range resistance between each other. This leads me to believe they are "balanced", correct?

So we want to take this signal and somehow convert it to a USB input into a computer, which is must be digital so an A to D conversion must take place, correct?

Going off the 4 solder joints for the RJ9 Port on the back of the PCB:
I took a 1/8" stereo jack cable for the microphone, cut an end off, and soldered the tip wire to MIC+ and the sleeve wire to MIC-. The ring wire is not connected to anything.
I took a 1/8" stereo jack cable for the speaker, cut an end off, and soldered the tip wire to EAR+ and the sleeve wire to EAR-. The ring wire is not connected to anything.

I plug them into the USB sound card adapter I linked and the computer recognized the sound card and we can interface with it through Linux. When we try to talk back and forth between the computer and the modem (using various phones to eliminate any other common problem), we get this really bad echo. I'm not positive but I believe when we unplugged the speaker the echo went away.

The only thing I can tell that could be the cause is that on the USB sound card adapter the mic and speaker 1/8" female connectors tie GND the sleeve wire together to a common ground. So what it seems to me is happening is then microphone and speaker's return line combine through that common GND and just screw up the signal.

Any thoughts here? Any glaringly obvious thing I'm doing wrong or misinterpreting.
Is there any easier technology to simply plug an RJ9 connector into the jack, convert the audio to digital, and it outputs to a USB plug and acts as a sound card that Linux can communicate with?
 

AnalogKid

Well-Known Member
Most Helpful Member
After two posts, why is the "datasheet" still a secret?
What modem?
What dev board?
What headset?

A modem is a modulator-demodulator, a device that takes a digital data serial or parallel bit stream and converts it to multi-frequency, multi-amplitude, and/or multi-phase tone bursts. You say it "sounds great", but that is not how it should sound. It should make your fillings hurt. What you are calling a modem might in fact be a hybrid or network interface, a *very* different thing. A hybrid takes the full-duplex (bidirectional) phone line audio and separates it into independent incoming and outgoing analog signals. This can be done with active electronic circuits, or with a single multi-winding transformer. What do you have? Manufacturer? Part number? Datasheet?

And, what is it you are trying to achieve? How about a block diagram that shows the signal types and flow among your system elements?

ak
 

jjmalove

New Member
Sure here is the modem in question:
http://www.kowatec.com/download/3G/SIM5320_Hardware_Design_V1.07.pdf

And the dev board we are using that it is installed on:
http://www.kowatec.com/download/3G/SIM5320_EVB_kit_User_Guide_V1.01.pdf

As for what we are trying to achieve there are some limits on explanation due to specific work reasons, sorry, but I will obviously share whatever I possibly can.

The "headset" we used that sounded great is just a coil cable phone with an RJ9 jack. The specific brand I think is HDphone. When we plug that in and listen to the call through the modem it sounds like you would expect, nice and clear.

The desirable function however is to take that audio signal from the modem, and instead convert it to something that a Linux operating system can see and manipulate in various ways. Likewise the signal we want to send to the microphone of the modem would be a manipulated signal from a Linux operating system.

Edit: Also, full disclosure. The team is, and has been, typically software guys not hardware. I am a senior electronics technician but junior level electrical engineer just trying to help get them a hardware fix to solve this issue. The project itself is about 95% software. We have a major lack of true EE's in our group and so just doing my best to figure out something maybe a little over my head.
 

audioguru

Well-Known Member
Most Helpful Member
You keep saying you hear an echo. Did you know that an echo is caused by a delay in the sounds? The delay can be caused by the sound coming out the speaker and traveling through the air to the microphone. With a headset replacing the speaker then there is no speaker for the microphone to hear so there is no echo.
 

AnalogKid

Well-Known Member
Most Helpful Member
For either the line out or speaker out, the P signal can be picked off, low-pass filtered, capacitor coupled, and single-ended amplified or attenuated as necessary to drive an A/D input. To drive the mic input, you probably need only drive the MIC P input single-ended with a signal that is capacitor-coupled and attenuated down to mic level (about 5 mVrms).

Edited to change mic level to millivolts.

ak
 
Last edited:

audioguru

Well-Known Member
Most Helpful Member
attenuated down to mic level (about 5 Vrms).
Mic level is a range of millivolts, not volts. A normal conversation level at 1m from a mic is 1 or 2mV. Normal conversation level at 10cm from a mic is about 10mV and "eating the mic and screaming into it" is a few hundred mV.
 

spec

Well-Known Member
Most Helpful Member
The audio input/output are described in paragraph 3.4.2 of the hardware design manual referred to in the documentation: http://www.kowatec.com/download/3G/SIM5320_Hardware_Design_V1.07.pdf

The audio output is class D, full bridge, like many speaker/phone outputs on portable devices these days. It is thus not surprising that an 'echo' is on the sound.

There are other complicating factors:
(1) The modem speaker output is mono and the sound card input is stereo.
(2) The voltage output level to the speaker may be too high for the sound card input.

My initial feeling is that a solution to the problem is quite simple. I will give it some thought.

spec
 

audioguru

Well-Known Member
Most Helpful Member
Ask Nigel Goodwin about the hifi sound from his Behringer class-D amplifier. Of course there is no echo.
My son-in-law has a very small Bluetooth portable speaker that amazes me. It is very heavy for its small size and it produces high quality music and voice sounds. It receives signals from his smart phone recordings. No echo.
 

jjmalove

New Member
I am thoroughly confused at this point haha.

Nigel
I tried connecting EAR- and MIC- to common ground but it still sounds the same.

I don't think the problem can be that the microphone is picking up the speaker's audio. We get the same "echo" problem listening to the call on a phone or through a laptop's speaker.

Nigel you said I have connected my output and input directly together? How have I done that and what am I doing wrong? Is my choice of sound card simply not going to work with these balanced signals coming from the modem?

I'm basically good to just rig whatever possible hardware wise to make this work, I'm just confused how to do so.
Bottom line, the end goal:
Modem talks to a Linux operating system as a sound card, and we can communicate back and forth.
 

spec

Well-Known Member
Most Helpful Member
You can get some very strange effects from class D and from overloading if you are not careful which the OP may be interpreting as 'echo'.

I do not know what Nigel's Beringer amplifier or any other amplifier, class D or not, has to do with this post. AG you seem to have a private mission about class D. I am not implying for a moment that there is anything wrong with class D, or that class D has any sort of echo per se.
But class D requires more care, especially with the filter-less chips which can have a load of HF on their output that may cause EMI.

I should have said in my post #12 that the output from portable devices: mobile phones, laptops etc, is very often by a full bridge class D driver running off 5V typically, and is used to drive the internal speakers/transducers.

Many of the semiconductor manufacturers have chips specifically for this purpose. http://www.analog.com/media/en/technical-documentation/data-sheets/SSM2518.pdf. I guess this is the type of output from the modem as indicated in the schematic of the referenced document in post #12

(The output (earphones) for portable devices is often a half bridge for each of the left and right channels, but using the same chip.)

When coupling the modem to the sound card it is important to configure the ground return correctly to eliminate problems. This and a suitable low pass filter was the approach I was considering.

I had missed your post #9 AK. Theses were my thoughts too.

spec
 
Last edited:

Nigel Goodwin

Super Moderator
Most Helpful Member
Nigel you said I have connected my output and input directly together? How have I done that and what am I doing wrong? Is my choice of sound card simply not going to work with these balanced signals coming from the modem?
I was under the impression that you had connected the two -ve connections together, which would feed the output back to the input.

I would suggest trying ignoring the negative pins entirely, connect just the +ve's to the sound card, and the gnd connection between the two gnd's.

Bear in mind the sound card output will be massively to high for the mike input on the modem, so will need attenuating.
 
Status
Not open for further replies.

Latest threads

EE World Online Articles

Loading
Top