Troels Gravesen project with ScSp 18W8545
Amish revisted
I notice now its the old 6" ... I once had a speaker with those
remember I got real silly 'custom xo' made by experts ... which was total crap ... sounded better with completely random 6db filter
I suppose it has potential, and maybe little need for a very complicated crossover
Amish revisted
I notice now its the old 6" ... I once had a speaker with those
remember I got real silly 'custom xo' made by experts ... which was total crap ... sounded better with completely random 6db filter
I suppose it has potential, and maybe little need for a very complicated crossover
Sorry - I didn't mean raw drivers out of the box - I meant raw drivers *in* the box assuming your box is similar in dimension and driver placement to another proven designers speaker with raw measurements available.
If you can only measure above 500Hz (as indicated in your gating marker on one of your FR plots above) then baffle size doesnt' matter when comparing your box with another designers. Driver placement and baffle dimensions will affect the ripple (baffle signature) on the response - but again, most 2 way "box" designs follow a very similar baffle size / driver placement on baffle so baffle diffraction will be similar.
All I'm saying (as others have said) is:
1. set mic position and levels (do not change). Set these low enough so you don't blow the tweeter (I make sure the voltage output is <= 0.5v on the amplifier output terminals). you won't need a protective cap on the tweeter (or put a 30uF or larger one on for peace of mind if you insist)
2. measure tweeter in box (gated - but do not move the gate when measuring the woofer)
3. measure woofer in box (gated)
The trick with gating is to set it based on the driver that is furthest away (to eliminate first room reflections). Usually for a 2 way - you will measure on axis to the tweeter (listening height), so set your gate (impulse) based on the woofer, since it is furthest away from the mic.
Are you absolutely sure you aren't changing anything between measuring one driver and another in the box? because your levels are different between the woofer and tweeter measurements in your graphs.
your woofer should measure about 3 - 6dB lower average than the tweeter on average.
So for example - if your reference level (doesn't matter what this is) - shows the tweeter measurement at -15dB, your woofer should be measuring about -21db to -18db on average. your tweeter is more sensitive so you would expect a higher average FR curve
compare the above graphs to the manufacturers, others designs etc....
Once your graphs are within say 10% - you know you are in the ballpark. Any variation will be measurement setup (mic not recording flat unless you have had it calibrated), driver variations, baffle variations etc...
No worries. I'm a novice, but I've found problems compound. If the basics arent right (raw in box measurements), then you will carry over any mistakes into crossover design. You basically send yourself down the wrong path.
Hence why - although frustrating, take the time / patience to setup each step and get it working correctly (ie. measurements) which will save time and frustration in the long run.
yeah, it is tricky
and you will always tend to believe your measurements more than your ears
I know this from other technology area
it is indeed very difficult to not trust the measurements
yeah, it is tricky
and you will always tend to believe your measurements more than your ears
I know this from other technology area
it is indeed very difficult to not trust the measurements
yeah, it is tricky
and you will always tend to believe your measurements more than your ears
I know this from other technology area
it is indeed very difficult to not trust the measurements
this is off topic, but yes, I realise a "flat on axis FR curve" tells you limited information:
a) what is the power response like - pertinent to real world listening positions
b) what is the in room response like - pertinent to how reflective / absorbent / big / small a room is
c) what frequence response sounds natural to you - pertinent to frequencies you are more or less sensitive to, affected in turn by what parts of the spectrum your hearing has deteriorated in
d) what non-linear distortion is present - what even and particularly odd order harmonics have not been dealt to
measurements do tell almost everything, but not everything.
get it reasonably flat and with good phase tracking then you are on the right track, wether you like to have a certain character to the sound its different matter.
my system responds to reasonably flat response, and after 5 significantly different speakers that i built last year, i concluded that flat, even power response and good phase tracking is 3 essential factors to make a good speakers.
get it reasonably flat and with good phase tracking then you are on the right track, wether you like to have a certain character to the sound its different matter.
my system responds to reasonably flat response, and after 5 significantly different speakers that i built last year, i concluded that flat, even power response and good phase tracking is 3 essential factors to make a good speakers.
Designing a passive crossover is actually a total pain in the butt, if you want it to be at all accurate.
First you've got to decide what frequency range each driver can do well. Looking at published frequency response graphs is better than nothing. Then you have to decide if one of the drivers needs to be attenuated to match up well with the other driver(s). Again published specs are better than nothing, if you don't have the equipment to measure these things yourself. It's usually the tweeter that will need a resistor of several ohms in series with it. With 8 ohm drivers, a 1 ohm resistance in the coil can throw off your calculations by more than 10%, so measure that and include it in the math.
But here's the frustrating part; they call it an 8 ohm driver but it may well not be. If you look at published impedance curves for most drivers, you find that the impedance varies over frequency. An 8 ohm five inch driver I was using was actually 15 ohms at the frequency I wanted to cross it over at (around 3kHZ). This is why "off the shelf" store bought crossovers are not likely to be anywhere near accurate. This should be measured if you know how, or at least refer to the published spec.
The Seas Millenium tweeters (which I have in my main system) are some of the few 1 inch dome tweeters that can perform excellently down to almost 1kHZ, but Linkwitz and I agree that 1.4kHZ is a good place to limit them with a 4th order highpass filter. If the filter is 2nd order, I'd say 2.5kHZ is about as low as I'd take them. If the crossover filter is 1st order (1 pole) I'd stay at 3kHZ or higher, that being a tradeoff with off axis response of the total system. It's not good to have an abrupt change in dispersion at a crossover frequency. If the driver below it in frequency is five inches, I wouldn't choose a crossover frequency above 3kHZ. The "room response" will suffer. You'd want to attenuate the tweeter a few dB more than the on-axis calibrated mic suggested, to reduce the "room response" harshness this would likely cause.
Linkwitz and I also agree that a slight dip at 3kHZ is a good thing, so I'd consider using a 1 pole crossover where you adjust the -3dB points such that the woofer rolls off at 2.5kHZ, and the -3dB point for the tweeter is at maybe 3.5kHZ. You would want to verify which phase on the tweeter works best using pink noise, a calibrated mic, and any variation of a spectrum analyzer. The tweeter would probably need to be wired with reverse phase, relative to the woofer with this arrangement.
Personally, I would rather use an active crossover ahead of the poweramps, not only because it's much more predictable and accurate, with much steeper slopes (4th order), but once you've got the chassis and power supply in place for that, it's easy to add active EQ to make a closed box woofer be acoustically flat to 30HZ, which I do and love.
You would need to have a SPICE type circuit modelling and analysis program in your computer to verify operation of the active filter circuits, and knowledge of how to scale existing circuits to get the frequencies you want. I've been doing this for decades with great success.
I hope this helps.
First you've got to decide what frequency range each driver can do well. Looking at published frequency response graphs is better than nothing. Then you have to decide if one of the drivers needs to be attenuated to match up well with the other driver(s). Again published specs are better than nothing, if you don't have the equipment to measure these things yourself. It's usually the tweeter that will need a resistor of several ohms in series with it. With 8 ohm drivers, a 1 ohm resistance in the coil can throw off your calculations by more than 10%, so measure that and include it in the math.
But here's the frustrating part; they call it an 8 ohm driver but it may well not be. If you look at published impedance curves for most drivers, you find that the impedance varies over frequency. An 8 ohm five inch driver I was using was actually 15 ohms at the frequency I wanted to cross it over at (around 3kHZ). This is why "off the shelf" store bought crossovers are not likely to be anywhere near accurate. This should be measured if you know how, or at least refer to the published spec.
The Seas Millenium tweeters (which I have in my main system) are some of the few 1 inch dome tweeters that can perform excellently down to almost 1kHZ, but Linkwitz and I agree that 1.4kHZ is a good place to limit them with a 4th order highpass filter. If the filter is 2nd order, I'd say 2.5kHZ is about as low as I'd take them. If the crossover filter is 1st order (1 pole) I'd stay at 3kHZ or higher, that being a tradeoff with off axis response of the total system. It's not good to have an abrupt change in dispersion at a crossover frequency. If the driver below it in frequency is five inches, I wouldn't choose a crossover frequency above 3kHZ. The "room response" will suffer. You'd want to attenuate the tweeter a few dB more than the on-axis calibrated mic suggested, to reduce the "room response" harshness this would likely cause.
Linkwitz and I also agree that a slight dip at 3kHZ is a good thing, so I'd consider using a 1 pole crossover where you adjust the -3dB points such that the woofer rolls off at 2.5kHZ, and the -3dB point for the tweeter is at maybe 3.5kHZ. You would want to verify which phase on the tweeter works best using pink noise, a calibrated mic, and any variation of a spectrum analyzer. The tweeter would probably need to be wired with reverse phase, relative to the woofer with this arrangement.
Personally, I would rather use an active crossover ahead of the poweramps, not only because it's much more predictable and accurate, with much steeper slopes (4th order), but once you've got the chassis and power supply in place for that, it's easy to add active EQ to make a closed box woofer be acoustically flat to 30HZ, which I do and love.
You would need to have a SPICE type circuit modelling and analysis program in your computer to verify operation of the active filter circuits, and knowledge of how to scale existing circuits to get the frequencies you want. I've been doing this for decades with great success.
I hope this helps.
Last edited:
This is a wise piece of advice. AFAIK the tweeter has a very flat frequency response up to 20 kHz, but the measurement shows a rise of 4 dB after 5.5 kHz and a peak of 9 dB at 13.5 kHz. Taking this as reality will be very misleading...
Can be the mic ECM8000?
Yes, it's the frequency response of the mic. You need to install a correction file.
Attached are measurements of the Behringer ECM8000 determined by a german site: 200 Mikrofonkalibrierungen - eine Übersicht
Attached are measurements of the Behringer ECM8000 determined by a german site: 200 Mikrofonkalibrierungen - eine Übersicht
Attachments
I bought my calibrated microphone at hifi-selbstbau.de
Other addresses for calibrated mics:
lautsprecherkauf.com
ibf-acoustic.com
Other addresses for calibrated mics:
lautsprecherkauf.com
ibf-acoustic.com
Wich mic did you bought ECM-40? I see in the same shop have calibration service link:
http://lasip.hifi-selbstbau.de/index.php?page=shop.product_details&flypage=flypage.tpl&product_id=19&category_id=9&option=com_virtuemart&Itemid=64
I don't speak Deustch, do you know if shipping is included?
http://lasip.hifi-selbstbau.de/index.php?page=shop.product_details&flypage=flypage.tpl&product_id=19&category_id=9&option=com_virtuemart&Itemid=64
I don't speak Deustch, do you know if shipping is included?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.