Headphone Out to Amplifier -- is there a way to modify the circuit to do this safely?

Status
Not open for further replies.

lemmings

Member
Hey everyone,

Is there a way to modify the circuit of the Audio Shield Headphone Output in order to be able to plug it into an amplifer? I know as it is it's only meant to go into headphones only, as it uses a virtual ground. What would it take to modify this?

Thank you!
 
Just make sure your amplifier ground isn't in any way connected to the teensy ground.
 
I'm making a teeny synth to eventually be available as a kit and it's out of my control what sort of configurations people end up connecting it to. I'd rather just fool proof the design so someone doesn't end up damaging the audio shield or external equipment. When I first got into synths when I was younger I did not understand the difference between headphone out or line out, so it's understandable that this would occur from time to time.

I've been reading the SGTL5000 manual as well as other forum posts or resources I could find. The manual has an application example with the 20 QFN version of the SGTL chip which just puts two 200uF capacitors on each of the headphone amp outputs and then the headphone jack is grounded, the virtual ground is left unconnected. Would this work? And would there be any considerations regarding phasing or distortion? Also, could this configuration have any negative impacts on the SGTL's headphone amp outs?

If it does work, would anyone possibly be able to explain to me why this works? I know this creates essentially a high pass filter circuit on each output, but does that manage to create a DC Bias also? Thank you for any insight you can provide.
 
You could use the lineout pin header which uses teensy GND

I'd like to have the ability to simultaneously use the headphones and line outs with their individual volume control the device. Also line outs won't drive headphones as well, it's better to use the headphone amp circuit.
 
The manual has an application example with the 20 QFN version of the SGTL chip which just puts two 200uF capacitors on each of the headphone amp outputs and then the headphone jack is grounded, the virtual ground is left unconnected. Would this work? And would there be any considerations regarding phasing or distortion? Also, could this configuration have any negative impacts on the SGTL's headphone amp outs?
I suspect there are implications for distortion and noise, especially for low-impedance load like a headphone - basically this is not
what the chip was designed for. Perhaps you need a different DAC. Stereo separation is probably compromized too, thinking about it,
the only way to find out is to try it and measure the performance with test-equipment...
If it does work, would anyone possibly be able to explain to me why this works? I know this creates essentially a high pass filter circuit on each output, but does that manage to create a DC Bias also? Thank you for any insight you can provide.
Capacitors block DC. You choose the caps to be so high in value that the high-pass action is of no consequence (say at 6Hz or so)
1000uF is a much better choice than 220uF BTW, for 32 ohm headphones.
 
If you're worried about damage, the problem isn't using headphone outputs to drive something you're not supposed (people do this all the time in the real world. Sometimes you break stuff, sometimes you don't). The real problem here is the virtual ground used by the SGTL5000.

If you really want to make sure someone doesn't blow up the Teensy Audio board by accidentally connecting a real ground to the headphone ground (something most people would feel is reasonable to do), then you would need to do something like put a transformer in to electrically decouple the headphone output. But this doesn't seem appealing as you would need two transformers (one for each channel). This would provide a safe ground on the output transformer that can connect to grounds elsewhere.

Have you considered switching to another CODEC? The wm8731 has line out and headphone output, and the HPOUT can use true ground. I used it on the board I created. For reference, including schematic, look at the datasheet here: TGA Pro

There are also other codecs with built in headphone outs and use conventional grounds.
 
would anyone possibly be able to explain to me why this works?

You can think of DC voltage as zero frequency. A high pass filter blocks low frequencies and zero is as low as frequency can get. (ignoring overly abstract mathematical concepts)
 
Thanks for all the input so far everybody, this is very helpful. Both Blackaddr and MarkT you are saying I should just use another codec that was designed specifically for this, I'll definitely look into other options for the next project I do, these are good points and suggested options.

However, setting aside what is ideal for best quality and avoiding any possible scenario that could damage the SGTL, but now more aiming for the "it's probably safe and good enough" territory: can I add two appropriately sized capacitors on the L and R headphone amp outputs, leave the virtual ground floating, and then connect the headphone jack ground to the 0V teensy ground like this diagram for the 20 QFN version of the SGTL5000? Is this mostly safe to do for protecting the 30 QFN SGTL on the teensy audio shield if headphones are plugged into external gear with different ground references? Or is the 30 QFN just a different beast and will not handle this configuration at all? I also don't really mind if there is some more audio degradation, just protecting the SGTL is mainly what I care about.

20QFN.jpg

Thanks again everyone!
 
Status
Not open for further replies.
Back
Top