lm3876TF hots up too much

[peranders]

Any more then a really thin layer of compound actually reduces the thermal transfer to the heatsink.

Also ensure both surfaces and completely flat and free from any manufacturing defects.

Exactly In this case less is more but not too little. If you take too much the grease will creep and collect dust and durt and things get real messy.

As we say in sweden: You should take "lagom" with grease.

"lagom" is not too much and not too little, just enough. a in lagom is like a in chart.
http://www.websters-online-dictionary.org/translation/lagom

 

[quickshift]

As we say in sweden: You should take "lagom" with grease.

"lagom" is not too much and not too little, just enough. a in lagom is like a in chart.
http://www.websters-online-dictionary.org/translation/lagom

I'm going to adopt that as a special technical term for the exact quantity or grease that should be used. I'll run a very quick training session for the people that look after the hardware and explain it to them

 

[sagarverma]

hi rune,
u have raised a brilliant point about dc at i/p.one very imp thing(that might shed some light)when the amp is powered up but until its not connected to computer for audio signals speakers give out very horrible sound( i dissmissed it as charging of caps).but when i connect it to pc(for audio signals)(no matter after how many mins i connect it to pc after power ON),the amp lies dead silent until the signal is given to it to get to work.
i cannot test with oscilloscopes before monday(when i go 2 college),so is there any other way to test dc at o/p.
please tell me properly what is the method without oscilloscope bec. mon. is too far

for perander::
i have earlier posted the link to the pcb and circuit that i have used.i have also mentioned the changes i have made(gain res etc etc)so please check it out.i m uloading two pics of my finished amp.quality regreted bec of webcam.two pics,one to show the heat sink properly and other is closeup .

thanks in advance.

 

[peranders]

The shape of the heatsink is not particulary effective when it comes to cooling. If the area around the IC is pretty hot then you have pretty high "deg/watt" between the flanges and the mounting area. If this is the case it's not much to do I'm afraid.

 

[sith]

Well it seems u'we stuck till oscilloscope.... Mebe u can try implement some zobel into circuit... try to put some capacaitor between +/-input legs of chip cca.220p...
Do both chips(channels) have same problem???

 

[sss]

sss:lz. explain properly how to remove oscillations(ic is jus warm at no load ,i suppose u gettin me)

sorry

i've looked at your schematic , u got a zobel network at the output , right?
if so try to remove it , connect the speaker directly to the chips output .
if u got no zobel network then probably thats your broblem .

u can check DC at the output with a simple multimeter , no need for a scope .just connect the multimeter and measure the DC.

if u got no multimeter and u wanna know if theres DC at the output :
power up the amp , connect the imput signal to the amp , but put the music on pause , now when u connect the speaker do u hear a pop sound and u see the speaker muving a little ? if so , u got dc at the output

 

[sss]

changes in circuit from me.
no 10k trimmer.
r1 changed to 47k
r4 changed to 47k
10ohm 5watts res. used
12 turns of wire

ok, u got no 10k trimmer , u should use some kind of resistor
10K<R<30K to the ground .
yup , the gain is too high + u got no DC blocking capacitor at the input , thats bad , u should use one , with that gain the smallest dc at the input will be couple of volts at the output .

 

[sagarverma]

hi sss
what is zobel??is it the parallel lr network?if its that then i have it at o/p.
i have a multimeter and will use it to check the dc.anyways i think its(dc) the culprit.
i will change rf to 22k or midway from 47k and use 100uf cap at i/p signals.
will posts the results soon after.

hi sith
i have 100nf caps and 2200/50v caps from both rails to ground.

Do both chips(channels) have same problem???

what u tryin to say.i have only mono version here.plz see the pics posted by me carefully

peranders::::the same heat sink kept tda series(155q and 8563)warm.so i dont think that transfer factor of h.s is doubtable


will post the results soon.thanks to everyone.i am going to

 

[sss]

this is how your input should be :

An externally hosted image should be here but it was not working when we last tested it.

i've added the components "C" and "R"
C should be bigger or equal to 1uF and.....
10K<R<47K

 

[peranders]

sagarverma, did you have silicone grease between the IC and the heatsink? I can't see any white "kladd" in the pictures.

 

[runebivrin]

yup , the gain is too high + u got no DC blocking capacitor at the input , thats bad , u should use one , with that gain the smallest dc at the input will be couple of volts at the output .

Not quite so. C3 makes sure the DC gain is 1, so the output will follow the input volt for volt.

Rune

 

[sss]

Not quite so. C3 makes sure the DC gain is 1, so the output will follow the input volt for volt.

Rune

ohh , didnt notice C3
you are right

 

[sagarverma]

hi
THANK U ALL FOR EXTENDING SO MUCH OF HELP.THE PROBLEM IS RECTIFIED.

i applied grese on both surfaces and used quick glue to ensure that the ic is flat with the heat sink to max possible extent.this ensures that heat sink heats up in proportion to ic.i did all the changes to the amp as suggested by all,change the Rf,used 10uf i/p cap,and 22k i/p res. to gnd(as sss said,sss::thanks for dia. but i got what u meant to tell me.i tried all but to no avail,so i unscrewed the heat sink ,drilled requisite hole and remounted it on the ic with proper method of application of grease as told by quickshift.EUREKA!!the heat conduction was perfect.theamp doesnt heat up much now and is normal as i want it to be even at full volume..

now the final shape of the amp is just like as in the schematics with the changes that i told that i made earlier at first.
peranders:: in the pic,no grease is there.

i m posting the pics of final structure.thepcb has a bit of nail enamel too that was applied to get more track thickness.

THANKS TO ALL FOR HELP IN SOLVING THE PROBlEM


-------------------------------------------------------------------------
THANKS TO DIYAUDIO FOR ITS EXISTENCE.

 

[peranders]

So of this we see that the grease is important when much power must be transported away. Uneven surface is like insulation instead of conduction.

 

[sagarverma]

yeah it really is,i applied a thin layer on both surfacs that must have filled up the unforseen air pockets and made the conduction effecient.the circuit otherwise is perfect

------------------------------------------
THANKS TO DIYAUDIO FOR ITS EXISTENCE

 

[sagarverma]

an earthquake struk jus a min back.mild but terrifying!!!!!!!!

 

[Nordic]

Hey, as long as your amp survived!!!!!!!!!!

 

[sagarverma]

this was my state when earthquake stuk,mag. 3.2 here in delhi.lasted a good 1min.
AMP IS OK N FINE