Crossover question.

[Andre_Db66]

I have a questiin for anyone who understands the matter and is willing to take the time to explain something about the AR crossover, probably usefull for general understanding of crossovers to me or others.

It came to my attention there are a few different versions af the AR6 crossover, one without inductor coil before the woofer. Basically these coils in series with the woofer i have seen in all my AR speakers and i have tried out changing this in one of my sets of AR6 with dissatisfaction. I cant put my finger on it why i disliked the sound.

My question is actually simple, what does this inductor do, if basically a bipolar cap is enough to create high pass signal?

Bose did that in a daughter brand that i had, simple cap, resistor and warning lights for overload, comprising the entire XO.

If i were to lets say want to create a low pass @60hz butterworth and high pass to broadband, so above 60hz is unfiltered, how would i go about doing this?

Andre

[ReliaBill Engineer]

Inductor attenuates high frequencies. Capacitor attenuates low frequencies. When used in series with the driver.

Some manufacturers just let the woofer play, allowing the voice coil to act as an inductor to naturally attenuate high frequencies on its own. 
 

60 Hz is a low frequency, requiring a substantial inductor on the woofer to roll off all higher frequencies. But your midrange and/or tweeter will be completely unprotected from being driven by too-low frequencies; distortion and likely damage to both will result.

 

Below is a general purpose chart, that’ll get you in the ballpark for frequency using various components and standard driver impedances.

 

It indicates a 20 mH inductor for 62 Hz @8 ohm woofer. That’s a big inductor if air core; smaller if an iron core.

 

[Andre_Db66]

Thank you for taking time to answer.

First, the high pass will go to fullrange philips broadband 12 inch ad1255 so no harm done if there is all the freq range.

But... hypothetically using a 20mH in series on the plus side of the 15 inch JBL woofer after which it is branched off to the philips with a 250uF in series on the plus would give me a 1st order butterworth xo @ approximately 60hz with 6db on the woofer rolloff and 12db on the philips @55?

What freq does the ar3a cut off after the 2.85mH inductor? The woofer sounds like a subwoofer so i am curious about that. The values of caps and coil are not that high there... ( perhaps this is what i didnt like in the inductorless ar6 experiment, too much mid overlap/presence )

[ReliaBill Engineer]

The 15” and 12” would need to be parallel wired. The JBL would roll off at 60 Hz and above using the 20 mH inductor, about where the Philips would pick up naturally. The Philips is a 1969-1971 “full range” which will only make it to about 8,000 Hz; it’s a whizzer cone accordion surround woofer. It shouldn’t need an inductor nor capacitor. 
 

 

 

 

[Andre_Db66]

The 20mH inductor placed in series with the JBL on the plus side? Perhaps simply a 100 watt lpad between the JBL and Philips to adjust the level opposing each other?

To understand where all this is coming from. I have an open baffle project with 3 drivers(also peerless tweeters) involved and until recently an active crossover 3 way and 3 amps. I am intrigued by the apparantly simple crossover  of the AR speakers in terms of parts used, i have 4 available pots and plenty of insulated inductor wire gauge 17 to possibly create a similar crossover like the AR speaker has with 1 amp and a passive crossover.

Btw i think the information on the Philips is incorrect somehow, afaik it goes up to 18k but probably has strong rolloff early on. So it doesnt necessarily need a tweeter, especially to my 57 y old ears that hear very little above 12k, yet it is noticeable and pleasant to add them from 12k up. Super tweeters are not required here 😁

I still would like to know at which freq the ar3a cuts off the low end with only 2.85mH? Is it a second order filter and which type?

[ReliaBill Engineer]

2.8 mH puts the AR woofer roll off between 400-500 Hz @ 8 ohms. But that’s kind of low for the midrange. 
 

AR liked gradual XO slopes on their early speakers. More modern speakers, like Vandersteen and VMPS, also like gradual (6 dB/octave) slopes. 
 

That 12” woofer cone, with its attached whizzer cone, would have a very difficult time producing audible output at 18 kHz. Too much mass and too much inductance in the VC. Typical usable high frequency output for that is 8 kHz.

[Andre_Db66]

https://www.hupse.eu/radio/speakers/AD5200.htm

The ad1255/m7 is the succsessor to the ad5200. Check out the specs...

Norelco in US

It is with good reason these speakers are sought after and very expensive. Your pictures exhibit a pair on ebay, broken bolts, and sold as parts for the incredible amount of plus 600$. This is no joke, these speakers are amazing...

The broken bolts is a flaw (after many years, they are from 1969) and should be replaced with stainless steel ones with expertise.

[Andre_Db66]

17 hours ago, ReliaBill Engineer said:

2.8 mH puts the AR woofer roll off between 400-500 Hz @ 8 ohms. But that’s kind of low for the midrange. 
 

AR liked gradual XO slopes on their early speakers.
 

 

What type of xo are they considered to be?

How is it that there is a factory ar6 without coil? That seems an extreme deviation from AR standards...

[ReliaBill Engineer]

I guess AR let the 6’s woofer run open ended, with no coil. Apparently AR just designed around the natural roll off at the woofer’s high end. Butterworth is the natural 6 dB roll off using a coil or inductor, in most cases.

[Andre_Db66]

24 minutes ago, ReliaBill Engineer said:

I guess AR let the 6’s woofer run open ended, with no coil. Apparently AR just designed around the natural roll off at the woofer’s high end. Butterworth is the natural 6 dB roll off using a coil or inductor, in most cases.

This is true in some ar6. I think if i recall correct it ws RaRa who brought this up. Anyway i read about this here on CSP and tried it for myself. I have however 3 pair of ar6 and 2 are alnico woofers and they also differ in xo. This is however apparantly consistent with either the alnico or ferrite version of the woofer. I tried the inductorless experiment with ferrite magnet and was not impressed. Others have said it makes a very good sounding ar6. I wonder what i am missing here...

Doesnt butterworth make a rather more sudden roll off?

If i were to try and make these coils of 20mH, what air gap should i be using approximately?

[Andre_Db66]

I have a 2ax woofer laying around and a corresponding cabinet. If i were to make that into say a supporting sub to a set of ar6, i would only have to increase the value of the original coil, right?

(Dreaming of 5.1 ar6 surround 🙃)

[Andre_Db66]

I created an inductor of almost 37mH with 3 #7 ar inductors and iron core. It works but gives a lot of high bass/ low mid. The roll off is very long. What is responsible for shorter rolloff, thicker gauge wire or larger iron core or both?

[ReliaBill Engineer]

6 dB/octave first order low pass network is quite “long” in its roll off. You’d have to go to a 2nd order filter to get a steeper slope of 12 dB/octave. 
 

Here is another chart that shows capacitor and inductor values to get that 12 dB/octave roll off.  Inductor in series, capacitor in parallel with the woofer terminals. Top of chart shows wiring for low pass woofer.

[Andre_Db66]

Those charts are nice... thanks a lot.

I had some quirky readings on my inductor coils. Eventually i discovered that 3 x #7 wound on a little bit bigger iron core did not give me 37mH but 17 smthn. I added in series another 2 coils making a sum of approximately 21.5mH. Still having a long rolloff i first chose the bass speaker which was loudest and nicest. I had the jbl 2226H@8R E140@8R and the one i decided on K145@4R. Then i found out that according to the xo calculator my inductor related to a 60hz Linkwitz xo (4R woofer), second order but it was missing a 330uF cap. I didnt have anything at hand so the cheapest way to try without immediately paying top dollar for bipolar, i decided on 2 NOS 680uF polar elcos in series, minus attached to each other creating a bipolar cap. It works...! I have now a strongly diminished mid/high so at least 12db linkwitz rolloff and it suits my purpose.

All i have to do is give it a listening session as soon as i have properly connected both channels. Little pressed for time but sunday i will do this for sure. If it is satisfying i shall most likely order Jantzen or Mundorf 330uF BP caps.

I am still going to build a subwoofer from the spare 2ax chassis/case i have but not quite sure what xo point i should use...

[giovanni56]

On 4/9/2024 at 11:54 PM, ReliaBill Engineer said:

La rete passa basso del primo ordine da 6 dB/ottava è piuttosto “lunga” nella sua attenuazione. Dovresti usare un filtro del 2° ordine per ottenere una pendenza più ripida di 12 dB/ottava. 
 

Ecco un altro grafico che mostra i valori del condensatore e dell'induttore per ottenere un'attenuazione di 12 dB/ottava. Induttore in serie, condensatore in parallelo ai terminali del woofer. La parte superiore della tabella mostra il cablaggio per il woofer passa basso.

Those tables give capacitance and inductance values that are derived as if the drivers were pure resistors, in reality the drivers have a resistive and reactive component that varies (by a lot, especially at the Fs of the component), depending on the frequency, so they give a very rough indication with respect to the correct values for making a proper crossover.

[RoyC]

3 hours ago, giovanni56 said:

Those tables give capacitance and inductance values that are derived as if the drivers were pure resistors, in reality the drivers have a resistive and reactive component that varies (by a lot, especially at the Fs of the component), depending on the frequency, so they give a very rough indication with respect to the correct values for making a proper crossover.

Very true...tables will only provide some general guidance. The mechanical properties of drivers are a huge variable.

[Andre_Db66]

I used a  calculator from DIY, not the tables. Perhaps i should experiment more with the inductor value. I cranked that one up so much because i was still hearing so much high pass which obviously ceased when i implemented a capacitor. All materials used were laying around so i did not spend any money as it is not the definitive solution. Lack of time prevented more tweaking. I will do more listening and fiddling soon.

According to the chart the 17mH may well be enough as long as i include a cap...

[giovanni56]

8 hours ago, Andre_Db66 said:

Ho usato una calcolatrice fai da te, non le tabelle. Forse dovrei sperimentare di più con il valore dell'induttore. L'ho alzato al massimo perché sentivo ancora così tanto il passa alto che ovviamente è cessato quando ho implementato un condensatore. Tutti i materiali utilizzati erano in giro, quindi non ho speso soldi perché non è la soluzione definitiva. La mancanza di tempo ha impedito ulteriori modifiche. Presto ascolterò e giocherò di più.

Secondo la tabella i 17 mH potrebbero essere sufficienti purché includa un limite...

With the calculator, the results are obtained by applying formulas that are applied for resistive loads while the loudspeakers behave as reactive loads as a function of frequency, moreover, many other parameters are needed for a proper implementation of a crossover, for example: 
the measurement of the impedance and T&S parameters of the transducers, the frequency response of the transducers, the nominal diameter of the transducers, the measurement of the distance of the emission centers between the speakers, the measurement of the offsets of the emission centers of the speakers, the resistance value R of the inductances that will be used in the crossover, the altitude and distance of the measurement microphone from the emission center usually of the midrange for a three-way speaker, etc. etc, finally, good simulation software and a measurement system in a reverberant or otherwise anechoic environment (near-field measurements). I see it difficult to make a good crossover with a simple calculator or table, also because you could get a potentially critical load, getting a very low impedance with deadly phase rotations, for the amplifier to which you connect the system.

[giovanni56]

26 minutes ago, giovanni56 said:

With the calculator, the results are obtained by applying formulas that are applied for resistive loads while the loudspeakers behave as reactive loads as a function of frequency, moreover, many other parameters are needed for a proper implementation of a crossover, for example: 
the measurement of the impedance and T&S parameters of the transducers, the frequency response of the transducers, the nominal diameter of the transducers, the measurement of the distance of the emission centers between the speakers, the measurement of the offsets of the emission centers of the speakers, the resistance value R of the inductances that will be used in the crossover, the altitude and distance of the measurement microphone from the emission center usually of the midrange for a three-way speaker, etc. etc, finally, good simulation software and a measurement system in a reverberant or otherwise anechoic environment (near-field measurements). I see it difficult to make a good crossover with a simple calculator or table, also because you could get a potentially critical load, getting a very low impedance with deadly phase rotations, for the amplifier to which you connect the system.

Forgot:
Many listening checks of the system for possible corrections and finally validation of the reproduction system made.

[Andre_Db66]

Old oak dining table, 2.5 inch thick, hollow and filled with bird cage sand. 100 pounds per speaker approximately.

Alnico JBL and Philips drivers

As you can see i am into reusable materials and technology. That also sparked my love for AR and Advent...

Potential critical load and deadly phase rotations, very severe descriptions for an open baffle system speaker that is used as nfm. I understand what you are saying and under normal circumstances where the speakers would be asked to deliver maximum power like in a computed box which is made with t&s parameters accordingly, this would certainly be an issue. As my open baffles are only used near field at 7ft, and the usage of drivers that are suitable but not specifically meant for this purpose i am mainly dependant on visual and auditorial inspection of the system. Read.... correction of the filters. When i play too loud the xmax limit becomes a problem for the K145 anyway, also for the AD1255. The latter being only 30 watt and highly sensitive will only allow moderate input from any given amp, tube or SS. The bass is turned up only so far to keep the balance right and carefully monitoring the xmax. The amplifier used for this is a professional amp which delivers a low impedance load without problems and also has clipping warning lights and overlaod protection. So i do not feel there is much to fear, i am safely playing music. A bump in the frequency range can be heard as it is unnatural and therefore eliminated simply by careful listening. The room is very small actually and the speakers are comparable to headphones this way, so anything that is not right will pretty soon give listening fatigue. I am on the right track, only need to tweak the lot a bit more...

As the philips is playing full range, the tweeter and 15" bass is only there to add the missing frequency ranges.