Crossover question

[Brian]

Blokes....

The gif is the crossover for the KLS3 MKII Gold, which uses the Audax HM130CO mid unit.

Anyway, which components in the mid section determine the lower and upper crossover points?

According to the text in the diy supplement, the HF unit cuts in at above 5Khz, and the bass unit works upto 400Hz, so the mid must come in at that point. Hence I assume it's operating from 400Hz to just over 5Khz.

The Audax HM100ZO can still be purchased, I've been told it's a better unit than the HM130CO in the KLS3. If I wanted to replace the mid unit with the 100ZO, would these same crossover components cause the upper and lower ends of the mid to be at the same frequency points? Basically, would it work, or would it need a total redesign.

The 100CO is used with the HD3P Hf unit in the KLS11 design and crosses over at 5.5Khz, and to the bass unit at 800Hz. Could I use this crossover in the KLS3?

Thanks

Brian

[Ianm2]

there is actually no text book formula for the way xovers are done.

the electrical roll off of the filter, adds to the acoustical rolloff of the drive units, its from this measurement the electrical values are calculated and adjusted, just adjusting the value without measuring is futile and doomed to failure, unless you understand the way the drive unit and filter interact.

The series component usually sets the cutoff approximately, and the shunt component will modify the 'knee' and roll off rate, but its more complex than that.

[Yellow]

Brian,

To answer your question about the "components" that affect the x/o point:

In this mid-range (band-pass) filter, it is configured as what is called a 2nd-order filter (sort-of!!!). In other words, the theoretical slope of the roll-off of the response will be 12dB/octave. The roll-in is 1st-order. There appears to be a cap (C5) in series, and just a balancing resistor across the terminals which simply changes the electrical characteristics of the DCR that the cap/amp sees!! This is a very strange configuration and to be honest I would not recommend it at all. The roll-in and roll-off should be the same "order", really, otherwise phase problems crop up big-time. So we have effectively two 50uF caps in series which is 25uF. The resistor across C4 will be damping its effectiveness somwhat, but I think you could ditch the C4/R3 combination and replace C4/5 with a single 25uF cap without too many issues. You may need to play with R2 to tweak its balance.

Look at the bass-unit x/o. you have an inductor (L2) in series and a capacitor (C3) in parallel across the driver. Obviously, these components result in the roll-off (Low-pass) off the response at a certain point.The mid-range is a combination of exactly this configuration and the opposite, a high-pass filter, where the cap is in series and the inductor is in parallel.

So, L1 & C2 roll-off your response for your mid-range and C4/5 is the feed cap which determins when the response feeds in.

One big misconception with drive-units is the cross-over point. The cross-over, is simply where the two frequency responces crossover! They don't STOP at these frequencies. A tweeter, with a 1st order filter could easily be operating usefully at 400Hz. Likewise, a bass-unit in a normal 2-way system will still be giving useful output upto and beyond 8-10kHz, even though its x/o may be at 2.8kHz. At the x/o the response is simply rolling off, so at 6kHz it may be 12dB down on the natural level, and at 12kHz, it'll be a theoretical 24dB down, which is still useful output.

Hope this helps a little......! PM me if you need further help.

Regards,

Neil.

[Brian]

Originally Posted by Yellow

Brian,

To answer your question about the "components" that affect the x/o point:

In this mid-range (band-pass) filter, it is configured as what is called a 2nd-order filter (sort-of!!!). In other words, the theoretical slope of the roll-off of the response will be 12dB/octave. The roll-in is 1st-order. There appears to be a cap (C5) in series, and just a balancing resistor across the terminals which simply changes the electrical characteristics of the DCR that the cap/amp sees!! This is a very strange configuration and to be honest I would not recommend it at all. The roll-in and roll-off should be the same "order", really, otherwise phase problems crop up big-time. So we have effectively two 50uF caps in series which is 25uF. The resistor across C4 will be damping its effectiveness somwhat, but I think you could ditch the C4/R3 combination and replace C4/5 with a single 25uF cap without too many issues. You may need to play with R2 to tweak its balance.

Look at the bass-unit x/o. you have an inductor (L2) in series and a capacitor (C3) in parallel across the driver. Obviously, these components result in the roll-off (Low-pass) off the response at a certain point.The mid-range is a combination of exactly this configuration and the opposite, a high-pass filter, where the cap is in series and the inductor is in parallel.

So, L1 & C2 roll-off your response for your mid-range and C4/5 is the feed cap which determins when the response feeds in.

One big misconception with drive-units is the cross-over point. The cross-over, is simply where the two frequency responces crossover! They don't STOP at these frequencies. A tweeter, with a 1st order filter could easily be operating usefully at 400Hz. Likewise, a bass-unit in a normal 2-way system will still be giving useful output upto and beyond 8-10kHz, even though its x/o may be at 2.8kHz. At the x/o the response is simply rolling off, so at 6kHz it may be 12dB down on the natural level, and at 12kHz, it'll be a theoretical 24dB down, which is still useful output.

Hope this helps a little......! PM me if you need further help.

Regards,

Neil.

That's pretty informative, Neil. Just what I was looking for. A bit of tweaking is in order.

Thanks

Brian

[petercom]

Originally Posted by Yellow

Brian,

In this mid-range (band-pass) filter, it is configured as what is called a 2nd-order filter (sort-of!!!). In other words, the theoretical slope of the roll-off of the response will be 12dB/octave. The roll-in is 1st-order. There appears to be a cap (C5) in series, and just a balancing resistor across the terminals which simply changes the electrical characteristics of the DCR that the cap/amp sees!! This is a very strange configuration and to be honest I would not recommend it at all. The roll-in and roll-off should be the same "order", really, otherwise phase problems crop up big-time. So we have effectively two 50uF caps in series which is 25uF. The resistor across C4 will be damping its effectiveness somwhat, but I think you could ditch the C4/R3 combination and replace C4/5 with a single 25uF cap without too many issues. You may need to play with R2 to tweak its balance.

Neil.

Neil, you are quite right to point out that slopes should be roughly equal in design, but it's acoustic slopes we are interested in, not electrical ones. The piezo treble unit has a limited acoustic bandwidth which, in combination with the first order electrical filter, gives close to second order acoustic roll off.

As for C4/R3 this was put in to achieve a flatter impedance curve. As you say there is room for experimentation here but don't necessarily throw away this advantage for valve amps.

[Brian]

Originally Posted by petercom

Neil, you are quite right to point out that slopes should be roughly equal in design, but it's acoustic slopes we are interested in, not electrical ones. The piezo treble unit has a limited acoustic bandwidth which, in combination with the first order electrical filter, gives close to second order acoustic roll off.

As for C4/R3 this was put in to achieve a flatter impedance curve. As you say there is room for experimentation here but don't necessarily throw away this advantage for valve amps.

I've been thinking a lot about crossovers, even dreaming about them. I have some options that all sound like educational and fun to me.

1. I'm wondering whether I should scrap the C4/R3 and go for 1 x 25uF cap as suggested. However, the question is what kind of cap'. Falcon sells 25uF Solen for £4.99 each, I'd need 2, so it's a cheap experiment by using these. However, I read time and again that Auricap are a vast improvement. To use these I could go to Parts Connexion and order 2 x 20uF and 2 x 5uF and parallel them for a further supposed advantage. They don't have 25uF. This is more expensive though, like ~$120 plus postage!

I'm trying to get a bit more 'sparkle' to the sound, and slightly less 'bloom' in the bass.

2. Really getting into the experimentation lark I could start changing the component values completely. Not sure how to work out the crossover freq from a value of a cap and an inductor, but I know that the inductor in the bass section for standard KLS3 was 12mH. This was changed to 2.7mH to raise the crossover point from 100Hz to 400Hz. Might shifting this back to 12mH/100Hz make much difference, given what I'm trying to achieve?

3. Should I be bumping the crossover point up even more, to give less work to the mid unit? I've read some text about a speaker called 'Polyarti' that uses the same bass and mid units as KLS3, this uses a bass to mid crossover of 800Hz, could I use those values without tampering with the higher end of the midrange crossover section?

Looking at other WAD designs like KLS11, I note that in the high pass crossover section there is a cap and inductor, in KLS3 there isn't. Is this something I could look at experimenting with? Perhaps change values to the ones used in that design, although the mid is a different unit.

Lots of things to think about. I suppose the best recommendation for me would be a good book, or website to learn more about crossovers. It's interesting to me that KLS11 uses the same treble unit but is using completely different component values on the HF and MF sections.

Brian

[petercom]

You can read as many books on crossovers as you like but it won't tell you what to do with your drive units.

There are only two ways to work out a crossover. The easiest is to measure the drivers and put these into a crossover software like LspCAD and then iterate between ideas tried out in a virtual crossover and doing the listening tests on them. But this is expensive.

The other is to try out your ideas based on someone elses design and trim them to suit your tastes. This needs a bucketload of components and a lot of time. But this is very frustrating unless you are extremely lucky and happen to hit it right.

You can't work out the crossover frequency from just the components because the drive units a) don't have a constant impedance and b) don't have a linear frequency response. That's why you need to measure the drivers.

To get more sparkle from the sound try bypassing each 50uF with a small value polypropylene, say 3u3. This won't alter the crossover point much but will allow a better path for hf. The Soniqs PXX range should be suitable.

As for bass bloom you can do quite a lot with capacitors and resistors across the bass unit, somewhere between 50 and 200uF in series with 10 or 15 Ohms would be worth trying. You're flying blind here a bit but at least you can't do much damage to the crossover point.

I'll try and look up the history of the design of the KLS3 if I can and see if I can make some more concrete suggestions.

[Richard]

Brian,

You're considering 2 approaches, one is to modify the xover by changing the component values and the other is to change to better quality caps of the same value.

Try whatever but if you do one at a time you'll find out what is doing what and what works. I can tell you making value changes without measurement can be very confusing.

Having done the change from big value electrolytics to the same value in polyprop I can say it works well and you'll like the difference, better in all areas and no downsides. This is true of the series caps and the shunt in the bass (which improves resolution in a similar way to the series in the treble).

Falcon are the people to use as they'll send exact selected values which is important.

Rich

[Greg.]

Hi Brian,

You've heard my KLS3 based system in a relatively small listening room so you know I don't have bass bloom and the sound is balanced throughout the range and it extends very well. It seems to me after all the work we have done on getting internal damping correct that you are still not sorted. I have also built the MkII version you have and my son uses. He does not have your problem either. I feel you are a bit transfixed on the crossover issue and this won't probably find you a solution. There is something we are missing here. We know the constructor of your speakers deviated from the original design. I think we need to look again at this.

Remember, you need at least 18'' spacing away from walls etc and other attention to room acoustics might be in order. As you know, I have lots of soft furnishings. I've filled my room with eight listeners and all regarded the sound as balanced. If you have the internal damping now correct and the speakers are properly positioned, we need to look again at the basic construction. I really can't believe you're not getting the right result. We might need to look at the rest of the system, after all, the speakers are the last link in the chain and much can be acheived by tuning the rest of the system to suit.

Best wishes,

Greg

PS. If you don't agree with my analysis of my own system, please say so. I won't be offended. It'll help me to understand what your issue is and possibly how to sort it.

[Brian]

Originally Posted by Greg.

Hi Brian,

You've heard my KLS3 based system in a relatively small listening room so you know I don't have bass bloom and the sound is balanced throughout the range and it extends very well. It seems to me after all the work we have done on getting internal damping correct that you are still not sorted. I have also built the MkII version you have and my son uses. He does not have your problem either. I feel you are a bit transfixed on the crossover issue and this won't probably find you a solution. There is something we are missing here. We know the constructor of your speakers deviated from the original design. I think we need to look again at this.

Remember, you need at least 18'' spacing away from walls etc and other attention to room acoustics might be in order. As you know, I have lots of soft furnishings. I've filled my room with eight listeners and all regarded the sound as balanced. If you have the internal damping now correct and the speakers are properly positioned, we need to look again at the basic construction. I really can't believe you're not getting the right result. We might need to look at the rest of the system, after all, the speakers are the last link in the chain and much can be acheived by tuning the rest of the system to suit.

Best wishes,

Greg

PS. If you don't agree with my analysis of my own system, please say so. I won't be offended. It'll help me to understand what your issue is and possibly how to sort it.

No problem with that, Greg. You may well be right, of course it might just be me hankering after something the don't do, and emphasing something that isn't really a problem. Everybody else in my house thinks they sound great!

Maybe I need to look at the CDP, it did give a 'fuller' sound at your place than your turntable so perhaps it's not matched that well. I'll have a think about it and also about the construction of the cabinets. I've already been in touch with Avon Plywood for more info about their veneered MDF and their cutting facilities.

More later........