You have to power digital part of csr module through galvanicaly isolated dc-dc transformer. I'm sure this is right way to eliminate digital noise in es9018 dac . Regarding to "pop" sound. I use symetrical output, all 4 signal lines are connected to gnd through 10k resistors to avoid hight impedance on amplifier input. Unwanted sound is more than noticable. POP apears when module state goes from a2dp to connected. I have to check signal routing options in Csr chip.
I had very loud pop sounds when I used only one phase per channel of CSR 8645 audio output. Since CSR 8645 audio output is differential like AK449x or ES9018 dacs, you have to use an active low pass filter to get unbalanced R L output.
With CSR 64215 I can hear very tiny electronic noise when it goes into standby, but it is not an issue, even though it is strange to hear that noise, since I am listening with ES9018 dac over I2S and why that noise is sent over I2S I do not really understand.
BAT and GND shorted
Hello, I finally get managed my CSR to work. I was playing music with it but suddenly FTDI turned itself off, my computer get bluescreened and i discovered that BAT and GND are shorted. I really need to get it to work to 04.07 so i don't have time to buy another one. It is my second because my 1st came damaged and did not work at all. FTDI was only power supply so i don't know what could make it happened.
Hello, I finally get managed my CSR to work. I was playing music with it but suddenly FTDI turned itself off, my computer get bluescreened and i discovered that BAT and GND are shorted. I really need to get it to work to 04.07 so i don't have time to buy another one. It is my second because my 1st came damaged and did not work at all. FTDI was only power supply so i don't know what could make it happened.
I believe Doctormord answered this many months ago, twice in fact... although I could be wrong. If you look at the output on a scope you'll note the output suddenly changes as the output amp is put into power saving mode (about 3 seconds after music has been stopped, or paused). If you disable 'low power codec' then I believe this will keep the DAC and output amp on, so no pop.
A better option would be to use a GPIO from the BT module to mute the output amp - most chip amps have pop suppression when muted/disabled/enabled.
Pop/clicks are always the bane of audio which is why manufacturers go to great lengths to provide soft mute functions, charge pump negative supplies (to eliminate large output caps on headphone drivers) and the like.
I am still having issues with 'electrical noise' on the outputs (32 ohm headphones), despite using a differential to single ended opamp for each channel (TS922), and powering the output amp with BT module separately (two batteries, not sharing a ground). It is of course far more noticeable when they share the same power supply, but I was surprised that even just with balanced lines between the BT module and amp, there is still some low level noise that is only noticeable between tracks.
It changes as my hand is moved near the balanced connections, so it seems it is more radiated that conductive - makes sense given its a crappy breadboard proto. I am going to try using a MAX97220B with short screened leads to see if that cures it. It needs a power amp anyway because its just a bit too quiet on 32ohm phones.
You say 'FTDI was my only power supply' - it isn't a power supply. It has an on board 3.3V reg but that is only rated for ~50mA. Without a schematic of your setup we can't know what has happened, or what, if anything, is damaged.
It sounds like you may have fried your FT232R, which may have taken out the bluetooth module, but really.. its anyones guess without a hookup diagram.
I've disabled 'low power codec' but it didn't help (I wrote about it some time ago). Output buffer/amplifier switches off after 3 seconds regardless of setting "low power codec".
I do not see any difference with the low power codec on or off. I've also tried to use mute in output amplifier (trigerred by a2dp stream) but the end of streaming state is in the same moment as pop sound (this solution unfortunately does not work). I recently tested the second F-3188 with CSR8645 (from another provider). "Pop sound" is much quieter but noise (white, not digital) is not acceptable (damaged module ?).
Why are you using differential to single ended conversion? Try to make the whole signal path balanced (bridged headphone amplifier configuration, you need 4 opamps with high output current). It should help.
I do not see any difference with the low power codec on or off. I've also tried to use mute in output amplifier (trigerred by a2dp stream) but the end of streaming state is in the same moment as pop sound (this solution unfortunately does not work). I recently tested the second F-3188 with CSR8645 (from another provider). "Pop sound" is much quieter but noise (white, not digital) is not acceptable (damaged module ?).
Why are you using differential to single ended conversion? Try to make the whole signal path balanced (bridged headphone amplifier configuration, you need 4 opamps with high output current). It should help.
For testing, I am using headphones which share a common ground, and haven't yet got PCB's made up for testing various headphone chip amps (most of which have differential input). So my options were to cut the headphone cord and connect the headphone speakers directly (probably the best way) or prototype a headphone amplifier with differential input. I have lots of TS922 to hand, and they seem reasonable for 32-ohm headphone amps, so I knocked up one on vero board. It's single supply (large output caps) so, not ideal but it was just for testing the BT module.
I suspect many of the problems people have is that they simply wire their amp/speakers to L+ and R+, and then use either L- or R- as the common ground which not only defeats the point of a balanced connection but also means one channel has the other channel's reference. Add to that poor power supply design and its no wonder people are getting interference/noise/static.
I do not have access to particularly expensive audio gear, and have no intention of buying any, I just wanted to test the quality with Apt-X over standard bluetooth (SBC) and despite the noise issues, I'm impressed. Certainly good enough for listening on reasonably cheap over-the-ear-headphones. I mostly listen to MP3's anyway, I haven't been spoiled by having 'ultra quality' music.
I just want the convenience of wireless/bluetooth headphones.. but having control over what headphone speakers to use (that is what lets down so many bluetooth headphones), what features, battery life, any bells and whistles etc.. DIY belutooth headphones ticks all those boxes, even if it is a pain to sort out.
I shall try that later on this evening. I haven't 'played' with the settings in a while, and there are so many it is easy to get lost. I vaguely remember from the datasheet/API documentation that there were several user config keys that controlled the DAC and output amp, where-as the 'config app' only has one tickbox for 'low power codec'. I'll see what changes.
Buriedsode. Which power rail are you using for powering balance-single ended converter (common with power amp ore BT module ?) What module are you using for DC-DC isolation? Some modules are not really isolated although in the description states "isolated". I also had digital noise problems but now there is complete silence.
I'm powering the BT module separately from the Headphone amp - both running from batteries. Amd I'm using the F-3188 V2.0 with breakout board (just has power-on enable and caps, nothing else).
Only connections between the two are the 4 audio lines. I do have some audio isolation transformers somewhere, and will try those as a last resort. It should be noted that when headphone speakers are connected directly to the module, I hear very little noise, perhaps only when powering up/down (acceptable). But as the volume isn't that high, it is hard to tell whether there really is any background 'bleeps'.
I'll try the 'low power codec' thing now, and post my results for you
I am hoping it is just my shoddy test circuit - which is a discrete differential to single ended high current opamp design for driving headphones (TS922, gain of 2). I didn't bother using precision resistors, hoping that 1% match would be enough to pretty much eliminate common-mode noise on the inputs.. but it doesn't seem to.
I'll test it with a MAX9722 board, again with its on power supply.
With headphones speakers connected directly to the bluetooth module - balanced, not sharing a common ground - I can't hear much interference, but.. those particular headphones don't have the greater high end response. I've got a very cheap pair of 'ear buds' that I'll cut and wire directly.
I'm not judging the sound quality with these, just that they are so 'tinny' that any high frequency interference will be audible. If its clear, then I know its my shoddy diff-SE circuit.
I am still willing to accept that it is the bluetooth module itself that is at fault - they are very cheaply made so I didn't expect 'perfectly quiet' devices and I didn't pay much so I don't expect much.
Well, after listening for about half an hour I think I can say that using earbuds and/or proper headphones when wired directly (speakers directly to module, differential, no common ground) I cannot hear the interference. If its there, its too quiet for me to hear even when the module is on max volume.
So, thats a good thing. But the output is still a bit low for my liking - thus the need for a headphone amp. Whilst I wait for the MAX9722 board to arrive, I'll be scratching my head as to why the interference (which I'm now sure is common-mode noise) manages to be so loud on the output of my little headphone amp.
I still hear - very faintly I might add - the pop that slavj is on about, but another advantage of using a chip headphone amp is that these often have pop/click suppression and a mute input, which would be handy. Frankly the pop when stopping/starting music doesn't bother me, it was the bleeping noises between tracks that were ticking me off.
So, thats a good thing. But the output is still a bit low for my liking - thus the need for a headphone amp. Whilst I wait for the MAX9722 board to arrive, I'll be scratching my head as to why the interference (which I'm now sure is common-mode noise) manages to be so loud on the output of my little headphone amp.
I still hear - very faintly I might add - the pop that slavj is on about, but another advantage of using a chip headphone amp is that these often have pop/click suppression and a mute input, which would be handy. Frankly the pop when stopping/starting music doesn't bother me, it was the bleeping noises between tracks that were ticking me off.
Hi,
I bought 2 of these modules (F3188 V2.0, CSR 8645) from China and wanted to test them before I use it in more complex project.
I can't get them to work. For my project I don't want to use battery, just constant 5V. So I'm connecting just 2 pins- GND (bottom left) and VBUS (6th from bottom left) to my lab power supply.
The module then takes about 18mA for a second or two and goes to 0mA.
When I short POWER_EN pin (3rd from bottom left) to HIGH it goes to 18mA for about 5 seconds and then returns back to 0mA. During these 5 seconds none of the LED blinks and it's not discoverable by my smartphone.
I tried to power it with 3.7V on BAT pin (4th from bottom) with exactly the same results.
I also tried to power it from both- VBUS and battery at the same time, connecting the battery to BAT pin and BAT_SENSE pin.
I tried to connect POWER_EN to 5V, 3.3V, 1.8V.
All these attemps lead me to the same results.
So either both of these modules are faulty (unlikely) or I'm doing something wrong.
I appreciate any help.
I bought 2 of these modules (F3188 V2.0, CSR 8645) from China and wanted to test them before I use it in more complex project.
I can't get them to work. For my project I don't want to use battery, just constant 5V. So I'm connecting just 2 pins- GND (bottom left) and VBUS (6th from bottom left) to my lab power supply.
An externally hosted image should be here but it was not working when we last tested it.
The module then takes about 18mA for a second or two and goes to 0mA.
When I short POWER_EN pin (3rd from bottom left) to HIGH it goes to 18mA for about 5 seconds and then returns back to 0mA. During these 5 seconds none of the LED blinks and it's not discoverable by my smartphone.
I tried to power it with 3.7V on BAT pin (4th from bottom) with exactly the same results.
I also tried to power it from both- VBUS and battery at the same time, connecting the battery to BAT pin and BAT_SENSE pin.
I tried to connect POWER_EN to 5V, 3.3V, 1.8V.
All these attemps lead me to the same results.
So either both of these modules are faulty (unlikely) or I'm doing something wrong.
I appreciate any help.