Active crossovers - opamps vs DSP

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
planet10 said:


From what i undersatnd there is a lot of room for improvement in this unit as well... The fellow from OverKill Audio says he has significantly improved the ones he uses with a better case, and an analog power supply.

dave

Would you really expect the folks at Overkill to sell their $70,000 speakers with a stock, over-the-counter crossover solution like a DEQX PDC? By the same token, would you really expect the [mentally insane] buyer of a $70,000 speaker system to be happy with an unmodified crossover? :whazzat:
 
frugal-phile™
Joined 2001
Paid Member
mac said:
Would you really expect the folks at Overkill to sell their $70,000 speakers with a stock, over-the-counter crossover solution like a DEQX PDC? By the same token, would you really expect the [mentally insane] buyer of a $70,000 speaker system to be happy with an unmodified crossover? :whazzat:

The fellow called me on the phone out of the blue... he wanted me to design a TL for his new woofer, i don't think he had a $70 k system at the time -- i seem to recall more like $20k (seems a bit steep for anything in the pics i've seen). He never did follow thru (but then his woofer isn't really suited to TL)

dave
 
^^ Tens of thousands of people use minidsp products, how about that ESP from year 1999?

They are totally different products, the verstatility of minidsp (alike most other dsp-xo boards) is ovewhelming. I have speakers that use minidsp 2x4HD, 4x10HD and Hypex FA123.

I can't see any use for that ESP LR4 electric circuit, because I don't know of any loudspeaker drivers that don't need equalization and delay to achieve LR4 acoustic slopes that sum well. Think about baffle loss/peak, diffractions, driver resonances etc. nonlinearities.
 
Last edited:
I've used the ESP LR crossover and have nothing to say against it. Very neutral sounding if you use quality parts, and cheap to manufacture (I didn't order the boards, I just adapted the project to my needs and did the PCB myself).
DSP is out of my league. MiniDSP is just too expensive, and programming something on ARM is pretty much unreachable.
 
A single Minidsp 2x4HD is 205$+cargo, software licence is 10$. You can use older mid-class AV receiver (that has multichannel analog input) as amp, fleamarket price around 50$.

If one uses only a computer/server as source, dsp programs can be found for free, then you need a multichannel soundcard/dac or multiple outboard dacs and amps.

Crossover components for a passive 3-way easily cost almost the same with quality coils and caps.
 
The miniDSP 2x4 HD isn't the best option for tri-amping, but would be ideal for bi-amping (one will need a 2 way passive crossover for the tweeter/mid if they wanted to make a 3 driver system). Two miniDSP 2x4 kits would be the cheapest route using their products for a 3 way design. Check out Stereo 3/4way Xover to see which units they recommend.
Dayton audio has the DSP-408 4x8 DSP on parts-express for $150 that could also be used for tri-amping.
 
IMHO DSP is great for development because it is quite and flexible, and one can switch back and forth without changing anything hardware.
But once the design is done, many choose to emulate the digital solution by analogue ones.
And that was a statement by some very experienced professionals developing their own digital solutions.
https://www.diyaudio.com/forums/the...owtorch-preamplifier-ii-9870.html#post5285486
(post #98695)
There are not many crossover designs that can only be done digitally.

The choice is of course not restricted to DSP vs opamp.
I have seen systems completely done with a PC in MatLab.
And in my own (analogue) attempt, I have now replaced most IC opamps (at unity gain) with discrete JFET followers.
A Modular Analogue Active Crossover Filter Solution
https://www.diyaudio.com/forums/pass-labs/140488-b1-turbo-chip-14.html#post5740480
(post #136)

And for the few modules where opamp is a must, a discrete dual opamp come to the rescure.
https://www.diyaudio.com/forums/analog-line-level/218373-discrete-opamp-design-334.html#post5711190
(post #3333)

There is a lot more fun doing your own, and they cost also less than any commercial offering.
The modular design makes it even more flexible.

:)


Cheers,
Patrick
 

Attachments

  • 181020 Duelund 3W Act XOvr.JPG
    181020 Duelund 3W Act XOvr.JPG
    310.7 KB · Views: 189
ive been wondering about power consumption. My interest is battery power setups.

Which is lower power? Old car dsp units (alpine h701/h700) seam to use 10w+~, i forget exactly. Old analogue car equalizers use quite a bit of power too... i need to measure and write down results...
there was a seller of analogue crossover board on ebay... been wondering what the power consumption of it would be? i can dig the link up if interested.

DSP can do time correction, which is quite an advantage.... something i only am just starting to play with. for 4.1 speaker setups.

for the price and power consumption of these dsp units, one could use a old laptop running gnu/linux and jack or Pulseaudio Crossover Rack - multi-way crossover design & implementation with linux
 
Passive XO = a good compromise, can be misunderstood by most diy's, this is like alchemy, the limit is your imagination.

OPamp XO = your system is officially a two dimension world with a single cross over point and theoretical curves, nothing will fix timing, nothing will address peaks, bass boost will be required and you will hear a very 1d sound = un-involving at best.

DSP = you lose most of your precious bits of your precious high resolution digital file to be transform with mathematical formulas into something completely different from the original, sensitivity is cut by at least 3 to 6 db, noise floor is way higher, you never heard digital sound until you turned on DSP filters, like having a marathon with one leg runners.
 
DSP = you lose most of your precious bits of your precious high resolution digital file to be transform with mathematical formulas into something completely different from the original, sensitivity is cut by at least 3 to 6 db, noise floor is way higher, you never heard digital sound until you turned on DSP filters, like having a marathon with one leg runners.

As Wolfgang Pauli once said: "That is not only not right, it is not even wrong."
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.