LM3886TF Gainclone prototype questions

[ObliVion]

I have finished a prototype of my amp using LM3886 chips. I now have some questions. First of all the mute resistor (the one that goes to pin 8) is too small because i did not have the calculated value around. In the datasheet it says that it should draw at least 0.5 mA and mine is drawing 0.7 mA will this break anything?

The amp has been idle for 4 hours and the heatsink is just about 40 degrees celcius, is this normal, and if not could it be the above mentioned resistor (don't think so myself). The reason I get all upset about the temperature is that I've read in serveral places that this chip does not get hot. - Just measured them again 37 degrees and the amp has been playing at low volume for half an hour.

I am impressed by the sound, real nice, and I've used no special components and standard mains wire. Nice! Can't wait till I get some quality cables and a case.

I've attached the schematics.

Thanks.

[ObliVion]

Here is a picture of the poor amp crawing a nice case to live out it's days in As can be seen the heatsink is a little small, but this I'll fix.

Thank you for this nice forum I very much enjoy reading the posts here. Without this forum I would never have gotten this far

Thanks.

[memorex]

hey oblivion,
I've built an amp using four lm3886. Have you had a chance to look at the application notes that national has? I used their BP design. Those notes might give you some more insight.

As for the mute circuit, .7ma wont hurt anything. Infact, i would increase it to 2ma for clearance space.

Your amp sounds like its producing a lot of heat at idle. That's not normal. That or your heat sink is too small. Could you post a picture of your heatsink?

[ObliVion]

Here's the picture. The heatsink is 25x4x2 cm. I think it is to small and will get a bigger one.
Still the chips are getting hot. I've read that they did'nt even feel hot after twelve hours of idle. Mine get hot with no signal, and probably a bit colder when playing at low volume.

Could the chips be oscillating?

[pro]

How many volts is the +- rail?
Ciao.

[squadra]

Originally posted by ObliVion
I have finished a prototype of my amp using LM3886 chips. I now have some questions. First of all the mute resistor (the one that goes to pin 8) is too small because i did not have the calculated value around. In the datasheet it says that it should draw at least 0.5 mA and mine is drawing 0.7 mA will this break anything?
I've attached the schematics.

Thanks.

Hello Oblivion,

Looking at the circuit the current would be more like 10 mA (30V / 2k7),
If you have used a smaller value then the current will be even higher.

I'm going to use 18k with a +/- 32V supply.

According to the docs a 220pF cap between pin 3 & 4 (the inputs) can be used to prevent oscillating, you can try that.

Peter

[ajitm]

Hi,

I have just put togather my LM3886TF chip. What I observe is after just half an hour of idle time, the heatsink heats upto around 40 degrees ( no measurement just a guess ). My circuit is based on http://www.diyaudio.com/forums/showt...950#post410950

I have added 100uF cap on the mute pin.

I have a 220pF between Vin+ and Vin-. Supply rail voltage is 30V. Grounding done as per guidelines in the data sheet. DC voltage at output is around 150mV. Now I want to know

a. How to detect and prevent any high frequency oscillation ?

b. Is the 150mV DC at output harmful ? If yes how to reduce it ?

The sound however is really good and almost no hum at zero input.

Thanks.
Ajit

[memorex]

Hello everyone,

So everyone here is using only one LM3886 chip right? If you are using one chip, then figuring out what is causing the heat should be straight forward. The datasheet says it should be drawing less than 50mA when idle with 30volt supply rails. That's 1.5 watts of heat. So either you have a really light heat sink that cannot dissipate 1.5 watts heats without raising 15 degrees, or you have oscillations, or you have a bad dc offset on the output or... _____, or a combination.

To figure out if there is oscillation, just hook an oscilliscope. Ground it on the speaker ground (at the speaker), set it to .5volt scale or less, touch the ouput and different ground lines throughout your circuit, and pan through the different time scales. Oscillations with this chip are normally caused by grounding loops. If you aren't star grounding and/or if you are mixing preamp and line level grounds with power grounds, then this could definitely be a problem. There are lots of things that can cause oscillations - bad grounding, insufficient power supply filtering, long thin connection wires, poor pcb layout etc etc.

I'm thinking 150mV DC output with only one chip is way off. This is a high powered op amp and should be precise. I'm thinking it shouldn't be more than 10mV max. Do you have a dc blocking capacitor in the feedback? Could you post your exact circuit as you have it right now? To check if the dc offset if causing some of the heat, disconnect the speaker and let the amp idle.

[ObliVion]

I've used the value stated (2K7) it is just that my calculations showed that I needed a higher value.

Can anybody else comment on how hot the chip gets when idle?

pro:
The voltage (as shown in the schematics) is +/- 30V

[ObliVion]

When measuring at the output with my meter set in the mV range it reads no higher than 1.0!
I will leave the amp with no speaker connections to see what happens.

About the ground arrangement. I simply connected it all in a star between the PSU smoothing caps. I understand that I should be dealing with two separate star grounds one for power (including speakers, and one for the signal, and that these stars should be connected by a single wire. Is this correct?

Thanks