Passive line level crossover?

[HighFidelityGuy]

Hi,

I'm having some issues getting my speakers and subs to play nicely together. To prevent inconsistent phasing issues I have to set the crossover frequency on my subs very low. However, this causes a gap in the frequency response between the speakers and subs which stops them from blending together seamlessly. I think this is due to my main speakers dropping down quite low but having a very gradual roll off. As quite a lot of the music I like has very low frequencies that need to be reproduced at the same level (if not louder) than the higher frequencies, I often run into issues where one track requires the subs to be in phase and the next requires them to be out of phase. After a year of putting up with this I'm starting to get pissed off with changing the phase every few minutes. :mental:

BTW, I run a stereo pair of subs connected to the second pair of outputs on my attenuator.

I've been giving this issue some thought recently and I think the simplest solution is to make my main speakers roll of quicker, or earlier (higher) or both. I realised that I could make a simple first order high pass filter by putting a pair of caps between my attenuator and power amps. My subs would still connect directly to the attenuator as they have on-board low pass filters.

I've made a diagram to illustrate one channel to show how I'm imagining this will connect up (top) and I've included a schematic of my speaker's crossover in case that helps (bottom):

http://i2.photobucket.com/albums/y48/DTM2000/Hi-Fi/Crossovers.jpg

So do you think this will work ok?

My main concern is that as I've done the passive mod on my Caiman that I'm going to end up with caps in series which will reduce their values and screw up the crossover point. (The 47uF coupling cap in the diagram is in my Caiman) However, seen as my attenuator (a 10K alps blue velvet) will be between the caps, I'm wondering if this will prevent that issue. :scratch: I'm also slightly concerned that adding an extra crossover like this will screw up the way the crossover in my speakers work. Also, I gather that this method passive filtering could introduce new phasing issues higher up the frequency range. So is that likely to cause an issue?

I'd also be interested to hear any alternative methods of solving this issue as I'd really prefer not to have a high pass filter constantly in the signal path as I usually don't have my subs turned on if the music doesn't require it. I guess you can't have everything though, :doh:

Thanks. :)

[John]

It might be worth playing with the sub inputs I use the HT lead from amp speaker terminals I found getting phase right essential to good sound but very different speaker system and bass system (mine active dipole bass rather than sub bass)
The Opus3 are well voiced think issue is more likekly to be the subs I would be reluctant to alter the vrossover on the Opus 3

[Welder]

I empathize with your problem Dave. 6 years of intermittently messing about with an active build and I still haven’t found a satisfactory solution to integrating a pair of subs. So, no, I don’t have an answer for you, except, I would be reluctant to use any attenuation before a passive crossover. To do this properly one needs to completely re calculate the crossover.

I have come to accept the best way to get bass is to include the required frequencies in the main speaker and forget about subs altogether.
Alternatively, if you must have a sub then just use one and get the positioning right.

[Reid Malenfant]

Hi Dave, for starters you haven't mentioned what type of subs they are :scratch: This is important as certain types will alter the phase of the sound over the whole of there range, where as others won't.

Looking at your top diagram i can see a rather large problem straight away ;) You are attempting to drive a 10K & 50K impedance in parallel via a 10K passive potentiometer. What you need is a buffer, something of low enough output impedance that'll drive that low a load. The capacitor idea won't work very well either simply because the passive pre amp has miles too high an output impedance, get that down to 100 ohms or less & it stands a chance of working :) Yep, you need a buffer or an active pre amp ;)

It's really not very difficult to integrate a sub or subs just as long as they are pretty close to the main speakers & the crossover frequencies match. Or in the case of your main speakers you'll want to find the -6Db point. You then need to set the sub so it rolls in at the same frequency & then there should in theory be a phase shift adjustment that allows for 0 - 180 degress of variable phase shift. The other 180 degrees is taken care of by a switch usually...

Do these subs have a speaker level input? What subs are they out of interest & where are they situated in comparison to the main speakers?

[Reid Malenfant]

I have just taken a look at your system in your signature, are you attempting to pair up the Wharfedale SW250s with the Opus3 or a different set of subs?

Just had a look at the SW250 specs & all i can see is a phase invert switch but no phase shift control :scratch: If this is the case let me know as that'll be the problem, you might also find that you can't set the crossover frequency low enough on the subs either :mental:

My guess is you'll need to set it to about 30Hz or a tad less ;)

E2A:- Just found out that the lowest xover frequency appears to be 35Hz & it definately looks like you only have a phase inverting switch :doh:

What you could do is get hold of an active low pass filter & add a phase shifting network to it. Then you could simply set your sub to it's maximum frequency & set the external xover to the desired one & get the phase right at the same time.

It looks like your speakers are probably reflex in design, this will complicate things as you are trying to match a reflex with a 24Db/octave rolloff with a sealed box & a probable (though i'm by no means certain) 12Db octave low pass filter. What you need is a 24Db/octave low pass to the subs to get the phase in line with the speakers or it'll change all the time with different frequencies as it appears you are discovering :(

[DSJR]

Ooooh, I wouldn't do it the passive way.....

Firstly, I'd wait for the gator line driver for your Caiman, as this should stop any reaction of the caps in the Caiman you've fitted, reacting with the cables and impedance of the amps it sees. This should help sort the bass out at a stroke - you get owt for nowt in this world and if the passive output mod worked universally, everyone would have done it, s'not rocket science you know :lol:

In the meantime, try moving the sub around as this will probably have more of a difference if it's driving room modes or not.

The problem with many subs in the past was that they had two jobs to do, neither mutually compatible I found. The first task is loud, boomy, tuneless rumbles for films and the second is an articulation of low bass notes for music. I remember REL making subs for both markets in their range and never the twain would meet. In fact, the old fashioned looking baby M&K sub of yesteryear was tons better on music than many bigger Rel and KEF subs.

My recommendation for a sub (but you'll have a very hard job getting a home dem sadly) is possibly the AVI one. The box is stock far-east, but the driver has been very carefully selected for boom-control (it has a huge magnet) and the amp has a versatile filter setting. Also, this sub has been designed for a speaker that naturally rolls off below 60Hz and which cuts off below 20Hz as the big ATC's do from memory.

[HighFidelityGuy]

Many thanks for all the replies. There are too many for me to start quoting so I'll just try to reply as best I can...

I use a pair of Wharfedale Opus3 speakers and a pair of Wharfedale SW250 subs. The subs are connected to my passive pre in stereo. So as my speakers are 3-way, I essentially have a 4-way setup.
The subs have crossover settings from 35Hz to 85Hz in 10Hz steps. There is also a full pass setting and a 0/180 degree phase switch. There is no variable phase control. The Opus3 are bass reflex and the subs are sealed.

I'd prefer not to mess with the crossover in my speakers as I'm sure I don't have the skill to design something better. I also get the feeling that the general consensus is that adding more passive filtering is not the best approach.

The reason I looked into passive filtering is that I'd really prefer not to have an active crossover in the signal path of my main speakers. I would consider using an outboard active crossover on the subs and I do actually have one I could try but it's lowest low pass setting is 60Hz. :(

My subs are located directly behind each speaker, so they can't get any closer, however that's the only place they will fit in my room so I can't experiment with placement. :(

I didn't realise that my subs had such a low input impedance until I made the diagram above and I hadn't thought about the fact that their input impedance was in parallel with the input impedance of my power amps. So I see how I could be getting all sorts of impedance miss-match issues due to me using a passive pre.

On the subject of buffers; I have looked into trying some but that was mainly to go between my DAC and preamp. It now sounds like I'd need one or two after the preamp as well. :doh: That seems to me like treating the symptoms rather than the cause and I don't like the idea of having so many extra circuits in the signal path. Doing the passive mod on my Caiman has proved to me that keeping the signal path short and un-cluttered yields the best results. :)

So I may need to experiment with an outboard low pass filter for my subs but it sound like it would be better for me to solve my impedance issues caused by my passive pre first. The simplest option is probably to get an active preamp but due to my requirements my options are quite limited. The best option I've come across so far seems to be the Wyred4Sound STP SE (http://www.itemaudio.co.uk/wyred_stp_se_preamp.html).
Could someone please check the specs of this and let me know if you think this would meet my requirements from a technical point of view? I know it has the features I need but I'm a bit unsure on impedances and stuff like that.

Many many thanks. :cool:

[Reid Malenfant]

Dave, all you really need is to get the speaker/sub interface sorted ;)

You don't need to buy a pre amplifier, all you need is to get a crossover sorted out for the subs & set the built in one to as high as it'll go.

What i'd suggest is find some kind of active crossover for bass duties & modify it to raise the input impedance to at least 100K ohm so it has virtually no effect on your passive pre amp, feed this into a mixer (one op amp would do this) to create a mono signal. Then set the low pass filter (24Db/octave) to 30Hz -6Db & make a phase shifting network & then feed both subs from a common output.

Take it from me that there is no way on this earth that you'll ever be able to tell that the subs were effectively in mono :lol: Not at those frequencies ;)

In fact the whole thing could be chucked together on a bit of vero board & cased up, maybe not by yourself, but trust me it's not rocket science :)

[HighFidelityGuy]

Dave, all you really need is to get the speaker/sub interface sorted ;)

You don't need to buy a pre amplifier, all you need is to get a crossover sorted out for the subs & set the built in one to as high as it'll go.

What i'd suggest is find some kind of active crossover for bass duties & modify it to raise the input impedance to at least 100K ohm so it has virtually no effect on your passive pre amp, feed this into a mixer (one op amp would do this) to create a mono signal. Then set the low pass filter (24Db/octave) to 30Hz -6Db & make a phase shifting network & then feed both subs from a common output.

Take it from me that there is no way on this earth that you'll ever be able to tell that the subs were effectively in mono :lol: Not at those frequencies ;)

In fact the whole thing could be chucked together on a bit of vero board & cased up, maybe not by yourself, but trust me it's not rocket science :)

Thanks Mark, that would be a s*it load cheaper too. :lol:
I've just been reading the manual for my subs and it seems they have 24db/octave filters already. I know the lowest setting is only 35Hz but that's pretty close to what you're suggesting. So perhaps I could get this working after all.

Here's some details on my speakers to help with calculations etc:

Crossover frequencies: 700Hz, 4KHz
Frequency Fb: 35Hz
3dB point: 30Hz
10dB point: 23Hz

I looked into making a variable phase control device a year or so ago and succeeded in finding a design but when I ask for some help with it on a DIY HiFi website everyone recommended that I didn't bother and said they did more harm than good. They said that each frequency that passed through it would get phase shifted by a different amount, so the results were very unpredictable. So I gave it up as a bad idea. What's your opinion on this?

I'm going to do some more testing with my subs set at 35Hz to see how that works. I tried it once before but as I mentioned before it made the subs sound separate so that the sound noticeably jumped between the speakers and subs. I'll let you know what I find this time.

Cheers. :)

[Reid Malenfant]

Go & give it another go but before you give up get yourself some open cell foam. I'll tell you why shortly :eyebrows:

It appears your speakers may well be -6Db @ 25Hz :eek: This would be the place to cross over from the speakers to the subs, it means that the subs will not really be doing a lot & frankly as they are only a 10" in a sealed box i should imagine they'll add very little tbh. What they will do though is fill in a bit of a hole that the main speakers can't do as they'll be reflex loaded. Effectively below the port resonance frequency they'll produce nothing of real value.

What you could try & do is plug the ports with some foam, this will give them a lower box Q & be somewhere in between a reflex & sealed box, it'll raise the low frequency cuttoff & possibly make it much easier to integrate with the subs. You'd probably need to set the subs to about 40Hz (at a guess).

The main speakers bass would then be a tad drier & they'd roll off from a higher frequency but the transient response would improve. In fact they'd probably match in with the subs much better as the loading (sealed box) would be similar.

It might be worth giving that a go before even contemplating electronics :eyebrows:

On the phase control, yes i'd agree but then the same happens in all crossovers so it's no big deal & in fact it's what you want.

[HighFidelityGuy]

Thanks Mark, that all makes sense.
BTW, I think I know what the foam's for. I have lots of THESE (http://www.studiospares.com/bass-traps/acoustic-corner-cube-bass-absorber-grey/invt/465020/) stuffed in every available corner. It was the first thing I tried after seeing an RTA plot of my room. Oh yeah, I have an RTA, so I can provide measurements. I forgot to mention that. My main acoustic problems are room modes at 125, 300ish and 600ish Hz. I suck these out with EQ. I also have a big dip at 400 but I don't touch that as EQ won't help. I need some diffusion on my rear wall but that's for another day.

If the foam's not for acoustic treatment then I'm interested to know what it's for. My wife might leave me if I buy more. :lol:

Anyway, back to crossovers; as you've noticed my main speakers do drop very low and I really cannot fault them on that front, it's just that because they roll of gradually I need to re-enforce some of the frequencies that they produce, so that's where I run into phase issues.

I like the idea of plugging the ports, I hadn't thought of that and I can see that helping. I'll test out a few things and get back to you.

Cheers. :cool:

[HighFidelityGuy]

Well it didn't take me long to run through the main tracks that used to cause me phasing issues. Whereas I used to have to change the phase setting on the subs depending on which of these tracks I listened to I can now successfully listen to them all with the subs set to 35Hz, 180 degrees out of phase. The speakers and subs seem to be integrating together much better than before as well, so I don't know why this didn't work the first time I tried it. :scratch: I have changed a bunch of other things in my setup since then though, so that could explain it. Either way it's all working much better than it ever has before so I'm a very happy chappy. :cool:I'll still play around with plugging the ports out of interest.

So now that that problem is solved it's just the issue of impedance matching to sort out. How difficult would it be to mod my subs to raise their input impedance?

Cheers. :cool:

[Reid Malenfant]

Probably not that difficult if you can get them apart & change a resistor or maybe two. You'll need to be able to solder :eyebrows:

I'm pleased things are improving :)

[HighFidelityGuy]

Things certainly are improving, thanks for the help!

I can solder and I've already had one of the sub's in bits to fix a rattle, so I've got an idea of how they go together. I'll take one apart again and see if I can get a photo of the circuit sorted. Cheers. :)

[DSJR]

Dave, all you really need is to get the speaker/sub interface sorted ;)

You don't need to buy a pre amplifier, all you need is to get a crossover sorted out for the subs & set the built in one to as high as it'll go.



If this reply was directed at me, I only suggested a Gator output board for the Caiman because it may have an easier time driving multiple amps (?). The volume control on the Caiman could then be used instead of the passive contol.

Apologies if the above suggestion confuses things.

[Reid Malenfant]

:lolsign: Hi Dave R, no, it was directed at Dave with the subs :eyebrows:

No worries anyway fella ;)

Actually you do make a very good point there, no need to mod the subs to increase input impedance :)

[HighFidelityGuy]

I'm interesting in trying the Gator output stage but I can't use the volume control on the Caiman as it's not remote controllable and I'm going to need analogue inputs soon, so a preamp will be required. So with that in mind, if I used the Gator with my passive pre would I still need to raise the input impedance on my subs? I'm guessing I will due to the 10K potentiometer in my passive pre. Cheers. :)

[HighFidelityGuy]

I've got the day off work today, so I've managed to get one of my subs dismantled.
I tried to take a photo of the PCB but my camera isn't good at macro shots so it came out too blurry to use. So I'll try to explain the input stage.....

The signal pin of the mono RCA input connects to a 10K resistor which I'll call R1. The other side of R1 connects to the inverting input of a TL071CN (http://www.st.com/stonline/products/literature/ds/2296.pdf) opamp. The inverting input also connects to another 10K resistor (R2) and the other side of this connects to the output of the opamp. The non-inverting input of the opamp appears to be connected to ground.

Does that provide enough info for you to work out what needs changing to raise the input impedance? I'm hoping it's just a case of swapping R1 or R2 or both.

Cheers. :)

[Reid Malenfant]

Yes, all you need to do to raise the input impedance is change both of those resistors :) I'd suggest 100K ohm which isn't too high which might cause a bit of noise & not too low. The Goldilocks zone :eyebrows:

[HighFidelityGuy]

Awesome, thanks Mark! :cool:

I did some thinking about the fact that I'm running my subs and power amps in parallel with the output of my passive attenuator. My power amp's manual says they are 50K input in the specs but says they are 47K on another page. :rolleyes: So based on standard resistor values I'm guessing they are 47K input. So once my subs have a 100K input will that mean that the overall input impedance my attenuator sees is roughly 31.9K? Cheers. :)

[Reid Malenfant]

Yep, bang on the money ;) 31K973 :)

[HighFidelityGuy]

Great, so I've figured out that bit. :)
So if my attenuator is seeing a 31K973 load and my attenuator is 10K, does that mean that my DAC is seeing a maximum load of 7617.5K?

If that's correct, then applying the 1:10 source to load ratio that means that my DAC needs to have an output impedance of 761.75 Ohms or less to drive that load properly? Is that right?

Thanks. :)

[Reid Malenfant]

Spot on again, you should find that if you are going to Gator your Caiman the output impedance will be likely to be no more than 100 ohms, maybe quite a bit less :)

[HighFidelityGuy]

Awesome, thanks Mark. :)

I'm interested in finding out what the output impedance of my Caiman currently is with the passive mod. I've read THIS (http://en.wikipedia.org/wiki/Output_impedance) and now understand how to take the measurement but it doesn't mention anything about if any signal should be being played. Should I play a test tone or anything?

BTW, I've bypassed the Caiman's muting circuit, so the output coupling caps are connected directly to the DAC's output pins.

Apologies for all the questions. :eek:

Cheers.

[Reid Malenfant]

Do you have a multimeter? If yes & you have some form of test CD or some way of injecting a constant digital signal into the DAC then it shouldn't be too difficult.

Get yourself a precision resistor, 1% metal films are cheap enough :eyebrows: I'd suggest 1K ohm should be enough to get a reliable measurement. You only need to do this on one channel ;)

Disconnect all outputs from the DAC & spin your test disc with say a 1KHz test tone & measure the AC voltage on the DAC output. Note it down as V1 ;)

Connect your 1K ohm resistor (call this resistor "R") onto the output from centre pin to screen & spin the disc again with the same test tone. Measure the output & see how much it has dropped, note it down again as V2.

To then calculate the output impedance "Z" do the following math:-

R x (V1 - V2)/ V2

[Reid Malenfant]

As an example, if V1 is 1V RMS & V2 is 750mV RMS then the output impedance is 333 ohms.

1000 x (1 - 0.75)/ 0.75 = 333 ;)

[HighFidelityGuy]

I have a multimeter but it's a bit cheapo. I'll borrow a Fluke from work and order a 1K resistor at the same time as the 100K's for the subs.

Many thanks for all your help. :cool:

[HighFidelityGuy]

As a side note, I realised last night that I could easily test out whether simply having a higher load impedance provided any sonic benefits by physically disconnecting my subs. So today I first listened to some tracks with my subs connected but switched off and then listened to the same tracks with the subs disconnected. I was quite surprised at the level of improvement I heard by disconnecting them. The treble was smoother and the bass dropped lower and had more punch too it! The bass now almost drops as low and loud as when my subs are connected and working properly but not quite. :eek: It seems the reduced input impedance has been holding back my main speakers from showing what they really have to offer.

I'm hoping that raising the input impedance of my subs will allow this level of performance when the subs are connected but switched off. However this has got me wondering if while I've got the subs in bits I can incorporate a relay into the input stage so that it gets disconnected when the subs are put into standby. That way I get all the benefits all the time and non of the down sides. :scratch::eyebrows:

Does anyone know what type of relay would be suitable for this? I'm not sure what voltages I'll have available for the coil yet. I'll take some measurements when I get the Fluke from work.

Cheers. :)

[HighFidelityGuy]

Well I think I've found a suitable relay for my input disconnecter idea but I've not had any luck finding a switched voltage source inside the sub. I'll see if I can get hold of a service manual.

Failing that I've had an alternative idea. When the subject the use of buffers to solve this was suggested I thought that I would need two, one for the subs and one for the power amps. However, if I just got one and connected this between my attenuator and the subs and amps would that still work ok? That way instead of the subs and amps being in parallel with the output of the attenuator they'd be in parallel with the output of the buffer. Or would having the buffer between my DAC and attenuator be better? :scratch:

Cheers.

[Reid Malenfant]

The type of relay that i'd use is ideal for audio & cheap ;) If you decide to go ahead get a couple of BT47 relays.

If you go with a buffer i'd connect it after the attenuator. You could connect it so it drives the subs & power amps or just the subs, or just the power amps :eyebrows:

I feel pretty sure that in standby the main power rails feeding the sub amps will be at zero volts, i honestly can't imagine that the power amps will be powered up whilst they are in standby :scratch:

[HighFidelityGuy]

Thanks for the suggestions Mark. I'll look into those relays and have a think about what to do about a buffer. I think the simplest solution that will tick the most boxes will be to install a buffer after my attenuator as you suggested and use that to drive the amps and subs. I'm guessing doing that will kind of isolate my DAC from the input impedance of the subs and amps and provide a higher overall impedance. That way I shouldn't need to worry about the subs messing up the impedance quite so much. I'll still raise the input impedance of the subs though.

I've not been able to figure out exactly what gets switched off when the subs get put in standby. There are two main PCB's and a display panel with LED readout. When in standby the LED readout switches of and no sound comes out the driver but the whole PCB the handles the audio input and filtering stays switched on. The other PCB is the amp but that's difficult to get a mulitmeter to due to cable length issues preventing me from powering the amp while it's out of the enclosure. To be honest I think my relay idea is a non-starter as it looks like it's going to be more bother than it's worth. I think using a buffer will give me more benefit all round, so I'll concentrate on that. I've also put my name down for a Gator board, so I'll see if that helps.

Cheers. :cool: