LM3886 kit from AM

Status
Not open for further replies.
Yep, I got almost the same kit, and this is the right schematic. You should consider adding a 220pF capacitor between pin 9 and 10 on the LM3886. I've read that this should filter out potential RF noise or something. You can also find this in the datasheet for the IC.
I soldered a poly cap directly to the pins, under the PCB. If you want to read more about improvements to this particular PCB design, read THIS POST!
 
Thanks a lot! I saw a post about putting a 220pF cap across R5, but every schematic showed that it was connected through pin 9 and 10!

The kit we both own doesn't have a zobel network and an inductor at the output. Is this really necessary? Of course I'll add it if it's necessary 🙂


So this is the picture of a 20kHz square wave with a 220pF across R5, right?
295957d1344886261-lm3886-effect-compensation-network-cc-rf2-cf-img_2061.jpg
 
Last edited:
Yes, this is the picture of a C accross R5.

I believe what every schematic shows untill I see something else and then I believe my own eyes.

I do have a Zobel network and an inductor at the output. The inductor does only influence the output when you make the output capacitive, say with 2 microFarad. Then you see ringing and that is normal and it prevents the amplifier from oscillating.

They say that you need a Zobel network for a complex loudspeaker. I have never seen the difference - maybe because my speaker is not so complex - but it doesn't harm either.

Marc.
 
these are my current loudspeakers. I won't call them complex, but they have a crossover inside. SEAS Mimir

I'm not worried about adding a zobel at the output. Thats seems like an easy job. The problem is in the input side. Your schematic (in the post you posted) looks a bit different than mine, and I don't think there is possible to make the same circuit as you have.
the 20k resistor and 47pF capacitor seems to crash with R3 in the schematic in the third post 🙁 I also don't have the R* resistor in your schematic.


295892d1344862386-lm3886-effect-compensation-network-cc-rf2-cf-lm3886-cc-cf-rf2-1-.jpg
 
Please don't interchange my schematic with yours because it has other purposes! Therefore you must read my thread more thoroughly.

The only thing you have to do is to take your schematic and place the C of 220 pF between pin 9 and pin 10 and then put it over R5 and see the difference.

Marc.
 
The RF attenuation is the same.

But the impedance seen by the +IN pin is set mainly by R5 and R6+C5 in parallel.
If C6 is installed parallel to R5 then +IN sees R5||[R6+C5||C6.
This lowers the HF impedance significantly.
Most BJT inputs benefit from a low source impedance.

The Reason for removing C6 from the -IN pin is that capacitance on the -IN pin can promote instability. R4 should be mounted very close to -IN pin.
 
Thank's Andrew..
So, what's your best advice for RF attenuation ?
If I change C5 (1 uF) by 3.3 uF, how calculate the new input impedance ?
Below the pcb from Analogmetric with explained componants :

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


I think R4 is very close to -IN pin..
I don't know were is C1 C2 and C3, so I have appointed A B and C..
Phil.
 
The RF attenuation parallel to R5 as posted by Marc and myself !!!!

Reverse engineer the PCB
And find where the Power Ground (between the +ve and -ve Power Input terminals)
meets the Signal Ground (to the left of the Signal Hot terminal).
Find also what other components share the trace between Power Ground and Signal Ground. Particularly WHERE they tap in. Draw out a MAP of the GROUND trace with all the tapping points.
The designer almost certainly will have this WRONG !
 
Status
Not open for further replies.