Hello CSandrelli,
Thanks for sharing.
Did you really use it without problems?Do you have pictures and/or description of your build?
I like to see them.
I am missing important correction networks, like explained by Tom at his site(Neurochrome).
And it is not a composite design with correction for lower THD+N.
Schematic From post 1 has only 5mV DC offset (one stage)so no adjustment necessary.
Bridged with two power stages DC offset below 1 mV. According to original builder.
Thanks for sharing.
Did you really use it without problems?Do you have pictures and/or description of your build?
I like to see them.
I am missing important correction networks, like explained by Tom at his site(Neurochrome).
And it is not a composite design with correction for lower THD+N.
Schematic From post 1 has only 5mV DC offset (one stage)so no adjustment necessary.
Bridged with two power stages DC offset below 1 mV. According to original builder.
Beware that a lot of "300 W" power supplies sold for audio use are really 100 W supplies that can handle 300 W overload.
Tom
Thats right, I had in the past the same experience. Best is to test it yourself.
And buy twice the current you need. So for this amp for one channel buying a 1000W +/- 35V smps is just enough when you are lucky. The cost are only 48 euro. Better is the 1500W version but there it starts at +/- 45V but that is to much for the lm3886. Another option is to buy two 700/800W single 36V smps. But that takes a lot of space in the housing.
And buy twice the current you need. So for this amp for one channel buying a 1000W +/- 35V smps is just enough when you are lucky. The cost are only 48 euro. Better is the 1500W version but there it starts at +/- 45V but that is to much for the lm3886. Another option is to buy two 700/800W single 36V smps. But that takes a lot of space in the housing.
According to LM3886 datasheet it should not oscillate.
As I have understand it, then is lower gain(more feedback) is lower noise and distorsion.
As I have understand it, then is lower gain(more feedback) is lower noise and distorsion.
I will start building the stereo unbalanced 2 x Parallel Lm3886. Because I have only 300W power supply. This has a gain of 21x.
Then I leave the LM3886 with 11x gain and the LME49710 does about 2x gain For the unbalanced 2 x LM3886 PCB.
Where did you read that?
The schematic I posted is a 4x3886 at x21 and no auto oscillation happens
I've been designing composite amplifiers with the LM3886 for over a decade now. I didn't read it. I based my statement on experience.
Of course not.
This is what we're talking about:
22k/1k yields a feedback factor of 1/22, corresponding to an inverting gain of 22 or non-inverting gain of 23. You're saying that the LM3886 is not stable below these values. Correct?
That statement is false. As indicated in the LM3886 data sheet, you can run the LM3886 at a gain of 10. We can dig a little further into the data sheet and find the open loop gain/phase curves. I've marked the closed loop gain of 10 (20 dB) and the corresponding phase margin (~65º).
65 degrees of phase margin is plenty. You won't even get an overshoot on the step response. Some may argue that you could push it even further, say, a gain of 4-5 (-> 45º PM), but I'd play it conservative here. Also, above curve is for a typical part. To allow for part-to-part variation, you really should not go below a gain of 10, so Rf = 10k, Rgnd = 1k, with the LM3886.
What is not covered in the data sheet is that the loop gain of the LM3886 tends to collapse as the VAS gets closer to saturation, which affects stability near clipping. There're ways to deal with this. See the AC Test Circuit #2 in the data sheet for example. Around here these components (Cc, Rf2, Cf) are known as the "optional stability components". If you want good performance with an LM3886 I recommend that you add those. However, in a composite amp, such as the one we're discussing here, those components can present a stability challenge, which is likely why they were omitted here. In my designs I chose to address this issue slightly differently so that I get both good stability and good performance near clipping.
Tom
I realize I'm late with this feedback: You really don't want the PCB to dangle by the LM3886s. You want mounting holes by the heat sink as well. Oh, well. Too late now.
Tom
Tom
Understood. I provided the feedback for the benefit of those who finds this thread years from now.
Some five-minute epoxy and either some standoffs or angle brackets attached to the heat sink would do wonders.
Tom
Some five-minute epoxy and either some standoffs or angle brackets attached to the heat sink would do wonders.
Tom