2 Way speakers - Suggestions on improvements?

[Sultanen]

Hi,
I have a pair of old speakers that i really like the look of and plan to keep. What im looking for is some suggestions for improvements to kick them up a notch
[IMGDEAD]http://www.faktiskt.se/phpBB3/download/file.php?id=4174[/IMGDEAD]
Box (closed):

ITT 4015
Measurements: H 35,0 - B 21,0 - D 22,0 cm
Volume: 8,6 liter
Weight 4,3 kg

Tweeter

Peerless CT62
http://spicaspeakers.com/pdfs/peerless-801730.pdf

Woofer

Peerless 833429 WF165
fs= 4.57000000000000E+0001
Vas= 2.07000000000000E-0002
bl= 7.50000000000000E+0000
ss= 1.30000000000000E-0002
Qts= 3.50000000000000E-0001
Re= 6.10000000000000E+0000
Le= 8.00000000000000E-0004
LeLoss= 6.50000000000000E-0001
dmax= 4.00000000000000E-0003
pmax= 1.50000000000000E+0002

Crossover

Highpass filter 4000Hz - 5uF bipolar cap

Yesterday i replaced the foam rings on the woofers =)
[IMGDEAD]http://www.faktiskt.se/phpBB3/download/file.php?id=4257[/IMGDEAD]
[IMGDEAD]http://www.faktiskt.se/phpBB3/download/file.php?id=4326[/IMGDEAD]

Measurement from 35cm with Tweeter and Woofer
[IMGDEAD]http://www.faktiskt.se/phpBB3/download/file.php?id=4327[/IMGDEAD]

Measurement from 35cm only Tweeter
[IMGDEAD]http://www.faktiskt.se/phpBB3/download/file.php?id=4330[/IMGDEAD]

Measurement from 35cm only Woofer
[IMGDEAD]http://www.faktiskt.se/phpBB3/download/file.php?id=4329[/IMGDEAD]

I appreciate all input and thoughts on improvements!
I really like the speakers but i'm open to do modifications to them if needed to improve the sound.
I got a tip from a guy about converting them to Bass Reflex, thoughts on that?

 

[eriksquires]

Were these measurements taken bypassing the crossovers?

If the only crossover component is a 5uF cap then it looks like whoever built these did not take baffle step compensation into account.

Get your data into Xsim. With either speaker at the listening location:

1 - Place microphone 3' in front of tweeter.
2 - Measure each driver without the crossover. To protect the tweeter use either put a 20uF bi-polar or film cap on the tweet or make sure you don't have any bass in your test signal. Test at no more than 2.83V (1W into 8 Ohms)
3 - Measure both together, still without the crossover.

Import into XSim. Alter your woofer delay until your combined tweet + woofer FR matches your real life response. This will let you get the driver settings correct.

If those measurements were taken with the crossover, bass reflex or not is not your problem at all. Let's figure out the correct crossover first, then you may not even care.

Best,


Erik

 

[eriksquires]

After that's done, find out what your crossover parts are and put them into Xsim so you know your starting point. Compare the XSim predicted to your actual full speaker measurements (tweet, woof and crossover all working)

Once you reach this point, you can go really nuts fast.

Erik

 

[eriksquires]

Simulation will go much faster and accurate if you have impedance data. DATS V2 from Dayton works great. You measure the woofer in the cabinet. Tweeter can happen in or out.

 

[Sultanen]

Thanks alot for your guidance!
I will look into Xsim and probably have some questions when i get going
I don't have access to a 20uF bipolar cap atm. How would i limit the output to the tweeter to "2.83V (1W into 8 Ohms)"?

I really did not know where to go so i really appreciate the help getting a decent crossover fitted!

 

[eriksquires]

Usually whatever test tool you use has a way of outputting a reference signal, or signal generator. Disconnect the speaker and set the signal generator to 50 to 60 Hz, where V meters are most sensitive. Adjust volume until meter says 2.83 volts.

I use OmniMic, and the choice of test signals includes "Bass Excluded" but I think you could probably do the same thing with something free like Room EQ Wizard.

Best,


Erik

 

[Sultanen]

Could be worth mentioning that the drivers is not the original, they were replaced by my uncle maby 15-20 years ago.

 

[eriksquires]

I should point something out. For measuring driver distance, it's best to put the speaker out in the middle of the room. This will give you the least amount of interference. While doing so make sure you do not move the speaker or the microphone, not even 1 mm.

However, for measuring the FR for the crossover, putting the speaker in the preferred listening location is best. This will give you the most accurate idea of the baffle step effect.

Baffle step is why you have measured that rise from 100 Hz to 1kHz. If you want to experiment for fun, measure your speaker out in the middle of the room on a stool, then back against the wall. You'll see how this changes.

Baffle step compensation is to use crossover components to flatten the woofer's response. If you look at your woofer chart, look at the 100 Hz point. Draw a horizontal line there to 10kHz. That's what we want to get to, more or less from the woofer.

When that's figured out, you'll need to bring the overall tweeter level down.

We'll take it step by step. The good news is that you can do most of this in XSim for free. What's going to throw things off is the lack of impedance measurements. The tweeter is not very inductive and about 7 Ohms, which is a very good thing. It's very close to an ideal resistor, so it's much easier to simulate even without knowing the actual impedance chart.

Once we are close, you'll have choices to make about the overall level, BSC circuit, whether you want to add a low pass filter to the woofer, stuff like that.

Lastly, you should expect that you'll need to turn up the volume on your stereo compared to before, but when you do you'll have a lot more to listen to than you did before.

Best,


Erik


Erik

 

[eriksquires]

Could be worth mentioning that the drivers is not the original, they were replaced by my uncle maby 15-20 years ago.

And to think that the replacements have already rotted. These speakers are probably haunted. It's possible the original driver's had much higher Qts or Vas, so that they produced more even bass back then, or that the crossover that was there originally has been completely removed.

I think it's best to use this as a learning experience for right now. Everything that you need to learn about speaker design and analysis is in your hands.

When we're done you'll be shopping for woofers, tweeters and parts like a pro.


Best,


Erik

 

[Sultanen]

Okey, thats the max level sorted =)
Xsim uses FDR response files as input for the drivers, what tool generates them? I usally use Tombstone(Tolvan Data) to get my curves but there i can't export that file format, .

 

[eriksquires]

Almost forgot, it's worth at some point soon measuring both speakers to make sure they are reasonably similar after all the history they've had. And figure out a way to seal off those missing screw holes. Hot glue guns and little blocks of wood behind them can help brace stripped faces.

I really like to use round, hex head screws for speakers. I think they look a lot cleaner. Like these:

 

[Sultanen]

Thanks alot for all the great info! If i don't respond to some detail i can assure you that i read everything like 10 times trying not to miss something =)

I should point something out. For measuring driver distance, it's best to put the speaker out in the middle of the room. This will give you the least amount of interference. While doing so make sure you do not move the speaker or the microphone, not even 1 mm.

However, for measuring the FR for the crossover, putting the speaker in the preferred listening location is best. This will give you the most accurate idea of the baffle step effect.

Baffle step is why you have measured that rise from 100 Hz to 1kHz. If you want to experiment for fun, measure your speaker out in the middle of the room on a stool, then back against the wall. You'll see how this changes.

Baffle step compensation is to use crossover components to flatten the woofer's response. If you look at your woofer chart, look at the 100 Hz point. Draw a horizontal line there to 10kHz. That's what we want to get to, more or less from the woofer.

When that's figured out, you'll need to bring the overall tweeter level down.

We'll take it step by step. The good news is that you can do most of this in XSim for free. What's going to throw things off is the lack of impedance measurements. The tweeter is not very inductive and about 7 Ohms, which is a very good thing. It's very close to an ideal resistor, so it's much easier to simulate even without knowing the actual impedance chart.

Once we are close, you'll have choices to make about the overall level, BSC circuit, whether you want to add a low pass filter to the woofer, stuff like that.

Lastly, you should expect that you'll need to turn up the volume on your stereo compared to before, but when you do you'll have a lot more to listen to than you did before.

Best,


Erik


Erik

I will use that measurement method, tomorrow i will get my tripod rigged so i can minimize the movement!
What a eureka moment with the explination off baffle step compensation! Sounds great!! Lets turn up the volume!

Step by step sounds awesome I guess that we need the impedance measurements at least on the woofer then? Maby we could DIY something with an Arduino to get something decent?

Im starting to like this hobby more and more, new thinks i never heard off pops upp all the time

And to think that the replacements have already rotted. These speakers are probably haunted. It's possible the original driver's had much higher Qts or Vas, so that they produced more even bass back then, or that the crossover that was there originally has been completely removed.

I think it's best to use this as a learning experience for right now. Everything that you need to learn about speaker design and analysis is in your hands.

When we're done you'll be shopping for woofers, tweeters and parts like a pro.


Best,


Erik

As you say the original drivers were probably way different from these, and nothing of the original crossover is left. The bipolar 5uF cap was fitted by me today(made from two 10uF "polarized" electrolytic caps mounted - to -), it was just a quick fix and i figured it was better than the two different(one on each tweeter) 5uF resp 4,7uF "polarized" electrolytic caps that has been in the speekers for the last 15-20 years...

It will be a great learning experience! Ill trust in your guidance!

Almost forgot, it's worth at some point soon measuring both speakers to make sure they are reasonably similar after all the history they've had. And figure out a way to seal off those missing screw holes. Hot glue guns and little blocks of wood behind them can help brace stripped faces.

I really like to use round, hex head screws for speakers. I think they look a lot cleaner. Like these:

[IMGDEAD]https://www.parts-express.com/data/default/images/catalog/240/081-312_HR_0.jpg[/IMGDEAD]

With speakers you mean the entire speaker? or part for part? Ill measure all that we need.

Hotglue a wooden in the back is a great idea! Maby fitting a similar black hex bult and a M6/M4 countersunk nut in the wood?

When speaking about sealing, is it overkill to have some kind of foam ring glued to the speaker to seal the "gap" between the wood and the speaker?

 

[Bob Richards]

I don't recommend going bass reflex. It's hard to get right, and has poor damping of the driver cone at the low bass end (below the tuned freq.). With a closed box woofer, you can add active EQ ahead of the power amp to make it acoustically flat to 25HZ.

The big rolloff from about 1kHZ on down is probably mostly "baffle step" issue. Published graphs are usually done with the driver mounted on a large wall facing into an anechoic chamber, so there's no baffle step, or room acoustics issues. Low bass can be pumped up with tone controls.

The biggest criticism might be the off axis response will have a big jump at 4-5kHZ, where the tweeter takes over. If it was mine to mess with, I'd add a 3 inch midrange driver in a totally non-resonant sub enclosure, and I'd let it do 500HZ to 7kHZ. The Peerless/Tympany TG9FD1008 from Madisound is my present favorite. But getting a passive crossover to be accurate is actually a big job. You have to measure driver FR and efficiency and impedance at the frequencies you want to crossover at, before you can design the circuit. You'd want to verify the result with a cal'd mic and pink noise (or equiv.) too. Personally, most of my crossovers are 24dB/octave active ahead of the power amps. They're much more accurate, better sounding, very predictable, and I'm not convinced that that route takes more time than getting a 3 way passive to be accurate. But then, I'm a speaker builder addict. I've been building them since 1968ish.

 

[Sultanen]

Were these measurements taken bypassing the crossovers?

Forgot to answer this. The woofer measurement is without any crossover, the tweeter measurement is with the 5uF bipolar cap so its probably not the full curve there.

Btw, sorry for the slow responses, the reviewing of new members comments takes while sometimes.

 

[Sultanen]

I don't recommend going bass reflex. It's hard to get right, and has poor damping of the driver cone at the low bass end (below the tuned freq.). With a closed box woofer, you can add active EQ ahead of the power amp to make it acoustically flat to 25HZ.

The big rolloff from about 1kHZ on down is probably mostly "baffle step" issue. Published graphs are usually done with the driver mounted on a large wall facing into an anechoic chamber, so there's no baffle step, or room acoustics issues. Low bass can be pumped up with tone controls.

The biggest criticism might be the off axis response will have a big jump at 4-5kHZ, where the tweeter takes over. If it was mine to mess with, I'd add a 3 inch midrange driver in a totally non-resonant sub enclosure, and I'd let it do 500HZ to 7kHZ. The Peerless/Tympany TG9FD1008 from Madisound is my present favorite. But getting a passive crossover to be accurate is actually a big job. You have to measure driver FR and efficiency and impedance at the frequencies you want to crossover at, before you can design the circuit. You'd want to verify the result with a cal'd mic and pink noise (or equiv.) too. Personally, most of my crossovers are 24dB/octave active ahead of the power amps. They're much more accurate, better sounding, very predictable, and I'm not convinced that that route takes more time than getting a 3 way passive to be accurate. But then, I'm a speaker builder addict. I've been building them since 1968ish.

Thanks for the input!
If i can keep the closed box and not modify it im happy, from what i understand, and you also confirm it, the closed box is a lot less complex than bass reflex.

About adding a midrange driver, i can't really say anything about that, i understand way to little about these complex questions at the moment With "totally non-resonant sub enclosure" you mean a separate box outside of my existing sealed box? But from what i understand you prefere not to have the crossover in the speaker but rather before the amp?

Might be worth mentioning that i have a surround amp atm, Yamaha RX-V459. I know its not the best but it is what it is at the moment =)

 

[eriksquires]

Thanks alot for all the great info! If i don't respond to some detail i can assure you that i read everything like 10 times trying not to miss something =)

I will use that measurement method, tomorrow i will get my tripod rigged so i can minimize the movement!
What a eureka moment with the explination off baffle step compensation! Sounds great!! Lets turn up the volume!

Step by step sounds awesome I guess that we need the impedance measurements at least on the woofer then? Maby we could DIY something with an Arduino to get something decent?

As you say the original drivers were probably way different from these, and nothing of the original crossover is left. The bipolar 5uF cap was fitted by me today(made from two 10uF "polarized" electrolytic caps mounted - to -), it was just a quick fix and i figured it was better than the two different(one on each tweeter) 5uF resp 4,7uF "polarized" electrolytic caps that has been in the speekers for the last 15-20 years...

It will be a great learning experience! Ill trust in your guidance!

With speakers you mean the entire speaker? or part for part? Ill measure all that we need.

Um, when I meant you should measure both speakers, I meant the entire speaker to find issues exactly like that mis-matched capacitor.

I've had a pair of Focal's where one woofer never worked due to a crossover part breaking. I spent a tremendous amount of time making it better assuming both speakers were identical, and they weren't.

Hotglue a wooden in the back is a great idea! Maby fitting a similar black hex bult and a M6/M4 countersunk nut in the wood?

Sounds like too much effort. Just make sure the wood you hot glue is a nice hard wood and not particle board. That will be plenty resilient for your needs.

When speaking about sealing, is it overkill to have some kind of foam ring glued to the speaker to seal the "gap" between the wood and the speaker?

Depends on the condition of everything. If in doubt, yes, you need foam strips to seal the drivers. When thinking about sealing the cabinet you also need to think about making sure all the screws are in place, all the drivers are in and any plates in the back air-tight. You don't need to act as if you are an Ebola worker, but open screw holes and warped wood need to be dealt with. You also need to consider in your process you'll probably take each driver out 3-4 times until you are done, so you want to make sure things are sturdy, air tight, and won't come apart when working on them.

So, make sure you hand tighten or use a low clutch setting on any power tools you use to help you.

With the wooden blocks you use to support your driver, use big pilot holes! Make sure you can screw into the hardwood without causing blisters, but not such big pilot holes that it's easy to strip the screws.

One thing to start looking for is a local supplier of speaker parts. A local store with parts is great, but otherwise it's time to at least start looking online for a Swedish or at least EU store you can turn to. I only know AudioHobby.EU because their prices on expensive Danish parts are so good, but they are a bit high end. You need a less expensive hobby supply store. In the US this is usually Parts-Express for us.

Brands you can use to search:

Audyn, Bennic, Mills, Mundorf, Jantzen, Goertz, Solen, Vifa, Peerless, Tymphany.

Find a couple of stores with good prices. If you can't find any of those, find a good general electronics store and we'll make do.

Best,


Erik

 

[eriksquires]

Woodworking tip: The harder the wood is, the larger a pilot hole you will need, so it pays to experiment on a scrap.

The smallest diameter pilot hole is the size of the inner shaft, without the threads. The maximum is of course the size of the screw with the threads. So somewhere between these two diameters is your ideal pilot hole size.

Best,

Erik

 

[eriksquires]

I used an online calculator for that speaker using specs I found here. It seems the ideal cabinet for that driver is around 0.22cuFt or 6.23 liters. With an 8 liter box you're going to end up with reduced Q (less bass) but, since you like to put the speaker up against the wall at the end of the day it may be a perfect fit.

Best,

Erik

 

[Sultanen]

Um, when I meant you should measure both speakers, I meant the entire speaker to find issues exactly like that mis-matched capacitor.

I've had a pair of Focal's where one woofer never worked due to a crossover part breaking. I spent a tremendous amount of time making it better assuming both speakers were identical, and they weren't.

Okey, i will do a full swipe like the first curv on each speaker when i have the tripod this afternoon.

When i started with this i actually noticed that the left tweeter + cable was not attached.. The end result of all of this can only be an improvement

Woodworking tip: The harder the wood is, the larger a pilot hole you will need, so it pays to experiment on a scrap.

The smallest diameter pilot hole is the size of the inner shaft, without the threads. The maximum is of course the size of the screw with the threads. So somewhere between these two diameters is your ideal pilot hole size.

Best,

Erik

Sounds like too much effort. Just make sure the wood you hot glue is a nice hard wood and not particle board. That will be plenty resilient for your needs.

With the wooden blocks you use to support your driver, use big pilot holes! Make sure you can screw into the hardwood without causing blisters, but not such big pilot holes that it's easy to strip the screws.

Haha, i like nuts and bolts I will find some wooden blocks and experiment a bit.

Depends on the condition of everything. If in doubt, yes, you need foam strips to seal the drivers. When thinking about sealing the cabinet you also need to think about making sure all the screws are in place, all the drivers are in and any plates in the back air-tight. You don't need to act as if you are an Ebola worker, but open screw holes and warped wood need to be dealt with. You also need to consider in your process you'll probably take each driver out 3-4 times until you are done, so you want to make sure things are sturdy, air tight, and won't come apart when working on them.

So, make sure you hand tighten or use a low clutch setting on any power tools you use to help you.

Okey, i think ill need something like that, the woofer is probably sealed quite well but i don't trust the seal around the tweeter.

I will do all mesurments with all screws in place.

One thing to start looking for is a local supplier of speaker parts. A local store with parts is great, but otherwise it's time to at least start looking online for a Swedish or at least EU store you can turn to. I only know AudioHobby.EU because their prices on expensive Danish parts are so good, but they are a bit high end. You need a less expensive hobby supply store. In the US this is usually Parts-Express for us.

Brands you can use to search:

Audyn, Bennic, Mills, Mundorf, Jantzen, Goertz, Solen, Vifa, Peerless, Tymphany.

Find a couple of stores with good prices. If you can't find any of those, find a good general electronics store and we'll make do.

As far as i know there are no local shops in my town, i will search for some physical store in some town close by, but usally i buy electronic components for other projects online.

AudioHobby.EU looks like a great store, with great prizes.
The price on Jantzen Audio Cross Cap is about 50% of a Swedish webshop i know:
Cross Cap - HIFI KIT
Jantzen Audio Cross Cap Capacitors - Fidelity Components Shop

I used an online calculator for that speaker using specs I found here. It seems the ideal cabinet for that driver is around 0.22cuFt or 6.23 liters. With an 8 liter box you're going to end up with reduced Q (less bass) but, since you like to put the speaker up against the wall at the end of the day it may be a perfect fit.

Best,

Erik

Sounds great! A thought on that; the boxes are filled with glass wool, as far as i have understood it the insulation will make the box act as it was bigger so maby i should remove it?

 

[eriksquires]

Hehe, see? I knew something would be off.

I'm glad AudioHobby has good prices for you too! I found them by accident. Next time I get into an expensive build they are getting all my money.

I live in San Francisco, one of the highest tech cities in the world and there's nowhere for me to get most of what I wan to buy tech-wise.

You are right about stuffing adding volume but leave your box alone until you've measured it in place and you have simulated the end results. Measure first, simulate, simulate simulate, then modify after.

I thought of something. Technically, if you are reasonably sure of your cap's value you can do the interferometry measurements with the cap in place, just have to add the cap to your simulation. However, if these are 15 year old caps, the likelihood of that being true is too small. Those are very robust tweeters though, so you should be OK without a cap.

Best,


Erik