Amp Design based on Spkr Impedance

[Moze]

I am building a LM3886l gainclone class AB amplifier to power my 2 front LR and Center channel speakers - all are 2 way speakers with 2x6.5" woofers and 1" tweeter so each speaker will be consuming 2 channels and will be crossed over with a minidsp. I am using ARTA Limp to measure the speaker impedance and although all my speakers are 8ohm nominal rated from manufacturer, they all spend most of their time at 4ohms based on my impedance measurements in ARTA.

Here are the measured results:

L&Rs (Polk Audio)
Fs: 34hz
VAS: 33liters - fairly high for 6.5" woofer they are very compliant
Impedance Response: 4.1 ohms from 100hz to 1,500hz
Re: 3.6ohms

Tweeter
Fs-1,200hz
Re-4.6ohms
Impedance - 4 ohms from 2,000hz to 10,000hz

Center channel B&W LCR 600 S3
Rated 8ohm nominal (min 3.2ohms)
Impedance Response Measured - 4.2ohms from 100hz to 1khz
Tweeter drops to 3ohms at 10khz

So, my design decision rests on whether to bridge my amp chips or parallel...my problem is that both manufacturers rate these at 8ohm speakers however if I design a bridged amplifier with the LM3886, its going to have a minimum acceptable impedance of 8ohms and these speakers to me look more like 4ohm nominal loads.

Then again, I could also parallel the chips to drive 2 to 4ohm loads-I am hesitant to do this as well because perhaps there is a capacitive or inductive aspect that I do not know about that makes the amp see these speakers as 8ohm loads?

Any opinions are greatly appreciated, what do you think is this a 4ohm speaker and should I design based on a 4ohm load or should I treat them as 8ohm speakers and Im missing something here that makes these an actual 8ohm load. I am used to using Dayton Audio speakers whose speakers - if rated at 4 or 8 ohms- will actually measure at 4 or 8 ohms respectively. Their 8 ohm speakers never drop below 8ohms and same for the 4s

 

[Fast Eddie D]

Don't bridge 3886 into a 4 ohm load.

There is a TDA chip designed for paralleling. You can use one input circuit that drives multiple parallel output buffers. You can bridge and parallel to your heart's content, theoretically.

 

[Moze]

Don't bridge 3886 into a 4 ohm load.

There is a TDA chip designed for paralleling. You can use one input circuit that drives multiple parallel output buffers. You can bridge and parallel to your heart's content, theoretically.

Yes thats my question - do my speakers represent a 4ohm load based on the measurements or are they 8ohms like the mfr says?

 

[Sonce]

4 ohm load.

 

[Dissi]

Two 4 ohm woofers wired in series result in a 8 ohm load. A series resistor like in a passive crossover would increase the tweeter impedance as well.

 

[Moze]

Two 4 ohm woofers wired in series result in a 8 ohm load. A series resistor like in a passive crossover would increase the tweeter impedance as well.

Actually the measurements I posted are for the speakers wired in series already

 

[JMFahey]

Actually the measurements I posted are for the speakers wired in series already

Nonsense.
How can you measure and post parameters for "two speakers in series"?

 

[Sonce]

Actually the measurements I posted are for the speakers wired in series already

You mean in parallel?!
B&W LCR 600 S3 loudspeaker is a 2.5-way design, so both woofers (with their filters) are wired in parallel, so it is a 4-ohm loudspeaker, regarding the 4.2-ohm minimum impedance in the most demanding bass/midrange region.
You didn't indicate which model are Polk L&R loudspeakers, but most likely they have both woofers in parallel, so again they are 4-ohm loudspeakers, regarding measured 4.1-ohm minimum impedance in the bass/midrange. Your indication that Re is 3 ohm is confusing. It is Rdc for both woofers (in parallel), isn't it?
Your intention to drive actively perfectly good brand-name passive loudspeakers, ripping of their passive crossovers is not justified. As B&W LCR 600 S3 is a 2.5 design, you must keep at least the passive filter (for lower frequency) for the woofer, and make a higher crossover point (with miniDSP) for both woofer and midbass, to integrate with tweeter. Woofer with his passive low-pass filter must be wired in parallel with the midbass unit. Otherwise you can't have a two-channel active system (one amp for tweeter, and another amp for woofer/midbass), but you need a 3-way active system.
Only if Polk loudspeaker is a genuine 2-way design (not 2.5-way) with both midbass units in parallel and only one passive filter for both units, than you can pull off passive crossover and implement 2-way active system, as you intended to do.

 

[Fast Eddie D]

What Sonce said.

If you intend to re-engineer a 2.5 way speaker, you better know what you're doing or else you're going to be really disappointed.

So unless you know the crossover frequencies and baffle step response of the speaker and intend to build an active crossover bi-amp or tri-amp system, leave the speakers alone and just configure your amplifier for 4 ohm load. With 3886 this means no bridge amp and reduced supply voltage for the chip. Refer to the data sheet.

A 2.5 way speaker can present a very demanding load for an amplifier. There is no way I know of to re-engineer them without using an active crossover, bi-amped system.

My recommendation is to configure the 3886 for 4 ohm load per the data sheet. After reading your comments, I'm not convinced that you're ready for anything more complex than that, yet.

 

[Moze]

You mean in parallel?!
B&W LCR 600 S3 loudspeaker is a 2.5-way design, so both woofers (with their filters) are wired in parallel, so it is a 4-ohm loudspeaker, regarding the 4.2-ohm minimum impedance in the most demanding bass/midrange region.
You didn't indicate which model are Polk L&R loudspeakers, but most likely they have both woofers in parallel, so again they are 4-ohm loudspeakers, regarding measured 4.1-ohm minimum impedance in the bass/midrange. Your indication that Re is 3 ohm is confusing. It is Rdc for both woofers (in parallel), isn't it?
Your intention to drive actively perfectly good brand-name passive loudspeakers, ripping of their passive crossovers is not justified. As B&W LCR 600 S3 is a 2.5 design, you must keep at least the passive filter (for lower frequency) for the woofer, and make a higher crossover point (with miniDSP) for both woofer and midbass, to integrate with tweeter. Woofer with his passive low-pass filter must be wired in parallel with the midbass unit. Otherwise you can't have a two-channel active system (one amp for tweeter, and another amp for woofer/midbass), but you need a 3-way active system.
Only if Polk loudspeaker is a genuine 2-way design (not 2.5-way) with both midbass units in parallel and only one passive filter for both units, than you can pull off passive crossover and implement 2-way active system, as you intended to do.

Sorry if I wasnt being clear - the speakers measured in series were the polks. That is how they are wired from factory. They are a traditional 2way with 2 woofers and a tweeter. The info I posted about the BW was with no modifications and with the input jumpers installed.

The Re measurement of 3.6ohms was also with both woofers in series. With woofers in parallell, it drops to 1.85ohms

 

[Moze]

What Sonce said.

If you intend to re-engineer a 2.5 way speaker,

A 2.5 way speaker can present a very demanding load for an amplifier. There is no way I know of to re-engineer them without using an active crossover, bi-amped system.

My recommendation is to configure the 3886 for 4 ohm load per the data sheet. After reading your comments, I'm not convinced that you're ready for anything more complex than that, yet.

I am not looking to modify any crossovers or re engineer - simply asking if based on the impedance measurements I posted, should I design my amp based on the actual impedance or the mfrs rated impedance..I did not indicate I wish to meddle with the speakers themselves

 

[Sonce]

Sorry if I wasnt being clear - the speakers measured in series were the polks. That is how they are wired from factory. They are a traditional 2way with 2 woofers and a tweeter.

The Re measurement of 3.6ohms was also with both woofers in series. With woofers in parallell, it drops to 1.85ohms

OK, now is clear. So, if you pull out passive crossover from Polks, then you can use miniDSP with two amps (one for tweeter and the other one for the woofers). Amplifier for driving woofers must be designed for a 4-ohm load.

The info I posted about the BW was with no modifications and with the input jumpers installed.

But driving B&W 2.5-way loudspeaker actively with a 2-way electronic crossover remains very problematic problem. High-pass filter for tweeter is easy to be done with miniDSP, but you can't toss away the composite 2-frequency-knee low-pass passive filter and replace it with a active miniDSP low-pass filter - it is impossible. You must retain at least the L3 (5 mH) inductor for the woofer and implement low-pass filter at about 4 kHz for both of woofers. Again, amplifier for driving woofers must be designed for a 4-ohm load.

 

[Sonce]

I am not looking to modify any crossovers or re engineer - simply asking if based on the impedance measurements I posted, should I design my amp based on the actual impedance or the mfrs rated impedance..I did not indicate I wish to meddle with the speakers themselves

I am lost on this... Are you trying to use miniDSP and two amplifiers for driving the loudspeaker, or not? If yes, you must meddle with loudspeakers themselves, open them and pull out their passive crossovers. If not, you don't need miniDSP and two amplifiers per speaker - you need only one amp designed for a 4-ohm load and use the loudspeakers as they are designed - with passive crossovers.

 

[Fast Eddie D]

I think the quick and easy answer is to use an amplifier that will drive a 4 ohm load. A 3886 can be configured for 4 ohm load; I'm not going to look at the datasheets but I think you use +/- 28/ volts.

I have measured impedance curves of several popular loudspeakers (Bose and Polk, etc) and any that are rated "4-8 ohm" are almost always 4 ohm; sometimes 6 ohm. Bose speakers in particular are very demanding of amplifiers and don't give much back in return for their demands. They do sound noticeably better with some heft behind them.

Always go with the measurement!