I will shortly be building an economy (I'm trying to avoid the word 'cheap'!) active crossover for a two-way system. (I’m on a very tight budget but already have a pair of good amps) I am trying to decide which kit to purchase. Marchand, HEL, Silicon Chip or something else out there. I’m very interested in any members experiences with any of these or anything else.
The Marchand XM1 has been mentioned a few times on these posts and has reasonable figures for noise and distortion. It is also quite reasonably priced at US$25 for board and parts (2 required for stereo)
http://www.marchandelec.com/xm1.html
Harmonic distortion at 1KHz: less than 0.01%
Signal to Noise ratio: better than 110dB
HEL (High Efficiency Loudspeakers) market a stereo kit for US $86 which includes the power supply.
https://secure11.websitecomplete.com/sharrongrp/shop/showDept.asp?dept=531
http://www.highefficiencyloudspeakers.com/SchematicDiagrams.html
They have responded to my email regarding figures that the distortion etc are too low for them to measure (?) and assuring me that extremely low noise and distortion are critical to their design.
The third option is the project from “Silicon Chip”Jan 2003. A kit is not available, however the cost of the board is AU$30 and parts would be under $50 or $60
http://www.siliconchip.com.au/cms/A_30278/article.html
Total harmonic distortion Typically .003% at 1V RMS
Signal to noise ratio -94dB unweighted (22Hz to 22kHz) with respect to 1V RMS
The Marchand XM1 has been mentioned a few times on these posts and has reasonable figures for noise and distortion. It is also quite reasonably priced at US$25 for board and parts (2 required for stereo)
http://www.marchandelec.com/xm1.html
Harmonic distortion at 1KHz: less than 0.01%
Signal to Noise ratio: better than 110dB
HEL (High Efficiency Loudspeakers) market a stereo kit for US $86 which includes the power supply.
https://secure11.websitecomplete.com/sharrongrp/shop/showDept.asp?dept=531
http://www.highefficiencyloudspeakers.com/SchematicDiagrams.html
They have responded to my email regarding figures that the distortion etc are too low for them to measure (?) and assuring me that extremely low noise and distortion are critical to their design.
The third option is the project from “Silicon Chip”Jan 2003. A kit is not available, however the cost of the board is AU$30 and parts would be under $50 or $60
http://www.siliconchip.com.au/cms/A_30278/article.html
Total harmonic distortion Typically .003% at 1V RMS
Signal to noise ratio -94dB unweighted (22Hz to 22kHz) with respect to 1V RMS
Check out the MOX group buy as well...
And if 1st order (or droopy 2nd order) is suitable a PLLXO is really cheap...
http://www.t-linespeakers.org/tech/filters/passiveHLxo.html
dave
And if 1st order (or droopy 2nd order) is suitable a PLLXO is really cheap...
http://www.t-linespeakers.org/tech/filters/passiveHLxo.html
dave
thanks for the info Richie00boy and planet 10.
Although having no active components appeals to me I am drawn by the steeper slopes that an active design can provide, so PLLXO is out.
I had a look at Rod Elliott's site and it's pretty interesting. I had a close look at the circuit on it for the 2way and then had a closer look at the circuits on the HEL and Silicon Chip sites and realised that the three are almost identical. Don't know why I was suprised! Anyway two differences are that SC and RE use an extra op amp for an output buffer. Possibly a bit unnecessary - I'd prefer a lower number of amps in the signal path if possible.
Another difference is that as the SC version is 3way, and even if I only use two channels, there are more op amp stages (TL074 )used for each mid and LP section (six each!). I would need to add a link to bypass them.
The Marchand also has a number of LF353 op amp stages (six for two channels) but the circuit is quite different to the other three (HEL, SC and RE) and does not have an output buffer.
I am starting to lean towards either the HEL option - single board, no output buffer (5 NE5532 op amp stages per channel), incorporating the power supply or a pair of Marchand XM1's.
If the links in the original post don't work, try these.....
http://www.highefficiencyloudspeakers.com/
and this one for Rod Elliott.....
http://sound.westhost.com/projects.htm
Although having no active components appeals to me I am drawn by the steeper slopes that an active design can provide, so PLLXO is out.
I had a look at Rod Elliott's site and it's pretty interesting. I had a close look at the circuit on it for the 2way and then had a closer look at the circuits on the HEL and Silicon Chip sites and realised that the three are almost identical. Don't know why I was suprised! Anyway two differences are that SC and RE use an extra op amp for an output buffer. Possibly a bit unnecessary - I'd prefer a lower number of amps in the signal path if possible.
Another difference is that as the SC version is 3way, and even if I only use two channels, there are more op amp stages (TL074 )used for each mid and LP section (six each!). I would need to add a link to bypass them.
The Marchand also has a number of LF353 op amp stages (six for two channels) but the circuit is quite different to the other three (HEL, SC and RE) and does not have an output buffer.
I am starting to lean towards either the HEL option - single board, no output buffer (5 NE5532 op amp stages per channel), incorporating the power supply or a pair of Marchand XM1's.
If the links in the original post don't work, try these.....
http://www.highefficiencyloudspeakers.com/
and this one for Rod Elliott.....
http://sound.westhost.com/projects.htm
despite reading many references)The HEL kit seems reasonably priced as far as such things go. I do think the extra buffer is a good idea if you are using the pots and driving lines...if the amps end up in the same box, then no worries.
I've also experienced nothing but good things from Rod Elliot. It's a bonus that you and he are both in Australia.
If you are interested in checking out MOX, there is a group buy open for the next week. The topology is very similar to the HEL...it offers a few more features, but it would be harder to setup than a complete kit.
http://www.diyaudio.com/forums/showthread.php?goto=lastpost&forumid=46
I've also experienced nothing but good things from Rod Elliot. It's a bonus that you and he are both in Australia.
If you are interested in checking out MOX, there is a group buy open for the next week. The topology is very similar to the HEL...it offers a few more features, but it would be harder to setup than a complete kit.
http://www.diyaudio.com/forums/showthread.php?goto=lastpost&forumid=46
The MOX looks very interesting...... took me a while to wade through the original thread, it's pretty long. However as I don't require any EQ I think you are right in that it is probably a bit of overkill and would take a bit more to set up.
The Marchand circuit is beyond my meager circuit knowledge.....I can see how the output of the chain of op-amp stages is fed back to the inverting input at the begining but the analysis is to difficult for me to grasp. I'll see if I can find some "state variable filters for dummies" site to satify my curiosity. Never-the-less the performance appears good.
I've also had a rethink on the desirability of an output buffer stage. Even though the interconnects will be pretty short and wont present a problem, the two power amps I'll be feeding are quite different, so perhaps a buffer is needed to avoid any changes in the response caused by the different power amp input impedances. Is this correct?
The Marchand circuit is beyond my meager circuit knowledge.....I can see how the output of the chain of op-amp stages is fed back to the inverting input at the begining but the analysis is to difficult for me to grasp. I'll see if I can find some "state variable filters for dummies" site to satify my curiosity. Never-the-less the performance appears good.
I've also had a rethink on the desirability of an output buffer stage. Even though the interconnects will be pretty short and wont present a problem, the two power amps I'll be feeding are quite different, so perhaps a buffer is needed to avoid any changes in the response caused by the different power amp input impedances. Is this correct?
I get lost trying to trace state-variable filter circuits as well. My 3rd-grade-level understanding of the basic concept is pretty easy though. The idea is you can derive the highpass signal by subtracting the lowpass signal from the full-range signal. That's good for adjustable XOs because it always keeps the highpass and lowpass synchronized as you change Fc.
Yep, they're convenient because a four-gang pot can accomplish an adjustable frequency capability. However, state-variable filters have a couple of problems. The "derived" nature means you can't have slightly different high/low pass crossover frequencies and they also won't support asymetrical (different sloped high/low pass) operation. For the most part neither of those would be a concern, but for those DIYer's that really want to optimize their designs a Sallen-Key topology would be better.
Cheers,
Davey.
Cheers,
Davey.
2 way crossover
Hi Kimbo and all,
There is one more interesting crossover still available (I think) in Australia from Altronics. It is based a design published in Electronics Australia in 1990s. It is a 24dB/octave design and one can choose between Bessel and Butterworth configurations by using different resistor values. One may also change cap values to get lower crossover frequency.
Bessel option is very useful as this configuration allows for best transient response at the expense of somewhat slower roll-off than in the case of Butterworth. 4th order Butterworth here is obtained by cascading two 2nd order filters so in the end we get a Linkwitz-Riley filter.
An all-pass filter is used to mimic the phase response of the high pass filter and then by subtraction the low pass filter is derived. As a result both outputs are phase consistent.
Altronics sells PCBs with all components but most of these are good for nothing. One has to replace all caps with polypropylene, resistors with preferably low noise ones and most importantly replace TL ICs with at least OPA132/2132P for lower noise.
I have most of the original article scanned so I can email you a copy so you could take a look at the schematics befor making any decisions. If want the copy email me on:
janusz.pradzynski@dpi.wa.gov.au
Cheers,
Hi Kimbo and all,
There is one more interesting crossover still available (I think) in Australia from Altronics. It is based a design published in Electronics Australia in 1990s. It is a 24dB/octave design and one can choose between Bessel and Butterworth configurations by using different resistor values. One may also change cap values to get lower crossover frequency.
Bessel option is very useful as this configuration allows for best transient response at the expense of somewhat slower roll-off than in the case of Butterworth. 4th order Butterworth here is obtained by cascading two 2nd order filters so in the end we get a Linkwitz-Riley filter.
An all-pass filter is used to mimic the phase response of the high pass filter and then by subtraction the low pass filter is derived. As a result both outputs are phase consistent.
Altronics sells PCBs with all components but most of these are good for nothing. One has to replace all caps with polypropylene, resistors with preferably low noise ones and most importantly replace TL ICs with at least OPA132/2132P for lower noise.
I have most of the original article scanned so I can email you a copy so you could take a look at the schematics befor making any decisions. If want the copy email me on:
janusz.pradzynski@dpi.wa.gov.au
Cheers,
I debated on whether to add an output buffer or not. I was torn between trying to minimize stages, and trying to cover all bases. Since no problems were encountered on any system I tested the crossover on, the decision was made not to add another stage. So far, no one has complained about any incompatibilities. I imagine there are some amplifiers out there that may need a buffer stage, but for 99% of the amps out there, one should not be required. The crossover has been used on a variety of both solid state and tube systems, and I have only received positive comments. Here is an e-mail I received just yesterday:
“Hey Ron. How are you. I thought I would let you know just how great I think this crossover of yours is. I have compared it with two other electronic crossovers and yours definitely sounds the best. You can really hear a difference. This thing must have very low distortion. The other ones are brittle sounding by comparison. For a two way, the 800 hz is the way to go. Very quiet too. regards.”
This is fairly typical of the feedback I get on the unit. My comments to Kim about not being able to measure the noise or distortion was more in reference to the limitations of my test equipment more than anything else. Below about 0.02%, I find it difficult to measure with any accuracy.
The circuit is fairly typical text book topology, so similarities with other units is not surprising. Overall, I don’t think anyone will find a better value. The part quality is extremely high, which is what sets the HEL kit apart from the others (plus the on board power supply, which is a huge advantage for most people).
Thank you, and enjoy the music
Ron Shaw
H.E.L.
www.highefficiencyloudspeakers.com
“Hey Ron. How are you. I thought I would let you know just how great I think this crossover of yours is. I have compared it with two other electronic crossovers and yours definitely sounds the best. You can really hear a difference. This thing must have very low distortion. The other ones are brittle sounding by comparison. For a two way, the 800 hz is the way to go. Very quiet too. regards.”
This is fairly typical of the feedback I get on the unit. My comments to Kim about not being able to measure the noise or distortion was more in reference to the limitations of my test equipment more than anything else. Below about 0.02%, I find it difficult to measure with any accuracy.
The circuit is fairly typical text book topology, so similarities with other units is not surprising. Overall, I don’t think anyone will find a better value. The part quality is extremely high, which is what sets the HEL kit apart from the others (plus the on board power supply, which is a huge advantage for most people).
Thank you, and enjoy the music
Ron Shaw
H.E.L.
www.highefficiencyloudspeakers.com
I purchsed a Marchand XM-44 four way crossover with 24dB/oct linkwitz-Reily slopes about a year ago for about $800US. Just for fun, I built my own four way version using Vector project boards (same as Radio Shack) at a cost of about $200US. The one I built does sound better, but I don't know why. It could be the XM-44 uses relays at the outputs. A-B comparison tests show a definte improvement with my version. The XM-44 is very, very good however. A description of the block diagram is as follows:
1. Input board: AC coupling capacitors are 1.0 uf ploypropelene, one for each channel, followed by balanced to single ended conversion (my stereo system is all balanced). The balanced to single ended op amps also serve as an input buffer which is required to isolate the x-over from your preamp. From here, you feed a common line to each of the filter sections, in my case there are four filter sections.
2. Low pass filter at 55 hz, then a Shiva active EQ, then Single ended to balanced. This is for the pair of 12 inch Shiva subs.
3. Bandpass filter: High pass at 327hz followed by low pass at 55hz filters for the 12 inch Lambda Acoustics mid bass speakers, then Se to Balanced.
4. Bandpassfilter: High pass at 3000hz foloowed by low pass at 327 hz, then SE to balanced, for the PHL 6.5 inch midrange speakers.
5. High pass filte at 3000hz, the SE to bal for the Raven R2 ribbon tweeters.
If you just want a two way system, all you will need is an input section to provide AC coupling ( not necessary, but recommended) and a buffer followed by a high pass filter for the tweeters and a low pass filter for the woofer. the outputs from each filter do not need buffers assuming your amp input impedences are reasonable at about 47K ohms. You must, however, include 49.9 ohm resistors for each output. Use only 1% metal film reistors and polypropelene caps, nothing else. I used the "equal component" approach which makes selecting the resistor values for a particular x-over point a breeze. All filter caps are Panasonic polypropelene 3300pf execpt for the 55hz low pass filter in which case I used 6800pf silver mica.
The ESP and Linkwitz Lab sites are good resources. DIY it and have fun.
1. Input board: AC coupling capacitors are 1.0 uf ploypropelene, one for each channel, followed by balanced to single ended conversion (my stereo system is all balanced). The balanced to single ended op amps also serve as an input buffer which is required to isolate the x-over from your preamp. From here, you feed a common line to each of the filter sections, in my case there are four filter sections.
2. Low pass filter at 55 hz, then a Shiva active EQ, then Single ended to balanced. This is for the pair of 12 inch Shiva subs.
3. Bandpass filter: High pass at 327hz followed by low pass at 55hz filters for the 12 inch Lambda Acoustics mid bass speakers, then Se to Balanced.
4. Bandpassfilter: High pass at 3000hz foloowed by low pass at 327 hz, then SE to balanced, for the PHL 6.5 inch midrange speakers.
5. High pass filte at 3000hz, the SE to bal for the Raven R2 ribbon tweeters.
If you just want a two way system, all you will need is an input section to provide AC coupling ( not necessary, but recommended) and a buffer followed by a high pass filter for the tweeters and a low pass filter for the woofer. the outputs from each filter do not need buffers assuming your amp input impedences are reasonable at about 47K ohms. You must, however, include 49.9 ohm resistors for each output. Use only 1% metal film reistors and polypropelene caps, nothing else. I used the "equal component" approach which makes selecting the resistor values for a particular x-over point a breeze. All filter caps are Panasonic polypropelene 3300pf execpt for the 55hz low pass filter in which case I used 6800pf silver mica.
The ESP and Linkwitz Lab sites are good resources. DIY it and have fun.
After much deliberation I finally decided to go with the Marchand XM1.
As much as I'd love a Behringer or the Marchand XM44, I'm on a tight budget (and anything would be an improvement over my current low-fi system)
The Silicon Chip design could have been used by modifying the board to reduce it to two way - I would also then have the option to by-pass the buffer and try without in order to simplify the circuit path. On board power supply is also attractive. However the difficulty in obtaining high tolerance capacitors put this option and the similar Elliot design (great website though) out of the running. P.S. do any Aussies know where to obtain 1% or 2% tolerance caps in Melbourne? - I don't have a capacitance meter to match my own.
This left the HEL and the Marchand. The HEL has lots going for it..... hi tolerance parts kit supplied, on board power (I could ask that the 120 v - 12v wall pack be left out to reduce postage and provide my own 240v version), the input/output RCAs on board, etc. and Ron Shaw's quick and informative responses to my queries. However I want to play around with the crossover frequencies. The HEL only comes in three freqs, and I would need to supply and solder different caps & resistors to the board in order to experiment. Many thanks anyway Ron!
That leaves the Marchand XM1.......easy to change crossover freq with plug in resistors, on a DIP plug and although the power supply is separate and an extra cost, it is 120/240v (yes I know, I could have easily cooked up my own +/- 15v supply and reduced postage).
Many, many thanks to everyone for their suggestions.
Just one more related question........many passive crossovers include an impedance equalisation network to cancel out the rising impedance of the woofer(s) (in my case 0.8mH x 2). Is this still required when using an active crossover? (I'm crossing at somewhere around 3kHz as it is a two way). Having passive components after the power amp seems to defeat the purpose of going active. If it is required, is it possible to put it just after the output of the active crossover? Or do I just use the eq on the preamp to try to smooth the response?
As much as I'd love a Behringer or the Marchand XM44, I'm on a tight budget (and anything would be an improvement over my current low-fi system)
The Silicon Chip design could have been used by modifying the board to reduce it to two way - I would also then have the option to by-pass the buffer and try without in order to simplify the circuit path. On board power supply is also attractive. However the difficulty in obtaining high tolerance capacitors put this option and the similar Elliot design (great website though) out of the running. P.S. do any Aussies know where to obtain 1% or 2% tolerance caps in Melbourne? - I don't have a capacitance meter to match my own.
This left the HEL and the Marchand. The HEL has lots going for it..... hi tolerance parts kit supplied, on board power (I could ask that the 120 v - 12v wall pack be left out to reduce postage and provide my own 240v version), the input/output RCAs on board, etc. and Ron Shaw's quick and informative responses to my queries. However I want to play around with the crossover frequencies. The HEL only comes in three freqs, and I would need to supply and solder different caps & resistors to the board in order to experiment. Many thanks anyway Ron!
That leaves the Marchand XM1.......easy to change crossover freq with plug in resistors, on a DIP plug and although the power supply is separate and an extra cost, it is 120/240v (yes I know, I could have easily cooked up my own +/- 15v supply and reduced postage).
Many, many thanks to everyone for their suggestions.
Just one more related question........many passive crossovers include an impedance equalisation network to cancel out the rising impedance of the woofer(s) (in my case 0.8mH x 2). Is this still required when using an active crossover? (I'm crossing at somewhere around 3kHz as it is a two way). Having passive components after the power amp seems to defeat the purpose of going active. If it is required, is it possible to put it just after the output of the active crossover? Or do I just use the eq on the preamp to try to smooth the response?
OK, I'll not bother with impedance compensation. If after extended listening it appears to be a problem, I'll temporarilly drop in an extra variable eq section at the low output of the xover and experiment.
I did consider using a large cap to protect the tweeter, but I'll wing it without for starters.....
* the tweeters are pretty low cost
* the high pass is being fec back into the main amp of the NAD that I'm using for the preamp, and it has relay speaker protection (but I'll check it on the scope before wiring up the speakers the first time!)
* I'd like to avoid any phase problems
* if I do eventually blow the tweeters, it's a great excuse to upgrade them! (and add that protection cap) ;-)
Cheers,
Kim
I did consider using a large cap to protect the tweeter, but I'll wing it without for starters.....
* the tweeters are pretty low cost
* the high pass is being fec back into the main amp of the NAD that I'm using for the preamp, and it has relay speaker protection (but I'll check it on the scope before wiring up the speakers the first time!)
* I'd like to avoid any phase problems
* if I do eventually blow the tweeters, it's a great excuse to upgrade them! (and add that protection cap) ;-)
Cheers,
Kim
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.