I took an old Radio Shack CS-5, gutted the old Linaeum tweeter, and substituted the GRS 2522. (The old tweeter was not working well and I could not fix it.) I chose the CS-5 because I had existing RS LX-5 speakers, and thought it might make a good tonal match, but time does not seem to be kind to the Linaeum tweeters. The GRS 2522 looked like it might be a reasonable substitute, so I decided to give it a try.
I tested with REW (generating white noise) in a non-ideal location (speaker propped on other stuff, and near a wall and corner), and got the below result. Except for the dip at 8k, this is better than I expected. I managed to guesstimate the match with the woofer pretty well, and it's eerily flat through the important midrange.
Woofer and tweeter use 1st order crossovers, which seems like it could be problematic, but it matches up OK, based on this. I estimate the woofer crossover to be starting at about 1500Hz, and the tweeter at 4k, so there's a lot of overlap.
I need to do a more complete write-up, but thought in the meantime, someone might find this interesting and/or tell me what my 8k dip is from. I did add a series 2.2 ohm resistor to try to match levels, so it's possible that it brought the higher treble down more than lower treble, I suppose.
(Edit: fixed attachment)
I tested with REW (generating white noise) in a non-ideal location (speaker propped on other stuff, and near a wall and corner), and got the below result. Except for the dip at 8k, this is better than I expected. I managed to guesstimate the match with the woofer pretty well, and it's eerily flat through the important midrange.
Woofer and tweeter use 1st order crossovers, which seems like it could be problematic, but it matches up OK, based on this. I estimate the woofer crossover to be starting at about 1500Hz, and the tweeter at 4k, so there's a lot of overlap.
I need to do a more complete write-up, but thought in the meantime, someone might find this interesting and/or tell me what my 8k dip is from. I did add a series 2.2 ohm resistor to try to match levels, so it's possible that it brought the higher treble down more than lower treble, I suppose.
(Edit: fixed attachment)
Last edited:
Nice cosmetics on your installation.
The dip at 8 kHz could be natural in the tweeter response, from the geometry around the tweeter (trim rings, nearness of the enclosure edge, etc.), or possibly an interaction with the woofer output.
If you want to investigate the dip further, you can try a very close measurement of the tweeter (mic about 0.25" away). That will give you a clearer picture of what the driver is doing without the influence of geometry and other drivers. You can do the same thing with the woofers. These measurements won't be exactly what you'll see at a larger distance, but they often more clearly show response trends/anomalies directly attributable to the driver output.
Something else to look at is off-axis response of the whole speaker at 1 or 2 meters (with first-order crossovers, sometimes you need 2 meters for proper integration). Sometimes dips like that fill in to the sides, so their contribution to the perceived frequency balance can be minimal.
Also, your SPL axis isn't calibrated (unless you live in an anechoic chamber), so I'm not sure how much background noise may be contaminating your measurement. The hump around 25 Hz makes me think there may be some noise in there (unless you had a subwoofer on while measuring). You can do a measurement of just the background noise to get a feel for where things are. You don't have to be at very high levels while measuring, but you want to be significantly above the ambient noise. How much above depends on how accurate you're trying to be in what frequency range. You can also average multiple measurements to improve this, but that's not as effective with constant noise sources.
The dip at 8 kHz could be natural in the tweeter response, from the geometry around the tweeter (trim rings, nearness of the enclosure edge, etc.), or possibly an interaction with the woofer output.
If you want to investigate the dip further, you can try a very close measurement of the tweeter (mic about 0.25" away). That will give you a clearer picture of what the driver is doing without the influence of geometry and other drivers. You can do the same thing with the woofers. These measurements won't be exactly what you'll see at a larger distance, but they often more clearly show response trends/anomalies directly attributable to the driver output.
Something else to look at is off-axis response of the whole speaker at 1 or 2 meters (with first-order crossovers, sometimes you need 2 meters for proper integration). Sometimes dips like that fill in to the sides, so their contribution to the perceived frequency balance can be minimal.
Also, your SPL axis isn't calibrated (unless you live in an anechoic chamber), so I'm not sure how much background noise may be contaminating your measurement. The hump around 25 Hz makes me think there may be some noise in there (unless you had a subwoofer on while measuring). You can do a measurement of just the background noise to get a feel for where things are. You don't have to be at very high levels while measuring, but you want to be significantly above the ambient noise. How much above depends on how accurate you're trying to be in what frequency range. You can also average multiple measurements to improve this, but that's not as effective with constant noise sources.
Last edited:
With that special faceplate, you can see why I wanted to try to get this tweeter to work by mounting sideways! Not shown are the modifications I had to make to the box underneath - a bit messy - and you can't see that the front plate doesn't clamp down as snug as before. There's a small gap on the top-middle. This does not affect the area under pressure by the woofers so it's just cosmetic and not noticeable unless examining the box.
I believe the above chart was made without the faceplate, but I did do a quick test with the faceplate, and the dip was still there.
While I bought a calibrated mic, I haven't figured out how to calibrate the SPL level in REW. I don't think it's actually real noise (below 60Hz or so) but some other artifact, such as a low mic level on the sound card. The rest of the frequency range seems reasonable enough. Perhaps there's a sensitivity setting that would help.
I'll have to give your other suggestions a try, but maybe I should try to get REW to work better first? Or try a laptop and see if that works better. (I should do that anyway, as I want to test my car system.)
Thanks again for your comments.
I believe the above chart was made without the faceplate, but I did do a quick test with the faceplate, and the dip was still there.
While I bought a calibrated mic, I haven't figured out how to calibrate the SPL level in REW. I don't think it's actually real noise (below 60Hz or so) but some other artifact, such as a low mic level on the sound card. The rest of the frequency range seems reasonable enough. Perhaps there's a sensitivity setting that would help.
I'll have to give your other suggestions a try, but maybe I should try to get REW to work better first? Or try a laptop and see if that works better. (I should do that anyway, as I want to test my car system.)
Thanks again for your comments.
As long as you feel confident that the midrange and up is correct and you're not down in the room noise, the readings you're getting should be fine. The absolute SPL isn't critical at the moment.
I don't use REW, so can't give advice on mic calibration beyond what's at the links below. There are many users and threads on diyAudio though, so I'm sure someone can help you if you want to improve its performance further.
https://www.roomeqwizard.com/help/help_en-GB/html/inputcal.html
https://www.roomeqwizard.com/help/help_en-GB/html/calfiles.html#top
I don't use REW, so can't give advice on mic calibration beyond what's at the links below. There are many users and threads on diyAudio though, so I'm sure someone can help you if you want to improve its performance further.
https://www.roomeqwizard.com/help/help_en-GB/html/inputcal.html
https://www.roomeqwizard.com/help/help_en-GB/html/calfiles.html#top
Last edited:
Another point I should have made earlier is that sometimes speakers don't sound like they measure. if they did, developing them would be easier.
If it sounds good to you, that's the most important part. There are highly regarded speakers with similar dips in response.
If it sounds good to you, that's the most important part. There are highly regarded speakers with similar dips in response.