Adam5022 Posted August 26, 2022 Share Posted August 26, 2022 ive been tinkering with this crossover for awhile. it looks pretty flat, but im not sure if its good enough? what do you think? any help would be great as im only just getting into this, cheers 1 Link to comment Share on other sites More sharing options...
davewantsmoore Posted August 26, 2022 Share Posted August 26, 2022 32 minutes ago, Adam5022 said: ive been tinkering with this crossover for awhile. The crossover looks ok. It's summing (+6dB) nicely and over a wide range (and overage and a bit above and below ish). Where you might go wrong, is "garbage in, garbage out".... remember that the actual result, rather than the one you see in your simulator.... will depend on your measurements, and whether they are truly representative of the speaker. Examples... What's with the broad drip between 200 and 700? Is that real, or just a gigo measurement? If it's real you definitely have to do something about it .... more likely you have measured, windowed, spliced, your data wrong, etc. If you apply some really aggressive smoothing by eye through the response.... it looks like 800 to 1600 is about 3dB higher than the next octave (1600 to 3200) if this is real, it will be quite audible (detail, harsh, etc.) .... is it real? don't know. Perhaps try using some more gating/windowing, or more smoothing, or comparing to some measurements taken closer to the drivers. All that being said.... it's +/- 2dB .... which is pretty good.... so if noting if iffy in your data, then it's pretty good (flat is probably not the right listening target through) (what is the distance from the speaker?) Link to comment Share on other sites More sharing options...
andyr Posted August 26, 2022 Share Posted August 26, 2022 Looks pretty good to me, Adam. The 'crossover is at ~1500Hz, right? And you seem to have connected the drivers the right way round (relative to each other) ... as you don't have a dip in the overall response, at the XO frequ. The only improvement I can think of is to put a small, 5w resistor in series with each tweeter - so the overall FR tilts down towards the rhs. Andy Link to comment Share on other sites More sharing options...
Adam5022 Posted August 26, 2022 Author Share Posted August 26, 2022 20 minutes ago, davewantsmoore said: What's with the broad drip between 200 and 700? i was thinking this curve was actually a good thing? like a "house curve" you see in sub builds? 20 minutes ago, davewantsmoore said: Perhaps try using some more gating/windowing, or more smoothing, or comparing to some measurements taken closer to the drivers. (what is the distance from the speaker?) im not too sure what gating or windowing is mate? any chance you could elaborate in a way a dumbass like myself could understand haha on the cross over design (jeff bagby) it was the listening position at 2.000 i assume thats 2 m? but not sure Link to comment Share on other sites More sharing options...
Adam5022 Posted August 26, 2022 Author Share Posted August 26, 2022 24 minutes ago, andyr said: The 'crossover is at ~1500Hz, right? The only improvement I can think of is to put a small, 5w resistor in series with each tweeter - so the overall FR tilts down towards the rhs. Andy yeah mate, 1500hz is correct, sweet. ill look at the resistor Link to comment Share on other sites More sharing options...
davewantsmoore Posted August 26, 2022 Share Posted August 26, 2022 15 minutes ago, Adam5022 said: i was thinking this curve was actually a good thing? like a "house curve" you see in sub builds? Try to get it flat.... narrow wiggles are ok, but broad peaks/dips are quite audible even if only 3dB. You may have made an error in the part of the procedure where you splice the near and far measurements together (assuming you are following jeffs method) 15 minutes ago, Adam5022 said: im not too sure what gating or windowing is mate? any chance you could elaborate in a way a dumbass like myself could understand haha Ah, ok. Sorry for the confusion. If you're following Jeff/Charlies procedures, then just do that (and ignore "windowing") .... but you're response looks quite wiggly...... which is likely much more to do with your measurements, than to do with "no polishing the crossover enough". 15 minutes ago, Adam5022 said: on the cross over design (jeff bagby) it was the listening position at 2.000 i assume thats 2 m? but not sure Yep. Is that how far the mic was from the speaker? (2m?). Link to comment Share on other sites More sharing options...
Janus77 Posted August 26, 2022 Share Posted August 26, 2022 Build thread please Link to comment Share on other sites More sharing options...
Adam5022 Posted August 26, 2022 Author Share Posted August 26, 2022 @davewantsmoore these are just "sims" I guess. Purely from frd/zma I very much assume the wiggly-NESS is from me haha both drivers are very flat on the charts online. Then turn to $#@ once I touch them haha Link to comment Share on other sites More sharing options...
Adam5022 Posted August 26, 2022 Author Share Posted August 26, 2022 ive had a couple wines and have been playing around more. i decided to change to woofer in my MTM. im getting better responses. but still not happy with the results. im new to the DIY game, but as a reference, the base layer (7 channels) in my HT is a M&Ks150mkII/s150T line up and krix atmospherix for my heights so pin point accuracy is my main goal, otherwise what am i doing? im watching and re watching "toid DIY" on youtube, that guy knows his ****. but if you can recommend others that can teach an idiot like me, that would be sick haha Link to comment Share on other sites More sharing options...
davewantsmoore Posted August 28, 2022 Share Posted August 28, 2022 On 26/8/2022 at 8:55 PM, Adam5022 said: @davewantsmoore these are just "sims" I guess. Purely from frd/zma You mean from the manufacturer supplied data? Link to comment Share on other sites More sharing options...
Adam5022 Posted August 28, 2022 Author Share Posted August 28, 2022 @davewantsmoore yeah that's correct mate Link to comment Share on other sites More sharing options...
Adam5022 Posted August 29, 2022 Author Share Posted August 29, 2022 so ive changed LF drivers from a Dayton to a B&C, the B&C owns the Dayton in all ways. and it has given me an awesomely flat response. only thing, is my impendence is going haywire? showing different things on different graphs. we have a very simple 2nd order XO in play. MTM config. x2 of the LF drivers both 8ohm and wired accordingly, so that should give me a 4ohm load hey? which matches the tweeters ohm also. i think its possibly just how i traced the graphs of the B&C drivers, maybe it didnt perfectly translate to a good ZMA file? what do you guys think? cheers fellas Link to comment Share on other sites More sharing options...
RoHo Posted August 29, 2022 Share Posted August 29, 2022 The woofer impedance trace is wrong-minimum at 25 ohms. Looks like you got the scaling wrong when you traced the driver impedances. I would retrace both impedance curves and check that the original matches what ends up in PCD. 1 Link to comment Share on other sites More sharing options...
Adam5022 Posted August 30, 2022 Author Share Posted August 30, 2022 Will do mate. Do you recommend a program to trace? I'm currently using FP graph tracer. But it's pretty useless. In saying that, I'm really happy with the crossover. knowing that it's just a graph misshap. Maybe I'll just order the speakers and get real measurements. Which opens up more Learning, how to make real sweep files haha Link to comment Share on other sites More sharing options...
RoHo Posted August 30, 2022 Share Posted August 30, 2022 I use that program too and it does the job for me. Double check you have set the boundaries correctly and added the correct scales. Link to comment Share on other sites More sharing options...
Adam5022 Posted August 30, 2022 Author Share Posted August 30, 2022 I double checked my x,y,z parameters and they seem good. I think it's just the trace, as it looks like this when I highlight it ( see pic) On the peak it's all dotty and Incomplete, then corrects itself on the arch up. I've tried clicking a bunch of different spots to find the best match, but no good. Be great if I could just trace the line, rather than clicking on it and hoping for the best Link to comment Share on other sites More sharing options...
RoHo Posted August 30, 2022 Share Posted August 30, 2022 19 hours ago, Adam5022 said: I double checked my x,y,z parameters and they seem good. I think it's just the trace, as it looks like this when I highlight it ( see pic) On the peak it's all dotty and Incomplete, then corrects itself on the arch up. I've tried clicking a bunch of different spots to find the best match, but no good. Be great if I could just trace the line, rather than clicking on it and hoping for the best The main problem is not the shape of the trace but that on PCD the minimum impedance from the trace you show above is 25 ohms (the broad dip at about 400 Hz). this should be 4-8 ohms. Something funny has happened to the scaling on the tracer program, not on PCD. Make sure you have set the scale to linear, not log, for the impedance. Double check the values of the boundaries you set on the acquisition window. 2 Link to comment Share on other sites More sharing options...
Allan Posted June 28 Share Posted June 28 this resembles a Proac curve, if it were me I would have tried it. Driver quality will come into play but certainly worth a try Link to comment Share on other sites More sharing options...
GlassAudio Posted June 29 Share Posted June 29 @Adam5022 Crossovers really need to be designed from measurements taken of the drivers in the cabinet including impedance measurements. You can get setup doing this for not much $$ For crossover design, this software is free and extremely good https://kimmosaunisto.net/ Calibrated measurment microphone https://www.cross-spectrum.com/measurement/calibrated_dayton.html Suitable USB audio interface https://focusrite.com/en/usb-audio-interface/scarlett/scarlett-2i2 1 Link to comment Share on other sites More sharing options...
Allan Posted August 2 Share Posted August 2 On 30/06/2024 at 6:59 AM, GlassAudio said: @Adam5022 Crossovers really need to be designed from measurements taken of the drivers in the cabinet including impedance measurements. You can get setup doing this for not much $$ For crossover design, this software is free and extremely good https://kimmosaunisto.net/ Calibrated measurment microphone https://www.cross-spectrum.com/measurement/calibrated_dayton.html Suitable USB audio interface https://focusrite.com/en/usb-audio-interface/scarlett/scarlett-2i2 buy yourself the DATs V3 measurement system, highly recommended! You'll need something like it to import accurate impedance measurements into your xover software. Link to comment Share on other sites More sharing options...
Recommended Posts