Re: Hi Noodles...
The Peerless 810653 tweeter is the WES catalogue is quite a surprise. Used it on quite a few speakers and there are no major problems with it and easy to work with. I still like the SEAS 27TFFC better, but it's closer than I thought possible as I never considered Peerless for good tweeters.
Andy Graddon said:
The Peerless 810653 tweeter is the WES catalogue is quite a surprise. Used it on quite a few speakers and there are no major problems with it and easy to work with. I still like the SEAS 27TFFC better, but it's closer than I thought possible as I never considered Peerless for good tweeters.
Zaph said:
Hi John,
I had a discussion last fall at the mad board (should still be in the archive) with somebody called David T (I believe), who wanted to outsource the design of a HT set.
He had taken apart a 850467, but failed to follow up on my question about the details of the motor configuration.
To my understanding, all CSC, CSX and HDS have an alu ring below the pole plate, probably not around the pole piece itself, but on the outside of the VC, hugging the magnet.
If the alu phase plug sat ot directly on top of the pole plate (which it probably does), it would provide enough shielding in itself. So where does the copper ring sit? Inside the gap? Probably not, because its Bxl is only slightly lower (8.0 vs. 8.9, so probably within tolerance).
I'd love to know, because I have a couple of CSC217 (#850128, probably an OEM version for the German Peerless distributor) that have pretty high distortion compared to the 850477. If putting a copper or alu ring or peg onto the pole piece is all that is needed, I'd probably venture to unglue the dustcap.
I have a few HDS182PP (again, an OEM version with softer suspension but same motor data), and I can confirm that stuff about the cone (no holes, former extended to dustcap).
As I said, its lower inductance is responsible for the more extended and seemingly more wiggly top end. Just try to roll it off at the same point the 850439 begins to roll of, and it will look very similar. The remaining difference is below 1-2 dB and can probably be attributed to the stiffer cone center.
Regards,
Eric
David did answer, the copper ring sits on the pole piece, not inside the gap.
http://www.madisound.com/cgi-bin/archive_discuss.cgi?read=324549
His answer also contains a link to the original thread.
http://www.madisound.com/cgi-bin/archive_discuss.cgi?read=324549
His answer also contains a link to the original thread.
So if there is no copper inside the gap, only between pole piece and phase plug (which seems to be made from solid Al), I wonder why they use copper at all. A ring of 1.5 mm wall thickness pure Al will have the same shielding effect as 1 mm Cu, which is a typical thickness for such a ring.
Of course, if they used some AlMg alloy for machinability, the conductivity might be lower, but the plug has ample thickness...
The sectional view provided by Peerless does not show the Cu ring, but the data sheet says it has one, and so does David who has dismantled a driver.
Of course, if they used some AlMg alloy for machinability, the conductivity might be lower, but the plug has ample thickness...
The sectional view provided by Peerless does not show the Cu ring, but the data sheet says it has one, and so does David who has dismantled a driver.
SUMMARY OF THE INVENTION
With the present invention it has been found that it is not necessary for the copper cylinder to extend completely over the working length of the voice coil in order to effectively fulfill its purpose. It is sufficient that a short cylinder piece is mounted about the root of the central part of the magnetic system, whereby the associated inductive coupling with the voice coil will ensure that the coil will represent the same impedance whether, at a given moment, it is moving towards or away from the iron filled area at the central core portion of the magnet system. However, it will still be required to stabilize the induction of the voice coil, such that the induction will be substantially the same, no matter where the voice coil is located relative to the air gap, and to this end it has been found that it is fully sufficient to mount a compact copper ring immediately at either side of the air gap, i.e. copper rings which, axially, are substantially shorter than the axial dimension of the voice coil, while being suitably thick in order not to substantially limit the current induced in these rings. Hereby it becomes sufficient that the rings occur as relatively short units, and the totality of three ring parts will thus be able to show a combined axial length considerably shorter than the operational length of the voice coil. This will condition a both lighter and cheaper design of the loudspeakers.
http://patft.uspto.gov/netacgi/nph-...txt&s1=scan-speak&OS=scan-speak&RS=scan-speak
You may need this browser plug-in to view the images....
http://www.alternatiff.com
With the present invention it has been found that it is not necessary for the copper cylinder to extend completely over the working length of the voice coil in order to effectively fulfill its purpose. It is sufficient that a short cylinder piece is mounted about the root of the central part of the magnetic system, whereby the associated inductive coupling with the voice coil will ensure that the coil will represent the same impedance whether, at a given moment, it is moving towards or away from the iron filled area at the central core portion of the magnet system. However, it will still be required to stabilize the induction of the voice coil, such that the induction will be substantially the same, no matter where the voice coil is located relative to the air gap, and to this end it has been found that it is fully sufficient to mount a compact copper ring immediately at either side of the air gap, i.e. copper rings which, axially, are substantially shorter than the axial dimension of the voice coil, while being suitably thick in order not to substantially limit the current induced in these rings. Hereby it becomes sufficient that the rings occur as relatively short units, and the totality of three ring parts will thus be able to show a combined axial length considerably shorter than the operational length of the voice coil. This will condition a both lighter and cheaper design of the loudspeakers.
http://patft.uspto.gov/netacgi/nph-...txt&s1=scan-speak&OS=scan-speak&RS=scan-speak
You may need this browser plug-in to view the images....
http://www.alternatiff.com
Noodle asked me in a private message about the FR response part. Maybe I was not clear enough, so I'll answer here.
The 850439 has an inductance of 1.3 mH according to the data sheet, and this seems roughly right.
The 850467 is spec'd with 1.2 mH, but this seems to be a typo. Looking at the published impedance curve, it's more like 0.2 - 0.5 mH.
So add a series inductor of 0.8 - 1.1 mH, and the response curve will look a lot like the curve of the 850439, i.e. it will begin to roll off the slope with 6 dB/oct beginning at about 1.5 kHz.
The 850439 has an inductance of 1.3 mH according to the data sheet, and this seems roughly right.
The 850467 is spec'd with 1.2 mH, but this seems to be a typo. Looking at the published impedance curve, it's more like 0.2 - 0.5 mH.
So add a series inductor of 0.8 - 1.1 mH, and the response curve will look a lot like the curve of the 850439, i.e. it will begin to roll off the slope with 6 dB/oct beginning at about 1.5 kHz.
What i am confused about is would adding the inductor in series make it a similar difficulty to crossover as the phaseplug-less version?
Using the peerless data-sheet 6db/oct from 1.5khz would roughly bring the graph back down to flat from 1.5-5khz and make it more similar to the other driver in response.
If that is the case I will go for the phase plug and a Seas 27TDFC at this current stage.
Using the peerless data-sheet 6db/oct from 1.5khz would roughly bring the graph back down to flat from 1.5-5khz and make it more similar to the other driver in response.
If that is the case I will go for the phase plug and a Seas 27TDFC at this current stage.
Andy, I'm curious to know what was done to your D25s to damage them like that! Mine seem pretty tough. Given that they are protected by a fairly strong phase shield, I don't know how you could damage them, short of piercing them with something sharp. There's nothing I can see that fingers would do, unless the intent was to damage them, in which case you need grilles! Or locks on doors!
JJ,
How about an intermediate step? Have a go at designing a crossover for one of your existing speakers, see how you go. I see the temptation here to get into something that you might find it difficult to pull off. Or going with drivers for which you have a xo design to fall back on. Another thing to try is tweaking an xo you have and see if you can improve it, and do some measurements to confirm things.
I can certainly see the appeal of the HDS phase plug driver. Linkwitz called it a value for money champion in distortion tests, placing it in between the P17 and excel and SS drivers.
JJ,
How about an intermediate step? Have a go at designing a crossover for one of your existing speakers, see how you go. I see the temptation here to get into something that you might find it difficult to pull off. Or going with drivers for which you have a xo design to fall back on. Another thing to try is tweaking an xo you have and see if you can improve it, and do some measurements to confirm things.
I can certainly see the appeal of the HDS phase plug driver. Linkwitz called it a value for money champion in distortion tests, placing it in between the P17 and excel and SS drivers.
That might be a sensible move, get the measurement equipment, have a fiddle with my existing stuff or some cheap drivers then go onto better things.
However the drivers in my current set of speakers would definately be classed as cheapies, and LSK themselves say it does have a bit of a midrange dip at crossover (3db or so) and that the bass is a bit boomy. However assuming i picked the 27TDFC and either of hte peerless woofers (still un decided) i doubt it would be impossible for me to get something workable going, and if i struggle i will probably just ask for help.
However the drivers in my current set of speakers would definately be classed as cheapies, and LSK themselves say it does have a bit of a midrange dip at crossover (3db or so) and that the bass is a bit boomy. However assuming i picked the 27TDFC and either of hte peerless woofers (still un decided) i doubt it would be impossible for me to get something workable going, and if i struggle i will probably just ask for help.
Peerless 850488
Rabbitz,
Without a steep roll-off, the 850488 has some screaming, shouting sound on the higher frequency range which is quite irritating to my ears and makes me feel fatigue. This happens with both parallel and serial networks of low orders. I used Speaker Workshop to design my XOs and was sure it was not due to frequency irregularities. You can test it by running the 850488 alone and see how it sounds above 3-6k. It doesn't sound that good in my opinion. I might be the interaction with my box. My box is heavily damped and stuffed so this shouldn't be a problem. The only thing I have not done and yet to try is to apply some damping materials on the edges of the panels immediately surrounding the cones.
I agree with you that first order serial XO with the 850488 sounds very special. I found that the sound was very satisfying from this driver with this XO, better than almost anything I have heard. It was sweet, sweet and sweet. Many people believe that the Rogers LS3/5a has been unbeatable in female voice reproduction regardless of speaker prices and makes. When I completed a proper 1st order serial XO with the 850488 and T330D, I had the feeling that the sound on female voices were so accurate and sweet that it must be at least comparable to the 3/5a which I listened to for many times. My impression was that it was better than the full range FE206E I listened to a few months ago housed in a Lowther horn. I am now convinced that if multi-way speakers are designed properly, they can sound as coherant as or better than full-range drivers.
However, I regret that even so I still could not avoid some screaming, shouting problems I had with low order XOs with this driver. There was also a bad lobbing problem I experienced. Slightly moved my head a few inches away, the sound was different. More importantly, I worried about the limited protection on my tweeter. Even with a 13 gauge inductor, there would still be low frequencies getting into the tweeter. The tweeter I use is the Dynaudio T330D that was retailed at $1,000 a pair at Speaker Works. I can't afford to blow it up with the 1st serial XO. So I abandoned the XO. My new design uses a very deep roll-off, yet it still preserves at least 90% of the transient and accuracy of the 1st order serial XO. Now I have maximum protection for the tweeters, a very wide dispersion without any obvious lobbing problems, and it is overall a much better XO, though female voice reproduction is no longer as good as the 3/5a or my previous XO (but only slightly less good).
Part of the reason that the 850488 sounds so good was due to my steep roll-off at the HP as well.
I still have a centre speaker to make and I have 4 x 850528 - shielded version of the 850488 on hands. The tweeter will be different - I am thinking about SS2905-9700. A modified XO will be needed. Any more info on the 850488 is helpful.
Regards,
Bill
Rabbitz,
Without a steep roll-off, the 850488 has some screaming, shouting sound on the higher frequency range which is quite irritating to my ears and makes me feel fatigue. This happens with both parallel and serial networks of low orders. I used Speaker Workshop to design my XOs and was sure it was not due to frequency irregularities. You can test it by running the 850488 alone and see how it sounds above 3-6k. It doesn't sound that good in my opinion. I might be the interaction with my box. My box is heavily damped and stuffed so this shouldn't be a problem. The only thing I have not done and yet to try is to apply some damping materials on the edges of the panels immediately surrounding the cones.
I agree with you that first order serial XO with the 850488 sounds very special. I found that the sound was very satisfying from this driver with this XO, better than almost anything I have heard. It was sweet, sweet and sweet. Many people believe that the Rogers LS3/5a has been unbeatable in female voice reproduction regardless of speaker prices and makes. When I completed a proper 1st order serial XO with the 850488 and T330D, I had the feeling that the sound on female voices were so accurate and sweet that it must be at least comparable to the 3/5a which I listened to for many times. My impression was that it was better than the full range FE206E I listened to a few months ago housed in a Lowther horn. I am now convinced that if multi-way speakers are designed properly, they can sound as coherant as or better than full-range drivers.
However, I regret that even so I still could not avoid some screaming, shouting problems I had with low order XOs with this driver. There was also a bad lobbing problem I experienced. Slightly moved my head a few inches away, the sound was different. More importantly, I worried about the limited protection on my tweeter. Even with a 13 gauge inductor, there would still be low frequencies getting into the tweeter. The tweeter I use is the Dynaudio T330D that was retailed at $1,000 a pair at Speaker Works. I can't afford to blow it up with the 1st serial XO. So I abandoned the XO. My new design uses a very deep roll-off, yet it still preserves at least 90% of the transient and accuracy of the 1st order serial XO. Now I have maximum protection for the tweeters, a very wide dispersion without any obvious lobbing problems, and it is overall a much better XO, though female voice reproduction is no longer as good as the 3/5a or my previous XO (but only slightly less good).
Part of the reason that the 850488 sounds so good was due to my steep roll-off at the HP as well.
I still have a centre speaker to make and I have 4 x 850528 - shielded version of the 850488 on hands. The tweeter will be different - I am thinking about SS2905-9700. A modified XO will be needed. Any more info on the 850488 is helpful.
Regards,
Bill
I have a Peerless 850439 mounted in a diy TL box which had some problems, which I could sense when listening but I didn't know what to fix. Now that I can measure frequency response and impedance, there is some possibility of making improvements to the system and realize the full potential of the driver.
Here is the manufacturer's graph of the 850439...
Here is my own measurement of the unmounted driver with modest home instrumentation [disregard the note upper left corner, it is unmounted in the measurement shown]. It shows a large dip between 3k and 4k that I did not expect to see, it may be real or a result of measurement conditions. It was made by nearfield sine sweep, I will try a few other things tomorrow to see if it makes any difference.
If the dip is a measurement gremlin I will be relieved, because the rest of the graph shows it to be a well-behaved speaker, which is why I bought it.
Does anyone have a measurement for the phase-plug version 850467 ?
Here is the manufacturer's graph of the 850439...
An externally hosted image should be here but it was not working when we last tested it.
Here is my own measurement of the unmounted driver with modest home instrumentation [disregard the note upper left corner, it is unmounted in the measurement shown]. It shows a large dip between 3k and 4k that I did not expect to see, it may be real or a result of measurement conditions. It was made by nearfield sine sweep, I will try a few other things tomorrow to see if it makes any difference.
If the dip is a measurement gremlin I will be relieved, because the rest of the graph shows it to be a well-behaved speaker, which is why I bought it.
Does anyone have a measurement for the phase-plug version 850467 ?
Attachments
Hifinutnut, I wonder if you would be better using a little eq and then going with the first order xo if you like it the best. I would think you should not have a problem with blowing up the tweeter with a passive xo that sounds good. The tweeter should start to lose fidelity and sound hard before there is any risk of damage. Have you considered trying a notch filter on the mid?
I can't say I really have an opinion on first order vs other crossovers where the drivers can handle it, but if you like that better it seems to me the better approach would be to use filters to get the response smooth rather than have to use steeper slopes. Ymmv
That dip appears very strange. Something must be wrong. I would try running a tone sweep through starting from say 2 octives either side, see if you can hear a dip from the position you are measuring in. A little trial and error to remove possible sources of this error. Perhaps you could also measure another speaker to see if you get the same dip.
I can't say I really have an opinion on first order vs other crossovers where the drivers can handle it, but if you like that better it seems to me the better approach would be to use filters to get the response smooth rather than have to use steeper slopes. Ymmv
That dip appears very strange. Something must be wrong. I would try running a tone sweep through starting from say 2 octives either side, see if you can hear a dip from the position you are measuring in. A little trial and error to remove possible sources of this error. Perhaps you could also measure another speaker to see if you get the same dip.
HDS
Paulspencer,
I did the speaker design properly based on Speaker Workshop, my ears and numerous hours of experiments. However, I have not used the measurement capability of SW therefore I wouldn't be able to know what sort of notch filter required. I always believe there must be some real risk of blowing up a tweeter using 1st order serial XO (1OSXO). Looking at a 1OSXO circuit you know this can happen because an inductor can not have 0 DCR. When I tested it, I used only medium sound level and was too worried to turn the volume up, although the Esotar T330D can stand 1000+W surge current.
Yes I am convinced that it is easy to have good transient response from 1st order network, but numerous other problems with this topology make it not practical - unless you can find such drivers, of course. It is very hard to preserve good transient response for a high order network, but it may still be done perhaps. In my case, I believe the higher order I am using now is at least 90%-95% as good as the 1OSXO as far as transient response is concerned - which is normally difficult to achieve - that is why I have given my inputs on the HDS here. I might have been lucky. The sound is superb and I simply don't know how to improve it so I have no intention to alter it. The speaker throws out a very wide, deep sound stage with good dynamics while preserving sweetness without a hint of harshness - what more could I expect? I have only one remaining small problem to deal with - the bass with the HDS 8". My speaker has a total of 8 of them. After much work, the bass is now OK but is as good as the mid/high, which my friends said as good as any commercial speakers at any prices. Part of the reasons must be the fantastic T330D I am using.
If anybody can give me hints on what to use to design active circuits / XO based on the driver's response, phase and impedance like what I have been doing with SW with passive design, I would be very greatful. The sound of the speaker (especially the mid and high) now should beat or compete with the top of the range B&W, JM Lab, Avalone, etc. If the bass can be improved then it will definitely be a reference grade speaker.
Paulspencer,
I did the speaker design properly based on Speaker Workshop, my ears and numerous hours of experiments. However, I have not used the measurement capability of SW therefore I wouldn't be able to know what sort of notch filter required. I always believe there must be some real risk of blowing up a tweeter using 1st order serial XO (1OSXO). Looking at a 1OSXO circuit you know this can happen because an inductor can not have 0 DCR. When I tested it, I used only medium sound level and was too worried to turn the volume up, although the Esotar T330D can stand 1000+W surge current.
Yes I am convinced that it is easy to have good transient response from 1st order network, but numerous other problems with this topology make it not practical - unless you can find such drivers, of course. It is very hard to preserve good transient response for a high order network, but it may still be done perhaps. In my case, I believe the higher order I am using now is at least 90%-95% as good as the 1OSXO as far as transient response is concerned - which is normally difficult to achieve - that is why I have given my inputs on the HDS here. I might have been lucky. The sound is superb and I simply don't know how to improve it so I have no intention to alter it. The speaker throws out a very wide, deep sound stage with good dynamics while preserving sweetness without a hint of harshness - what more could I expect? I have only one remaining small problem to deal with - the bass with the HDS 8". My speaker has a total of 8 of them. After much work, the bass is now OK but is as good as the mid/high, which my friends said as good as any commercial speakers at any prices. Part of the reasons must be the fantastic T330D I am using.
If anybody can give me hints on what to use to design active circuits / XO based on the driver's response, phase and impedance like what I have been doing with SW with passive design, I would be very greatful. The sound of the speaker (especially the mid and high) now should beat or compete with the top of the range B&W, JM Lab, Avalone, etc. If the bass can be improved then it will definitely be a reference grade speaker.
Sounds like a nice project that has turned out well.
Re: Peerless 850488
Interesting stuff and you certainly have made the effort to get the best out of the 850488. Maybe a higher Zeta might have help to tone down that higher frequency zing. For what it's worth I crossed over at 3450Hz with a Zeta of 0.9 which makes the speaker a tad forward but not in your face. A thing about series xo's is that it hard to correct for any nasties in the driver by compensation for this and that. Drivers must be chosen wisely.
I have found by using inverted drivers with the woofer on top of the tweeter gives me better results on 1st order xo's most of the time (polar tilt and ZDP) and I think doesn't have the phenomenom of sound varying as you move your head.
The 850528 is great in centre and I use it in a TM centre with vertical alignment with a Vifa D26NC tweeter which helps to give a low profile of 212mm.
Cheers
HiFiNutNut said:
Interesting stuff and you certainly have made the effort to get the best out of the 850488. Maybe a higher Zeta might have help to tone down that higher frequency zing. For what it's worth I crossed over at 3450Hz with a Zeta of 0.9 which makes the speaker a tad forward but not in your face. A thing about series xo's is that it hard to correct for any nasties in the driver by compensation for this and that. Drivers must be chosen wisely.
I have found by using inverted drivers with the woofer on top of the tweeter gives me better results on 1st order xo's most of the time (polar tilt and ZDP) and I think doesn't have the phenomenom of sound varying as you move your head.
The 850528 is great in centre and I use it in a TM centre with vertical alignment with a Vifa D26NC tweeter which helps to give a low profile of 212mm.
Cheers
bzdang,
close measurements will be accurate only for the low range of any driver. As you go up in range you will get very diffrent results depending on where at the cone you measure. Those differences thend to even out at distance.
Hifinut,
I can easily imagine your speaker sounds wonderful. I have not used the HDS 5" but the Esotec 5" with D260 and also T330D. The series x-over you refer to will not have the transient respons of a true first order filter though. I assume you talk about a filter with basically a coil and cap wher tweeter and mid is in series? This will give something in between 6 and 12dB slopes, close to 9dB. This will be added to the inherent rolloff of the drivers for a summed filter response and acoustic slope of 3rd to 4th order or something like that. IOW you should not need to worry about blowing up your Esotars. The Sonus Faber Extrema use a series resistor of 7.5ohm or so in series with T330D and a 0.2mH inductor to ground after the resistor, which give less protection than a typicall first order series filter.
/Peter
close measurements will be accurate only for the low range of any driver. As you go up in range you will get very diffrent results depending on where at the cone you measure. Those differences thend to even out at distance.
Hifinut,
I can easily imagine your speaker sounds wonderful. I have not used the HDS 5" but the Esotec 5" with D260 and also T330D. The series x-over you refer to will not have the transient respons of a true first order filter though. I assume you talk about a filter with basically a coil and cap wher tweeter and mid is in series? This will give something in between 6 and 12dB slopes, close to 9dB. This will be added to the inherent rolloff of the drivers for a summed filter response and acoustic slope of 3rd to 4th order or something like that. IOW you should not need to worry about blowing up your Esotars. The Sonus Faber Extrema use a series resistor of 7.5ohm or so in series with T330D and a 0.2mH inductor to ground after the resistor, which give less protection than a typicall first order series filter.
/Peter
For what it's worth I just have to add this - I'm not as advanced as most of you guys so with me it's a bit more hit and miss.
I have spent 2 years upgrading the main 2 pair of speakers and found a point that I am more than happy with, there is a deep beautiful symmetry between them that I have finally achieved.
First pair- tweet scanspeak d2905-97, mid scanspeak 13m8640, base peerless 12 inch 831857. X overs 1200hz & 5khz, 12db/oct.
amp - arcam A85.
Second pair - tweet peerless 811827, mid peerless 850488, base 20 year old refurbished 10 inch drivers from a pair of boston acoustics A100's. X overs 700hz & 3500hz. Power amp - Nad C272.
I found the same problem with the peerless mid 850488 that any higher x.o than 3500hz and some unwanted noise starts to appear.
The peerless and scan speak drivers just sound so good togeather.
I have spent 2 years upgrading the main 2 pair of speakers and found a point that I am more than happy with, there is a deep beautiful symmetry between them that I have finally achieved.
First pair- tweet scanspeak d2905-97, mid scanspeak 13m8640, base peerless 12 inch 831857. X overs 1200hz & 5khz, 12db/oct.
amp - arcam A85.
Second pair - tweet peerless 811827, mid peerless 850488, base 20 year old refurbished 10 inch drivers from a pair of boston acoustics A100's. X overs 700hz & 3500hz. Power amp - Nad C272.
I found the same problem with the peerless mid 850488 that any higher x.o than 3500hz and some unwanted noise starts to appear.
The peerless and scan speak drivers just sound so good togeather.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.