Yes and the gating represents measuring time.
This measuring time in creases when you want to measure low frequency's.
2msec is great to measure high frequencies in home with out being influenced by room acoustic. when the speaker is more then 68cm free from floor and wall.
2msec = 68cm 500hz
5msec = 172cm 200hz
20msec= 688cm 50hz
When a measurement is done and your looking at the impulse response for gating then you can see if there are reflections involved, then it can be that the gating is placed for that moment to measure as low as possible.
see attachment. you see a perfect impulse measurement. It is the impulse of the sound-card input connected to output.
So after the impulse there is no noise measured a flat line.
see attachment. you see a perfect impulse measurement. It is the impulse of the sound-card input connected to output.
So after the impulse there is no noise measured a flat line.
Attachments
Mike sent me a PM this morning asking if I could measure my Onkens response. Unfortunately I'm just a bit too busy at the moment to do the measurements, but might be able to do so within a couple of weeks. The recent computer rebuild necessitated by a catastrophic motherboard failure means I don't yet have the capability to measure anything.. 
A couple of thoughts however. GM does have a point about comb filtering - this is a well known issue of the design, and previous measurements I did also provided evidence of surprising levels of port radiation at unexpected frequencies, however they were typically 20dB or more below the direct radiation. (Some additional damping might help this, particularly in the upper octave of the pass band.)
One of the justifications of this design is that the port velocities are so low that port noise even at extremely high spls is non existent. This is definitely the case in practice, and high port velocities IMHO give rise to high levels of distortion in my very limited experience.
The efficiency achieved with this design is what appealed to me and a horn with the extension I wanted would have been difficult for me to design and build with the knowledge I possessed at the time.
I have managed to achieve something relatively close to critical damping. Amplifier source impedance and X-O dcr has to be taken into account.
Those response curves bear relatively little resemblance at all to the response I measured in room with mine (single box driven) where all of the the biggest deviations were below 100Hz, with two moderately high Q resonances below 38Hz. To the very best of my recollection above 100Hz to their crossover at 800Hz I measured significantly better than +/-2dB flatness (1/3rd octave smoothing) so I am wondering what else might be going on to account for those peaks and dips in the response. Without knowing all of the details of the particular Onken box tested it is hard to know what sort of mistakes might have been made in their design and construction. And as a beginner I might have gotten very lucky in their design and construction, they're not as simple as it appears to get right. (My drivers are placed asymmetrically on the baffles and not at all near the floor, driver centers somewhat below the center line of the baffle board.) I spent a lot of time thinking and asking questions before building these, and expected if the performance was not satisfactory to built a more conventional bass reflex instead.
My vent area was >85% of SD and most Onken design calcs recommend this number as a minimum.
I did a couple of thd% measurements on the system early on, and nothing stood out as being particularly odd. The worst I saw IIRC was a couple of % at fairly high spls, well above my normal listening levels, but unfortunately I did not record the levels used or distortion measured - it just wasn't an issue, and given the amp it is not unlikely that some of it came from there. It certainly sounds very clean, I am not aware of an particular issue with distortion below 800Hz where I cross to a JBL 2440. I wouldn't however dismiss Greg's comments out of hand, it may well be that my (our's as a species) hearing is not that sensitive to the nature of the distortion generated by this design, but I don't know for sure.
Greg's anti-Onken stance is somewhat surprising to me since he was absolutely instrumental in helping me to figure out how to build mine. I am not sure I could have gotten what I now consider to be an extremely successful result without his assistance..
Regardless of Greg's comments I am delighted with mine, and am not aware of any massive deficits in their performance, and have heard a lot of horns and direct radiators that to my ear were at best no better and in a lot of cases far worse than what I listen to at home.
My SE amplifiers generate significant levels of distortion at power levels in excess of a couple of watts, the Zu/Denon phono cartridge on my TD-124 MKII generates significant distortion as well. Wish it weren't so, but in the end my ears are pretty satisfied. There is of course always room for improvement, but what I've got sounds pretty good. In my engineering dotage I measure, try to understand, fix clearly blatant issues, and then forget it if I am happy enough with the sound. I made a huge mistake in the original X-O design, but got it right the second time around, and haven't felt compelled to do anything other than mess with increasingly expensive compression drivers for the mids and top end.
Until I can actually do a few measurements please take this all as anecdotal, the measurements are too far in the past and no longer available for publication. (You can refer to the old Onken thread though for some idea of what I did.)
A couple of thoughts however. GM does have a point about comb filtering - this is a well known issue of the design, and previous measurements I did also provided evidence of surprising levels of port radiation at unexpected frequencies, however they were typically 20dB or more below the direct radiation. (Some additional damping might help this, particularly in the upper octave of the pass band.)
One of the justifications of this design is that the port velocities are so low that port noise even at extremely high spls is non existent. This is definitely the case in practice, and high port velocities IMHO give rise to high levels of distortion in my very limited experience.
The efficiency achieved with this design is what appealed to me and a horn with the extension I wanted would have been difficult for me to design and build with the knowledge I possessed at the time.
I have managed to achieve something relatively close to critical damping. Amplifier source impedance and X-O dcr has to be taken into account.
Those response curves bear relatively little resemblance at all to the response I measured in room with mine (single box driven) where all of the the biggest deviations were below 100Hz, with two moderately high Q resonances below 38Hz. To the very best of my recollection above 100Hz to their crossover at 800Hz I measured significantly better than +/-2dB flatness (1/3rd octave smoothing) so I am wondering what else might be going on to account for those peaks and dips in the response. Without knowing all of the details of the particular Onken box tested it is hard to know what sort of mistakes might have been made in their design and construction. And as a beginner I might have gotten very lucky in their design and construction, they're not as simple as it appears to get right. (My drivers are placed asymmetrically on the baffles and not at all near the floor, driver centers somewhat below the center line of the baffle board.) I spent a lot of time thinking and asking questions before building these, and expected if the performance was not satisfactory to built a more conventional bass reflex instead.
My vent area was >85% of SD and most Onken design calcs recommend this number as a minimum.
I did a couple of thd% measurements on the system early on, and nothing stood out as being particularly odd. The worst I saw IIRC was a couple of % at fairly high spls, well above my normal listening levels, but unfortunately I did not record the levels used or distortion measured - it just wasn't an issue, and given the amp it is not unlikely that some of it came from there. It certainly sounds very clean, I am not aware of an particular issue with distortion below 800Hz where I cross to a JBL 2440. I wouldn't however dismiss Greg's comments out of hand, it may well be that my (our's as a species) hearing is not that sensitive to the nature of the distortion generated by this design, but I don't know for sure.
Greg's anti-Onken stance is somewhat surprising to me since he was absolutely instrumental in helping me to figure out how to build mine. I am not sure I could have gotten what I now consider to be an extremely successful result without his assistance..
Regardless of Greg's comments I am delighted with mine, and am not aware of any massive deficits in their performance, and have heard a lot of horns and direct radiators that to my ear were at best no better and in a lot of cases far worse than what I listen to at home.
My SE amplifiers generate significant levels of distortion at power levels in excess of a couple of watts, the Zu/Denon phono cartridge on my TD-124 MKII generates significant distortion as well. Wish it weren't so, but in the end my ears are pretty satisfied. There is of course always room for improvement, but what I've got sounds pretty good. In my engineering dotage I measure, try to understand, fix clearly blatant issues, and then forget it if I am happy enough with the sound. I made a huge mistake in the original X-O design, but got it right the second time around, and haven't felt compelled to do anything other than mess with increasingly expensive compression drivers for the mids and top end.
Until I can actually do a few measurements please take this all as anecdotal, the measurements are too far in the past and no longer available for publication. (You can refer to the old Onken thread though for some idea of what I did.)
Last edited:
Onken 285GMF
Might be a good time to post again with a pic of the first finished Onken for Supravox 285GMF . Looks great now fully trimmed and varnished , but currently doesn't sound great ..(
I've done a first round of bracing some panel resonances ( eg. 300Hz on the upper back panel ) with some internal structure , but I think I need advice on the damping/stuffing .
As it is now, I have full cover of the rear panel , inside, with 1/2" wool felt ; and all of the (8) vents stuffed lightly with long-haired wool at the back .
I'm thinking I ought to remove stuffing from 2 or 4 of the vents to see how it changes, and will probably do this in next day or two . Sound is lacking tone and dynamics in the mid-bass .
Plan B if it doesn't sound much better will be to call a halt on the assembly of the 2nd speaker and consider the options ..
I might want to ask GM what would be the approx volume for the MLTL design he suggested as a modern alternative to the Onken, some way back in this thread.
This box is about 7 cu ft . The design is - http://www.supravox.fr/kits/jensen.pdf
Mark
Might be a good time to post again with a pic of the first finished Onken for Supravox 285GMF . Looks great now fully trimmed and varnished , but currently doesn't sound great ..
(I've done a first round of bracing some panel resonances ( eg. 300Hz on the upper back panel ) with some internal structure , but I think I need advice on the damping/stuffing .
As it is now, I have full cover of the rear panel , inside, with 1/2" wool felt ; and all of the (8) vents stuffed lightly with long-haired wool at the back .
I'm thinking I ought to remove stuffing from 2 or 4 of the vents to see how it changes, and will probably do this in next day or two . Sound is lacking tone and dynamics in the mid-bass .
Plan B if it doesn't sound much better will be to call a halt on the assembly of the 2nd speaker and consider the options ..
I might want to ask GM what would be the approx volume for the MLTL design he suggested as a modern alternative to the Onken, some way back in this thread.
This box is about 7 cu ft . The design is - http://www.supravox.fr/kits/jensen.pdf
Mark
Attachments
Modeling your box as a simple BR, there's a gradual roll-off starting at ~400Hz, with F3 at ~85Hz. This could explain the mid-bass issues you're experiencing. For a flatter response, the vent length needs to be shortened up quite a bit. This is confirmed by the Onken Excel calculator.
That's probably a good idea. You'll definitely want opinions from more knowledgeable folks here. Nice looking box BTW.
jeff
The room has a lot to do with how a LF speaker performs and how hard the driver is pushed around/at Fb determines how much driver modulation/comb filtering occurs from the vents, so when you did the measurements did you feed it a tone burst near/at Fb? Also, I don't remember how long your vents are, but the shorter they are, the higher in frequency their pipe harmonics and since sound intensity falls at 1/f, they wouldn't be nearly as intense if much shorter than the 416A's and some others that show high levels of vent harmonic distortion.
Yes, vents are open pipe resonators with both even and odd harmonics, so damping them same as TL will tone them down and of course will roll off Fb a bit too. Not many 'free lunches' in audio design.
Right, small vents get all choked up.
Well, I do seem more sensitive to the various types of 'ringing' than most folks I've been around, but based on your comments I'm thinking your system's vents don't get enough 'excitement' to generate much distortion. If true, this basically proves my 'broken record' assertion that a reflex should be tuned below the lowest notes the system is likely to 'feel' at any useful SPL.
It's not so much than I'm anti-Onken as I am pro accurate reproduction (high SQ) which to me doesn't include the Onken alignments unless the vents are damped, especially when a large, single vent MLTL has historically sounded better overall to those relatively few of us that have owned both.
You're welcome, I think.
As far as helping folks with them or any type of alignment for that matter, it's not for me to tell them what they should want or try to convince them that my way is 'best' as some folks on the various forums seem prone to do, but I will tell what I know or think I know, so that folks can hopefully make a more informed choice or at least get as much performance as practical out of whatever they want. GM
Here's a MathCad sim I did of it ~ 4 yrs ago with minimal damping and none in the vent. The only difference of course being I had to sim one large vent instead of several, so only has the one major phase flop instead of multiples. Notice the vent gain is enough to ~flatten it out down to Fb in half space with one ohm of series resistance using published specs, so sounds like some may need to be added once the stuffing is removed from the vents if there's not enough room gain.
GM
Attachments
Last edited:
Thanks GM, that's how I remember it , I have that plot stored.
I will remove some or all of the vent stuffing this evening and see how I go.
Currently this is running from a SE 300B amp with a source Z of somewhere between 1 and 2 ohms, but soon I will be able to run it from a Pass Labs F4 which will be more like 0.1 ohm . Once I get some suitable chokes (on order) I will apply a 750Hz low-pass , which is the system intention, and it will be interesting to hear that, I expect it'll be easier to judge the low bass, though I guess the mid-bass may lose some tone as the midrange is removed. I must say the 285GMF is almost a full-ranger , it's a shame to roll-off the midrange !
I will remove some or all of the vent stuffing this evening and see how I go.
Currently this is running from a SE 300B amp with a source Z of somewhere between 1 and 2 ohms, but soon I will be able to run it from a Pass Labs F4 which will be more like 0.1 ohm . Once I get some suitable chokes (on order) I will apply a 750Hz low-pass , which is the system intention, and it will be interesting to hear that, I expect it'll be easier to judge the low bass, though I guess the mid-bass may lose some tone as the midrange is removed. I must say the 285GMF is almost a full-ranger , it's a shame to roll-off the midrange !
Ok , now we're getting somewhere !
The ports must have been pretty heavily damped. I took out the wool on 6 out of the 8 ports and now I have bass tone, dynamics and more low bass output . I was playing some Duke Ellington and Jaco Pastorius last night with a big grin across my face .
It got even better when I rigged-up the Horn ( GPA 288-H and Azura AH-425 ) on the right channel and pushed the two speakers together . Sounds half-baked but it gave me a strong impression of the pleasures to come.
I shall be playing a fair bit of mono this evening !
It's fair to say the 285GMF doesn't sound as clean and open as it did on OB, and never will do, but it seems like a good compromise to get the extension while keeping this nice driver with its sweet midrange .
MJ
The ports must have been pretty heavily damped. I took out the wool on 6 out of the 8 ports and now I have bass tone, dynamics and more low bass output . I was playing some Duke Ellington and Jaco Pastorius last night with a big grin across my face .
It got even better when I rigged-up the Horn ( GPA 288-H and Azura AH-425 ) on the right channel and pushed the two speakers together . Sounds half-baked but it gave me a strong impression of the pleasures to come.
I shall be playing a fair bit of mono this evening !
It's fair to say the 285GMF doesn't sound as clean and open as it did on OB, and never will do, but it seems like a good compromise to get the extension while keeping this nice driver with its sweet midrange .
MJ
I designed and built a pair of Onken boxes almost 5yrs ago now and have very little inclination to listen to anything else. Damp the back, bottom, and one inside wall of the box with some acousta-stuff or better still some lambs wool. You will need to do some testing, but you don't want to over do it, an inch to 1 1/2" of acousta-stuff seems like enough, and lambs wool may require a bit less.
I run 8 port Onkens with no stuffing in the ports, so I'd say take out the stuffing in the last two ports. Top end is handled by 2440s on 2311 horns and 2308 diffusers (tried JBL 2397 and early Edgar Salad bowls - preferred the snap and directivity of this setup best). The highs are handled by 2402s.. Lots of detail, good imaging, very articulate sound, and very good bass performance. IMHO I have not heard anything significantly better amongst the large diy crowd I run with..
You're welcome!
Gees, that's low IME, I can't recall < 2.4-2.5 with older amps much higher.
No experience, but I've been told it's enough to work well with a ~500-800 Hz/1st order XO. That's wide alright if true.
GM
FWIW I agree, my current 300B SE amplifier with 3K OPTs is right around 2.5 ohms, but uses no negative feedback.. Winding dcr of the primary and secondary contributes a significant % as well. Can't remember the exact numbers, but combined probably nearly an ohm..
I'm using 2nd order butterworth at 800Hz mainly because of the horns used on the midrange drivers, and it made sense I thought at the time to use the same slopes on both - why I thought that exactly I am not sure, but in practice it works very well.
When designing the XO you do need to take the amplifier's source impedance into account as it's not a perfect voltage source..
I guess I'm showing my age since to me SET = zero nfb and frankly it didn't occur to me it might have any. I mean what's the point of SET once you dilute it with nfb? Oh well, different strokes.........
Nothing wrong with 'textbook' filters as long as you adhere to its requirements which unfortunately is often at odds with what the drivers (and amps if high output impedance) want.
GM
Nothing wrong with 'textbook' filters as long as you adhere to its requirements which unfortunately is often at odds with what the drivers (and amps if high output impedance) want.
GM
Thanks Kevin for the extra advice, next time I stop listening to music and unscrew the 12 long screws from the top, I will remove the remaining wool and add some wool felt onto the interior of the top panel .
GM :
Amp is a custom job so output impedance a bit lower than normal . KR300B XLS ( 670R Zout approx ) into Dave Slagle 3K:5R amorphous equals about 1.1R + winding resistance ( 0.3R as best I can measure ) so call it 1.5R .
Supravox 285GMF measures almost flat up to 6.5kHz and sounds very sweet through that range except for a bit too much emphasis around 2.5k . I was lucky to get a nearly-new pair from a friend for a substantial discount .
MJ
GM :
Amp is a custom job so output impedance a bit lower than normal . KR300B XLS ( 670R Zout approx ) into Dave Slagle 3K:5R amorphous equals about 1.1R + winding resistance ( 0.3R as best I can measure ) so call it 1.5R .
Supravox 285GMF measures almost flat up to 6.5kHz and sounds very sweet through that range except for a bit too much emphasis around 2.5k . I was lucky to get a nearly-new pair from a friend for a substantial discount .
MJ
There are a significant number of Japanese 300B SE designs that use some limited global feedback, the general consensus is a few dB is OK.. Personally I don't like to use NFB in my SE amps..
Yeah, designing the XO was a none trivial exercise, but seems to work pretty well. Nothing obviously wrong in the measurements or listening..
'Onken' done right (acoustically) IME: http://community.klipsch.com/forums/storage/4/815128/butterfly.jpg
GM
GM
OK, thanks!
Oh yeah! Definitely recommend trying it run WFO with just a cap and damping resistor to 'taste' on the horn.
GM
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.