What to choose ?

ToG · 2004-11-12 2:12 pm

Well, hello all. First, please ignore my poor english, and just imagine i'm writing in a very good way ... :xeye:

I'm trying to find how i can make my speaker (4 ohms) fall in love with the new lmxxxx based power amp i want to build.

I've searched a lot on this forum but still can't decide wich one is the one : lm3875 (buffered unbuffered .....) lm3886, lm3886 in parallel .... and the psu that goes with it.

The main problem is the curve below, i'm not sure it will fit with the tons of different lm3875 i've seen ... don't want to burn anything

and on the other way, i'm not sure to need something else.

So please take a look at the impedance curve of my speakers and let me know, what can i do ??

all your suggestions will be very appreciated.

ToG

carlmart · 2004-11-12 2:43 pm

ToG said:
Well, hello all. First, please ignore my poor english, and just imagine i'm writing in a very good way ...

I'm trying to find how i can make my speaker (4 ohms) fall in love with the new lmxxxx based power amp i want to build.

I've searched a lot on this forum but still can't decide wich one is the one : lm3875 (buffered unbuffered .....) lm3886, lm3886 in parallel .... and the psu that goes with it.

The main problem is the curve below, i'm not sure it will fit with the tons of different lm3875 i've seen ... don't want to burn anything

and on the other way, i'm not sure to need something else.

So please take a look at the impedance curve of my speakers and let me know, what can i do ??

To start with I would not think of the 3875 to drive a 4 ohm speaker to rather high levels: think 3886.

Start with one 3886 per channel, with a +30/-30 supply and then see what happens. If you don't get any clippings or strange noises for the levels you listen to then question is solved.

If you do then you can do two things:

1) Parallel two 3886 for each channel.

2) Go for active amping, using the two sides of the LM4780 to drive each channel bass speaker, keeping the 3886s you had for the tweeters or mid/tweeters (if your box is 3-way).

That should take care of it.

Carlos

ToG · 2004-11-13 12:32 am

Your explanation is ok for me, very clear and direct.

Thank you, i have now to study the forum to find the shematics i will use.

Do you have one in mind ? maybe the datasheet one ?

carlmart · 2004-11-13 2:50 am

ToG said:
Your explanation is ok for me, very clear and direct.

Thank you, i have now to study the forum to find the shematics i will use.

Do you have one in mind ? maybe the datasheet one ?

Now you got me. There should be a few schematic options to consider.

Nuuk's site is a very good place to start, particularly because he has been following many alternatives that came up on this forum and applied them. Have a look:

http://www.decdun.fsnet.co.uk/gaincloneindex.html#top1

Even if he uses the LM3875, practically the same things apply on the 3886. The latter only has an extra resistor and capacitor for the mute.

He is also very old fashioned and uses point to point wiring... but that is not catching.

Study your choices and implement them carefully.

Carlos

soundNERD · 2004-11-13 2:55 am

I would use a parallel 3875.

Although in parallel, much simpler due to lack of mute circuitry.

Also, better sound.

Parallel would drive a 4ohm load no problem at full power

Nuuk · 2004-11-13 9:48 am

He is also very old fashioned and uses point to point wiring... but that is not catching.

Listen to Carlos' advice. He is also very old fashioned and recommends pots instead of stepped attenuators - but that is not catching.

Evilsizer · 2004-11-15 2:31 am

what kind of speakers are they(diy or store bought, 2-way or 3-way).

If a 2-way/3-way look below...

looking at the curve you may want to make a "Impedance Equalization Circuit " for the tweeter from what i can tell.

linky for the circuit:
http://www.mhsoft.nl/spk_calc.asp
then click on "Impedance Equalization Circuit".

hope that might help you out some!

ToG · 2004-11-15 8:27 am

Nuuk said:
Listen to Carlos' advice. He is also very old fashioned and recommends pots instead of stepped attenuators - but that is not catching.

ok

what kind of speakers are they(diy or store bought, 2-way or 3-way).

it's a 2-way store bought 94db Sensitivity with a 21cm woofer and dome tweeter.

looking at the curve you may want to make a "Impedance Equalization Circuit " for the tweeter from what i can tell.

that's an idea.

peranders · 2004-11-15 8:56 am

ToG said:
Well, hello all. First, please ignore my poor english, and just imagine i'm writing in a very good way ...
ToG

There is no need for excuses. We understand you pretty well. :nod:

carlosfm · 2004-11-15 10:05 am

soundNERD said:
I would use a parallel 3875.
[snip]
Also, better sound.

Are you sure?

carlosfm · 2004-11-15 10:09 am

ToG, your speaker doesn't fall below 5 ohms in the impedance curve.
I've seen 8 ohm speakers much worse than that.
You can start with single LM3886, as Carlos said.

roibm · 2004-11-15 10:16 am

Re: Re: What to choose ?

carlmart said:
Start with one 3886 per channel, with a +30/-30 supply and then see what happens.

Just a note: that voltage should be after the diodes bridge, not your secondaries. I'd recommend(based solely on the datasheet) some 20-21V secondaries for driving 4Ohm loads. 22-23 should be ok in your case.

carlmart · 2004-11-15 10:21 am

Re: Re: Re: What to choose ?

roibm said:
Just a note: that voltage should be after the diodes bridge, not your secondaries. I'd recommend(based solely on the datasheet) some 20-21V secondaries for driving 4Ohm loads. 22-23 should be ok in your case.

22v + 22v secondaries should become +/-30v after rectification.

Carlos.

roibm · 2004-11-15 10:31 am

Re: Re: Re: Re: What to choose ?

carlmart said:
22v + 22v secondaries should become +/-30v after rectification.

Not always. If we take the 1.44 coefficient it is +/- 31.68.
In my case a so called highend 2x22 toroid from schuro gives me +/- 33.5 after the bridge.
There can be some toroid who would provide 30 I don't doubt it.

carlosfm · 2004-11-15 10:52 am

Re: Re: Re: Re: Re: What to choose ?

roibm said:
In my case a so called highend 2x22 toroid from schuro gives me +/- 33.5 after the bridge.

If you measure AC are do you have 22V on each secondary?
You must have more.
That is because mains voltage must be higher than normal, very common.
Mains voltage here is 230V, but most of the times I measure it is between 235~239V.
Trafos here have 230V primaries, so the output is normally a little higher.

roibm · 2004-11-15 11:00 am

Re: Re: Re: Re: Re: Re: What to choose ?

carlosfm said:
If you measure AC are do you have 22V on each secondary?
You must have more.
That is because mains voltage must be higher than normal, very common.
Mains voltage here is 230V, but most of the times I measure it is between 235~239V.
Trafos here have 230V primaries, so the output is normally a little higher.

My secondaries are(but of course) over 22V.
I never measured the mains tho, I'll do it today. It is also possible to get a higher voltage because of the windings.

carlmart · 2004-11-15 11:20 am

Re: Re: Re: Re: Re: Re: Re: What to choose ?

roibm said:
Not always. If we take the 1.44 coefficient it is +/- 31.68.
In my case a so called highend 2x22 toroid from schuro gives me +/- 33.5 after the bridge.
There can be some toroid who would provide 30 I don't doubt it.

roibm said:
My secondaries are(but of course) over 22V.
I never measured the mains tho, I'll do it today. It is also possible to get a higher voltage because of the windings.

As a general rule I compute ACV x 1.4 - 1.4v. Usually turns out quite exactly on the transformers I measured.

It will depend mostly on three things:

1) There should be a load at the end, even if a small one.

2) Transformers are usually 5/10% accurate from spec.

3) AC wall voltage may vary.

Carlos

soundNERD · 2004-11-15 10:34 pm

Carlos,
Parallel allows it to drive 4ohm at full power. And, without the mute circuitry, the sound is better (and I can tell, i've listened to both).

carlosfm · 2004-11-15 11:15 pm

soundNERD said:
Carlos,
Parallel allows it to drive 4ohm at full power.

I'm not saying that paralleled LM3875s isn't a better option than single LM3886.
But check the datasheets and look for the max. current output of both chips, and also compare the graphs for the power at 4 ohms.
Talking about current, one LM3886 is roughly the equivalent of two LM3875s.
For current-starving speakers there's no other option, but I would avoit paralleling chips if one single chip does the job fine.
There's much more than a chip in the equation of the whole amp.

soundNERD said:
(and I can tell, i've listened to both).

Me too.

ToG · 2004-11-16 3:10 pm

Don't know, i'm not sure which solution is easier for me. It will be my first electronic project.

Maybe one lm3886.

Maybe it's easier to deal with mute than to deal with DC offset ???

There is no need for excuses. We understand you pretty well.

I try to be clear and as i don't speak (or write) english everday (hmm hmm) i rewrite all my sentences nearly 20 times. :smash:

So probably it's not really spontaneous.

Just a note: that voltage should be after the diodes bridge, not your secondaries. I'd recommend(based solely on the datasheet) some 20-21V secondaries for driving 4Ohm loads. 22-23 should be ok in your case.

ok for that.

Does a toroid + diodes + caps is sufficient or do i need more like a regulated one (lm338 or else) ?

What to choose ?

Member

Attachments

Member

Member

Member

Member

Member

Member

Member

Member

Banned

Banned

Banned

Member

Banned

Banned

Banned

Member

Member

Banned

Member