Didn’t look at the others, but Fs, Qts, Vas all look really close. The other important one, sensitivity is MIA.
dave
dave
I suspect they thought most people would find a tolerance variation of 0.6822Hz in their Fs value (rather less than you can often get with local climatic variations) to be an acceptable pair match. However, if you feel that is an unacceptably wide deviation in production tolerance, I suspect you will be able to return them to the dealer.
Ah that's good. Sensitivity is MIA? Hmm, I was assuming dBSPL was a sensitivity measurement.
I'm still in the dark about what it means for the drivers to be 'matched within a .5% tolerance'... And it would seem inconsistent with the brand for this to be a glittering generality.
I'm still in the dark about what it means for the drivers to be 'matched within a .5% tolerance'... And it would seem inconsistent with the brand for this to be a glittering generality.
Ehh, I'm not sure. I figured I'd run it by you guys, who have more knowledge than myself. Is Fs so heavily weighted in the matching that it is considered the primary parameter for comparison? If so, I suppose I'm a little disappointed, but this could be, again, my own lack of knowledge. The guy who had the mismatched pair some time ago had closer Fs values than my 'matched' pair, for example, which gives me reason for pause.
Missed that completely.
OK, but not great. Better than average tolerances.
MA matches the Fs. I was more thorough. The other 3 mentioned are quite good.
dave
OK, but not great. Better than average tolerances.
MA matches the Fs. I was more thorough. The other 3 mentioned are quite good.
dave
Okay, quickly coming back to this (and with the reminder that I'm paid a small retainer as a consultant to MA so people are free to thing 'bias' if they wish, although part of my role is to give them honest feedback, including critique, so I'm far from being a yes-man)
-The 0.5% is frankly a little silly, since the metric is Fs and it's nigh-on impossible to achieve that regularly across driver ranges -especially as the permissible variation inevitably becomes smaller as the nominal drops, so where for a driver with a 75Hz nominal Fs you'd have 0.75Hz deviation permissible, for one with a 40Hz nominal, you've got 0.4Hz permissible variation. In theory, lower should be 'easier' -if you're dealing with woofers. With a wideband, the converse tends to be true. I wouldn't say 'they've made a rod for their own back' with that one, but it's ridiculously close -meaninglessly so, since in practice such tiny variations are totally swamped by enclosure & room effects, to say nothing of simple local climatic variations -temperature, pressure, humidity etc. Note that none of these are 'excuses' -they're just statements of fact.
-To have a pair of drivers with only 0.6Hz between them is basically the definition of 'matched'. Contextually speaking, that's extremely close by pretty much any sane measure. Remove the decimal point (which is basically irrelevant) and they both have an Fs of 40Hz, and given that you could probably get more than that variation by a simple couple of degrees of change in temperature when measuring -you can't reasonably ask for much more; that's about as good as it gets and most would kill to have that sort of a tolerance / match in a pair of production units.
-Clio's SPL (which is what you've got printed there -that's a straight Clio table) is a mathematical derivation from the T/S parameters, so short of significant differences, i.e. well over 1dB I wouldn't pay it too much mind except as a quick 'yes, they're the same basic value' -the FR graph is more useful as the actual average. Note that Evan sets Clio for 1m/w values -as an aside, I'm trying to get them to change this to 2.83v so there is better consistency & meaningful comparisons across different coil impedances; Mark has made that request too.
Short version -you have a pair-match, and a very good pair match by most production-driver standards. The variation is extremely small, so much so that multiple external factors will utterly swamp it in any possible use outside a laboratory, where the differences will still be essentially inaudible. Use & enjoy is my advice.
-The 0.5% is frankly a little silly, since the metric is Fs and it's nigh-on impossible to achieve that regularly across driver ranges -especially as the permissible variation inevitably becomes smaller as the nominal drops, so where for a driver with a 75Hz nominal Fs you'd have 0.75Hz deviation permissible, for one with a 40Hz nominal, you've got 0.4Hz permissible variation. In theory, lower should be 'easier' -if you're dealing with woofers. With a wideband, the converse tends to be true. I wouldn't say 'they've made a rod for their own back' with that one, but it's ridiculously close -meaninglessly so, since in practice such tiny variations are totally swamped by enclosure & room effects, to say nothing of simple local climatic variations -temperature, pressure, humidity etc. Note that none of these are 'excuses' -they're just statements of fact.
-To have a pair of drivers with only 0.6Hz between them is basically the definition of 'matched'. Contextually speaking, that's extremely close by pretty much any sane measure. Remove the decimal point (which is basically irrelevant) and they both have an Fs of 40Hz, and given that you could probably get more than that variation by a simple couple of degrees of change in temperature when measuring -you can't reasonably ask for much more; that's about as good as it gets and most would kill to have that sort of a tolerance / match in a pair of production units.
-Clio's SPL (which is what you've got printed there -that's a straight Clio table) is a mathematical derivation from the T/S parameters, so short of significant differences, i.e. well over 1dB I wouldn't pay it too much mind except as a quick 'yes, they're the same basic value' -the FR graph is more useful as the actual average. Note that Evan sets Clio for 1m/w values -as an aside, I'm trying to get them to change this to 2.83v so there is better consistency & meaningful comparisons across different coil impedances; Mark has made that request too.
Short version -you have a pair-match, and a very good pair match by most production-driver standards. The variation is extremely small, so much so that multiple external factors will utterly swamp it in any possible use outside a laboratory, where the differences will still be essentially inaudible. Use & enjoy is my advice.
Thanks for all the input here. Scottmoose -- that's exactly the perspective I was seeking to gain... What a great, thorough post. I learned a few new things from it -- Much appreciated.
This thought had crossed my mind (that the tolerance would have to be tighter and tighter as the numbers dropped if it were to maintain a certain percentage)... So it doesn't seem that percentages in general are the ideal form of measurement here. Perhaps rather just a raw Hz range would be more suitable (e.g., +/- 1Hz).
Ehh, maybe it doesn't really matter but for the more granularly focused among us, i.e., picky (me). Without having greater perspective initially, the numbers didn't seem to align favorably... In my case, the percentage difference in FS measurements between two drivers is about 1.6%, which is enough to trigger a, "wait, what now?" reaction when expecting less than .5%. Then there is always the possibility that my math is fuzzy, or the percentages are calculated using some weighting that I'm not aware of... Shrug.
On another note, I wanted to mention something I noticed about the frequency response charts. They are exactly the same between each of my drivers, and upon checking online, also seem to be exactly the same to other MAOP 10 charts as well. When the charts are overlayed, there appears to not even be the slightest variation in FR, which would be nearly physical impossible to achieve. I'm inclined to think that the same measurement is being copy-pasted onto every data sheet, rather than representing an independent test with results custom to that specific driver. Please correct me if I'm wrong here (?).
Maybe this is common knowledge, and my assumptions of independent tests are off... Yet, I've read other posts on here where people are under the same assumptions.
I believe I will heed your advice to "use & enjoy", and will do so with considerable relief (dodging the whole process of return shipping overseas).
This thought had crossed my mind (that the tolerance would have to be tighter and tighter as the numbers dropped if it were to maintain a certain percentage)... So it doesn't seem that percentages in general are the ideal form of measurement here. Perhaps rather just a raw Hz range would be more suitable (e.g., +/- 1Hz).
Ehh, maybe it doesn't really matter but for the more granularly focused among us, i.e., picky (me). Without having greater perspective initially, the numbers didn't seem to align favorably... In my case, the percentage difference in FS measurements between two drivers is about 1.6%, which is enough to trigger a, "wait, what now?" reaction when expecting less than .5%. Then there is always the possibility that my math is fuzzy, or the percentages are calculated using some weighting that I'm not aware of... Shrug.
On another note, I wanted to mention something I noticed about the frequency response charts. They are exactly the same between each of my drivers, and upon checking online, also seem to be exactly the same to other MAOP 10 charts as well. When the charts are overlayed, there appears to not even be the slightest variation in FR, which would be nearly physical impossible to achieve. I'm inclined to think that the same measurement is being copy-pasted onto every data sheet, rather than representing an independent test with results custom to that specific driver. Please correct me if I'm wrong here (?).
Maybe this is common knowledge, and my assumptions of independent tests are off... Yet, I've read other posts on here where people are under the same assumptions.
I believe I will heed your advice to "use & enjoy", and will do so with considerable relief (dodging the whole process of return shipping overseas).
Last edited:
I took advantage of what seems to be some good pricing on KJF audio for 2 pairs of MAOP 10.2. I could use one pair or both pairs in the design. I mentioned in earlier post about wanting to make an on wall due to toddler and storing my current main speakers for a while.
Poplar design would work for a one driver. I messaged Scott at Woden design who gave some great advice about some potential issues. It seems if I want to use both drivers I should go for a custom design which is fine no problem with that as I would need plans for machining. But trying to get a sense of what is feasible in shape and dimensions before deciding on 1 or 2 drivers per side.
What about 1 sealed driver 1 ported or 2 ported separated with an internal driver or 2 side by side like lowther 124
Or is 2 drivers per side just not a good idea? I always like mid-tweeter-mid arrangement so maybe I am biased to 2 or 3 drivers over 1
Poplar design would work for a one driver. I messaged Scott at Woden design who gave some great advice about some potential issues. It seems if I want to use both drivers I should go for a custom design which is fine no problem with that as I would need plans for machining. But trying to get a sense of what is feasible in shape and dimensions before deciding on 1 or 2 drivers per side.
What about 1 sealed driver 1 ported or 2 ported separated with an internal driver or 2 side by side like lowther 124
Or is 2 drivers per side just not a good idea? I always like mid-tweeter-mid arrangement so maybe I am biased to 2 or 3 drivers over 1
Attachments
As far as I know, they're individual sweeps. You can see the kind of tolerance deviation that exists between the drivers in the tabulated data, and while functionally speaking T/S values only apply at the bottom of the range, it gives you an idea how much variation you can expect to see in an FR plot, unless you change the graph scale to something meaninglessly coarse. In other words, with tolerances this close, you're unlikely to see much at that scale.
I think a little perspective would be valuable here, so for the sake of interest: most 'quality' drivers these days are manufactured with a nominal production QA / QC of 10% to Fs. It's quite normal to see variations of 10Hz or more in many midbass or wideband units from a nominal advertised figure depending on the measurement technique -and you can sometimes see significantly more between the drivers themselves if that 10% is taken as a +/- value rather than one way. Not many QA match to any spec. beyond that either. For instance, I've had multiple pairs of midbass and HF units from a certain well-known and highly respected manufacturer which in several instances only just scraped the Fs criteria with my most generous application of that tolerance, while in a number of instances (not all, to be completely fair) they exhibited even greater deviation elsewhere: in one case with a 35% deviation on Qt to advertised and about 26% between each other. I still use drivers from that company (I won't name them) albeit warily for obvious reasons. They produce a number of units in both 8ohm and 4ohm versions; interestingly it was always the former which showed the widest deviation.
Here, the difference between the drivers is 0.6622Hz, which is so small a value that it could be caused if it started raining outside resulting in a local pressure differential which came into play when the individual units were measured. That's the kind of scale we're dealing with here. And that 0.6622Hz does not even take the units over the whole number threshold, and exists within the same single value: in other words, they both have a 40Hz Fs with just a few tenths between them. As noted though, if you feel that is unacceptably large, then you absolutely should return them. It's functionally meaningless, the variation being essentially in the 'non-existant' category for any usage requirements, and you will spend a very long time indeed trying to find a pair of units of this type with that kind of tolerance, but it is your consumer right TTBOMK in most regions. I suspect MA would encourage you to do so also since you are uneasy with the quality of the product you have received.
I think a little perspective would be valuable here, so for the sake of interest: most 'quality' drivers these days are manufactured with a nominal production QA / QC of 10% to Fs. It's quite normal to see variations of 10Hz or more in many midbass or wideband units from a nominal advertised figure depending on the measurement technique -and you can sometimes see significantly more between the drivers themselves if that 10% is taken as a +/- value rather than one way. Not many QA match to any spec. beyond that either. For instance, I've had multiple pairs of midbass and HF units from a certain well-known and highly respected manufacturer which in several instances only just scraped the Fs criteria with my most generous application of that tolerance, while in a number of instances (not all, to be completely fair) they exhibited even greater deviation elsewhere: in one case with a 35% deviation on Qt to advertised and about 26% between each other. I still use drivers from that company (I won't name them) albeit warily for obvious reasons. They produce a number of units in both 8ohm and 4ohm versions; interestingly it was always the former which showed the widest deviation.
Here, the difference between the drivers is 0.6622Hz, which is so small a value that it could be caused if it started raining outside resulting in a local pressure differential which came into play when the individual units were measured. That's the kind of scale we're dealing with here. And that 0.6622Hz does not even take the units over the whole number threshold, and exists within the same single value: in other words, they both have a 40Hz Fs with just a few tenths between them. As noted though, if you feel that is unacceptably large, then you absolutely should return them. It's functionally meaningless, the variation being essentially in the 'non-existant' category for any usage requirements, and you will spend a very long time indeed trying to find a pair of units of this type with that kind of tolerance, but it is your consumer right TTBOMK in most regions. I suspect MA would encourage you to do so also since you are uneasy with the quality of the product you have received.
Last edited:
Note that there is already a twin Poplar for A10x. Also the Twin Frugel-Horn XL.
dave
The MAOP 10.2 CGR box is ready. I have never made a DIY speaker, I am satisfied with my two-way speakers. Curiosity brought me to it, I never had a full-range speaker. I've listened to full-range speakers a few times elsewhere. My impressions were mixed.
It was a big surprise when my new speakers started playing. I've been listening for about 50 hours and I really like the sound throughout the range. The bass is particularly surprising.
Thanks to Dave for the plans and advice.
It was a big surprise when my new speakers started playing. I've been listening for about 50 hours and I really like the sound throughout the range. The bass is particularly surprising.
Thanks to Dave for the plans and advice.
Very pretty.
At least another 50 hours of breakin needed. They will get better :^)
dave
I would have to do this one yet. As per all the miniOnken, they full-on trapezoid is best.
dave
I'll report back after another 50 hours of listening to music. I will try it with several amplifiers.