potential of the ksn1005?

[Low Rider]

Just to share some thoughts about xover for the motorola/csn ksn1005




I'm planning some PA monitor with this piezo and bg20 driver from visaton where i would cross them at 5k since bg20 altough tehnically a fullranger isn't very well behaved above this and cuts off at about 10k anyway.

So I'm collecting some input here regarding the crossover for the 1005 - could the above scheme work in practice?

Normally the 1005 is thought of as very harsh sounding what can be credited to that nasty 5dB peak at 5k due to cone resonance.

The above graph shows response w/ and w/o crossover simulated in BoxSim software (kudos to UweG btw.)

I'm gonna try this out anyway when I will be building the boxes (can't really object the $10 cost if it doesn't work out) but I would like to hear your thoughts about this (has anyone attempted something similar)

[freddi]

looks good - nice software. a stepup transformer can be used to improve sensitivity. IIRC KSN1016 had ~ 1/4 of 1% HD with 5029Hz sine and ~100-103dB output

can you make a nice "helper tweeter" crossover for coming in around 7-11Khz which removes the 5Khz bump?

I didn't try a 2nd order highpass - that might help protect things

[Low Rider]

I've been fidling around with LRC values in Boxsim and came up with this design
1. 30 Ohm paralel
2. series resonant circuit L=0.47mH C=3.3uF R=2Ohm connected PARALELL to the tweeter
3. 0,68 uF cap series to all this

It does not remove the bump completely but pushes it 15 dB down.
-3dB point is around 8500 Hz
around 10k is 0 dB

after 8k it rolls of 12 db/oct
and after 5k around 25 db/oct due to combined natural roll off

the sensitivity is 94,5 dB, you can reduce this 6dB by adding series 0.13uF cap to the tweeter - with bigger values you reduce it less -this doesn't affect overall response



All of this is assuming this simulation works correctly, than we will both be happy with our motorolas

[sreten]

Hi,

A thorough understanding of the wiki on piezo's will help with c/o design.

http://www.diyaudio.com/wiki/index.php?page=piezoXO

Crossing over capacitive loads is not intuitive, e.g. using capacitors as attenuators.

/sreten.

[Low Rider]

Originally posted by sreten
Hi,

A thorough understanding of the wiki on piezo's will help with c/o design.

http://www.diyaudio.com/wiki/index.php?page=piezoXO

Crossing over capacitive loads is not intuitive, e.g. using capacitors as attenuators.

/sreten.

Yes, thanks but I've read that allready. I have that motorola's pdf about piezos usage. But it doesn't say anything about using inductors in piezo x/o. I suppose a piezo output depends on rms voltage aplied to it, that is absolute value of complex and real parts of it or V*cos(phi).

Anyway, the resistor voltage divider im my circuit makes the piezo act as purely resistive load with very little reactance so I guess that from that on you can use any standard crossover designs on it. And 1005could use some.

[sreten]

Hmmm......

Just because you've read it doesn't mean you thoroughly understand it,

/sreten.

[Low Rider]

What? What's up with the offensive attitude?

I'm no loudspeaker acoustics specialist, but as a electronics engineer i think i have some insight into LRC networks which include crossovers.

Anyway, inductors are nothing new to the piezos, even some factory models have transformers in them akin to what freddi showed.

I'm looking here for what other people's expeirience with the 1005 or similar were like and whenever has someone tryed x/o like this and with what results.

[sreten]

Hi,

I did not intend to come across as offensive.

How to deal with the peak before roll-off is decribed in the Wiki,
and the expense of inductors / transformers is not really needed.

A "standard" c/o would need 8 ohm loading - and presumably someway
of adjusting driver level - neither of which the shown circuit offers.

/sreten.

[Cal Weldon]

A very interesting thing happens when you run a 2 mH inductor in series, yes series, with a piezo. Try hooking one up and you'll see what I mean. It's a little different than you'd expect.

[Low Rider]

Originally posted by sreten
Hi,

I did not intend to come across as offensive.

How to deal with the peak before roll-off is decribed in the Wiki,
and the expense of inductors / transformers is not really needed.

A "standard" c/o would need 8 ohm loading - and presumably someway
of adjusting driver level - neither of which the shown circuit offers.

/sreten.

I tryed running the sim with various x/o mentioned in the link, and while with corect cap values i could make the 5k bump smaller, it never quite did go away completely. But I'll trust you on this, and since i have no means to measure the actuall response with different x/o, so i must relly on simulation to get something usable, and get it cheap.

Looking at the various csn piezo's response curves i've come to conclusion that 1005 can be made the most flat above 5k.

Can you suggest me another piezo with decent response from 5k and if possible lower?. I'm not looking at a serious output here, about 50W max. so Powerline series would be an overkill for this applic.

Originally posted by Cal Weldon
A very interesting thing happens when you run a 2 mH inductor in series, yes series, with a piezo. Try hooking one up and you'll see what I mean. It's a little different than you'd expect.

Are you trying to get me to fry my amp?
C'mon thats not even funny.