The effective alternative to smaller input (Cin) cap is:
Larger NFB (CI) cap value or larger feedback shunt (RI) resistor value.
In my opinion you do need both warm (boom) and thunder (thud) simultaneously (for a wealth of differences at the bass and minimized amplifier sonic signature), so it would be erroneous to make a setting that would monotonously boom-only or monotonously thud-only. We want to aim for an attractive balance while avoiding extremes. On that schematic I would try a range of input (Cin) cap values from 0.68u to 4.7u to see what value is the most useful.
It is possible that a larger NFB (CI) cap would benefit that schematic, but I don't know if it would benefit that amplifier. 220u parallel with 220u of exactly the same model cap as used previously would allow a rather clean test for 440u value--perhaps deeper bass without much change to the tone, is what should happen.
Compensation?zakman35 said:
Feedback current can make a sort of intensity adjustment that is most noticeable at ear sensitivity peak of 2.2kHz, probably affecting the presentation of female vocals on music. You can adjust resistor values until the amplifier's harmonic distortion is either pushed down into the baritone or higher than the audio band. . . in either case avoiding the ear sensitivity peak.
Adjusting feedback current, adjusting gain, and RF filtering, when done for the purpose of reduced distortion, those are all an introduction to compensation adjustment.
Last edited:
Hi Daniel,
From the first paragraph of my quote to your post I can understand that when you decrease the RiCi time constant (or increase the corner frequency), you get more warm bass (lower/stroner bass?)
Then you said that when you make Ci larger, it passes more bass, wich looks more reasonable to me cause you let more bass frequencies pass.
But that means you increase the RiCi time constant, which is the oppostie as you said in the first paragraph.
I may have not understood you correctly.
There is also another thing i would like to understand better, if in zakman's schematic Cin is set to 10uF and the Rin is set to 10k, then the corner frequency is 1.5Hz witch means it passes all the bass frequencies, so why is it set for "clean (natural?)"
tnks daniel!
He's fine tuning it to fit system symmetry. That takes some exploration. I think that, "exploration" is the answer to most of your question. LM3886 is a television amplifier, and therefore textbook specs make a television amplifier. He wants to use it for a hi-fi instead.
And he succeeded!! Kudos!!! Despite appearances, that's not easy.
First, it is good to start with decent specs. And then we explore.
The datasheet's 22u NFB cap blocks bass and midbass by folding it back up into the midrange, which is going to get ever so useless. Instead of that problem, he started with 220u, a passable minimum size. It is slightly undersize. But the foldback from that 220u won't INCREASE the midrange. Instead, this bigger cap will increase lower-midbass. This allows one to try different sizes of input cap for different amounts of warm. If the NFB cap were technically correct size or larger (3hz, probably huge), then different size input caps could not greatly affect h1/h2 bass harmonic balance.
Please refer to AndrewT's posts on NFB cap size so you can get the technically correct size. Use for a road map. It is good to start with technically correct prior to exploring, else lost. After all, it is more fun to fine tune when you begin with high performance.
P.S.
Decreasing input cap size to get cleaner bass doesn't always work as we'd like, since we get a lesser amount of bass that way. An high performance alternative is output caps. With output caps, the amplifier is blind to subharmonic load and a well chosen size will encourage some very low bass in combination with increased headroom. That's the progressive current drive effect. Unfortunately, it is speaker specific and should probably be counted as a speaker crossover component, not an amplifier component. It is also easy speaker protection. Perhaps this alternative technique is best suited to active speakers or wherever the amplifier and speakers are made to match each other. I mention it for situations where you need more control over bass tuning, or perhaps there's some plaster that needs shattered with huge bass uncommon to chip amplifiers.
And he succeeded!! Kudos!!! Despite appearances, that's not easy.
Observe that his NFB shunt doesn't pass as much bass as the input on his design. That's the warm setting. More in than out is current bottleneck causing bass foldback (2nd order, warm). A tiny amount of this error may be nice, but a huge amount isn't.
First, it is good to start with decent specs. And then we explore.
The datasheet's 22u NFB cap blocks bass and midbass by folding it back up into the midrange, which is going to get ever so useless. Instead of that problem, he started with 220u, a passable minimum size. It is slightly undersize. But the foldback from that 220u won't INCREASE the midrange. Instead, this bigger cap will increase lower-midbass. This allows one to try different sizes of input cap for different amounts of warm. If the NFB cap were technically correct size or larger (3hz, probably huge), then different size input caps could not greatly affect h1/h2 bass harmonic balance.
Please refer to AndrewT's posts on NFB cap size so you can get the technically correct size. Use for a road map. It is good to start with technically correct prior to exploring, else lost. After all, it is more fun to fine tune when you begin with high performance.
P.S.
Decreasing input cap size to get cleaner bass doesn't always work as we'd like, since we get a lesser amount of bass that way. An high performance alternative is output caps. With output caps, the amplifier is blind to subharmonic load and a well chosen size will encourage some very low bass in combination with increased headroom. That's the progressive current drive effect. Unfortunately, it is speaker specific and should probably be counted as a speaker crossover component, not an amplifier component. It is also easy speaker protection. Perhaps this alternative technique is best suited to active speakers or wherever the amplifier and speakers are made to match each other. I mention it for situations where you need more control over bass tuning, or perhaps there's some plaster that needs shattered with huge bass uncommon to chip amplifiers.
Last edited:
OH! If you're trying to calculate this, you're going to get a caveat similar to percentage versus margin, which are different.
Here it is with caps:
At the filter point of the input cap, which is a roll-off, the NFB cap must have a flat response with no roll off. That's clean. It is the place to start.
Here it is with caps:
At the filter point of the input cap, which is a roll-off, the NFB cap must have a flat response with no roll off. That's clean. It is the place to start.
Last edited:
Daniel I finally got it, i also made calculations according to Andrew's posts.
You two helped me alot, I will come back soon to ask your opinion about my final design.
Thank you very much!
I'm curious to see your final design. I'm currently working on a 3886 design and this thread has some good information in it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.