Passive Time Delay Network

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I need some basic Help to get the answer to this question:

I want to create a time delay passive network and I cant fined any diagrams on how to build it or the formula on how to calculate the parts.

For example if I want to delay my tweeter 1.5 ms from the mid range and the tweeter is 4 ohms and crossed at 2500 hz. how that's done passively (diagram) and how do you calculate the values of the components (formula or calculator) and last can I daisy chain these to create a ladder, like for an array beam forming

Please help me Seriously I been looking the answer to these question for 3 weeks.
 
The problem for you is there is no "basic" way to implement a passive delay network. They are just not all that practical when it comes to loudspeaker crossovers, although they can and have be used. I advise you to reconsider using one.

Did you try to Google "passive delay line"? That would be a good place to start reading up on the subject...

Instead you might consider changing the crossover, using asymmetrical slopes, etc. This also has an effect on time and phase behavior and is a more common way to time align a system via a passive crossover network.
 
Last edited:
Here's how to make one:

Lattice phase equaliser - Expand Your Mind

As others have mentioned, passive delay is a bad idea these days. In the eighties, companies like thiele used passive delay in their speakers.

Nowadays, it's way way cheaper to do it with DSP. It's pretty easy to spend $200 on the components for a passive delay.

One thing you might check out is the crossover that JBL uses for their CBT array. It uses a novel topology to add a little delay to curve the array.

Also, you can use bill Waslo's xsim to simulate this. (It's free)
 
Each inductor has dcr of 0,35. Impulse response shows delay to be about 1,5 ms.
To time delay the tweeter, you'd do similar with high pass values. Happy playing!
 

Attachments

  • TD filter.jpg
    TD filter.jpg
    56 KB · Views: 906
  • FR delayed.jpg
    FR delayed.jpg
    184.5 KB · Views: 795
  • Impulse response delayed.jpg
    Impulse response delayed.jpg
    38.8 KB · Views: 692
  • Raw FR.jpg
    Raw FR.jpg
    152.7 KB · Views: 666
  • Raw driver Impulse response.jpg
    Raw driver Impulse response.jpg
    38.6 KB · Views: 794
Last edited:
The CBT crossover that JBL use is exactly what my goal is I have an straight array with 12 Dayton RS180-8 and 9 BG Neo 8 PDR per channel and I want to do a resistive virtual CBT by adjusting the timing on all the drivers. But it has to be done pasively because I only have 6 dsp outputs on my DEQX what I been looking for is the formula or a calculator I can use to calculate the components values to do just that with my drivers.
 
Delaying a tweeter by 1.5msec is quite difficult to do, and would require a very large number of components. That's about the same delay you'd see in 1.7 feet of sound wave travel in air! And at 20kHz, it would be equivalent to 30 cycles of a sine wave.

Think about where those 30 cycles would have to be within your proposed lumped network between the time they go in and the time they go out. If you took an oscilloscope and probed within that network, you could see a maximum of 180 degrees phase shift (one half cycle) at any frequency from a single series/shunt pair -- and that is with extreme signal loss at that frequency. You won't be delaying by even a cycle at any frequency by going through just a single component (RLC) network leg. So you are looking at more than 60 components (30 times 2 for each whole cycle) already, and you've lost nearly all the signal.... and still, you also need to do it without the huge attenuation, and you have to also tailor the network to delay lower frequencies by the (frequency/0.0015) cycles, so expect lots of additional components.

You can do a large delay like that (at least theoretically) over a narrow bandwidth by using a second or higher-order all-pass network (and very high quality inductors and capacitors), but a large, flat delay over say a 10:1 bandwith is not very practical.
 
Last edited:
Member
Joined 2010
Paid Member
Using a rough calculation,the Tannoy implementation of a lattice filter for band correction is set at around 10 kHz. The values of the components would need to be approx. 10 times larger (and carefully matched) for this delay and phase correcting network to work with a Tannoy DC speaker and second order xover, functioning at 1kHz. The lattice network is the same circuit used by Harry F. Olson for delay in his paper "Gradient Loudspeakers",see Loudspeakers Volume 1,JAES.
 
Member
Joined 2010
Paid Member
And I forgot to add for ericimage2 ,s benefit, that it is easier use the values that are the same as you would calculate for a second order Butterworth network.This does presume you are using such a crossover between your midrange driver and your tweeter and that you have taken steps to correct the impedances of the speakers to give a 4 ohm impedance at Xo.
 
The CBT crossover that JBL use is exactly what my goal is I have an straight array with 12 Dayton RS180-8 and 9 BG Neo 8 PDR per channel and I want to do a resistive virtual CBT by adjusting the timing on all the drivers. But it has to be done pasively because I only have 6 dsp outputs on my DEQX what I been looking for is the formula or a calculator I can use to calculate the components values to do just that with my drivers.

There's a thread on here about it, with schematics and sims.

One thing to note is that there's *barely* any delay on the JBL CBT. Enough to curve the array by about ten degrees iirc.

Also, you can buy the xover from JBL, just find the part# and order it over the phone.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.