[andyr]

I have recently had to relay the mid-panel wires on my right Magnepan IIIa … as, unfortunately, while in the process of replacing the MDF frames (which support the ribbon and bass/mid panel) with hardwood frames, I gashed the mylar sheet and broke a mid-panel wire when my screwdriver slipped.   (And, no … I hadn’t had a drink or three!  )

After making the repair, I find the repaired mid-panel has a resistance of 6.3 ohms … whereas the original mid-panel has a resistance of 3.1 ohms.  (I believe I know why this has happened and can fix it … but that’s not the subject of my post now.)

By way of further background, I have a 3-way active system, driven by LF monoblocks … so I can switch on or off, the individual amps powering each driver (ie. there are 6 mains switches in all – or 8 if you include my mono active XOs).  Also relevant to my post is the fact that Maggie drivers are almost entirely resistive – so “resistance” is an appropriate term to use for them … whereas with cone drivers, we talk about “impedance”, as they have an inductive component.

Switching off the bass and ribbon amps on each side, so I only have the mid-panel amps working and switching my tuner to 'mono', I can hear the soundstage is skewed to the left.  IE. when powered by identical LF55 amps, the LHS mid-panel (which has 3.1 ohms resistance) plays louder than the RHS mid-panel (which is 6.3 ohms).

Doing the same thing with the pairs of LF amps which power the bass panels and the ribbons, the soundstage with each pair is dead centre – as to be expected when the drivers’ resistances are the same.

Thus, a LF55 powering a 3.1 ohm driver plays it louder than when the driver resistance is 6.3 ohms.  Extending this idea, it means that when a driver’s impedance increases, the amp’s output will correspondingly decrease.  Hence a Zobel is required if the driver is to deliver a flat output over the frequency range which it operates in.  This obviously holds true in active setups, just as much as passive (it’s perhaps just a worse problem in a passive setup, as the overall speaker impedance typically has a couple of spikes across the frequency range, where the LP & HP filters cross over ... which will result in severely reduced volumes at those frequencies).

Regards,

Andy

[art]

I think you have a problem of efficiency, which as nothing to do with zobel networks. I would rather know why the two driver panels have vastly different resistance values.

Pat

[andyr]

I think you have a problem of efficiency, which as nothing to do with zobel networks. I would rather know why the two driver panels have vastly different resistance values.

Pat

You've completely missed my point, Pat. 

I have a RHS Maggie mid-range panel which has double the resistance of my LHS panel ... a perfect sitaution for an experiment!    Why this is so is irrelevant to the situation ... which is that the LHS mid-range panel (the original) plays louder than the RHS panel (the repaired one).

I am arguing that this is solely because of their differing resistances ... an AKSA 55 can deliver 55w @ 8 ohms; @ 3.1 ohms (the original mid-range panel) it can deliver 142w ... at 6.3 ohms it can only deliver 70w.

The result is that the lower-resistance panel plays louder ... extending this logic says to me that, with a conventional cone driver, its increasing impedance with higher frequencies means that these (higher frequencies) will be delivered at lower volume than the lower frequencies - which is not a good thing - unless a Zobel is used to control a cone speaker's impedance rise.

Regards,

Andy

[AKSA]

Andy,

I largely agree, but Pat is also correct.  The Zobel on a power amp is used to impose a generic load on the amp at very high frequencies to confer more stability, and this has to do with phase shift induced by reactive loads (like speakers) and its attendant effect on the global negative feedback loop........  The 6dB frequency of the 10R/0.1uF Zobel on the power amplifier is thus set very high, closer to the pole frequency of the amp, at 159KHz in fact.

A Zobel specifically designed for a driver uses very different values.  The 8" Peerless 884 on the VSonics uses 5.6R/39uF, and this is set to hold phase shift at the crossover point within certain limits, to deliver a seamless crossover between drivers and to enhance imaging.  In this case, the 6dB frequency of this Zobel is a mere 728Hz, a VERY different kettle of fish to the Zobel corner frequency used for the amp and mounted on the amp module.  The two Zobel networks, though identical in configuration, serve two very different purposes.

Thanks to Pat and Andy for your posts.

Cheers,

Hugh

[andyr]

Andy,

I largely agree, but Pat is also correct.  The Zobel on a power amp is used to impose a generic load on the amp at very high frequencies to confer more stability, and this has to do with phase shift induced by reactive loads (like speakers) and its attendant effect on the global negative feedback loop........  The 6dB frequency of the 10R/0.1uF Zobel on the power amplifier is thus set very high, closer to the pole frequency of the amp, at 159KHz in fact.

A Zobel specifically designed for a driver uses very different values.  The 8" Peerless 884 on the VSonics uses 5.6R/39uF, and this is set to hold phase shift at the crossover point within certain limits, to deliver a seamless crossover between drivers and to enhance imaging.  In this case, the 6dB frequency of this Zobel is a mere 728Hz, a VERY different kettle of fish to the Zobel corner frequency used for the amp and mounted on the amp module.  The two Zobel networks, though identical in configuration, serve two very different purposes.

Thanks to Pat and Andy for your posts.

Cheers,

Hugh

Hi Hugh,

All I was trying to do was show that, in my "experiment", I was able to show that 2 identical drivers (but with different resistances), have different output levels when driven by the same amp.

Regards,

Andy

[AKSA]

Andy,

No problem.  You showed that very effectively!   

How's the job?  Do you get time off for kafe?   

Hugh

[art]

If the resistance is different (and it shouldn't be in this case!), then of course the efficiency will be different. But then they really aren't identical, are they?

Why they aren't is what you should be focusing all of your attention on.

Pat

[andyr]

Why they aren't is what you should be focusing all of your attention on.

Pat

I know why they have different resistances, Pat.  But to get into an explanation of this would have meant a long diversion into the construction of Maggies ... whereas the subject of my post was about how differing resistances make the same amp produce different output levels - which I felt was a crucial issue with passive Xover design but also, to a lesser extent, with an active Xover.

Regards,

Andy

[art]

As a long-time Maggie owner, I would rather hear about that than your ideas on driver resistance and efficiency, and zobel networks. You are confusing cause and effect.

The amp is going to put out the same voltage, regardless of load z, zobel or lack thereof. Yes, the power dissipated in the load will obviously be different. But drivers are not spec'ed at power level; they are spec'ed at voltage level. Efficiency is then determined not by power level, but by drive voltage.

Your digression into zobels has nothing to do with one driver playing louder than the other.

Pat

[andyr]

As a long-time Maggie owner, I would rather hear about that than your ideas on driver resistance and efficiency, and zobel networks. You are confusing cause and effect.

The amp is going to put out the same voltage, regardless of load z, zobel or lack thereof. Yes, the power dissipated in the load will obviously be different. But drivers are not spec'ed at power level; they are spec'ed at voltage level. Efficiency is then determined not by power level, but by drive voltage.

Your digression into zobels has nothing to do with one driver playing louder than the other.

Pat

As you are a long term Maggies owner, it behoves me to adopt a civil demeanour! 

To go back to the beginning (and, yes, in retrospect, perhaps I shouldn't have extended my post to include mention of Zobels), what I was trying to get across was the fact that when 2 identical amps power one driver which is 3.2 ohms and another which has 6.4 ohms resistance - but is otherwise identical - the 6.4 ohm load does not play as loud.

To me this has implications when using inductive (cone) drivers as the generally increasing impedance with higher frequencies means the "noise" produced by the driver will be getting softer with higher frequencies (Maggie drivers are almost entirely resistive, so their load remains constant over the frequency range they are used for).  Plus, when a passive XO has an impedance spike or two (according to the modelling I have done in lspCAD, this typically occurs at the frequency(s) where a LP filter crosses over with a HP filter), this spike must mean that, at that frequency, the driver output is severely reduced.

I'm pretty sure the reason my repaired IIIa mid-range driver is showing about 6.4 ohms is the following (to be tested next weekend):
*  the mid-range panel consists of 12 runs of 31g Al wires running up and down the mylar.
*  Magnepan decided (for reasons of their own!) to make this 2 sections of 6 wires - ie. 3 up & down "loops", each run about 1.5m long - rather than 1 x 18m long wire in 6 loops.
*  each approx 9m length of 31g Al wire has about 6.4 ohms resistance.
*  2 x 6.4 ohms in parallel produces 3.2 ohms.
*  so the 2 loops in parallel produce the (correct) mid-panel resistance of about 3.2 ohms.
*  the Al wire doesn't "take" to solder well ... I think that one of the ends of one of the loops is not making electrical contact - hence I register 6.4 ohms on my meter, not 3.2.
*  reapplying a hot iron to the solder pads will hopefully cause the recalcitrant end to connect to the solder properly ... so I will have 2 x 6.4 ohm wires resulting in a 3.2 ohm driver.

Regards,

Andy

[rabbitz]

At a common voltage, a lower impedance driver (with every thing else equal) is going to have a higher SPL does it not? An example would be the Revelator 6.5" drivers that are made in 4R, 8R and 16R versions and SPL's at 2.83V do show the differences.

[andyr]

At a common voltage, a lower impedance driver (with every thing else equal) is going to have a higher SPL does it not? An example would be the Revelator 6.5" drivers that are made in 4R, 8R and 16R versions and SPL's at 2.83V do show the differences.

Aah, rabbitz ... interesting observation! 

I presume you mean that (with the same amp - and, before anyone jumps in by saying maybe the amp can't handle a 4 ohm driver, let's assume it can ... like an AKSA or LF) the same amp drives:
* the 4 ohm Revelator louder than the 8 ohm Revelator, and
* the 8 ohm Revelaotor louder than the 16 ohm version?

This mirrors my experience exactly. 

Regards,

Andy

[Rom]

Hi Andy

I think, Ohms law is the simplest explanation: Power is equal to E square times R .
So when the R change power change too.

Cheers
Rom

[rabbitz]

You've nailed it Rom and I can understand that even though I never got a handle on ohms law.

On the Revelator examples using 2.83V/1m we get:
4R = 90dB (when run 2 in series you get roughly 87dB which is the same as the 8R version)
8R = 87dB
16R = 83dB (when you run 2 in parallel you get 86dB approx which is similar to the 8R version)

All done to make dual woofers amp friendly.

I'm only using the Loudspeaker Cookbook simplistic info for the 3dB variations and in real life it does vary from that. Some say that 2 drivers in series gives the same SPL as the single driver (+3dB for doubling the cone Sd and -3dB for doubling the impedance) and 2 drivers in parallel will give +6dB (+3dB for doubling the cone Sd and +3dB for halving the impedance).

Andy, you're not hearing things as it's very real.

[rabbitz]

Andy

Zobels are relevant to this as they can be a tool to alter the frequency response of a driver by altering the driver's impedance.

An example of this is the much loved Peerless 810921 tweeter that has a rising response when you add it to a filter. A lot use an LCR network to overcome this, but an undersize (non textbook) zobel is cheaper and very effective as it alters the driver's impedance which tames the rising response.

[art]

The guy who started this has Maggies. Resistive load. Zobels not needed.

In your case, the zobel cancels out the inductance of the driver, which can effect response. But only in conjunction with its interaction with a passive crossover.

You guys are really going askew on this zobel stuff. Time for me to bow out, as my sanity is at risk.

Best wishes.

Pat

[Davey]

Sorry to interrupt here fellas/Hugh:  Power is equal to voltage squared (E^2) divided by R...not times R.

Zobels do NOT alter the drivers impedance.  They (can) alter the system impedance seen by the driving source.  If the output impedance of said driving source is very low then the frequency response of the driver will not be altered either.

As mentioned by Hugh, an amplifier Zobel is different than a driver Zobel seen in a crossover network.

Pat, you are right on with your comments.

Cheers,

Dave.

[Rom]

oppps, my bad  . Dave is correct P=EI and I=E/R, replace I on the power equation and becomes E square divided by R.

Thanks Dave.

Cheers
Rom

[rabbitz]

The guy who started this has Maggies. Resistive load. Zobels not needed.

In your case, the zobel cancels out the inductance of the driver, which can effect response. But only in conjunction with its interaction with a passive crossover.

You guys are really going askew on this zobel stuff. Time for me to bow out, as my sanity is at risk.

Best wishes.

Pat

I wasn't referring to using a zobel with the Maggies. My reference to zobels was it's in a passive xo and OK it doesn't alter the driver's impedance per say but alter's the system impedance as seen by the amp as Davey kindly mentioned. Just my wording was stuffed up.

Don't worry about your sanity as I'm out of here.