STM CircleOS forum / Audio part constructed wrong on primer2

fad_electric · 2009-10-13 03:05:35

I managed to put some basic waveforms on the audio output.

So i hocked it up on my oscilloscope. And it doesn't look to nice.
There is more hi-pitched noise then there is signal.

If I look at the schematics it looks like the engineers have skipped
some of the decoupling capacitors.

The most important one (1uF) recommended by ST, is missing.

fad_electric · 2009-10-13 03:39:51

The signal is about 0.15V peak to peak, and the jitter is about 0.23V peak to peak.
So the signal with the noise layered over the signal is about 0.38V

How can we patch the circuit board so this enormous jitter disappear?

Francis · 2009-10-13 07:46:26

There are a 100nF and a 10uF capacitors on Vcca. Onmy Primers, the sound is good with a headphone, but quite bad with the (small) loudspeaker.

fad_electric · 2009-10-13 14:50:37

Of course you cant hear it, I guess you ain't a bat, with ultrasound hearing.

But this signal is useless to send to other equipment. This isn't sound!
This is something to fry eggs on your tweeters with.

Look at the frequency response on the chart for the speaker...
http://www.kingstate.com.tw/eng/Product … ItemId=131

I guess we can signal to bats with it :-) The frequency response travels upward when it hits 20khz.

Probably the speaker pics up the Jitter. So thats why I want to puke when I use Primer2 :S

I think this may drain the battery. And thats probably why its so sensitive and wants to sound like an angry Asian giant hornet http://en.wikipedia.org/wiki/Asian_giant_hornet when the signal is to high.

The question still remains, how do I fix it? Does Raisonance pay for the hi-fi equipment I blow up then I use it. I have a hi-fi store next to me. They have the best hi-fi in my country in the $100 000 range. Maybe I can take this with me and blow out their windows. And cook some french eggs on the top of the amplifier.

Anyone who is hungry who wants to come with me?

fad_electric · 2009-10-13 16:48:50

Francis, If you look at the datasheet. You can see that you probably have done everything wrong with the decoupling for the STw5094A.

On VCCA it should be 1uF to GND, and 100nF to GNDA.
On VCCP it should be 10uF to GND, and 100nF to GNDP.
On VCCIO it should be 100nF to GND.

1 right of 6 possible doesn't do it!!!

You cant mix the ground planes and power supplies on an audio codec. All the jitter will get all mixed up.
Thats why you have more jitter than signal!!!

I don't think engineers on ST designed the decoupling this way because it would look fun on the schematics, as you have done in your 'frying french egg' construction. Fun but ain't right. And probably ain't that fun either.

repzak · 2009-10-13 16:57:34

*Plong*

fad_electric · 2009-10-15 04:09:33

OK!
I'm getting tired of this jitter.
So I took a closer look at it.

It seems that it's about 2,4Mhz.
And as mentioned before it has an amplitude
that is higher than the actual sound.

I don't know if it's to any help for anyone
to know this. But that's how it is.

If I plug in headphones to it, could it work as an antenna?
Maybe I could open a new radio station in my basement?
That would be awesome!!!

hello all list....................if this radiobroadcast sound stran.........................
then this is bec...............can't record and transm................the same time.........
.....some one might fi.....................that would be........................Thanks!........
CU Next time ...............................bzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz

repzak · 2009-10-15 06:26:07

Hello,

When you are so fast in blaming everything you are completely sure your test setup is correct?, like your oscilloskobe are correctly grounded, your PC is correctly grounded (if you measure with USB cable plugged in) and so on.... Try measure when it runs on battery...

Instead of all this complaining you could try to do some about is also, like youi already mentioned about capacitors, it is not so hard to add 100nF across C31 to see it that helps!!

Do like the other of us, help to improve this evalutaion platform, or buy some more expensive products there are made for youre use.

I am happy about raisonance can provide cheap hardware to play with cool hardware!

Kasper

fad_electric · 2009-10-15 08:43:02

Cool! down... Im just reporting so that we can fix this. My radio show was very lighthearted.

The 2,4Mhz signal is still there when I disconnect it from the USB. But it changes. When it' s disconnected (running on batteries) then it looks like a saturated sinus waveform. It gets more distorted when connected to USB, but the amplitude is the same. And the signal is still there when you mute. It's hardware related, and independent of software.

There is nothing wrong with my oscilloscope. It's from the "cool hardware" all right.

Maybe I could test with some different capacitors. I don't know if I have the right ones at home.

repzak · 2009-10-15 10:59:16

The SMPS switch frequenxcy is 1.4Mhz... could be related to this.

fad_electric · 2009-10-15 11:33:50

I have no program to read gerber files. I tried some free programs but I could not read them. I guess that I can't be to much help then.

However, the via between the audio codec and the LCD (on the corner of the chip pointing towards the joystick, has a signal that is 5 times the clock rate of this unwanted signal in the audio path). The (~2,4Mhz) signal goes to the speaker also.

This problem is easiest solved by using the original PCB files.

Francis · 2009-10-15 12:40:41

The schematic had been validated by the ST guys working for the audio device (STw5094A). I don't think that it is so bad. Not perfect, but for sure not as bad as you describe. For sure the loud speaker saturates (it is more a hearphone than a loudspeaker) and the sound output on the loud speaker is poor (I guess it's why it has been disabled on the MP3 application). We tried with larger loudspeaker from the same manufacturer and the result is much better. Now, if you connect the headphone output to an external (PC-audio) audio equipment (PC-loud-speakers), and you launch the MP3 application (from Kasper), you should have a good quality sound (at my side, I don't distinguish between the PC music and the MP3 player running on the Primer).

Jitter could be generated by the regulator. For months, the main issue we are encountering is around the regulators. The L6928D regulators are very sensitive at the input, especially sensistive to ESD peak. The input voltage is limited to 6V max, and small peaks could alter them. It could be that for your primer, one of the regulators (probably U17 in your case) has been already altered. When 'altered', the regulators efficiency is dramatically low, and the regulator consumes half of the current. In such a situation, it would impact the audio signal.

repzak · 2009-10-15 12:53:29

Also i think, even then there are some 2.4Mhz noise, and you put it into an power amplifer, the bandwith is so low that it is filtered away.

If you want to use it for other purpose i suggest you on the cable add a first/second order filter, where you have plenty of room to filter that signal away (in case in generates problems)

http://www.pentalogix.com/Download/download.shtml

Free gerber viewer

Kasper

# 1 2009-10-13 03:05:35 Audio part constructed wrong on primer2

Audio part constructed wrong on primer2

# 2 2009-10-13 03:39:51 Audio part constructed wrong on primer2

Re: Audio part constructed wrong on primer2

# 3 2009-10-13 07:46:26 Audio part constructed wrong on primer2

Re: Audio part constructed wrong on primer2

# 4 2009-10-13 14:50:37 Audio part constructed wrong on primer2

Re: Audio part constructed wrong on primer2

# 5 2009-10-13 16:48:50 Audio part constructed wrong on primer2

Re: Audio part constructed wrong on primer2

# 6 2009-10-13 16:57:34 Audio part constructed wrong on primer2

Re: Audio part constructed wrong on primer2

# 7 2009-10-15 04:09:33 Audio part constructed wrong on primer2

Re: Audio part constructed wrong on primer2

# 8 2009-10-15 06:26:07 Audio part constructed wrong on primer2

Re: Audio part constructed wrong on primer2

# 9 2009-10-15 08:43:02 Audio part constructed wrong on primer2

Re: Audio part constructed wrong on primer2

# 10 2009-10-15 10:59:16 Audio part constructed wrong on primer2

Re: Audio part constructed wrong on primer2

# 11 2009-10-15 11:33:50 Audio part constructed wrong on primer2

Re: Audio part constructed wrong on primer2

# 12 2009-10-15 12:40:41 Audio part constructed wrong on primer2

Re: Audio part constructed wrong on primer2

# 13 2009-10-15 12:53:29 Audio part constructed wrong on primer2

Re: Audio part constructed wrong on primer2

Board footer