Need to change passive xover frequency

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I have a high pass 2nd order @ 125hz, it has (series)100uf with 5mh to ground the impedance is unknown as it is a electrostatic.
I need to shift that 125hz frequency to 200hz, can anyone work out for me the simplest way?.
As the 100uf is made of multiple poly-props, could I just say take off about 40uf and make it 60uf instead of 100uf, would that shift the xover up from 125hz to around 200hz?

Cheers George
 
Thanks rabbitz,I take it from that you worked out what impedance (ESL) the crossover is seeing? Can you let me know?
Sorry I should have posted the impedance graph before

Cheers George
 

Attachments

  • monolith impeadence graph.gif
    monolith impeadence graph.gif
    89.3 KB · Views: 167
Hi George

In my original calcs I assumed impedance was equal which is close to true on a normal woofer.

Two formulae are used for 2nd order:
C=V/(Rf)
L=R/(Vf)
C cap value, V is a variable depending on type such as butterworth and LR but these cancel out in the calc so are ignored, R is impedance, f is frequency. Since V is the same, then for example for a cap, V=CRf then CRf for 200Hz = CRf for 125Hz which gives C(200Hz)=CRf(125Hz)/Rf(200Hz) which becomes:
C=(100x8x125)/(4x200) = 125uF

That impedance graph makes a huge difference and at 200Hz would be 125uF (yes goes up) and 1.56mH. As the impedance drops, cap value increases and inductor value reduces for a set frequency and 125Hz and 200Hz are not very far apart.

phew... :faint:
 
Martin Logan Monolith III with IIIp (passive xover)

Thaks rabbitz, sorry for the headache, that's what scared me when it saw the impedence graph, I asume that the graph is of the whole speaker as a complete product with the standard passive xover and bass driver insitu, and not just of the panel itself, I have no way of finding out.
But also the fact that now the bass part of the passive xover is gone becuse it's now active, would that change the graph much at the 200hz region of the els part of the graph.

The reason I'm going this way is that i tried to active xover the esl with the 24db LR active xover and did not like what even very good opamps did to the sound (sterilized it) and the passive correction components were left in place, so the balance was the same, just the 100uf and 5mh were removed. the active sounds two dimentional and more (mechanical/sterile) than the passive.

I need to shift that xover up to 200hz because the the active bass can go up that high (since a driver change) and if esl have a weakness its in the lower to mid bass in that they need more slam in that region, this is why the shift from 125hz to 200hz.

I think I should somehow measure the impedence @ 200hz of my esl's with just the passive corection networks in place and not the 100uf and 5mh but I've never done this, need to learn.

Cheers George
 
Managed to plot the impedence from 50hz to 1100hz, the red trace is planel only with the standard ML high pass xover intact, the blue trace is the way I want to use it with the ML high pass xover minus the (C2A & L3)100uf and 5mh but with everything else in circuit even the LCR is in circuit.

I need to cross over at 200hz with different values of the 100uf and 5mh, the blue trace is the one I believe is the one to calculate the new values of C2A and L3,
To me the impedence at 200hz (blue trace) looks as though it maybe hard to have an effective 2nd order at the impedence of 14ohms with it climbing and falling so steep after the 200hz point, what do you think?


Cheers George
 

Attachments

  • ml impedence.jpg
    ml impedence.jpg
    79.4 KB · Views: 109
An interesting challenge. The downturn after 400Hz appears offputting.

I wonder whether it wouldn't be worth also plotting the impedance phase and then putting this into a crossover simulator to do visually.

You'd want, then, to have an acoustic response plot too. You could sim the response from this both with and without the original L and C, thereby getting a visual on the slope of the original crossover, and hence something to mimic as you go up in frequency.

Even if you don't want this slope, it may assist you in visualising what it was you had been listening to in the past, and hence make a more informed decision.
 
georgehifi said:
Thanks lndm, then again Martin Logan used a 2nd order at the 125hz (C2A & L3) and it's just as steep on both sides, so maybe it would work fine at 200hz also,


I don't believe I suggested it wouldn't work, if I gave you that impression. Using a crossover sim, I have been able to do things with crossovers that previously I would have considered impossible without a complex crossover.

A textbook crossover for 14 ohms and 200Hz could be 33uF and 20mH giving a theoretical bessel slope. In your case, though, this would not work. The effect of the impedance on either side of the centre point shouldn't be ignored as the crossover has a heavy influence within a range of frequencies.

Generally speaking though, you could start with the spirit of Rabbitz' first suggestion of reducing the values so the resonance frequency is in proportion by the same as the proportion of 125Hz vs 200Hz.

Then, to compensate for the increased impedance you could reduce the value of capacitance, and increase the value of inductance by the same amount as each other. This way, the frequency is the same, but the Q is adjusted to suit the new impedance.

Going this way, and taking a stab in the dark, and wanting to make life easy, you could start by keeping your 5mH inductor, and changing your capacitor to 39uF. Go from there :)

BTW, it is interesting that 100uF and 5mH resonates at 225Hz.
 
lndm said:



BTW, it is interesting that 100uF and 5mH resonates at 225Hz.

This is throwing me, as the specs clearly show that with these components it's 2nd order at 125hz, I'm scratching my head, also being new at this xover voodoo.

Mine is the IIIp not the IIIx

I wonder if the L3 @ 5mh could be a typo on the the circuit I posted and it's really 15mh?


Cheers George
 

Attachments

  • ml specs 1.jpg
    ml specs 1.jpg
    79.3 KB · Views: 58
dunno but since you use an active sub system, isn't it more easy to create a passive line level filter(as used in very high end pre's):
http://www.t-linespeakers.org/tech/filters/passiveHLxo.html
http://www.marchandelec.com/xm46.html
I did, and it sounds very good to my opinion...

I use two subs with 17cm(5 inch?) woofers combined with quad esl-57's. Pre's are also passive with ladder attenuators...

Just use high qualiy caps and resitors, but the end result is only a few components in the signalpath, as any signal benefits from simplicity....
And you don't need any BIG and expensive components between the amp and the speaker. So you can fully benefit from the dampeningfactor from your amp.
 
v-bro said:
dunno but since you use an active sub system, isn't it more easy to create a passive line level filter(as used in very high end pre's):
http://www.t-linespeakers.org/tech/filters/passiveHLxo.html
http://www.marchandelec.com/xm46.html
I did, and it sounds very good to my opinion...

I use two subs with 17cm(5 inch?) woofers combined with quad esl-57's. Pre's are also passive with ladder attenuators...

Just use high qualiy caps and resitors, but the end result is only a few components in the signalpath, as any signal benefits from simplicity....
And you don't need any BIG and expensive components between the amp and the speaker. So you can fully benefit from the dampeningfactor from your amp.

This cross my mind a few times, it would be so much easier to have a passive 12db filter at the input of the power amp.
But correct me if I'm wrong, when you series two 1st orders to make a second order together of identical components, don't you get a hump, or is it a smooth 12db per octave roll of at the frequency that they are calculated for, ( 200hz) in my case.
The input impedence of my amp is a steady 68k

Cheers George
 
to make a second order filter this way(I use this for the sub in my car) in the second series(as is stated in the theory in the linkhttp://www.t-linespeakers.org/tech/filters/passiveHLxo.html)
C2 should be one tenth the value of C1 and R2 ten times the value of R1.
No "hump" should appear...
At least I don't experience one, the only down side to a filter like this is slight insertion loss, but that is easy solveable with more power or more efficient speakers....
Or you might allready have plenty...

Max. cone displacement can be severall meters only once...
(joke).
 
To overcome amplifier impedance effects you can use a pot for R2, rest of the values I would buy cheap first to experiment with in different values.

Make it sound exactly the way you want it and buy the values in a good quality(polypropene caps will mostly do just fine...).

Can be a lot of fun and revealing to make such a filter....
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.