Passive crossover design |
Post Reply | Page 123 5> |
Author | |
Modeller
Registered User Joined: 03 March 2016 Location: UK Status: Offline Points: 31 |
Post Options
Thanks(0)
Posted: 23 June 2017 at 3:54pm |
Thanks, that's very sound advice.
I'll do some testing when the samples arrive next week. Before I commit to an injection-moulded enclosure I'll be playing with some high density 3D printed prototypes. I've got a couple different designs planned, but I am very open to having one suggested. I'll even send you the prototype unit once it's been tested. A genuine offer for someone with a flair for small design challenges.
|
|
odc04r
Old Croc Joined: 12 July 2006 Location: Sarfampton Status: Offline Points: 5483 |
Post Options
Thanks(0)
|
It'll produce sound under 200Hz sure, no driver has a brickwall response. But depending on where its resonance is the excursion under that point for a given power level compared to above it will be large, = dead driver with much less power than its quoted rating. So that all depends on how you plan to drive it. Measure a units impedance over the audio bandwidth to figure out where resonance is. Very easy to do with an external soundcard and one resistor + REW.
As it is rated to perform to 20KHz why do you want to put another tweeter above it? Simple is your friend when designing. Surely if it is just reproducing machine noises anyway the bandwidth required is going to be quite limited. Nothing higher than a few KHz I would imagine? If you must have another driver I am sure there are plenty of small dome HiFi tweeters that would fit the bill. Again you can often go lower than their suggested bandwith, if you are easy with the drive signal. |
|
Modeller
Registered User Joined: 03 March 2016 Location: UK Status: Offline Points: 31 |
Post Options
Thanks(0)
|
Thanks.
The speaker was tested from 200Hz, but looking at the signal response it actually rolls in at about 150 (http://www.tymphany.com/peerless/driver-search-results/driver-detail/?id=1079). I was hoping to reinforce and extend that with enclosure design. With a sealed enclosure I would just replicate the tested results. My problem, which I'm all too aware of, is a very limited volume. The audio source is full range, recorded with professional equipment, so the only limitations are the output devices. My thoughts, crazy as they are, would be to use this unit to reproduce frequencies up to somewhere between 500-1000Hz and then roll in a similar but separate driver for up to 20kHz. Don't worry, I know this is insane overkill for what is a scale model, but I'm already pushing boundaries with design so why not audio and electrical performance too? Cheers
|
|
odc04r
Old Croc Joined: 12 July 2006 Location: Sarfampton Status: Offline Points: 5483 |
Post Options
Thanks(0)
|
You would simulate that enclosure as a straightforward reflex. Where a reflex stops and a TL begins is another subject of discussion, but if the port is not long enough to support a 1/4 length bass cut off standing wave then it is a port assisted design, or reflex.
The driver you have chosen has a quoted bandwidth of 200-20KHz, is this the only driver in the system? It will not produce any bass or low mid but will be alright to reproduce vocals and most instruments. Think of any portable radio and you are not far off. I cannot find any TS specs for the driver but because they are such tiny little things there is no design here to be had really. Stick them in anything and they will sound the same. Make a compact sealed box for them and keep it as simple as possible to aid the overall design. There is no point at all in porting them at all. You will need to filter sub 200hz content (if it exists in your sources) before you apply it to them else they will die very quickly at any kind of power. If you are playing only known prerecorded audio, then you can filter the tracks using a PC before applying them to the system to save you having to do it with physical components. |
|
Modeller
Registered User Joined: 03 March 2016 Location: UK Status: Offline Points: 31 |
Post Options
Thanks(0)
|
I'd really appreciate some feedback on the enclosure design please.
I've tried a very simplified transmission line arrangement, although I'm not sure about the parallel escape channel.
|
|
Modeller
Registered User Joined: 03 March 2016 Location: UK Status: Offline Points: 31 |
Post Options
Thanks(0)
|
It's been a while as I've been hunting down appropriate drivers. I think I've found something perfect for the application – the Peerless TC55FD00-08. It has a good frequency response, and if I use them in pairs and tune the response of each I can use one for lower frequencies and one for higher. That is a very loose plan anyway.
This got me thinking about enclosure designs, and I've attached my current favourite. I'm a little worried by phase alignment from the port, but I'm hoping your combined wisdom can help out in that regard, plus any design enhancements you suggest. The design is totally open, except that I would like to keep within these rough limits (51mm wide FIXED, 63mm long, 62mm tall). Internal baffling is also open – I will be injection-moulding these especially for the model from a high density polymer. The floor is yours. |
|
odc04r
Old Croc Joined: 12 July 2006 Location: Sarfampton Status: Offline Points: 5483 |
Post Options
Thanks(1)
|
You can't implement a can't implement a crossover with just resistors, they are not a frequency dependent component in terms of their resistance. What you can do with them is implement a level matching network such that the SPL of different drivers fed from the same amplifier is compensated such that they work well together. E.g. a mid driver that is more efficient than a sub but you only have one amplifier channel to run them both from and you want to match the levels - so the mid is going to need a resistive network in front of it to pad the output down.
Your speaker and amplifier combo will be fine. Edit: Sorry I just realised you said resistor network after the crossover itself. With the addition of a Pi network to any driver you can make the amplifier 'see' any load you wish. But power will be dissipated in the resistors. When a crossover is operating in its passband, the load the amplifier sees is just the driver as the crossover is in series with it and contributing minimal resistance. At the 3dB crossover point between drivers you will see a little bit of extra resistance from the network added to each driver, and then those in parallel. So for a 4 and 8 ohm driver ignoring the extra crossover resistance in series the amp will see a load of 8//4 ohms (approximately). Read this web page thoroughly and all the answers you need are in it - http://sound.whsites.net/lr-passive.htm Section 6 deals with attenuation of signal post-crossover whilst maintaining a desired impedance. If you are unsure of your designs, simulate them using LTSpice, TINA, or any other electronics simulation package before you build. Edited by odc04r - 14 May 2017 at 7:09pm |
|
Modeller
Registered User Joined: 03 March 2016 Location: UK Status: Offline Points: 31 |
Post Options
Thanks(0)
|
Thanks again for all the help. I'm narrowing down driver selection, but I'm stuck on basic principles.
Let's say I choose 3 drivers all rated at 8 Ohm. After the crossover (depending on its design) one feed will go straight to the sub, so this feed is still rated at 8 Ohm, and another feed will be split between two drivers, so these will be rated at 4 Ohm. Is this correct? How will a basic resistor-based crossover change this perception? Also, I think I have found a really good driver for the sub, but this is rated at 16 Watts, and I only have a 3 Watt amp. What will happen to either the amp or the speaker if I attempt to drive it?
|
|
odc04r
Old Croc Joined: 12 July 2006 Location: Sarfampton Status: Offline Points: 5483 |
Post Options
Thanks(0)
|
Ah yes, sorry was being dumb there. Don't think about crossovers and do other work at the same time. That would have ended up being an resonant LC series notch filter with regard to voice coil current.
I'd still go active. Basically if you have an amplifier in the mix already, you have what you need for active filtering in terms of power rails. Current draw is no concern. Even if it has to be single voltage rail with coupling capacitors involved. |
|
studio45
Old Croc Joined: 16 October 2007 Location: United Kingdom Status: Offline Points: 3863 |
Post Options
Thanks(0)
|
You would want a series inductor, not a cap. If you wanted to make a second-order filter, the cap would go to ground *after* the inductor. If you just use a cap to ground, the impedance will drop to near zero at high frequencies, which will make your amp unhappy and suck all the drive signal away from the tweeter. So you only have to deal with a few watts. Perhaps the passive is not such a terrible idea then. Possibly simpler than designing an active one, *if* your drivers are well-behaved. For a start, a second-order Linkwitz-Riley at 8 kHz for 8 ohm drivers requires 320uH and 1.24uF. Here is a very small surface-mount inductor rated for 500mA and you could use tiny 10v electrolytics for the caps. For 4 ohms, use 160uH and 2.5uF. To go down an octave, double all values. Do you have an impedance measuring rig? It will tell you how well-behaved your drivers are. If the impedance graph has big features on it that are not to do with the fundamental resonant frequency of the drivers when mounted in your box, or the impedance has risen more than 25% above nominal at your chosen crossover frequency, a passive network will probably not be a good solution. Getting a flat transfer function out of a simple passive filter requires a pretty flat impedance and frequency response from the terminating driver. To deal with impedance variations you need to start adding different types of filter to the passive, which will quickly double or triple its physical size. Whereas an active filter can be made vanishingly small using surface-mount tech, and wouldn't need the extra networks to provide a flat transfer function - the amplifier's voltage-drive characteristic takes care of that.
|
|
Studio45 - Repairs & Building Commotion Soundsystem -Mobile PA
|
|
odc04r
Old Croc Joined: 12 July 2006 Location: Sarfampton Status: Offline Points: 5483 |
Post Options
Thanks(0)
|
Really? An active processing board would only take mA of power for a few op amps. I find it hard to believe that cannot be spared unless you are on a super tight power budget? If low power was a concern you could probably get that down to microamps with the right op amps. Stereo amplifier for your sub/top split and off you go.
|
|
Modeller
Registered User Joined: 03 March 2016 Location: UK Status: Offline Points: 31 |
Post Options
Thanks(0)
|
In an ideal world an active crossover would be great, but I have run out of power sources already because of moving parts, lights, etc. In fact, I need more! I'm afraid it has to be passive.
I'll work on a selection of drivers over the next few days and post them here. Thanks
|
|
Post Reply | Page 123 5> |
Tweet |
Forum Jump | Forum Permissions You cannot post new topics in this forum You cannot reply to topics in this forum You cannot delete your posts in this forum You cannot edit your posts in this forum You cannot create polls in this forum You cannot vote in polls in this forum |