LM4780 : good choice for 3 way active speaker ?

Status
Not open for further replies.
Jump to Latest
R

RickDangerous

Member
Joined 2013
#1
Hi,
I am planning to make a 3 way active speaker (inspired by the PMC IB1S).
I have been looking around for cheap chip amps that would suit such a speaker and i am currently considering the following (in each speaker) :
- 1x stereo LM4780 (2x60 W RMS) : tweeter + mid
- 1x bridged mono LM4780 (120 W RMS) : bass

Questions :
- Is this a good choice?
- Would it be better to use a 3883 on the tweeter & mid ?
- I've read that the bridged LM4780 gets real hot & needs a massive radiator + a fan. Any way to avoid the fan ?
- Can i use one single power supply for both the bass + tweeter & mid ? (1 PSU per speaker)

Thanks a lot.
 
#5
why would same impedance and efficiency mather at all?
one of the best features of an active 3 way x-over is the easy handling of sutch.
just before every poweramp a buffer is never a bad idea, and supposedly this buffer stage may not be a real buffer (not unity gain) to help overcome the difference between drivers.
Go for stereo only setup, that is good down to 4 ohm rated drivers.
Most probably it will supply more power than Your drivers can handle.
And match the sound levels just before the poweramps.
 
#8
Arty said:
why would same impedance and efficiency mather at all?
one of the best features of an active 3 way x-over is the easy handling of sutch.
just before every poweramp a buffer is never a bad idea, and supposedly this buffer stage may not be a real buffer (not unity gain) to help overcome the difference between drivers.
Go for stereo only setup, that is good down to 4 ohm rated drivers.
Most probably it will supply more power than Your drivers can handle.
And match the sound levels just before the poweramps.
Click to expand...

It is indeed my impression that going active helps to not suffer from impedance differences but i could be wrong.
But how do you use one stereo amp with two 3 way active crossovers? I imagine it then has to be an analog active crossover ?
I was thinking of using a WAF-Audio Najda digital active crossover (at least for R&D, and possibly in the final product, not sure yet), which has 2 inputs and many outputs.
So for R&D i need one mono amp per channel (6 therefore).
I have no objection to using one mono amp per speaker in the final product (or 1 stereo amp for both) with an analog active crossover, but is that possible? Does such a crossover have just one input and one output (both of which must be at line level because they must be before the amp) ?

Thanks

PS Any good books to recommend?
 
#9
If you are proposing a 3way active then you need 3 amplifiers.

If the sensitivities are 92, 88.5 & 90 as has been shown, then I would use the same amplifier for bass and treble and either a half power amp for the treble requiring a pair of lower voltage secondaries, or just use the same power on the same secondaries.

To me it seems that a single transformer with a single dual polarity PSU is the simplest way to achieve a 3way active solution.
 
#10
That 90dB seems a touch high, even for a 4 ohm unit. How far down does it go? Do you have the Thiel Small figures.

But if you are going to use it anyway - and yes the impedances shouldn't much matter - I would suggest you use the second 4780n in parallel mode. It won't give you the peak headroom that you would get with a bridged arrangement but it will have the advantage of each section being driven by effectively the same amplifier. You may also be able to hoik up the voltage on this one since it won't be working that hard (now seeing 8 ohms effectively, and on a pretty narrow bandwidth). In fact, while you can be conservative on the others, 30 to 35V, you get a good deal more latitude with the two in parallel. And no, you will not need a fan.

So, those are the three amplifiers. Effectively identical, but one able to deliver twice the current.


Edit. IIRC you actually do have the possibility of using two different supplies on the one chip (which if I'm right is where I would go) but you can use just one to begin with. I would definitely use a separate supply for the bass section. None of these have to be huge and the choice is mostly about what you think your system deserves. 160VA or 220VA would be around my starting point. But i would use the same total reservoir capacitance on each.
 
Last edited:
#12
ChristianThomas said:
That 90dB seems a touch high, even for a 4 ohm unit. How far down does it go? Do you have the Thiel Small figures.

But if you are going to use it anyway - and yes the impedances shouldn't much matter - I would suggest you use the second 4780n in parallel mode. It won't give you the peak headroom that you would get with a bridged arrangement but it will have the advantage of each section being driven by effectively the same amplifier. You may also be able to hoik up the voltage on this one since it won't be working that hard (now seeing 8 ohms effectively, and on a pretty narrow bandwidth). In fact, while you can be conservative on the others, 30 to 35V, you get a good deal more latitude with the two in parallel. And no, you will not need a fan.

So, those are the three amplifiers. Effectively identical, but one able to deliver twice the current.


Edit. IIRC you actually do have the possibility of using two different supplies on the one chip (which if I'm right is where I would go) but you can use just one to begin with. I would definitely use a separate supply for the bass section. None of these have to be huge and the choice is mostly about what you think your system deserves. 160VA or 220VA would be around my starting point. But i would use the same total reservoir capacitance on each.
Click to expand...

Supposedly it goes down to 30 Hz, but you know the manufacturers...
PRODUCT SPECIFICATIONS Power Handling (RMS) 150 Watts Power Handling (max) 290 Watts Impedance 4 ohms Frequency Response 30 to 1,500 Hz Sensitivity 90 dB 2.83V/1m Voice Coil Diameter 2"
THIELE-SMALL PARAMETERS Resonant Frequency (Fs) 29.7 Hz DC Resistance (Re) 3.6 ohms Voice Coil Inductance (Le) 0.6 mH Mechanical Q (Qms) 3.30 Electromagnetic Q (Qes) 0.86 Total Q (Qts) 0.68 Compliance Equivalent Volume (Vas) 4.22 ft.³ Mechanical Compliance of Suspension (Cms) 0.31 mm/N BL Product (BL) 8.48 Tm Diaphragm Mass Inc. Airload (Mms) 93.2g Maximum Linear Excursion (Xmax) 3 mm Surface Area of Cone (Sd) 525.9 cm²
MATERIALS OF CONSTRUCTION Cone Material Poly Surround Material Rubber Basket / Frame Material Steel Magnet Material Ferrite
MOUNTING INFORMATION Overall Outside Diameter 12.00" Baffle Cutout Diameter 11.00" Depth 5.8125" # Mounting Holes 8
OPTIMUM CABINET SIZE * Sealed Volume 5.99 ft.³ Sealed F3 39 Hz Vented Volume 11.87 ft.³ Vented F3 20 Hz * Enclosure volume/F3s based on BassBox "optimum" calculations -
As for the amplifiers, i don't understand everything (electricity is a mystery to me), but i like the idea of the one in parallel for the bass with an extra power supply (if i understood correctly).

Thanks, i think that does help me move forward.
I now need to decide on some drivers and order them.
 
#13
AndrewT said:
If you are proposing a 3way active then you need 3 amplifiers.

If the sensitivities are 92, 88.5 & 90 as has been shown, then I would use the same amplifier for bass and treble and either a half power amp for the treble requiring a pair of lower voltage secondaries, or just use the same power on the same secondaries.

To me it seems that a single transformer with a single dual polarity PSU is the simplest way to achieve a 3way active solution.
Click to expand...

Yes - I did the latter with a trio of Phillips TDA1514s a few years ago and it worked very well. Now starting out on a trio of 3886s to see how that will sound
 
#15
I see. The bass has quite a high Q and it seems you are going for BIG boxes!

Incidentally, on the midrange, I think Peerless do a 5" version which may have a more extended and flatter response. It will certainly have better dispersion and might even be cheaper. Personally I like those PPB drivers. You don't need to worry about excursion in general, depending on where you are going to cross them over. Mostly the driver won't move even a couple of mm.

Yes, you have understood correctly vis the paralleled bass amp. One of the many advantages of an active setup is that you don't have the power supply on one band moving because you have a current draw in another band. Ie. that a big bass thud doesn't move the supply for the tweeter amp. This is equally true, though possibly less important, between mid and treble.

Personally I would want to have two separate supplies there as well, though the scale of movement is vastly less. You'd need to do some research as to exactly which connections serve which amp on the chip (4780). Alternatively this is an argument for using 4x 3886 and what you then do is build 4 of the same amp (8 in total) and parallel a pair of them up (preferably after testing them individually). One caveat; mono the two amps after the input capacitors - their tolerances can be quite wide and you won't necessarily have the same high pass function on each. They will effectively be monoed beforehand, because they are being sent the same signal, but this is worth doing in addition (before the input resistors btw). You will also need a current sharing resistor on each of the outputs before you join those up. 0R1 to 0R22, 3W or 5W. Value depends on how closely you match the gains (min 1% tolerance but ideally hand selected afterwards) and whether you are going to have the R take the place of the inductor//R on the output. And I would say that on the bass amp at least you must have a zobel on each amp part. Normally the 3886 can do without these things but in parallel it is not worth taking any chances. 100nF + 10R is a pretty standard config and worked for me on a paralleled sub amp that I did.
 
#16
man...
maybe its just me. i do not know. but isin't this thread driftning away a bit?
Hi,
I am planning to make a 3 way active speaker (inspired by the PMC IB1S).
I have been looking around for cheap chip amps that would suit such a speaker and i am currently considering the following (in each speaker) :
- 1x stereo LM4780 (2x60 W RMS) : tweeter + mid
- 1x bridged mono LM4780 (120 W RMS) : bass

Questions :
- Is this a good choice?
- Would it be better to use a 3883 on the tweeter & mid ?
- I've read that the bridged LM4780 gets real hot & needs a massive radiator + a fan. Any way to avoid the fan ?
- Can i use one single power supply for both the bass + tweeter & mid ? (1 PSU per speaker)

Thanks a lot. "

-> go with 4 monoblocks, LM4780 set on voltgae safe for 4 ohm loads.
-> no need to bridge them.
-> You can use a single supply for all 4 amps, but it is advised to have the bass amps on theyr own psu.


Linkwitz-Riley Electronic Crossover
there You go.
allso in the same article:
An externally hosted image should be here but it was not working when we last tested it.

that is how You can deal with the differences in the efficiency of the drivers.

Idealy, I would assume that going for 6 monoblocks and active 3 way crossover would not add considerable costs, and can potentialy provide better quality.
In that case You would need 6 monoblocks, but no need for any kind of passive crossover element. You can run the mid and tweeter amps o the same psu, and leave the bass amps on theyr own psu.
That would be my advice.
From economy view it can happen that the active 3 way and 6 amps will be quite close in price to an active 2 way with 4 amps + 2 passive xover setup.

There are allso a gazzilion of chips that can be used.
Tweaking a passive xover is never easy. Dealing with the impedance of the drivers, accounting the losses yielded by the passive components are a challange alone. Not to mention when You might need to change any component. The costs are higher, and sometimes changin one components will force to change others too. Not cheap when we talk about inductors.
On the otherhand, 3 way active has the advantage of easy(er) tweeking, one could in case use sockets to enable quick swapping of caps and resistors to setup a new filter. Once built and set for an xover point changing the drivers will have no effect on the crossover freqvency. Speaker efficiency can be matched via a simple pot. These are features that most passive xovers do not have. In my opinion for experimenting it is a well worth investment.

You should first do a simulation of the drivers to see what kindof power requirements would the speaker it self have, maybe paralelling amps would simply be an overkill.
 
#17
I might be wrong but I'm sure I read somewhere (Loudspeaker Design Cookbook?) that if you use amps after an active crossover their efficiently at delivering power increases 3 fold compared to a standard amp followed by passive crossover. So putting 120w onto your bass driver in an active set up would equivocate to using 360w amp in a passive system. Are you trying to power a sound system for Notting Hill Carnival lol? In the first instance I would just go with 3 x 3886's, one on each driver, all powered from the same PSU. It will get you up and running in the simplest and most cost effective way. Then sit down and listen to it. You will be amazed at how powerful this will be with just 40-50w on each amp. I would take an educated guess that this kind of set up is /was the primary market for 3875/88 chips in the first place. You can easily 'upgrade' later to separate PSU's and tinker with increasing the bass if you will. I don't think you'll want or want to.
 
#20
I didn't see it drifting away - until someone came in and suggested passive components. 🙂

No 85dB btw.

All the efficiencies are well within quite narrow adjustment ranges and adjustment is something you want anyway.

No, there is not a 3 times increase in power in active systems. What can happen is that you get extra headroom because the two drivers can sum at various frequencies, but each is under the threshold of a single amp's rail. So you can then get greater than the peak available from the rails. It probably is about 3x the power when it happens but it doesn't happen all the time and no both sides are at peak at the same time when there is a crossover. (Incidentally, this phenomenon is worth looking at in differential amps too.)

The suggestion to parallel is relly to up the voltage and not have the bass section be the limiting factor in peak output. I would have preferred it bridged into a higher load in some ways, but this way has a number of advantages like keeping the chips cool. (Yes it could work with a single chip, almost certainly, but you won't be able to up the voltage, and you will have it as your weak point in the system, addressing the toughest part.)
 
Last edited:
Status
Not open for further replies.
Top Bottom