What to aim for in an amplifier designed for controlled bass

Jump to Latest
B

bdef

Member
Joined 2006
#1
What is it in an amplifier that makes it better to control the base?

I'm building an active three way, and currently I use LM3886, but I what my speaker to do better. I have a clear idea on how I want to do the amplifier for both midrage and tweeter, but what about the subwoofer.

I havn't measured the impedanse of the subwoofer, but it's a 4 ohm driver (SB23MFCL45-4) with two passive radiators, so I dont expect it to be easy.

I have four pairs of MJ15003/4 and I intent to use them in the amplifier. I have a choise between 2*18V/6A or 2*25V/6A trafo, but I rarely play loud.

Questions: Do I go for the bigger trafo to get more headroom?
Do I aim for lots of open loop gain and thereby lots of feedback, to get very low output impendance?
Or do I aim for the ability to (in theory) drive 1 ohm, to be shure to have current enough?
Og do I aim for low bandwith in order to make the fase margine in the "safe" aren to make it more stabile?
Do I aim for higher idle current, in order to output stage more linear?
 
#2
Chip amplifiers 3886 do not provide high current (see datasheet).
The current can be increased by additional transistors (class B). It will take a pair or more MJ15003/4.
You can use a Class B or D amplifier for the subwoofer.
 
#11
nironiro said:
"Do I aim for lots of open loop gain and thereby lots of feedback"
That's contradictory.
Click to expand...

Not really. It's just idiosyncratic terminology. Douglas Self's book, and many other sources in the field of audio amplifier design, often uses the expression "lots of feedback" to mean "lots of excess gain". The difference between open loop gain and closed loop gain (what EE textbooks would call "excess gain") is called "the amount of feedback" in audio amplifier discussions. When the excess gain is large, i.e., when open loop gain is much greater than closed loop gain, there is "lots of feedback".

For example an uA741 opamp has an open loop gain of about 77dB at 100 Hz. If it's used to make an amplifier with a closed loop gain of 20dB (i.e. 10x), then "the amount of feedback" at 100 Hz is said to be 57 dB. Figure below.

I recommend performing a mental substitution when you hear or read the phrase "lots of feedback". Replace it by "lots of excess gain, available to provide negative feedback"

_
 

Attachments

  • ua741.png
    ua741.png
    430.8 KB · Views: 217
#15
...the subwoofer, but it's a 4 ohm driver (SB23MFCL45-4) with two passive radiators -
Passive radiators make the subwoofer smaller. Good for the movie. Friendly for easy placement. But they have worse dynamics than a closed box. Less (don't)musical.
 
#16
Mark Johnson said:
Not really. It's just idiosyncratic terminology. Douglas Self's book, and many other sources in the field of audio amplifier design, often uses the expression "lots of feedback" to mean "lots of excess gain". The difference between open loop gain and closed loop gain (what EE textbooks would call "excess gain") is called "the amount of feedback" in audio amplifier discussions. When the excess gain is large, i.e., when open loop gain is much greater than closed loop gain, there is "lots of feedback".

For example an uA741 opamp has an open loop gain of about 77dB at 100 Hz. If it's used to make an amplifier with a closed loop gain of 20dB (i.e. 10x), then "the amount of feedback" at 100 Hz is said to be 57 dB. Figure below.

I recommend performing a mental substitution when you hear or read the phrase "lots of feedback". Replace it by "lots of excess gain, available to provide negative feedback"

_
Click to expand...

Exactly.

If one has an amplifier with a high amount of open loop gain, when the loop is closed, you end up with more available feedback you can apply without instability.

For tight controlled bass, if one were to pursue such a thing, you'd want a low damping factor this comes with lots of feedback. Beyond this you want the amplifier to be able to swing enough volts for your requirements and to be able to provide the required amount of current at the given output swing.

Even if you don't listen at high volumes the 2*25V transformer is the logical choice.

25V transformer = rails of 34V minus a couple of diode drops. Factor in some rail sag and output stage loss and you'll lose around 5V on the available output swing, perhaps a bit more so we're looking at 29V rails. Roughly 100 watts into a 4 ohm load.

One pair of output transistors would do this an actively driven subwoofer is not a difficult load to drive. Simulating a little the SB will only dip down to a hair under 4 ohms at the tuning frequency. Here the phase angle is close to zero, so easy to drive. I'm using a 20 litre box tuned to 30Hz. The worst you're looking at is around a 5-6 ohm load at a phase angle of 30-40 degrees. This is not a terribly hard load to drive.

To get the most out of the driver you would need more than 100 watts. A peak power of 300 watts would make more sense with an rms output of 200.

Then again you don't listen very loud so this is up to you.

To make the most of the hardware you've got available you could build a bridged amplifier. This would simply provide you with the voltage swing for transients should you ever need it. Two pairs of output transistors per side. If the voltage rails held steady with 34V you'd be looking at 400 watts into a 2 ohm load. This pushes the output transistors a tad too hard for my liking when you derate for a decent temperature rise. Luckily I don't think that would ever be a problem. The transformer you've got is only rated for 6A, the voltage rails will fall significantly as you approach higher output powers and will self limit the continuous output power as a result. You'll get short bursts of higher power but that's about it.

Given you don't want loud you could go the sensible route.

A standard, non bridged, power amplifier, use the 25V transformer. For increased reliability use two pairs of output transistors per side, and keep the Re resistor values as low as you feel comfortable going. Go for a complementary feedback amplifier to maximise efficiency and only use single pole compensation. The amplifier is only going to be used for low frequencies so two pole, or any other elaborate compensation techniques won't give you any more feedback where you need it.

This will leave you with two pairs of output transistors left over, giving you spares in case of an accident.

Place the amplifier close to the subwoofer to keep the cable run, from the sub to the driver, as short as possible.
 
#20
nironiro said:
"Do I aim for lots of open loop gain and thereby lots of feedback"
That's contradictory.
Click to expand...
Not at all.
High **open loop* gain means you will apply more NFB to reach the final amp gain.
Suppose final amp gain (as used, with NFB) is 20X.

You have 2 circuits, one has 100X open loop gain, the other 500X

When tamed down to 20X, the higher open loop gain one will be using 5X higher NFB than the other.
As in 14 dB more 😱
 
You must log in or register to reply here.
Top Bottom