I think that part of the problem lies in the fact that many manufacturers of drivers and other experts in the field have their own ideas on how to test, while at the same time remaining vague about the specifics of their test method. Mark has now disclosed (a good deal of) the relevant aspects of his driver measurements, but other manufacturers might do things differently again.
This certainly explains why there are so many people reporting drivers having different parameters than specified, beyond reasonable tolerance. IEC60268-5 also leaves considerable freedom in choosing the impedance measurement, although it does state that constant voltage drive is preferable over constant current drive, that a drive level should be chosen that gives consistent results, and it requires a one hour break in before tests are started. The additional efforts as described by Mark are not in the standard.
Anyway, I'm sufficiently convinced about the indicative nature of measurements done at home, even when I am technically able to use a method that complies with IEC60268-5.
What is reasonable tolerance.? Attached are measures of FE127eN that i did (the software was very consistent), and then the exact same driver by Mark using his kit, and Fostex factory specs for reference.
dave
Attachments
I have no idea in terms of percentage, but I think it would be reasonable to expect a driver to work in an enclosure which was dimensioned using factory specs. Did you have to change your plans based on these measurements?
On a different note: we're getting off-topic. The discussion is relevant to a wider audience than only those exploring the 7p and 12pw combo. Shouldn't we split the thread?
The cabinet -- the original Fonken -- worked fantasically well using Fostex numbers for the design.
dave
Hi Timpert,
A reasonable point. Which is why most driver makers spend allot of money on their production, anechoic and T/S measurement.
What is equally reasonable, is that Diyers realise their home testing won't likely be accurate (unless they've turned their home into an acoustic lab). However, I rarely see any such sensible acknowledgment from the Diyaudio side; Only from the senior experienced members like Bob, Dave etal who've been in the business (professional and amatuer roles) longer than most. You're receiving allot of their wisdom, please take note of it.
Yes I would normally split this thread but having done similar in the past, it usually brings out those members who don't trust makers, I get it kicked around in fall-out. I'm happy to facilitate posts like yours while they remain sensible and open-minded. Anything else belongs in the "Lounge" section of Diyaudio. Should you wish to explore the issue of measurement in greater depth, please start a new thread in the Lounge or similar section. Or you might wish to post here:
http://www.diyaudio.com/forums/lounge/200865-sound-quality-vs-measurements.html
What I no longer tolerate is an inquisition and/or an imposition posting style from those members who's dogma is peddled as "authoritative" when they've never designed or built drivers in their lifetimes. My health remains variable, sadly a few members have knowingly and deliberately taken advantage of my situation to turbo-charge their dogmas, venting onto the only driver maker willing to engage with them. Sadly its a standard of communication which is largely deemed acceptable on Diyaudio.
Thanks
Mark.
Last edited:
There is a Dutch saying, it goes : "de beste stuurlui staan aan de wal", which means, loosely translated, "the best captains are ashore". It means that those who pretend to know best are usually those who aren't directly involved in something.
I still feel the need to explain my "reasonable tolerance" remark. With that I mean normal product variance and measurement error. If I take a resistor and connect it to my multimeter, I can expect the value on the display to match the nominal value, within an error margin determined by resistor tolerance, multimeter error and contact resistance.
Now, if I measure the impedance characteristic of a loudspeaker with a setup that is known to give correct impedance measurements with resistors and capacitors, I would at first expect the outcome to match the manufacturer's specifications within a certain error margin, BUT, as it turns out here, the largest source of error is the fact that the measurement method chosen strongly affects the outcome. This is unlike the case of the resistor, which will always measure the same over a wide range of circumstances. As a result, you can only expect speaker measurement results to match the manufacturer's results if you follow the exact same procedures and rituals as used by the manufacturer.
This sensitivity to method remains unaccounted for with home measurements, and it is a trap which catches many DIY people. I certainly got caught.
Now, if I measure the impedance characteristic of a loudspeaker with a setup that is known to give correct impedance measurements with resistors and capacitors, I would at first expect the outcome to match the manufacturer's specifications within a certain error margin, BUT, as it turns out here, the largest source of error is the fact that the measurement method chosen strongly affects the outcome. This is unlike the case of the resistor, which will always measure the same over a wide range of circumstances. As a result, you can only expect speaker measurement results to match the manufacturer's results if you follow the exact same procedures and rituals as used by the manufacturer.
This sensitivity to method remains unaccounted for with home measurements, and it is a trap which catches many DIY people. I certainly got caught.
Timpert,
One thing too many assume is that T/S parameters are scalar (ie fixed numbers). This is not the case. They are curves that are a function of (at least), voltage, temperature, and air pressure.
dave
One thing too many assume is that T/S parameters are scalar (ie fixed numbers). This is not the case. They are curves that are a function of (at least), voltage, temperature, and air pressure.
dave
I might point out that this discussion of manufacture's/DIY T/S measurements is not unique to this forum. There is a current thread going on a Parts Express where a customer is incensed that he can't recreate Dayton Audio's numbers (LMS at 2.83V IIRC) with PE's DATS (~$100US and mV) system.
The key to this whole issue is to understand which manufactures post honest numbers and which manufactures post prototype numbers, or simply wishful thinking. Mark Audio, Fostex, Dayton, Seas are among those I an not going to measure. I trust the manufacture's numbers.
Bob
The key to this whole issue is to understand which manufactures post honest numbers and which manufactures post prototype numbers, or simply wishful thinking. Mark Audio, Fostex, Dayton, Seas are among those I an not going to measure. I trust the manufacture's numbers.
Bob
I think I found that thread. I just want to make clear that I have never been incensed or whatever, I just wanted to know why my results were off. I am very satisfied with the responses that I got.
Attachments
![IMG_0004[1].JPG](/community/data/attachments/410/410911-062a5411ec4edc06c78e91770eff0d35.jpg?hash=BipUEexO3A)
Nice job Chris I look forward to hearing them, maybe this weekend if I can make it down there to pick up the FF165's.
Bernie
Bernie
Chris, Dave,
The Dual TL MTM looks like it means business...
Besides the usual refined fare can you hit them with something heavier also? 🙂
The Dual TL MTM looks like it means business...
Besides the usual refined fare can you hit them with something heavier also? 🙂
Actually, been working in both the shop weekends and some bookkeeping for the past few evenings - haven't even turned the TV on since last weekend - and I watch a lot more than listen these days.
need a hand cart to move these puppies - yup, they'll be very interesting.
Last edited: