TDA8950th still start up noise

Status
Not open for further replies.
Some years ago I bought the connex NXP 8950th modules with 6800 uF PSU caps. I use a 2x25 volt 300 VA trafo and use the amplifier in stereo.

These modules a known to have noise untill they settle in. That's why I havent used them for long. I know have soldes a 100 uF and later a 470 uF capacitor to mode pin/gnd as suggested by Cristi. But no luck. The noise are still there (5 seconds or so).

What next? Is the trafo simply high voltage or is the PSU caps to small - should I use 10.000 uF.

When the module have settled the noise is gone - also when I turn off and then on.
 
From the other TDA8950 thread - Cristis answer
"One of my customers just emailed me and told me what's going on here and I consider that I had to chime in.
First of all I thought that this issue was fixed 4-5 years ago, when you first told me about. We spoke both on the forum and by email and at a point you told me that the issue was fixed and was just improper wiring causing the noise. (I have the email conversations and there are not few)
here you can see that dozens of posts were dedicated to this issue only Class D Audio Amplifiers

Now before I will review the last developments and why is again not working, I just want to know why are you spreading negative information on the forum about our products and my person ?
You made an affirmation that these modules are known to be bad and many customers complain about them. Please bring some solid prof before the matter become serious. at that time I sold over 300 modules identical with yours and beside you, there are no other customers reporting any such issue. and this was the very first TDA89xx amplifier which I sold ! If they were so bad, I couldn't sell at least 1/10 of them and the forum would be full of such threads."


I have tested some more. I tried mute for several minutes, then unmute (with a switch), and as soon as I unmute the noise is there. If I mute again for some minutes and unmute the noise comes back for 5 seconds and then it's fine. The longer the amp is on, the less noise (1 sec). I have made a video that I will upload somewhere. The module runs completely cold, besides two capacitors. It's two small capacitors near the input pins. There are four capacitors in total, but only the two closest two the middel (next to the two metal screws) get very hot (untouchable). When I mute they get cold again.


I know other have same problem, I bought 10 modules and sold to others (group buy) They all had this problem. In the original thread Akilleuz also complaint about this, and there is another one at this forum (with no relation to me), and he also was told to try a capacitor to ground - with no luck.

I have build one class AB amp and four class D amp, and the only one behaving like this is your TDA8950th.
 

Attachments

  • TDA8950th small.jpg
    TDA8950th small.jpg
    81.5 KB · Views: 375
Last edited:
You want me to prove that these modules are known to be bad (not sound quality - just the small noise - I have never said this modules was bad sounding - actually the opposite)

In you own thread post #70 where you answer Akilleuz you point out that you will implement this. My conclusion is that the rev1 is a bad design, because it should not have this noise issue.

About your person - I really think you are very agressive to persons who feel that they have been treated bad by you.

http://www.diyaudio.com/forums/connexelectronic/158601-class-d-audio-amplifiers-2.html#post2437780
 
In you own thread post #70 where you answer Akilleuz you point out that you will implement this. My conclusion is that the rev1 is a bad design, because it should not have this noise issue.

I wanted to hear the opinion from users which bought them directly from me not from you, the ones which you resold. From ~300 modules if they were so bad, why the forum isn't full with complaints. One thing I'm sure of. The modules were working without any issue when were sold. I couldn't sell them other way.
All the issues reported plus possible issues which I suspected that can occur were fixed with the next batch available since end of 2010. TDA8950 Ampv3 Connexelectronic Are there any complaints about this product so far ? But you still insist that "these modules are known to be bad" although this issue was remote and very possible caused by poor installation. And the noise wasn't there from the beginning. with every new update in your installation got worst. Am I right ? check post#22.
Now that I saw the picture of the installation I'm almost sure that the cause of the noise came up when the modules were installed on the heatsink. The way how were mounted could cause the damage. I sold modules with and without heatsinks and the ones which I sold with heatsinks, installed by me didn't had this issue. And yes, I suspect which parts from the board can cause the issue.
 
I have just measured the DC offset (why didn't I do that before?).

Right channel = 20 mV - ok
Left channel = 117 mV - in my book not so good

I put a pot in between SGND and Mode pin. I can slowly adjust from 0 volt to 4,66 volt. At 2 volt the noise kicks in and at 3,2 volt the music starts at low volume and slowly goes to full volume at 4,66 volt.

If I play music for 10 minutes or so, I can leave the amplifier off for a minute without the noise at power up. If I leave the amplifier off longer, the noise is there again (also when I very slowly turn the pot from 0 volt to 4,66 volt).

C9 and C16 get very very hot. The coils get hot, but I can still touch them. I guess 45 degrees celcius or so. The left channel (high DC) output coil gets hotter than the right channel (low DC).

The heatsink only gets slightly warm - below 40 degrees celcius. The heatsink is NOT connected to chassis.

I have also tried mono but same result.

Cristi: Is it ok for C9 and C16 to get that hot? I doesn't seem to affect the sound.
 
Last edited:
...And the noise wasn't there from the beginning. with every new update in your installation got worst. Am I right ? check post#22.
Now that I saw the picture of the installation I'm almost sure that the cause of the noise came up when the modules were installed on the heatsink. The way how were mounted could cause the damage. I sold modules with and without heatsinks and the ones which I sold with heatsinks, installed by me didn't had this issue. And yes, I suspect which parts from the board can cause the issue.

The noise was there from the beginning. 50 Hs hum I got rid off, and the faint noise on piano notes was subtle. The only noise that have been there from the beginning and on all modules so far, is still there. I have tried three different heatsinks. My initial test setup was correct mounted. The final setup was wrong because i used plastic screws and heatsning was connecteded to chassis. My current setup is correct - all metal screws and heatsink is isolated from chassis (chassis is a peace of mdf). You write that you suspect which parts - but you don't specify these parts. You say you are almost sure that it's when the modules where installed to heatsink - but you don't specify further. I use the isolation (between chip and heatsink ) that came with the modules, I use all metal screws in all 7 holes on the PCB.

I noticed two very hot capacitors (I know it doesn't have to be an issue).

Please be more specific - I still have one unused module - and I'm preparing a heatsink for it. So I really wan't to di it right this time.
 
No help, so I tried today. The very last module, have almost no start up noise. The first second it's silent, then comes very little noise for very short time (I have a capacitor on mode pin to signak gnd). This is the best module so far. I have also bought some cheap chinese modules. They are completely silent, and I just swap out the modules. All modules have little distortion that is audible at solo piano (also reported by other using TDA8950). If only right speaker is attached and only left input (and vice versa) I have little signal going through. This signal is heavily distorted. I have tried with another transformer 2x12 Volt to see if the modules was overloaded. But same distortion. Very low input level also give this distortion, so it's not to high input either.

I have read about one module with TDA8950 chip, that did not have the low distorion on piano. Now I will try to put it in a metal cabinet to eliminate outside noise to get in.
 
Last edited:
I can't eliminate the faint noise on piano (the new modules), it's not audible on rock recordings anymore.

The best solution I have tried is to solder signal ground from the RCA to gnd on PSU capacitors (as seen in the LC audio cook book). Also RCA signal ground to chassis works but not as well as the first option. It seems that this chip is very sensible to noise.

I have never experience this kind of distortion from any other amplifier I have build.

In the LC audio cook book you can read that Lars Clausen also experience this kind of distortion on piano on first revisions of the ZAPpulse 2 modules. But I can't figure out if this trick can be done in either of my 8950 modules.

L C Audio Technology / ZAPpulse2 tips

Right click and translate to english gives a fine translation.
 
The heat is only from the two electrolytic capacitors. Everything else runs cool.
I only have diagram for TDA8920/24 but it shoould be equal to my TDA8950TH, only layout differs. I could try with a 12 volt transformer - and se if the capacitor still gets hot.

http://www.connexelectronic.com/documents/TDA892x_Audio_Amplifier_Module.pdf

I also have other modules with same chip. These have no noise or heating issue (except for the piano noise others have also reported)
 
It would be nice to hear from Cristi. I think he is the only one who can sort this out.

Especially this

"The way how were mounted could cause the damage"

How are they mounted wrong? 5 metal screws, heatsink isolated from chassis, amplifier chip isolated from heatsink.

And why does the mentioned capacitors run very very hot, and is it ok or not.
 
Couldn't resist to test once more with 2x12 volt - now with speakers. The last time it was only to check how hot the capacitors got with 2x12 volt transformer.

Big surprise - The start up noise is now completely gone. And what I never realised before. The PSU caps are rated 35 volt. So with a 2x25 volt transformer, we are on the edge here. It should have been 50 volt. The other TDA8950 modules I have also use 50 volt.

I do have some 50 volt i the drawer somewhere. I hope they fit, then I will test if it's simply underrated PSU caps that's causing all this fuss.

Well the faint distortion on piano is still there, but it will be a great subwoofer amplifier.
 
Status
Not open for further replies.