The AccuGroove Situation - I'd like an answer: Anyone else?

[alexclaber]

http://www.acmebass.com/forum/acmeswitch.htm

 

Interesting...

 

Alex

[getz out]

Very interesting.

 

I have only had good dealings with Andy; pleasant man to do business with. Not surprising he would let physics get in the way of common sense.

[Eric VB]

For the sake of melody, I say ...

 

what if you had 4 speakers?

 

Muhahahaha!

[saxofunk]

He's passionate, that's for sure. Is he passionate enough to take the action he suggests?

[getz out]

Ric,

 

I thought about that, too. I thought about having dual-voice coil speakers and some funky switching. However, you always end up with the same problem... you can't have it both ways.

 

Take four 8-ohm speakers:

 

You could wire it for 2 ohms or 8 ohms. How would you wire it for 4 ohms?

[Eric VB]

Well, of course I was just being facetious, but then I actually did sit down and try to figure out all the possible combinations. I came up with 8, but I'm probably missing some: 32, 20, 13 1/3, 10 2/3, 6, 4 4/5, 3 1/5, 2 ohms. (Of course I'm thinking resistance, not impedance, because that makes my head hurt too much.)

 

For 3 1/5 ohms, first combine two 8 ohm speakers in series for 8+8=16 ohms. Now put that in parallel with the remaining 8 ohm speakers for 1/(1/8 + 1/8 + 1/16) = 16/5 = 3 1/5 ohms. But now I think you have uneven current distribution -- two speakers see V/8 amps and two see V/16 amps -- and therefore uneven power distribution (V^2/8 Watts and V^2/16 Watts).

 

(Feel free to correct my physics if it's wrong.)

[shmolous]

Hey remember those old Marshall Double stack cabs with 4 12's...Those were 8 (or were they 16)

ohm cabs wired series parallel. The end result-total impedence the same as the individual drivers

[Eric VB]

Well, the solution is simple.

 

If you want to use two configurations that see equal power distribution to all speakers, you could use all-parallel as one and all-series as the other.

 

Now, how many identical speakers do you need, and what impedance do they need to be, in order to provide 4 ohms in one configuration and 8 ohms in the other?

 

A non-linear pair of simultaneous equations with two unknowns solved by substitution tells us ...

 

[edit: well, I made a booboo. Here's the correct answer.]

 

(a) you need to use 4*SQRT(2)=5.65... ohm speakers, and

(b) you need to use SQRT(2)=1.41... speakers.

 

 math check:
1/Rt=[(1/R1)+(1/R2)+...+(1/Rn)]
   =[(1/R)+(1/R)+...+(1/R)]
   =n*(1/R)
   =[sQRT(2)]*[1/(4*SQRT(2)]
   =(1/4)
so Rt=4 ohm for the parallel case

Rt=R1+R2+...+Rn
 =R+R+...+R
 =n*R
 =[sQRT(2)]*[4*SQRT(2)]
 =4*2=8
so Rt=8 ohm for the series case

[getz out]

I can get my actuaries at work to get a model going if necessary, Ric.

 

[getz out]

Andy has cred, but I wonder if any other speaker builders have tackled this issue.

[jeremy c]

I hope that Andy is not making a veiled put-down of Mark Wright's AccuSwitch.

 

There are not only two speakers in my AccuGroove speaker....there is a midrange speaker and two tweeters in addition to the two 10" speakers and there are two crossover networks. All of these obviously allow for more combinations of impedances.

 

But I'll let Mark defend himself if indeed he is the one that this screed is aimed at.

[Luke73]

Very interesting article.

 

It certainly comes across as a pointed message.

[getz out]

Originally posted by jeremy c:

There are not only two speakers in my AccuGroove speaker....there is a midrange speaker and two tweeters in addition to the two 10" speakers and there are two crossover networks. All of these obviously allow for more combinations of impedances.

I don't think the midrange and tweeters would matter if they are on a crossover circuit.

[alexclaber]

Originally posted by jeremy c:

There are not only two speakers in my AccuGroove speaker....there is a midrange speaker and two tweeters in addition to the two 10" speakers and there are two crossover networks. All of these obviously allow for more combinations of impedances.

Unfortunately they don't.

 

Originally posted by jeremy c:

But I'll let Mark defend himself if indeed he is the one that this screed is aimed at.

I'd like to see that. Because my concern is that a curious individual (with a lot of speaker building and electronics expertise) has taken an AccuGroove apart to find out how the AccuSwitch works and has found that it's nothing more than a marketing scam. Which thus leaves me thinking that everything the manufacturer has ever written about AccuGrooves, not least of which are the amazing physics-defying frequency response and sensitivity specifications, is at worst a downright lie and at best a well massaged version of the truth.

 

Alex

 

P.S. I don't dispute they're some some of the best cabs you can buy - I just suspect that they don't quite live up to their exorbitant specs. And the AccuSwitch is a travesty.

[Prague]

I must have missed the point of it all. All that is shown is a crossover. Was there some Clash of the Titans thread I missed?

 

Who would determine an impedance with a voltmeter? A signal generator and a decade box would be needed at minimum.

 

A "DC impedance" [sic] is of no consequence for speakers.

[rumpelstiltskin.]

but that's kind of the point here, prague. since everyone specifies their speakers with a DC resistance (which doesn't have much correlation with real-world performance), he's showing how it's easy fool the bass cabinet-buying public.

 

the whole point is the the DC resistance changes, but the actual AC impedance changes minimally. so those employers of the mythical AcmeSwitch think they're getting something when they're not.

 

the only other way i could think of it working is by changing the crossover network entirely, so that the amplifier "sees" a different effective load than with another crossover network. but that has to be prohibitively expensive on a commercial product that doesn't change price with the added feature.

 

the clue in this is that he thanks a guitar store, presumably for providing him with an inside look.

 

robb.

[Gruuve]

Just a couple thoughts...the impedances of all the drivers do actually matter in a crossover network. For instance, in a multi-driver system, if you take out an 8-ohm driver and replace it with a 4-ohm driver, you've changed the crossover frequency by doing that change out. Also, the impedance a driver *sees* does change with frequency. The impedance is highest when the driver is reproducing it's resonant frequency. In fact, that's how you empirically find it's resonant frequency...by pumping different frequencies through it and identifying the frequency that results in the max impedance on a scope.

 

Whether Andy's commentary was actually targeted at Accugroove or not, I'd really like to hear Mark's comments on the AccuSwitch and how it works. You can't defy the laws of physics...there's no magic. The only thing I can think of is that somehow the different crossover networks provide the impedance difference. I honestly can't think of how that would actually work, but then I'm not an electrical engineer.

 

Hmmm...I may ping a friend who is an electrical engineer to get his thoughts.

 

Btw, anyone got an AccuGroove cab with an AccuSwitch plus a scope and tone generator (or have a local speaker shop)? We could actually answer this with empirical data ourselves. I'd be willing to chip in on paying a speaker shop to test one just to get the answer!

 

Dave

 

 

Dave

[Tom Capasso]

Interesting. Let's start with a post Mark made on the thread ACME and Accugroove:

 

Originally posted by Whappo (on September 20):

Good questions. Ive talked with Andy before & hes a great guy & builds nice cabs.

 

it's not possible (due to physics) to have a driver that has extreme low-frequency response AND still have reasonably high sensitivity.

 

The issue of sensitivity is multifaceted & not an easy one. The first drivers we had developed sounded great, but they were not very loud. Most of the time there is a trade off between sensitivity and the sound you are looking for. If you get one, you may have to compromise on the other. (In keeping with the quote above)

 

However, after testing 1,200 different drivers from 6 different countries, we finally got things nailed down where we were able to get the best of both worlds. It wasnt easy. But if it was, everyone would be doing it. There are over 21 different parameters that can be changed like a Rubiks Cube to make the driver do what you want. What a puzzle?

 

After worrying about the driver, its the cabinet design that also contributes. This includes volume, porting, tuning, material, etc. By the time you make many small changes to multiple items, it can add up.

 

The bottom line is that like us, Andy makes his cabinets unique & different compared to what everyone else does. We builders choose a path we think is best & hope that players tend to agree.

 

Are our specs correct? You bet your low B string!

 

Whappo

AccuGroove

Next, this month's BP has a review of Tri 210L and Bill Dickens 212. I consider the review to be a positive one, but Herrera spent some time on the AccuSwitch. Jon didn't feel it did much (that's my summarization), though it seems that he went to some lengths to test it. BP also printed a sizable response from AccuGroove which stated that the switch has allowed users to go to multiple cabs who could not before (me: presumably because they'd go below the min resistance of the amp). The response also sttes that they've had no complaints, and that BP's testing was inconclusive because the amps used were protected from shutting down at 2 ohms.

 

I can't address this well, because I haven't played an Accu cab, and because I'm not an engineer. Based on what I've read and what I've heard from other players (like Jeremy) that the strength of these cabs is in their sound and construction.

 

Tom

[Gruuve]

I have no doubt that the AccuGroove cabs sound great. BUT, that's not really the issue that's been brought to light. Not to be a "devil's advocate" here, the question IS a technical one. However, this technical question potentially brings Mark's credibility into question as well. I make it a point to keep an open mind myself, so I'd really like to hear an explanation on the AccuSwitch from Mark.

 

Dave

[getz out]

Originally posted by C.Alexander Claber:

P.S. I don't dispute they're some some of the best cabs you can buy - I just suspect that they don't quite live up to their exorbitant specs. And the AccuSwitch is a travesty.

Good point, Alex.

 

They might be great speakers, and that is what one should buy based on; however, if the above is true, I would by reluctant to do business with a company that would employ such marketing tactics.

[Eric VB]

WARNING! Boring non-political math ahead (because I know getz and his actuaries were interested).

 

Originally posted by jeremy c:

There are not only two speakers in my AccuGroove speaker....there is a midrange speaker and two tweeters in addition to the two 10" speakers and there are two crossover networks. All of these obviously allow for more combinations of impedances.

Certainly, there are more ways to connect 5 speakers than there are to connect only 2.

 

I played with the series parallel model schmolous mentioned. I seemed to convince myself that it didn't matter if it were series parallel or parallel series; as long as all the speakers were identical they would see the same current (power).

 

I took it a step further. As long as the array of speakers is a perfect square (2x2, 3x3, etc.), then the total impedance will be the same as the impedance of just one of the speakers. Neat.

 

If you start using rectangles, e.g. 2x3, things start to change. The total impedance ends up as a ratio of the rows (2) to columns (3) times the impedance of just one speaker ®: (2/3)*R. Change the symmetry and you change the total impedance; a 3x2 has (3/2)*R. (This is the more general case of the square, where the ratio is always n/n=1.)

 

Looking back at my earlier post, I wondered if an array could be used to simulate 1.41... speakers. Knowing that 10/7 is a close approximation for SQRT(2), I looked at a 10x7 array. If each row is in series, then 10 speakers of impedance R would together have impedance 10*R; if 7 of these rows were connected in parallel the total impedance would be (10*R)/7, or (10/7)*R. Since we want this to be 8 ohms, we solve for R: (10/7)*R=8 => R=56/10=5.6 ohm. So, 70 identical 5.6 ohm speakers wired in this fashion produces a total impedance of 8 ohms.

 

But what if you connected them as a 7x10 array? Now the total impedance is (7/10)*R. Solving for R this time for a total impedance of 4 ohms yields: (7/10)*R=4 => R=40/7, or about 5.7 ohms. Not quite, but very close to the 5.6 ohms required for the 10x7 array.

 

I decided that I needed to take an average of these two values for R. Seeing as how I was dealing with multiplications and such, I decided to take the geometric mean: SQRT[(56/10)*(40/7)] = SQRT[4*4*2] = 4*SQRT(2), or about 5.65 ohms. Gee, this number sounds familiar.

 

So, with 70 identical 5.65 ohm speakers, two wiring networks, and a 70-throw switch, you can have a cabinet with either 8.07 ohm or 3.96 ohm nominal impedance, and all the speakers will see the same current (power).

 

 

Since I'm not an EE, I can only conjecture what is possible with 5 speakers and 2 "crossover networks". I don't see a simple solution that doesn't involve sending lows to the midrange or mids to a 10, and neither sounds like a "good thing".

 

Another simple idea from looking at g****r rigs is the power brake concept, i.e. using a resistor capable of dissapating a lot of power (heat), but I can't see how that would be viable.

[Gruuve]

Ya know, we are making one invalid assumption...we are assuming that Mark used either 4 or 8 ohm drivers in the AccuGroove cabs, and that may not be the case. There are drivers with other impedances (although they're not commonplace), and it's theoretically possible to put them together in a way that would allow you to switch between a 4 ohm and 8 ohm configuration by changing which ones are serial or parallel. That would certainly explain the need for two different crossovers since changing the impedance would change the crossover frequencies. Anyone know how to do the math to figure out what impedances might work? I think we only need one example to prove that it's possible.

 

That said, it sure seems like changing the impedance would change the sensitivity of the cabinet as a whole, so for those cabs that have the AccuSwitch, which impedance is the sensitivity rating based on? I'd imagine the 4-ohm setting since that would give the higher sensitivity.

 

Dave

[getz out]

Originally posted by Dave Sisk:

Ya know, we are making one invalid assumption...we are assuming that Mark used either 4 or 8 ohm drivers in the AccuGroove cabs, and that may not be the case. There are drivers with other impedances (although they're not commonplace)

No so uncommon; 16-ohm and 32-ohm drivers can be found for pro-audio and music instrument applications. An Ampeg SVT has (or at least did in the day) 32-ohm drivers wired parallel to get to 4-ohms.

 

Originally posted by Dave Sisk:

, and it's theoretically possible to put them together in a way that would allow you to switch between a 4 ohm and 8 ohm configuration by changing which ones are serial or parallel.

How? The cabinet with the switch has two drivers. Say they are 2-ohm drivers. Say they are 16-ohm drivers. Say they are 32-ohm drivers. Say they are 6-ohm drivers.

[alexclaber]

The only way to switch impedance in a 3-way cab with two woofers is through:

 

A: Dual voicecoils - i.e. 8 and 16 ohm, paralleled for 4 or 8 ohm. Issues with power handling and negative effects on T/S specs etc; and Mark has told be they don't use dual voicecoils.

 

B: Input transformer - too big and heavy.

 

C: Series/parallel switching - can only quadruple or quarter the impedance - i.e. 4 to 16 ohms and vice versa.

 

If you try and get the mid and tweeters involved then the tone and power handling will change depending on setting - and it won't take much of the wrong frequency to upset the smaller drivers.

 

Tom & Dave: The implication in Andy's webpage is that he's taken apart an AccuGroove and has found out exactly how it works. And I'd be very surprised if he's misinterpreted the circuitry - he knows his stuff!

 

Here's a question for you: Why do the single woofer cabs not have an AccuSwitch when they need it the most as they're likely to be used in multiples? If they have got dual voicecoils to work then it doesn't matter if they have 1 or 10 woofers, the principal is the same. Likewise if they're using an input transformer.

 

The AccuSwitch design appears to rely on having two woofers so that the capacitor can hide one of them from DC in the alternate impedance setting. If you haven't got two woofers you can't do this.

 

And then continuing on from this, it therefore follows through that the El Whappo and Whappo Jr are not true 4-way cabs, otherwise the smoke and mirrors trick won't work. That stacks up with speaker design logic, as the 15" or 12" sub and 12" woofer cannot be that far apart in T/S specs (especially as the 12" sub is used as the sole woofer in the Tri-112L) and are therefore probably sharing quite a substantial portion of the low and midrange frequency spectrum, if not just working in tandem.

 

Again, as I've said before, I've heard enough anecdotal evidence to prove the excellence of these cabs. But the marketing and technology claims behind the designs appear to less than accurate.

 

Alex

[getz out]

Originally posted by C.Alexander Claber:

And then continuing on from this, it therefore follows through that the El Whappo and Whappo Jr are not true 4-way cabs, otherwise the smoke and mirrors trick won't work. That stacks up with speaker design logic, as the 15" or 12" sub and 12" woofer cannot be that far apart in T/S specs (especially as the 12" sub is used as the sole woofer in the Tri-112L) and are therefore probably sharing quite a substantial portion of the low and midrange frequency spectrum, if not just working in tandem.

Agree, this would not be a true 4-way cabinet. However, you can have the "quasi 4-way design" by having the "cabinet within a cabinet design." That is nothing new or earth-shattering; my PA cabinet (Peavey TLS) has two 15" drivers and a horn and is a quasi 3-way design since the two woofers are tuned differently and have an internal baffle that changes the volume for each.

[SteveC]

I don't have anyting to add. It is an interesting thread. I noticed the same thing in the BP review. Something seemed "fishy" to me as well.

 

Whappo, chime in. Tell us your thoughts.

[alexclaber]

Originally posted by getz76:

Agree, this would not be a true 4-way cabinet. However, you can have the "quasi 4-way design" by having the "cabinet within a cabinet design." That is nothing new or earth-shattering; my PA cabinet (Peavey TLS) has two 15" drivers and a horn and is a quasi 3-way design since the two woofers are tuned differently and have an internal baffle that changes the volume for each.

This is quite popular in hi-fi and PA design - 2 1/2 way designs where there are two (or more) mid/bass drivers and a tweeter and all the mid/bass drivers share the lows but only one of them carries the midrange for better dispersion and phase coherency. A good move by AccuGroove if that's what they're doing (and the only way I can see the Whappo cabs putting out as much bottom as they appear to do) but it's not the 4-way design they claim, it's 3 1/2 way!

 

Alex

[BenLoy]

I'm stayin' way, way out of this one.

 

For the record, I have one of Andy's Acme cabinets and love it to death. It's been with me to hell and back and people always comment on how great it sounds.

 

I played through Jeremy's AccuGroove cabinet in Berkeley and it sounded incredible as well. Huge, ballsy and agressive.

[Pernax]

Interesting indeed, a quick scan of other boards left an impression that the whole bass playing world is holding its breath a bit...

 

C.Alex, you left out option D on your previous post:

 

D: Alien technology - How else would you explain that weird low and funky sound originating from Area 51?

 

-P

[Gruuve]

So here's a feeble attempt at answering my own question. (Yes, I'm an enganeer by educashun.)

 

Assumption: all drivers are the same impedance

 

x = individual driver impedance

y = number of drivers

 

Goal:

Series impedance = 8

Parallel impedance = 4

 

Series impedance = x + x + x ...

or x*y

 

Parallel impedance = 1/[1/x + 1/x + 1/x ...]

or 1/[y * 1/x]

 

We know we want the series impedance to be 8 ohms.

Equation 1: 8 = x*y

 

Solve for x (we know we want the parallel impedance to be 4 ohms)

4 = 1/[y * 1/x]

4 = 1/[y/x]

4 = x/y

 

Equation 2: x = 4y

 

Solve for y by plugging in x:

8 = (4y) * y

8 = 4y^2

2 = y^2

y = 1.4142

 

Solve for x using the value for y:

8 = 1.4142x

x = 5.6569

 

Check:

8 ?= x * y?

8 ?= 5.65 * 1.41

8 ?= 7.97, close enough ignoring rounding error

 

4 ?= x/y

4 ?= 5.65/1.41

4 ?= 4.01, close enough ignoring rounding error

 

The problem is that you can't have 1.41 drivers.

 

So, I think it's theoretically possible have a switchable deal. It might end up being a ridiculous number of drivers. In practice, it could certainly be possible to find a driver with the right T/S parameters, then have the mfg build some with custom impedances.

 

Is this what Mark did? I dunno...Mark?

 

Dave