audio-talk

so for the last 2 months i have been slowly building a new set of speakers. there were 2 reasons for bothering. firstly my son damaged one of chp 70s by poking it, and secondly, i had started to get frustrated by them not being able to quite handle louder listening, they seemed to crack up on peaks. investing time into the amp power supply and fettling the circuit was not yeilding the results i wanted, and hearing the amp on other speakers showed the amp was doing fine.
these new metronomes are ones that have been designed new from the ground up, and are 99% there. there is a little more fettling needed to the crossover, and we are talking slight changes in values.

the speakers use an sb acoustics tweeter and a faital 6" pro mid bass driver, crossover is a first order that has been abit of a pig to get right but is now pretty much there. the are currently bare mdf, i didnt want to finish them before i got them right, i dont know wether to veneer them or go with my trademark white.

i'll get some photos on when i can remember my password for photobucket

im currently throwing everything i can at them thats abit difficult to reproduce after having a breakthrough with the crossover earlier today, this was after having to take a walk yesterday to refrain from throwing them through a window due to frustration with the crossover.
they are doing well

Nice work.

Not bad at all.
What're the plans for finishing these, and which Faital driver did you go for? - I'm gonna be using some of their 10"s soon.

I'm not sure I'll be going back to full-range drivers, for the reasons you've found out. They're good for low volume stuff, but get overloaded way too easily when you want Bon Jovi on a Saturday night.

Chris

Hi chris, im quite impressed with the faital drivers, build quality is good and these ones which are the 6fe100's sound quite neutral. They have not had many hours on them, probably about 40 hours but are loosening up nicely. They are quite flat up to about 2 k ish but there is a ruddy great peak at about 2.5k which has made getting the crossover right abit of a challenge with it being a first order.
In terms of finish, i think im going to go with my usual metronome white
The tweeter will come down pretty low so has integrated pretty well

Good effort Ant .
Don't talk to me about treble peaks in midrange drivers - aarg !

Just acquired these, which I'm hoping for good things from-

http://theartofsound.net/forum/showthre ... ht=crysler

after a couple of weeks faffing and listening ive managed to pretty much remove the nasty peak in the response, florence welch no longer gives a headache. theres still some work to be done which will involve slightly tweaking the inductor value and adding a small amount of stuffing above the driver to tame a very slight honk. i used the spectrum analyser app i used when we looked at my dads metronomes and a frequency sweep burned to a cd to identify the peak. i assumed it was the tweeter causing the issue, but disconnecting the bass driver and running the sweep with the tweeter only on its part of the crossover showed it was coming in well above, running the bass driver only proved it was the culprit. it was actually at a lower frequency than i thought it was. the app is useful to get a general idea of where to go, i wouldnt use it for anything requiring precision measurement, but as a tool to give a visual representation of the response its rather useful and has steered me in the right direction

Sounds like you need more components in the crossover.

If you use a first order low-pass to kill a peak, by the time the peak is suitably attenuated, you might've lost half or even a full octave below it. For example, a narrow 6dB 2kHz peak. If you use a first-order crossover that starts rolling off at 1kHz, then you'll get a hole in the response between 1 and 2kHz, but the peak will be at the nominal level of the driver.
If you wanted the peak to be 6dB down (which isn't much), you'd have to start rolling off at 500Hz. You get the idea.

Two solutions, so far as I can tell...
1 - use a steeper crossover. This would kill the peak, but leave the range below more intact
2 - notch filter. If you can get the driver peak flattened, a first order electrical crossover will be a first order acoustic crossover, and everything will act how you want it to.

I was using some 15" drivers in my PA system as midbasses. They had a horrendous peak at 1.6kHz, but the tweeter was only good for 1kHz. I used a deep notch filter and a 4th order crossover and everything was just fine.

HTH
Chris

Finally got the crossovers put in boxes.
Ive put a 4 pole speakon socket on the back of the cabs then the boxes have a captive lead to plug into it. That way the tweeter cant get the wrong signal and get blown.
A couple of binding posts on the other end of the boxes and jobs done.
Only taken 18 months to get round to it.

Sounds better too without all the dodgy clip lead connections and tarnished wire

ive been doing a bit of faffing with these on and off for the last few months.
firstly, i recently got around to doing some finishing work after nearly 2 years and covered them in an off white suedette fabric

DSC_0569 by anthony cresswell, on Flickr

i also had a little faff with the stuffing. there isnt much in them at all, some above the driver at the small end of the line, and now a little in the base of the line too. just a covering of the base and a little bit spread over the top of the port.

smoothed them out nicely. this is the frequency sweep and rudimentary app plot

Spectrum_2017_08_12_16_44_26 by anthony cresswell, on Flickr

the upward tilt at the bottom of the graph looks like a room mode, and some panel sound. In practice its not as bad as it looks, it seems nice and even

happy with that

The LF is a bit misleading as its a linear frequency axis.

The sweep I used was a log sweep, there is a linear sweep too, might try that and see if there is any difference in the plots

Nick wrote: ↑Sat Aug 12, 2017 6:08 pm The LF is a bit misleading as its a linear frequency axis.

totally agree, theres something amiss there Ant, and the rising response in the bass must be an error starting at some point. I've never seen a room mode that pronounced at the very end.

generally, but not always, I get room modes like in the following example.....the centre at 60 and 90 hz.....I would say about 90% of the time if the speaker is within 6' of 2 walls. They can be programmed out but it's truly trixy.

this is more like it, found a setting to change it to a log axis

Log freq sweep by anthony cresswell, on Flickr