I would like to make the LXmini and integrate a stereo chipamp in the base. I assume I would just directly connect a channel for each driver since the crossover functions are taken care of with the miniDSP. My question is that one of the drivers is 4 ohm while the other is 8 ohm. Would this impedance imbalance lead to one driver being much louder then the other? Am I suppose to have some sort of volume equalization network in between the amp and the drivers or does the DSP take care of this too?
http://www.linkwitzlab.com/LXmini/Introduction.htm
http://www.linkwitzlab.com/LXmini/Introduction.htm
You must design the bass/mid amplifier for the 4ohm nominal loading.
That sets the voltage on the supply that can be used.
If you use the same supply for the treble amplifier, then since it sees a higher load impedance it will be safe to use and not blow up due to excessive dissipation.
A 3886 can run @ ~+-28Vdc +- mains tolerance and drive a 4ohms speaker if you keep it cool.
The second 3886 running on ~+-28Vdc driving an 8ohms tweeter will just be cruising. This one can have similar cooling the the B/M amp, or a bit smaller if room is tight.
That sets the voltage on the supply that can be used.
If you use the same supply for the treble amplifier, then since it sees a higher load impedance it will be safe to use and not blow up due to excessive dissipation.
A 3886 can run @ ~+-28Vdc +- mains tolerance and drive a 4ohms speaker if you keep it cool.
The second 3886 running on ~+-28Vdc driving an 8ohms tweeter will just be cruising. This one can have similar cooling the the B/M amp, or a bit smaller if room is tight.
one key point that i don't see mentioned in this thread: the amps must have the same voltage gain.
Not for active.
When bi-amping a speaker that expects to have the same voltages applied to the speakers input terminals, then one MUST use the SAME gain.
For an active speaker one would expect to use different gains to compensate for the different driver sensitivities.
Hi,
If you building in accordance to SL's drivers specs. you need more power if you want to drive them to loud level. Both drivers are rather ineffcient.
If you building in accordance to SL's drivers specs. you need more power if you want to drive them to loud level. Both drivers are rather ineffcient.
it depends on the active network's design.
as to this particular design, Linkwitz says:
"The difference in power requirements for woofer vs. mid/tweeter opens the door to Class-A power amplifiers for the higher frequencies, where amplifier differences are usually noticed first. But the voltage gain of such amplifier must be made identical to the woofer amplifier's gain in order to stay within the dynamic range of the mini DSP 2x4" [emphasis added]
Reading the site a bit more I see that Linkwitz uses a Emotiva 5 channel amp and there is no reason to think that the different channels have different voltage gains. I checked the specs for the drivers and I see that the full range driver has a sensitivity of 85.4 dB at 2.83V and the woofer has 84dB at 2.83V. That seems pretty well matched to me and leads me to believe that the drivers are directly driven by the amps.
@sfthurber
My plan was to use a stereo amp for each speaker so there shouldn't be any difference in gain.
@sfthurber
My plan was to use a stereo amp for each speaker so there shouldn't be any difference in gain.
Last edited:
As ttan98 mentioned, a single LM3886 is not powerful enough for woofer duties in the LXmini system. A bridging configuration utilizing two chips would work.
There is some leeway for amplifiers with disparate voltage gains in the LXmini system, but not much. An amplifier with a bit lower voltage gain could be used on the tweeter channel and a relative adjustment made in the miniDSP 2x4 output.
Dave.
There is some leeway for amplifiers with disparate voltage gains in the LXmini system, but not much. An amplifier with a bit lower voltage gain could be used on the tweeter channel and a relative adjustment made in the miniDSP 2x4 output.
Dave.
yes, that's the most common approach.
the gain issue comes into play if you use a choose to use different amps designs for woofer vs tweeter.
a two chip LM3886 bridge may be marginal with the 8 ohm woofer load because each chip in the bridge will see effectively one-half the actual load impedance (ie: four ohms).
Last edited:
🙂 The woofer is a nominal 8 ohm (not 4 ohm) driver so is perfectly capable of being driven by a bridged LM3886 configuration. The existing Pluto 2.x design that has been time-proven by many users uses exactly this configuration. I know of what I speak on this topic. 🙂
Dave.
i edited my post to call it "marginal" based on the 4 ohm dissipation curves in the datasheet and the following from Linkwitz:
"For the two woofers I suggest a range from 50 W to 200 W max per channel into 8 ohm"
even at a +/- 25 V supply you'll dissipate about 60 watts to get anything over about 40 watts out.
it's a bit better at +/- 20 V.
by comparison a parallel LM4780 (2 LM3866 in one package) is nearly trivial to build and will end up at a much safer operating point.
"For the two woofers I suggest a range from 50 W to 200 W max per channel into 8 ohm"
even at a +/- 25 V supply you'll dissipate about 60 watts to get anything over about 40 watts out.
it's a bit better at +/- 20 V.
by comparison a parallel LM4780 (2 LM3866 in one package) is nearly trivial to build and will end up at a much safer operating point.
Last edited:
I'm not sure how you got sidetracked on dissipation. I thought we were talking about maximum performance of the chip at the highest rated Vcc and with adequate heat-sinking.
In that case, bridged LM3886's can deliver approximately 150W peaks into an 8 ohm load. That's right in SL's recommended 50W-200W range and perfectly acceptable for either the Pluto or LXmini design. (The Pluto design utilizes a +/-33V power supply.)
You can define that as marginal or "marginal" if you like....whatever. 🙂
Actually, a single LM3886 operating at a lowish Vcc might even be acceptable for the LXmini. It all depends upon how loud a user likes to play the system. 🙂
Dave.
In that case, bridged LM3886's can deliver approximately 150W peaks into an 8 ohm load. That's right in SL's recommended 50W-200W range and perfectly acceptable for either the Pluto or LXmini design. (The Pluto design utilizes a +/-33V power supply.)
You can define that as marginal or "marginal" if you like....whatever. 🙂
Actually, a single LM3886 operating at a lowish Vcc might even be acceptable for the LXmini. It all depends upon how loud a user likes to play the system. 🙂
Dave.
The specs for the LM3886 say that for 8ohms I get 50W on a +-35V supply. While that is at the low end of Linkwitz's recommendation, it still corresponds to 101dB on an 84dB@1watt speaker. Assuming I sit 10 feet away this drops down to 91dB which is still loud enough to make my wife yell at me. Of course closed spaces don't really follow the inverse square law for power so it's likely to be much louder then that.
If I push the supply to +-40V then I can reach 70-80W which is about the same power as the 70watt/channel Emotiva amp that Linkwitz used to test this system. I don't plan on rocking any arenas with this so I don't think I am in danger of running out of power even without bridging the amp. Besides, I'm sure I'll find an excuse to gut this and upgrade later. 🙂
If I push the supply to +-40V then I can reach 70-80W which is about the same power as the 70watt/channel Emotiva amp that Linkwitz used to test this system. I don't plan on rocking any arenas with this so I don't think I am in danger of running out of power even without bridging the amp. Besides, I'm sure I'll find an excuse to gut this and upgrade later. 🙂
well i was talking about engineering for safety of the chip with 4 ohm loads and that it's not trivial (and not an engineering sidetrack).
indeed, much of the datasheet section on application information is devoted to thermal considerations. there is good reason for this. see figures 4, 34 and 35. they establish thermal considerations as a (or the) principal constraint in driving low impedance loads.
you're talking about a single LM3886 (not bridge or parallel)?
it may well be loud enough for you and too loud for your wife 🙂
certainly it would be loud enough for me.
the top curve of datasheet figure 36 (and the 80V column in the table in figure 34) define the heatsinking you'll need at +/- 40V assuming you have an actual 8 ohm load. (speakers typically don't give you that so YMMV)
Last edited:
Oh I am not going to 40V. I can live with 50 watts and the lower requirements for heat dissipation and transformer size that come with it. After living with it for awhile I will see if I really need MORE POWER! 😀
- Status
- Not open for further replies.