GTA02A5: Headset Amp circuitry, max output power to headset
joerg.twinklephone at gmx.de
Sun Mar 30 11:47:40 CEST 2008
Am So 30. März 2008 schrieb TIM_LEE at fic.com.tw:
> Thank you for your kindly suggestion!!
> Lat week we just discussed how to improve the stable issue during GSM
> download stages. and the root cause is that R4405(33R) influences download
> and audio max volume. IF we changed R4405 from 33R to 0R, that would
> un-stable during GSM download.
Ooops, so much on estimating possible side effects in 1h of time at Saturday
late in the night. You're absolutely right: if we can't get to set output of
LM4853 to highZ by setting AMP_3.3V, AMP_SHUT, MP_IN to a decent state, it's
really only R4405 to save function of GSM-download. As mentioned in my
followup mail it's probably best to keep these 33R.
> C4110 & C4111 will also impact the low frequency response, base on 0603
> normal dimension, the typical capacitance value will be 4.7uf maximum.
I suspected this. Well, 4u7 will give us a somewhat acceptable f0(-3dB) of
<70Hz at least for connecting to highZ e.g. line in of home stereo.
For improving sound of headset connection though, maybe these C may be chosen
some special formfactor/specifications (goldcap? Lower max voltage?), or may
be mounted some non-standard way (small side down, PCB backside or piggyback
a second C...). Some 22uF are *absolute_minimum* for usage of Freerunner as a
MP3-player with selected headphones, 47uF a reasonable goal to be able to use
most standard headsets.
If this is absolutely not feasible at all, at very least we should consider
making it as easy as possible for user to field change a second C by him
tapping both connection rails.
> I will ask someone make some experiments using acoustic chamber in factory.
> Tim Lee
> > -----Original Message-----
> > From: openmoko-kernel-bounces at lists.openmoko.org
> > [mailto:openmoko-kernel-bounces at lists.openmoko.org] On Behalf Of joerg
> > Sent: Sunday, March 30, 2008 5:33 AM
> > To: openmoko-kernel at lists.openmoko.org
> > Subject: GTA02A5: Headset Amp circuitry, max output power to headset
> > Urgent Suggestion:
> > make C4110, C4111 100uF or at least as high as whatever my fit on PCB
> > (currently it is 1uF! in circuit diagram A5_1220)!!
> > make R4405, R4407 1R or 0R (is 33R now)
> > reason:
> > Avoid max volume of headset being too low.
> > Allow for usage of low impedance headsets.
> > Have better sound by not cutting bass with typical headset.
> > For multimedia usage, a good and loud sound is desirable.
> > Don't waste power.
> > Better adjust max volume by use of Wolfson mixer (+driver) than by series
> > (this is about FREEdom to kill your ears. If there are legal issues,
> > them by sw, not by hw design)
> > details:
> > Application info and datasheet do not demand any R, but clearly say that C
> > to be a certain minimum size.
> > with VDD=3,3V, the LM4853 has a typ. Po of 40mW to 33R in SE mode
> > (http://cache.national.com/ds/LM/LM4853.pdf page 4). However due to
> > R4407 we need highZ-headphones (>>33R) thus further reducing max Po.
> > Even with Z of earpiece = 32R there were only 20mW of 40 left to each,
> > half is heating device via R4405, R4407 (energy loss!)
> > LowZ-headsets won't work at all, 8R will get as less as 8mW with no bass.
> > With 1uF we might get weak bass anyway, even with highZ-headphones.
> > From datasheet, p5:
> > > Output coupling capacitor which blocks the DC voltage at the amplifier’s
> > > output. Forms a high pass filter(!!!) with the single-ended load RL at
> > > 1/(2π RLCO).
> > With e.g. *typical* (->ds, p.12) 32R-headphones, RL = 65 Ohm! Could
> > check the Z for OM-headset?
> > Please calculate f0 for this value!!!!!!
> > From application notes of LM4853, there's no indication for the defice
> > sensible to output shortcircuit, so any series R in output is useless.
> > This change should not affect any aspect of tested device stability, and
> > hopefully does not need a layout change (depending on size of C)
> > I really hope this will make it into MP!
> > cheers
> > jOERG
More information about the openmoko-kernel