diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Digital Line Level (http://www.diyaudio.com/forums/digital-line-level/)
-   -   Can you use a digital active crossover to design a passive analog crossover ? (http://www.diyaudio.com/forums/digital-line-level/232996-can-you-use-digital-active-crossover-design-passive-analog-crossover.html)

RickDangerous 28th March 2013 10:23 PM

Can you use a digital active crossover to design a passive analog crossover ?
 
I believe the title says it all.

And, as the title will tell you, i am totally new to all this DIY speaker crossover lark.
I have gathered, however, that the hardest part is getting the crossover right.
So i was wondering if i could maybe buy a used cheap active digital crossover (Behringer DCX2496 is most common, and can be modded easily & cheaply to sound better), get the crossover frequencies & slopes right for my drivers+cabinet, and then probably get a helpful soul to assist in the design of the analog passive crossover.
I'd imagine it wouldn't be perfect, but i'd also imagine it would be closer than what i would achieve without the digital active crossover to experiment with.

Now, if it can help, here is my aim:
- tri-amped 3 way speaker that goes down to 25 Hz or so
- flat, neutral
- detailed, analytical (PMC IB1S / MB2 are the inspiration here)
- used as monitors in a professional recording studio
- the amp is a hybrid tube amp, the DARED DV6C, a 6 channel amp.

First, i will probably go for an open cabinet 3 way design with cheap drivers, but the project above is the ultimate aim.

Thanks to all !
Rick

picowallspeaker 28th March 2013 11:02 PM

Quote:

Originally Posted by RickDangerous (Post 3431904)

Now, if it can help, here is my aim:
- tri-amped 3 way speaker that goes down to 25 Hz or so
- flat, neutral
- detailed, analytical (PMC IB1S / MB2 are the inspiration here)
- used as monitors in a professional recording studio
- the amp is a hybrid tube amp, the DARED DV6C, a 6 channel amp.

First, i will probably go for an open cabinet 3 way design with cheap drivers, but the project above is the ultimate aim.

Thanks to all !
Rick

That's not a 6 ch amp, it's not even half.
It doesn't feature a single analog part.

For good money & study I would search for some bargain old 2 channel amplifier ( integrated) and a 3 way speaker .
A 30/40 W amp is enough to power a 3 way in a medium sized room at
adeguate level .
For 20 Hz flat you really got to get into it, mathematically and phyisically...
and the power delivered could be always not enough ( or the drivers won't over-radiate :confused::rolleyes: )

RickDangerous 28th March 2013 11:08 PM

Quote:

Originally Posted by picowallspeaker (Post 3431951)
That's not a 6 ch amp, it's not even half.
It doesn't feature a single analog part.

For good money & study I would search for some bargain old 2 channel amplifier ( integrated) and a 3 way speaker .
A 30/40 W amp is enough to power a 3 way in a medium sized room at
adeguate level .
For 20 Hz flat you really got to get into it, mathematically and phyisically...
and the power delivered could be always not enough ( or the drivers won't over-radiate :confused::rolleyes: )

Well, it must have something analog in it, it uses tubes.
It has 6 speaker outputs, each of which can be adjusted for volume (and possibly delay, i'm not sure).
So are you saying that tri-amping with that amp would be the exact same as plain amping ?

picowallspeaker 28th March 2013 11:41 PM

Quote:

Well, it must have something analog in it, it uses tubes.
They ( Dared Company ) wanted you to think and answer like that :mad:;)
Consider also that in the 6Moons review there's mentioned the possibility to put 2 channels to Bridge Tied Load mode, but that option wasn't made for this model.
What you are going to make, I 've read along the pages on the forum, it's
possible and it's doable, but first you have to define 'cheap' . Because
you can make a good speaker with cheap parts, and a good speaker doesn't really need expensive drivers, or crossover parts .
And then there's the system, as the sound is made by the various parts ( processing ) ;the amplifier gives strenght to the signal , it gives gain to a low Z load : for a correct and safe operation the 8 Ω load is the standard :o

GoatGuy 28th March 2013 11:47 PM

OK, I read the whole darn DARED DV6C review, and I took the last hour to carefully go through the online schematics.

[1] It is an amplifier that has 3 - 12AX7 dual-triodes.
[1a] of which 5 triodes are in use (the "subwoofer" config doesn't use 6th triode)
[1b] where all triodes are NOT amplifying at all - but are cathode followers
[1c] ... which of course just follow the input like dumb bunnies.
[1d] but hey... the 3 tubes are bottom lit with bright red LEDs. Lots of pretty casework.

[2] The actual amplifiers are a trio of dual-channel "Class D" amplifier chips.
[2a] each has a "left" and "right" (or #1, and #2) channel
[2b] they're all hooked up to bypass caps and output inductors "by the book"
[2c] and they're powered to produce approximately 65 watts per channel pretty-free from distortion

[3] The power supply to the 12AX7's is laughable. Well, they only charge $695 list.

[4] The power supply to the Class-D amplifiers is fine - they use huge electrolytics, a bit of bypassing with smaller caps, and (if totally by the book) probably good circuit layout.

Which means that these will be fine for your experimentation, at least to more-or-less get some crossover performance measured "by ear".

So, some recommendations (and answers to your questions)

first, get an "instrumentation reference mic" of some sort. As lousy a name that Behringer generally has, their instrumentation reference mics are surprisingly good. Be aware though... they're "professional active condenser mics", so will need phantom power. This may cause you to climb for weeks down an unknown (but ultimately satisfying) rabbit hole.

second ... or ... get a USB condenser microphone of similar type for your computer. You're going to need something to record the findings (since trusting your ears is asking for trouble - through conflicting results)

third Yes, you may be able to use the Behringer DCX2496 to tune things. It is NOT a bad idea at all. You can tri-amp the 3 cones in each speaker box, and frankly, you shouldn't put a lot of uber-hi-end wire or anything else into the experimentation. You're looking for qualitative results.

fourth ... working with one speaker at a time, and any of dozens of free signal-generator applications for your iPhone, Android, PC or MacBook ... generate either a sweep of frequencies, or, pink noise, or, step through closely spaced frequencies to see how your cones really respond. Don't make this test a "loud" test, just pretty good [75-80 dB] at the microphone, to override any room noise.

fifth, then RECORD what you find, and use FFT (fourier transform) software to convert the response to a graphic form. You won't have access to phase, but it doesn't matter at this stage.

sixth Look at the peaks, look at the valleys. Try it again placing the microphone in a different position. See if it changes. Keep at this. You may want to put the microphone just 1 or 1.5 meters in front of (directly) the speaker. It will give the best baseline response.

seventh Now you have some idea of what you want to counteract either by "cross-over" frequency selection (which allows you to "dip" response if needed, without components!) or by component choice (number of poles, etc).

OK

eighth now it is time to run two speakers at a time. The EASIEST will be the tweeter and midrange cone. You know now what frequencies you'd like to use for cross-over, so try them. It'll not be very musical, but listen carefully ... trying to eliminate the 'hills' in response that make for such poor listening. Then, [ironically] cut the tweeter, and work with the woofer cone and the midrange. Get the crossover right for them. Add the tweeter in.

How does it sound? Don't touch ANY of your controls until you've spent a couple days listening to a wide variety of music. You need that so that you will be able to say - with some conviction - what you want to change about the system. Go ahead though, and move around the physical speakers a lot. You're now in the data phase.

Tenth ... Now remember that white/pink noise and microphone-response thing we did earlier? Time to do it again, now with all three speakers working. Make several recordings, do FFT spectrum analysis. Look again for peaks (that may surprise you for being there), and so on. You can now tweak the 2496 to maybe get them to drop, or perhaps to get the valleys to rise and fill back in. Listen again to music. Better?

Lastly ... you're now armed with enough information (and own opinion) that you should be able to constructively modify the response to achieve what you're looking for. It takes time ... and you use subtle changes, not big ones in the end.

Write down all the parameters, cross-over frequencies, db/octave rates, and whatever else comes from all this, and begin designing your fixed, higher quality crossover.

But do not be surprised if the one you build from components sounds somehow both better ... and yet different ... because it will have its own notion of "Q" and resonance, and decibels per octave, and phase shift. That, unfortunately is way, way, way beyond what I can write, or care to.

GoatGuy

picowallspeaker 28th March 2013 11:51 PM

Some bargain old 2 channel integrated amplifier and a 3 way speaker pair ....
:)
If any....

RickDangerous 29th March 2013 12:12 AM

Quote:

Originally Posted by GoatGuy (Post 3431999)
OK, I read the whole darn DARED DV6C review, and I took the last hour to carefully go through the online schematics.

[1] It is an amplifier that has 3 - 12AX7 dual-triodes.
[1a] of which 5 triodes are in use (the "subwoofer" config doesn't use 6th triode)
[1b] where all triodes are NOT amplifying at all - but are cathode followers
[1c] ... which of course just follow the input like dumb bunnies.
[1d] but hey... the 3 tubes are bottom lit with bright red LEDs. Lots of pretty casework.

[2] The actual amplifiers are a trio of dual-channel "Class D" amplifier chips.
[2a] each has a "left" and "right" (or #1, and #2) channel
[2b] they're all hooked up to bypass caps and output inductors "by the book"
[2c] and they're powered to produce approximately 65 watts per channel pretty-free from distortion

[3] The power supply to the 12AX7's is laughable. Well, they only charge $695 list.

[4] The power supply to the Class-D amplifiers is fine - they use huge electrolytics, a bit of bypassing with smaller caps, and (if totally by the book) probably good circuit layout.

Which means that these will be fine for your experimentation, at least to more-or-less get some crossover performance measured "by ear".

So, some recommendations (and answers to your questions)

first, get an "instrumentation reference mic" of some sort. As lousy a name that Behringer generally has, their instrumentation reference mics are surprisingly good. Be aware though... they're "professional active condenser mics", so will need phantom power. This may cause you to climb for weeks down an unknown (but ultimately satisfying) rabbit hole.

second ... or ... get a USB condenser microphone of similar type for your computer. You're going to need something to record the findings (since trusting your ears is asking for trouble - through conflicting results)

third Yes, you may be able to use the Behringer DCX2496 to tune things. It is NOT a bad idea at all. You can tri-amp the 3 cones in each speaker box, and frankly, you shouldn't put a lot of uber-hi-end wire or anything else into the experimentation. You're looking for qualitative results.

fourth ... working with one speaker at a time, and any of dozens of free signal-generator applications for your iPhone, Android, PC or MacBook ... generate either a sweep of frequencies, or, pink noise, or, step through closely spaced frequencies to see how your cones really respond. Don't make this test a "loud" test, just pretty good [75-80 dB] at the microphone, to override any room noise.

fifth, then RECORD what you find, and use FFT (fourier transform) software to convert the response to a graphic form. You won't have access to phase, but it doesn't matter at this stage.

sixth Look at the peaks, look at the valleys. Try it again placing the microphone in a different position. See if it changes. Keep at this. You may want to put the microphone just 1 or 1.5 meters in front of (directly) the speaker. It will give the best baseline response.

seventh Now you have some idea of what you want to counteract either by "cross-over" frequency selection (which allows you to "dip" response if needed, without components!) or by component choice (number of poles, etc).

OK

eighth now it is time to run two speakers at a time. The EASIEST will be the tweeter and midrange cone. You know now what frequencies you'd like to use for cross-over, so try them. It'll not be very musical, but listen carefully ... trying to eliminate the 'hills' in response that make for such poor listening. Then, [ironically] cut the tweeter, and work with the woofer cone and the midrange. Get the crossover right for them. Add the tweeter in.

How does it sound? Don't touch ANY of your controls until you've spent a couple days listening to a wide variety of music. You need that so that you will be able to say - with some conviction - what you want to change about the system. Go ahead though, and move around the physical speakers a lot. You're now in the data phase.

Tenth ... Now remember that white/pink noise and microphone-response thing we did earlier? Time to do it again, now with all three speakers working. Make several recordings, do FFT spectrum analysis. Look again for peaks (that may surprise you for being there), and so on. You can now tweak the 2496 to maybe get them to drop, or perhaps to get the valleys to rise and fill back in. Listen again to music. Better?

Lastly ... you're now armed with enough information (and own opinion) that you should be able to constructively modify the response to achieve what you're looking for. It takes time ... and you use subtle changes, not big ones in the end.

Write down all the parameters, cross-over frequencies, db/octave rates, and whatever else comes from all this, and begin designing your fixed, higher quality crossover.

But do not be surprised if the one you build from components sounds somehow both better ... and yet different ... because it will have its own notion of "Q" and resonance, and decibels per octave, and phase shift. That, unfortunately is way, way, way beyond what I can write, or care to.

GoatGuy

Wow, thanks a LOT, GoatGuy !
Pretty much exactly what i need for now.

One question, though:
When you say that the tubes "just follow the input like dumb bunnies", does that mean that they do not alter the sound at all ?

Thanks a lot.
Rick

GoatGuy 29th March 2013 12:41 AM

Quote:

just like dumb bunnies
means that ideally, they alter the sound the least of any amplification topology. The designers (wisely) used a -12 volt supply for the cathode, via resistor R79, R84, R87, etc for each triode. The triodes therefore will find a bias-point of about 1.3 milliamps (13 volts across the 10K cathode resistor), and will track that in a small range. If the input is nominally 1.0 volt, then the cathode will swing almost exactly 1.0 volt, which swings cathode current from 1.2 to 1.4 milliamps. This keeps the tube running in its very linear mode ... thus hardly coloring or influencing the input at all. Quite unlike a real "class A" amplifier where the plate on the top of a triode's schematic diagram is really doing amplification!

You're welcome.

PRINT the advice, post it on a wall so you can glance at it from time to time, to remind yourself of all the steps. They're in a certain order ... for reasons!

GoatGuy

counter culture 29th March 2013 12:44 AM

Quote:

Originally Posted by RickDangerous (Post 3432028)
When you say that the tubes "just follow the input like dumb bunnies", does that mean that they do not alter the sound at all ?

We can only hope not.

RickDangerous 29th March 2013 12:50 AM

Quote:

Originally Posted by GoatGuy (Post 3432070)
means that ideally, they alter the sound the least of any amplification topology. The designers (wisely) used a -12 volt supply for the cathode, via resistor R79, R84, R87, etc for each triode. The triodes therefore will find a bias-point of about 1.3 milliamps (13 volts across the 10K cathode resistor), and will track that in a small range. If the input is nominally 1.0 volt, then the cathode will swing almost exactly 1.0 volt, which swings cathode current from 1.2 to 1.4 milliamps. This keeps the tube running in its very linear mode ... thus hardly coloring or influencing the input at all. Quite unlike a real "class A" amplifier where the plate on the top of a triode's schematic diagram is really doing amplification!

You're welcome.

PRINT the advice, post it on a wall so you can glance at it from time to time, to remind yourself of all the steps. They're in a certain order ... for reasons!

GoatGuy

Heck, i'm gonna tattoo it on my forhead (backwards, i'm not stoopid).

Thanks again, that was exactly the sort of advice i needed.
I'll admit i didn't understand anything about the catholic swinging, though.


All times are GMT. The time now is 10:56 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2