Hi,
I'm just new here and I wanted to ask questions regarding my concern in my amplifier. I'm using TDA8950 as my power amp, my question is is it normal
that the temperature on the output filter goes up to 85degC?
The jetmelt melting point is around 85degC. There's no problem in the resistor / other components. Resistance didn't change a bit. But there are instances that I can hear sweeping noise(low-high freq) during stand-by.
https://s25.postimg.org/y3hytizkf/Amp_Output.jpg
Hoping for your insights on this.
I'm just new here and I wanted to ask questions regarding my concern in my amplifier. I'm using TDA8950 as my power amp, my question is is it normal
that the temperature on the output filter goes up to 85degC?
The jetmelt melting point is around 85degC. There's no problem in the resistor / other components. Resistance didn't change a bit. But there are instances that I can hear sweeping noise(low-high freq) during stand-by.
https://s25.postimg.org/y3hytizkf/Amp_Output.jpg
Hoping for your insights on this.
Not at all familiar with the Class D amps, but my suspicion from what you have said is that you have an oscillation. If you have (or have access to) an oscilloscope I would check the output for high frequency oscillations.
Tony.
Tony.
Hi Sir Tony,
I have 8 sample boards but only 3 of them have that specific stand-by noise. And ofcouse all of them have same output filter condition (melted silicon adhesive).
The noise was certainly one of the things that led me to the conclusion of oscillations, the lack of noise on some of the boards may not exclude it as a possibility either.
I've had output zobel resistors go up in smoke when an oscillation occurred in my class AB amp. a low level one I'm guessing could cause heating..
Perhaps you could post a schematic showing the configuration and values of the output filter.
Tony.
I've had output zobel resistors go up in smoke when an oscillation occurred in my class AB amp. a low level one I'm guessing could cause heating..
Perhaps you could post a schematic showing the configuration and values of the output filter.
Tony.
Actually all 8 units undergone 100 hours continuous ageing / operational test and the output filters are still okay even the silicon completely melted. My suspicion in the noise is perhaps the chemical reaction of the jetmelt on the components on top / bottom PCB.
Components with red mark in attached schematic are provision only for thru-hole.
https://s25.postimg.org/u8ekqygen/Output_Schematic.jpg
Components with red mark in attached schematic are provision only for thru-hole.
https://s25.postimg.org/u8ekqygen/Output_Schematic.jpg
OK I think I'll leave this for someone else, as that filter appears a bit strange to me (maybe normal for Class D) and is out of my range of experience!
I had been assuming it was the resistors that were getting hot, but on looking at the schematic I'm now wondering if it is the 680nF or 100nF caps??
Tony.
I had been assuming it was the resistors that were getting hot, but on looking at the schematic I'm now wondering if it is the 680nF or 100nF caps??
Tony.
Hehe based on the initial picture that i attached the excessive heat comes from the 2W resistors. Those 680nF / 100nF are SMT caps.
On the picture C206,207 cannot be seen, but I suspect they are ceramic capacitors. The designer wanted them to form a low pass filer with corner freq below switching freq, but instead of they started to resonate exactly at switching freq, this way output ripple was increased to a level that fried the resistor. The cause of bad resonance freq can be: decreased capacitance due to bad quality (bias voltage dependency), placement mistake, or wrong (less) switching freq.
Solution: mount C207B(C206B) also, let it be a poliethilene type.
Solution: mount C207B(C206B) also, let it be a poliethilene type.
The 2W resistors didn't burn since the resistance that was read still at 10 ohms.
There's only discoloration from melted adhesive and resistor color coating.
There's only discoloration from melted adhesive and resistor color coating.
Youre welcome! 
Do you really think this crap you wrote was an appropriate answer to my pelpful explanation?
Do you think an oxidized surface cannot be called burnt, and calling it burnt is reason to ignore everithing else?
Do you really think this crap you wrote was an appropriate answer to my pelpful explanation?
Do you think an oxidized surface cannot be called burnt, and calling it burnt is reason to ignore everithing else?
What I don't get is why there is a 100nF cap in parallel with the 100nf -10 ohm series combo... Also jumper 2001 seems to basically short out the top half of the amp...
My basic level of electronics knowledge says that most of the current will flow through the 100nf cap bypassing the higher resistance path provided by the same value cap with 10 ohms in series....
Tony.
My basic level of electronics knowledge says that most of the current will flow through the 100nf cap bypassing the higher resistance path provided by the same value cap with 10 ohms in series....
Tony.
Tony,
Youre right, the 100 nF should bypass the R+C at switching freq, therefore it is the main element in the LC LPF. 10 ohm + 100 nF is there only to damp resonance at corner freq.
J2001 must be not fitted in this configuration.
Youre right, the 100 nF should bypass the R+C at switching freq, therefore it is the main element in the LC LPF. 10 ohm + 100 nF is there only to damp resonance at corner freq.
J2001 must be not fitted in this configuration.
- Status
- Not open for further replies.