View Full Version : High-quality headphone audio solution for a Propeller app?

Lord Steve
12-13-2009, 03:32 AM
Currently I am planning on using a Microchip MCP4922 DAC to sit between the Prop and the audio amp stage.· However, the amplifier IC I wanted to use has gone into the "Last Time To Order" phase.· All I need is high-quality stereo·headphone output.· Is there a well-known beloved amp out there that will run from a single 3.3V supply?· Do I really need the DAC for high-quality?· What are some high-quality alternatives?

Thoughts are appreciated.· I'm·not an analog kinda guy.

Phil Pilgrim (PhiPi)
12-13-2009, 04:45 AM
The stereo audio quality emanating from the Propeller Demo Board is really pretty good, and it doesn't use a DAC. You might consult the Demo Board schematic for part names and component values.


Lord Steve
12-13-2009, 06:22 AM
That was a good tip.· Thank you.· That amplifier IC only comes in a "space-saving", impossible-for-Steve-to-hand-solder package.· Can anyone recommend another similar IC which comes in a little bigger SMD package (or one with leads)?

Is there a thread or documentation about how the DAC part of the audio path works?

Lord Steve
12-13-2009, 06:51 AM
Is it just using PWM?

Phil Pilgrim (PhiPi)
12-13-2009, 07:02 AM
It uses DUTY mode output from the counters, which is similar to PWM, but not the same thing.


12-13-2009, 07:35 AM
You may want to check out the MAX4410: www.maxim-ic.com/quick_view2.cfm/qv_pk/3459 (http://www.maxim-ic.com/quick_view2.cfm/qv_pk/3459), it comes in a 14-pin TSSOP package.


Piranha are people too.

Lord Steve
12-13-2009, 08:10 AM
Hey, lonesock.· Thanks for that tip, too.· It's even in production.· :D· Now to wrap my head around the differential duty cycle stuff.

Is the DIFFERENTIAL submode of the DUTY counter mode used in Chip McGracey's microphone_to_headphone.spin solely to provide TWO audio outputs (cuz the electret mic on the Demo Board is, after all, monaural)?· If there is an already-written explanation about this, please point me to it.· This just looks like PWM to me, but it's some different DUTY thing...idk.

12-13-2009, 11:50 AM
DUTY mode is PWM, but not in the usual form.

Normal PWM has a constant frequency and only the High-Low ratio changes.

At DUTY mode, the shorter Pulswith (High or Low) is always 1 clock cycle, and the other state (Low or High) is extended,
so that you get the right pulswith. That is the frequency changes with the pulswith.
On 50%:50% you have the highest frequency (40 MHz at 80 MHz system clock!). At 10%:90% you have a frequency of 8 MHz
and at a 16 bit resolution the end values +$7FFF and -$7FFF have a frequency of 4.15 kHz.

This makes this mode ideal for Audio, because quiet signals are around 50% with high frequencies, and loud signals cover the lower PWM frequency.

But there are also downsides: A DUTY output produces a lot of RF noise in the air and on the print.
And High-Quality is very relative. If you have single notes and low signals you will hear some digital noise. If you play music this is normaly not noticable.


Lord Steve
12-13-2009, 12:15 PM
I figured it out!· Thanks for your pointers.· You saved me ~3$ off my BOM.· :D

Those counters are too cool.· I have attached a .spin program (based off of Chip's microphone_to_headphones.spin) which samples the propeller's built-in sine tables at different angular steps to generate two samples.· These are added together and given to a DUTY mode counter's FRQB at a rate of 44100 Hz.· In my opinion, the playback is high quality.