2 way PA Crossovers

[kenwstr]

Hi, I want to ask here because sometimes in PA forums advice is well, the lowest common denominator. I look after a Church analogue PA system (mono), we do try to aim for true sound reinforcement so clarity and natural sound is important. By and large, we have achieved that but as always, there are still some issues remaining. I have found PA forums are not always the best place to assist achieving & maintaining this but audiophile ones can get a bit too essoteric and controvercial.

We are running a pair of 100 litre sealed front of house (FOH) speakers and a pair of bass reflex stage or fold back (FB) speakers. Both pairs of speakers are 2 way with "the good for PA" Motorola piezo tweeters mounted in horns. The FOH has Jaycar kit crossovers with an additional resistor network to emulate a tweeter load. The FBs do not have any crossover at all and are built by a PA company we sometimes use. Some years ago, the PA company told me we don't need crossovers with Piezo tweeters. I was dubious, so did an AB test with the FOH where one speaker had the crossover bypassed and decided it sounds better with the crossover in place. The FBs are a floor wedge bass reflex design but can be stand mounted. As FB floor wedges, they sound rather boomy with indistinct midrange. They do sound much more ballanced on stands.

My 1st question is whether those in the know would regard the use of crossovers with piezos desirable for clarity & rich mids or do you agree with the professional advice to loose them?

The next question relates the the fact that our FOH crossovers are now a few decades old and probably should be recapped at least. I don't know much about them in terms of component quality or filter order. As they are Kits, I ammume quality is not high. I may also decide to build and install crossovers in the FBs. So, where can I find a reasonable 2 way design, how to find the correct crossover frequency, modify the values accordingly and how to choose components, bearing in mind, this is for PA? I can use a soldering iron but know nothig much of crossovers. Anyone prepared to help me out with design?

Also I am told the resister network is necessay for the crossover to see the correct load so that it gives the correct crossover frequency. The network is in parallel with the tweeter, does the type of resister matter, they seem to be rectangular ceramic blocks, I assume it's a network to cope with the power handling?

 

[TLS Guy]

Well, Piezo tweeters are designed to be used without crossover. They present a capacitative load that decreases with increasing frequency and so they roll off as the resistance increases inversely to frequency. The problem is that makes them rough in the lower range.

Now Piezos are really only any good above about 3K Hz for the larger ones and 4 Khz for the smaller ones.

So they are usually a bad choice for a two way system as most woofers are badly breaking up by that range.

When I use a Piezo I will put a resistor in parallel, say 22 ohms, with the tweeter so I can roll it off before the roughness starts. The use a 1,5 mfd cap for the smaller piezos and 2,2 mfd for larger ones.

In all honesty though piezo tweeters are used in low end systems. As with all crossovers you should measure the speakers with appropriate measuring equipment.

If you want the same frequency response as stock then use a 4 to 4.7 mfd cap. This stops the bass burning out the piezo over time. If your tweeters have been used without crossover then they probably have some degree of burnout and loss of output by now and should be replaced.

But as I said piezos should be used as super tweeters really, if at all. They are not a good choice for two way systems as they are very rough at the lower end of their range.

 

[kenwstr]

Ta, TLS Guy

Yes, I agree piezo tweeters are low end. It's a small independent church, not a lot of funds and much of the system was originally home built. Been replacing, upgrading things as opportunity arises. From what I have read, there are only a handful of piezos considered suitable for PA, original production Motorola KSN 1142 being one of them. I have some data sheets on what model number parts that should be in the enclosures but I'm not sure it's totally correct.

The stated FOH crossover is Jaycar CX-2612 Crossover frequency 3.5 kHz roll off 12dB / Octave, so 2nd order. FOH drivers were replaced with Selenium 12PW3 12 inch https://www.lautsprechershop.de/pdf/selenium/selenium_12pw3.pdf some years ago as the originals coils got cooked. It does seem to me quite an ask for them to perform up to 3.5k as well but 2 way seems pretty normal for this size of PA system.

They certainly aren’t HiFi speakers but the space gets used for every activity including youth drop in centre so speakers have to be robust. I can’t imagine running ribbon tweeters in that environment. I guess a lot needs replacing but the speakers are not as dire as some other things. It’s all pretty old and I go around, testing, cleaning conditioning connections, faders pots, minor repairs etc. All in all, it’s not too bad, an audiophile friends tells me it’s the best church sound he has heard. My opinion is that the next biggest sound quality issue is the room acoustics. Not that the sound is terribly bad but it gets in the way of efficiently resetting EQ after maintenance or deploying effective foldback.

It’s just that I have been reading about crossovers with a view to doing my Polk 10Bs at home but then got to wondered whether it was worth considering the church speakers particularly as we do have difficulty getting useful foldback sound (without crossovers)

Maybe it’s better to just get new speakers but they cost so very much I doubt anyone will support that.

Ken

 

[highfigh]

Ta, TLS Guy

Yes, I agree piezo tweeters are low end. It's a small independent church, not a lot of funds and much of the system was originally home built. Been replacing, upgrading things as opportunity arises. From what I have read, there are only a handful of piezos considered suitable for PA, original production Motorola KSN 1142 being one of them. I have some data sheets on what model number parts that should be in the enclosures but I'm not sure it's totally correct.

The stated FOH crossover is Jaycar CX-2612 Crossover frequency 3.5 kHz roll off 12dB / Octave, so 2nd order. FOH drivers were replaced with Selenium 12PW3 12 inch https://www.lautsprechershop.de/pdf/selenium/selenium_12pw3.pdf some years ago as the originals coils got cooked. It does seem to me quite an ask for them to perform up to 3.5k as well but 2 way seems pretty normal for this size of PA system.

They certainly aren’t HiFi speakers but the space gets used for every activity including youth drop in centre so speakers have to be robust. I can’t imagine running ribbon tweeters in that environment. I guess a lot needs replacing but the speakers are not as dire as some other things. It’s all pretty old and I go around, testing, cleaning conditioning connections, faders pots, minor repairs etc. All in all, it’s not too bad, an audiophile friends tells me it’s the best church sound he has heard. My opinion is that the next biggest sound quality issue is the room acoustics. Not that the sound is terribly bad but it gets in the way of efficiently resetting EQ after maintenance or deploying effective foldback.

It’s just that I have been reading about crossovers with a view to doing my Polk 10Bs at home but then got to wondered whether it was worth considering the church speakers particularly as we do have difficulty getting useful foldback sound (without crossovers)

Maybe it’s better to just get new speakers but they cost so very much I doubt anyone will support that.

Ken

There's a whole industry within AV called 'House of Worship' and they have magazines, websites & forums for discussing it.

 

[TLS Guy]

Ta, TLS Guy

Yes, I agree piezo tweeters are low end. It's a small independent church, not a lot of funds and much of the system was originally home built. Been replacing, upgrading things as opportunity arises. From what I have read, there are only a handful of piezos considered suitable for PA, original production Motorola KSN 1142 being one of them. I have some data sheets on what model number parts that should be in the enclosures but I'm not sure it's totally correct.

The stated FOH crossover is Jaycar CX-2612 Crossover frequency 3.5 kHz roll off 12dB / Octave, so 2nd order. FOH drivers were replaced with Selenium 12PW3 12 inch https://www.lautsprechershop.de/pdf/selenium/selenium_12pw3.pdf some years ago as the originals coils got cooked. It does seem to me quite an ask for them to perform up to 3.5k as well but 2 way seems pretty normal for this size of PA system.

They certainly aren’t HiFi speakers but the space gets used for every activity including youth drop in centre so speakers have to be robust. I can’t imagine running ribbon tweeters in that environment. I guess a lot needs replacing but the speakers are not as dire as some other things. It’s all pretty old and I go around, testing, cleaning conditioning connections, faders pots, minor repairs etc. All in all, it’s not too bad, an audiophile friends tells me it’s the best church sound he has heard. My opinion is that the next biggest sound quality issue is the room acoustics. Not that the sound is terribly bad but it gets in the way of efficiently resetting EQ after maintenance or deploying effective foldback.

It’s just that I have been reading about crossovers with a view to doing my Polk 10Bs at home but then got to wondered whether it was worth considering the church speakers particularly as we do have difficulty getting useful foldback sound (without crossovers)

Maybe it’s better to just get new speakers but they cost so very much I doubt anyone will support that.

Ken

Well that crossover is no good for piezo tweeters. You have to approach them differently from moving coil ones.
Those Selenium drivers are well breaking up by 3.5 KHz and the design is not correct.

The other issue is that you can not replace drivers in an enclosure with different ones. The enclosure has to be designed to the driver's Thiel/Small parameters.

Now PA systems for a church are very different form home speakers. Usually the approach is to use a highly sensitive driver like those seleniums and mate them with compression driver mated to a horn with crossover somewhere between 500 and 1500 Hz. Since PA speakers take a lot of power electronic crossovers are used ahead of the amplifiers, so the bass driver and the HF compression drivers have their own power amplifiers.
This is done to prevent waste of power by passive crossover and to stop the passive crossover getting burnt out. Yes. PA systems are almost all 2 ways, but not crossed anywhere near 3.5 KHz. As they are crossed were there is a lot of power then that really demands an active crossover design with bi-amplification.

If you look at the frequency response of those Selenium drivers you can see why this makes sense. It also makes setting the correct levels between LF and HF a breeze.

In summary, using piezo tweeters with those bass drivers is not a correct design. Your crossovers are not being used in the correct application.

In all honesty you will not improve matters without a redesign and build from the ground up.

I suspect the pro who saw it noted that the crossovers were useless and said to himself this system is a mess, and will be better off without the crossovers to the tweeters, and I agree.

 

[kenwstr]

Hmm, that doesn’t jell at all with what everyone is hearing though I can see from the curves that what you say should be correct.

If the Piezo was connected without XO, the driver would be getting full range. With the XO, at least some HF is eliminated. Given the situation though, would it be better to just have a 2nd order low pass for the driver only to cross the piezo’s natural roll off at about 1.5 kHz to achieve -10dB for both piezo and driver where they cross?

I wonder if the guy who built these modified the XO, he certainly had the gear to do so and the actual caps don't look like the ones currently depicted on the Jaycar site, my memory is that they used to look the same (large silver cylinders).

Ken

 

[TLS Guy]

Hmm, that doesn’t jell at all with what everyone is hearing though I can see from the curves that what you say should be correct.

If the Piezo was connected without XO, the driver would be getting full range. With the XO, at least some HF is eliminated. Given the situation though, would it be better to just have a 2nd order low pass for the driver only to cross the piezo’s natural roll off at about 1.5 kHz to achieve -10dB for both piezo and driver where they cross?

I wonder if the guy who built these modified the XO, he certainly had the gear to do so and the actual caps don't look like the ones currently depicted on the Jaycar site, my memory is that they used to look the same (large silver cylinders).

Ken

Actually the crossover you posted will not attenuate those piezo drivers at all. The impedance is too high.

I agree with the pro to not cross the piezo drivers over in your situation.

 

[kenwstr]

Actually the crossover you posted will not attenuate those piezo drivers at all. The impedance is too high.

I agree with the pro to not cross the piezo drivers over in your situation.

Um, not sure you got my point, perhaps I was not clear. By driver, I was refering to the bass driver. So I was asking about feeding the Selenium bass driver from a LPF to be 10dB down at around 1500-1700 Hz. The point being that this would give some atenuation to the Selenium to supress it's 2 kHz problem area where there is a response dip and a distortion peak.

I am somewhat unwilling to implement this speaker without some kind of control on the Seleniums response curve because I have tried that and my ears and the rest of the sound team tell me that is worse than having the XO in place. Theory is fine and a most excelent guide but I would rather trust what we actually hear.

 

[kenwstr]

I'm pretty sure you completely misunderstood my last post because I in no way suggested attenuating the piezo. I consistently refereed to the Selenium bass driver as "driver" but refereed to the tweeter as "piezo". My suggestion was to leave the piezo completely unattenuated but to put a Low Pass Filter on the Selenium so that the LPF roll off crosses the piezo's natural roll off at -10dB which would be at a little above 1.5 kHz. This would be the lowest frequency I can get the Selenium to cross with the piezo and it also attenuates the worst areas of the Selenium which are the 2kHz distortion peak and response dip as well as the 3.3kHz response peak. Perhaps not HiFi ideal but, do you not think this would be a better implementation?

I looked up a Piezo manufacturer website which states that while a XO is unnecessary for a piezo, it is possible to use a XO provided a resistor is placed in parallel with the piezo. I have already stated that this is the current implementation. Certainly, sound level checks across tonal sweeps suggest the XO is attenuating both the Selenium and piezo. In any case, I agree 3.5kHz cross is too high prompting my suggestion above.

I have tried the speaker with no XO and it sounds worse. I am not a newbe, been doing this PA thing for 30 years and upgraded this system on a shoestring and some personal cost to the point that it compares very favourably with most other church systems now. I know it isn’t HiFi by any stretch but I know what I am talking about in saying it sounds better with the XO in place.

Just for background, it was the pro guy who (with full knowledge) suggested and supplied the Selenium’s to replace the original drivers. More recently, he demoed some FOH PA speakers with a view to replacing our current ones but admitted that ours sound better than his. I am always looking for improvements though.


Ken

 

[TLS Guy]

I'm pretty sure you completely misunderstood my last post because I in no way suggested attenuating the piezo. I consistently refereed to the Selenium bass driver as "driver" but refereed to the tweeter as "piezo". My suggestion was to leave the piezo completely unattenuated but to put a Low Pass Filter on the Selenium so that the LPF roll off crosses the piezo's natural roll off at -10dB which would be at a little above 1.5 kHz. This would be the lowest frequency I can get the Selenium to cross with the piezo and it also attenuates the worst areas of the Selenium which are the 2kHz distortion peak and response dip as well as the 3.3kHz response peak. Perhaps not HiFi ideal but, do you not think this would be a better implementation?

I looked up a Piezo manufacturer website which states that while a XO is unnecessary for a piezo, it is possible to use a XO provided a resistor is placed in parallel with the piezo. I have already stated that this is the current implementation. Certainly, sound level checks across tonal sweeps suggest the XO is attenuating both the Selenium and piezo. In any case, I agree 3.5kHz cross is too high prompting my suggestion above.

I have tried the speaker with no XO and it sounds worse. I am not a newbe, been doing this PA thing for 30 years and upgraded this system on a shoestring and some personal cost to the point that it compares very favourably with most other church systems now. I know it isn’t HiFi by any stretch but I know what I am talking about in saying it sounds better with the XO in place.

Just for background, it was the pro guy who (with full knowledge) suggested and supplied the Selenium’s to replace the original drivers. More recently, he demoed some FOH PA speakers with a view to replacing our current ones but admitted that ours sound better than his. I am always looking for improvements though.


Ken

No I did understand. The point is the piezo is not going to cross with those selenium drivers as you suggest without leaving a massive hole in the speech discrimination band. So that is why I agree with the pro about leaving it alone, unless you are prepared to redesign things. A piezo is not going to go down to where that selenium should be crossed, no matter what you do. I explained about the resistor shunt to the piezo before.

Now the next issue is you have an off the shelf crossover, which should never be used. All crossovers have to be designed for each individual set of drivers. I have no idea where that crossover you have is actually crossing. It will be nowhere near where is says it is and the order will be different than specified.

All I do know is that those drivers can not be crossed and mate with piezo tweeters at 1.5 KHz or even 2 KHz.

So although the response is peaked badly, if you actually did what you suggest speech discrimination would be poor.

So that is why I and the pro say leave it alone, unless you want to take the correct steps.

At a minimum those steps mean a custom crossover design and a totally new HF driver. The sensitivity of those selenium drivers would demand a compression driver and horn.

If you really want I can design a custom crossover for you and pick out a compression driver and horn.

As I said if I were doing this job properly, the selenium driver would be put in a box I designed for it, and mated with a compression driver and horn. The crossover would be electronic active, ahead of two amplifiers one driving the bass selenium drivers and the other the compression/driver horn.

That is how you design a proper church PA system and have done for about the last 40 years.

That is what it will take to really improve your system.