sreten said:
"Pureconjecture and uniformed"??? Where are you from mate? You might want to check your thesaurus...or maybe you were fumbling around in the dark for some smart phrases. You may want to fine-tune and re-tweak your "ill-thought out poorly designed" response. Sorry, couldn't resist.
Oh, and by the by, have you ever listened to the AR.COM DIY? Or were you pureconjecturing?
Sorry mate.
salas said:
Great advice, I will have him get those coils upright. We will also change the polarity of the tweeter as well as has been suggested. Thanks.
Response curves greatly speed up the design process. Forget what the ideal response curve should look like... but imagine you could draw the response ideal to your ears, then design a new speaker and using software simulate a crossover to match your intended response curve. You've done this much quicker than by trial and error listening and tweaking forwever to get it sound right.
Measurements are purely used to confirm your simulated design.
If you've checked everything, then are unhappy, then I think we've found your problem.
If you look at the modelled and measured FR curve for the BAMTM, you'll see a flat response.
If you prefer the AR.com DIY - then you need a speaker with a similar FR and distortion curves. I therefore very much doubt you would therefore prefer any of Zaph's designs.
Since most designers can't possibly know what we all prefer to hear, they design for whatever either sounds good to them, or for a target response. A commonly used target response is a flat response - which the BAMTM does well over a wide listening angle. A flat response can be a problem depending on what music you listen to (how it was mixed and the voicing preference put in at the studio).
I know the AR.com DIYs when tweeter connection is reversed will accentuate the upper midrange (cause a peak). This might lead you to think the BAMTM has a recessed vocal range - when in fact it is just producing the original signal going into it.
Another cause of brightness is a highly reflective listening room.
Can I ask how long you listen to the AR.com DIY's for? What volume levels? Do you find any part of them fatiguing over a long listening period?
It may take some time to get used to the BAMTM. Listen to it over a longer period. Do you find you get less listening fatigue than your other speakers?
PS: Before I could measure anything, I thought the KEFC30's had better vocals than the AR.com DIYs. I measured them purely to find out what part of the frequency spectrum was causing this. I found the KEF's had a flat response and good phase tracking through the crossover, whereas the AR.com had a dip and poor phase tracking through the crossover.
The measurements highlighted the differences. If I don't like a speaker - I don't like it regardless of what measurements might tell me.
I hope you follow what I am saying.
Cheers,
David.
SoundQuest said:
Only one of the two upright, and perpendicular to the other. Listen like that before you change polarity!
JimOfOakCreek said:
I've had 3.5r for the last 12 years and they are great. Not a bargain by anyones standard. They also have their own problems like any other speaker out there.
You guys flamed SQ pretty good, unfortunately I don't think he understands the technical aspects of designing speakers and building them. That's why he's suggesting moving the crossover point. It may also be that he is used to a pair of speakers that don't have flat frequency response and Zaph's design doesn't have that bump in midrange. Who knows? I bet SQ is going to find out that the joint on the inductor is bad or the tail on the capacitor is broken off.
SQ, have you scraped the wire varnish from the tails on the inductors? Could be as simple as that.
R-Carpenter said:
A little presumptuous, but okay, whatever. The original AR.COM had a small dip; Ed Frias rectified that with an updated crossover -- the same version I have used in the 9 pairs I have built. So no, my ears have not grown accustomed to a wiley sounding pair of speakers; I assure you, there is no bump in the midrange.
I will have him check the varnish though, this was his first pair and he did not wait for me to supervise him.
sreten raised a valid point. Apart from too much solder on the -ve return legs, it also looks like too much heat was applied for too long. Is it possible you have damages the capacitors?
I don't know if you can alter a capacitor's behaviour with overheated soldering. Obviously if the tweeter cap is broken, no sound will come out of the tweeter. IF the woofer's capacitor is broken - it should increase midrange levels (yet you mentioned these sounded recessed)....
Worth mentioning anyway.
David.
I don't know if you can alter a capacitor's behaviour with overheated soldering. Obviously if the tweeter cap is broken, no sound will come out of the tweeter. IF the woofer's capacitor is broken - it should increase midrange levels (yet you mentioned these sounded recessed)....
Worth mentioning anyway.
David.
For the record my AR.coms have the updated crossover - a dip is still there. I cannot say how less of a dip over the original version.
David.
SoundQuest said:
"Pureconjecture and uniformed"??? Where are you from mate? You might want to check your thesaurus...or maybe you were fumbling around in the dark for some smart phrases. You may want to fine-tune and re-tweak your "ill-thought out poorly designed" response. Sorry, couldn't resist.
Oh, and by the by, have you ever listened to the AR.COM DIY? Or were you pureconjecturing?
Sorry mate.
Hi,
Fairly obvious there is a unintentional space missing ..... I can spell.
No I've never listened to the ar.com. I have read the article when
it was available and I recall there was no real methodology used.
FWIW I stand by everything I said. I would take one of Zaphs
designs over any "halfbaked tuned by ear lash up" designed
with incomplete grasp of all the factors, which would have be
shown up by measurements.
My simple points made in post #47 is the ar.com is not IMO a
good reference, I cannot see the point of trying to insist it is.
However lets hope your problem is the inductor alignment issue
I described in post #56, rather than a different opinion on voicing.
🙂/sreten.
Dave Bullet said:
I was going to bring up this point last night. It does look like there was some serious heat applied. And unless he used a heatsink before the component, a lot of that heat went into his caps.
I keep my leads long I always attach an alligator clip before the component as a heatsink.
Edit: Just simulated it, looks fine as a result of the low crossover point. Peaking is seen with a smaller (incorrect) shunt cap on the woofer, perhaps check that value.
I have to say that I have a lot of respect for Zaph and have mostly looked at his driver data over the years. But I notice something that I have to question in this design. I've found over the years, that when the driver impedance is not fully compensated, I'm not promoting Zobles here, that the shunt cap on the woofer resonates (with a high Q) with the driver inductance and can produce significant peaking the the LP response.
A solution that I often use is to put a few ohms in series with the shunt cap, then of course reoptimze the values to get as close as possible to the target response. It saves on components and can sometimes provide a better response than with a Zobel. I've not built this exact system but the issue comes up more often than not.
Could someone measure the input to woofer electrical response, not simulated. I have also simulated it in SPICE and CALSOD so this effect is certainly real. It sometimes does not show up if the crossover point is very low, or if the driver has very low inductance. Curious?
Pete B.
I have to say that I have a lot of respect for Zaph and have mostly looked at his driver data over the years. But I notice something that I have to question in this design. I've found over the years, that when the driver impedance is not fully compensated, I'm not promoting Zobles here, that the shunt cap on the woofer resonates (with a high Q) with the driver inductance and can produce significant peaking the the LP response.
A solution that I often use is to put a few ohms in series with the shunt cap, then of course reoptimze the values to get as close as possible to the target response. It saves on components and can sometimes provide a better response than with a Zobel. I've not built this exact system but the issue comes up more often than not.
Could someone measure the input to woofer electrical response, not simulated. I have also simulated it in SPICE and CALSOD so this effect is certainly real. It sometimes does not show up if the crossover point is very low, or if the driver has very low inductance. Curious?
Pete B.
I have just completed the BAMTM and now you guys have me wondering
about tweeter polarity!
It seems clear on Zaph's website that the positive leg (after filtering) goes to the positive
terminal on the Seas tweeter. Is this wrong?javascript:smilie('😕')
about tweeter polarity!
It seems clear on Zaph's website that the positive leg (after filtering) goes to the positive
terminal on the Seas tweeter. Is this wrong?javascript:smilie('😕')
vaughn said:
Hi,
The crossover is 2nd order electrical, 4th order Linkwitz/Riley acoustic.
For 4th order L/R the drivers must be in phase. See modelled reverse null.
🙂/sreten.
Zaph BAMTM measurement
Hi All,
First off all respects go to zaph for his efforts and very good designs.
Just wanted to humbly say that I built the BAMTMs (using 7 ohm resistor option) and measured the response using Speaker Workshop. I see a constant slide from 100 hz to 2 Khz which levels after that with 10 dB difference from 2k to 100 hz. It manifests itseld as a too rich base sound when compared to ZD5. I checked the crossover components and everything was as per the design.
Next I did measure each driver individually and simulated the crossover in the software and got the same result as actual measured results. So either my tweeter is very insensitive (defective) which I doubt, or I am screwing something else up. changing the resistor to a lower value increased tweeter level a bit, and then as you reduce the resistor the whole Xover function collapses.
Can post the results of my tests if any one is interested.
Cheers and and respect to all of you
jetty
Hi All,
First off all respects go to zaph for his efforts and very good designs.
Just wanted to humbly say that I built the BAMTMs (using 7 ohm resistor option) and measured the response using Speaker Workshop. I see a constant slide from 100 hz to 2 Khz which levels after that with 10 dB difference from 2k to 100 hz. It manifests itseld as a too rich base sound when compared to ZD5. I checked the crossover components and everything was as per the design.
Next I did measure each driver individually and simulated the crossover in the software and got the same result as actual measured results. So either my tweeter is very insensitive (defective) which I doubt, or I am screwing something else up. changing the resistor to a lower value increased tweeter level a bit, and then as you reduce the resistor the whole Xover function collapses.
Can post the results of my tests if any one is interested.
Cheers and and respect to all of you
jetty
Go ahead and post your measurements, and, your data files with the crossover simulations for us to take a look at.
I've built a pair of the BAMTMs (sealed) and they measured extremely flat.
Where are you at? Are you near Columbus, OH?
I've built a pair of the BAMTMs (sealed) and they measured extremely flat.
Where are you at? Are you near Columbus, OH?
- Status
- Not open for further replies.