KP-302/3002 High Frequency Speaker - Intermittent & Scratchy Sound - Resolved, Bad Cap' Connection

[Deang]

No one around here has a sense of humor anymore. 

[Edgar]

1 minute ago, Crankysoldermeister said:

No one around here has a sense of humor anymore. 

Sorry, I'm never sure any more whether someone is joking or serious.

[Chief bonehead]

Never take Deano seriously…🤓

[Deang]

Half the time. 50 Cent. 

[captainbeefheart]

18 hours ago, Edgar said:

Turns out I had a ready-made example sitting right on my hard drive. A while back I designed a 1200 Hz 3rd-order LPF for a Philips AD12240W8 that I had measured. In the graphic below, the circuit on the left side, feeding VM2, is the woofer model and the filter constructed of real capacitors and inductors that I had on-hand. The circuit on the right side, feeding VM1, is the textbook 3rd-order Butterworth filter using ideal components. The effects upon the frequency and phase response are obvious, and are not limited to the crossover frequency (green is actual, red is textbook).

 

The second graphic shows the same, with the woofer inductance compensated with a Zobel network. The improvement is significant, in my judgment.

 

Can we see an impedance plot for your speaker model? It just doesn't look right to me. I looked up the specs for the Philips woofer and the numbers don't match. Can you show how you calculated the moving mass capacitance, inductance of suspension compliance, suspension resistance, and air load capacitance?

[Edgar]

8 minutes ago, captainbeefheart said:

 

Can we see an impedance plot for your speaker model? It just doesn't look right to me. I looked up the specs for the Philips woofer and the numbers don't match. Can you show how you calculated the moving mass capacitance, inductance of suspension compliance, suspension resistance, and air load capacitance?

 

I used measured values, not specs. I modeled the measured impedance with discrete components. You have all of the information you need to analyze the circuit model in whatever way you want.

 

Regardless of the actual values, the assertion was proved: woofer voice coil inductance can have significant effect upon the crossover filter under certain circumstances.

[jimjimbo]

18 hours ago, Crankysoldermeister said:

No one around here has a sense of humor anymore. 

It's global warming.

[captainbeefheart]

8 minutes ago, Edgar said:

 

I used measured values, not specs. I modeled the measured impedance with discrete components. You have all of the information you need to analyze the circuit model in whatever way you want.

 

Regardless of the actual values, the assertion was proved: woofer voice coil inductance can have significant effect upon the crossover filter under certain circumstances.

 

How can you take your simulation seriously when your speaker model is incorrect? You can't measure the variables I asked for, they need to be calculated from the T/S parameters of the speaker. For example your model says the resonant frequency is at 43.8Hz which is wrong, it should be 29Hz.

[Edgar]

8 minutes ago, captainbeefheart said:

 

How can you take your simulation seriously when your speaker model is incorrect? You can't measure the variables I asked for, they need to be calculated from the T/S parameters of the speaker. For example your model says the resonant frequency is at 43.8Hz which is wrong, it should be 29Hz.

 

It's in an enclosure.

[captainbeefheart]

13 minutes ago, Edgar said:

 

It's in an enclosure.

 

It should be a huge red flag that your "model" doesn't match what you are measuring, which tell us that your model is obviously flawed.

 

I looked up the T/S parameters and made a more accurate model and low and behold the simulation shows the same resonance and frequency response you measured.

[Edgar]

2 minutes ago, captainbeefheart said:

 

It should be a huge red flag that your "model" doesn't match what you are measuring, which tell us that your model is obviously flawed.

 

What on earth are you talking about? I said that I developed the model according to my measurements ... measurements taken when mounted in an enclosure.

 

Enough, already. This no longer has any bearing on the issue of voice coil inductance, and certainly not on the issue of a broken capacitor lead.

[billybob]

Speaking to original intent, hope the OP is happy now.

[captainbeefheart]

20 minutes ago, Edgar said:

 

What on earth are you talking about? I said that I developed the model according to my measurements ... measurements taken when mounted in an enclosure.

 

Enough, already. This no longer has any bearing on the issue of voice coil inductance, and certainly not on the issue of a broken capacitor lead.

 

We were discussing 3 way speaker crossover points and since voice coil reactance increases with frequency it was only natural for you to find a scenario where the woofer crossover is more than an octave higher than the discussion where the reactance is higher.

 

Moving goal posts to fit your argument is what I am talking about.

[CWelsh]

[captainbeefheart]

Here is the woofer network, with and without the added voice coil inductance. The RLC equivalent model of the T/S parameters are left out because they won't make a difference for this simulation as we only want to see the difference the voice coil inductance makes electrically in circuit. As expected not a discernible difference.

 

[Dave A]

I guess modeling things can be the end goal for some. Roy does this modeling thing too but he also stops to LISTEN and verify the academic calculations with audio perception by well trained ear and arrives at great conclusions backed by physical examples. Which are often sold by the hundreds of thousands to happy campers.

 

So once again I want to know what you do with these copious amounts of data. If you use this data to make things with lets see the practical real world examples and not just another chart. I can understand being fascinated with the process of designing from the I want to know part of the mind but to me the more important part of the  equation is what can and is done with the data. Until you build it all you have is predictive data. In the right hands this is a really powerful tool which cuts down on the amount of physical builds to verify the predictive behavior.   Analytical data derived from a built sample leads to does this idea really work and by analytical I mean trained ears and electronics and real tangible put your hands on them goods.

 

 So what do you do with all this stuff anyway?

 

  As an aside here. Edgar builds a lot of stuff and has a long track record of design builds. I always look forward to his informative real world posts.

[Dave A]

15 hours ago, billybob said:

Speaking to original intent, hope the OP is happy now.

Yes indeed and far more then he bargained for.

[Deang]

See what happens when you have a cap problem? Trouble, big trouble. 

[billybob]

As long as it comes out in the end.

[captainbeefheart]

8 hours ago, Dave A said:

I guess modeling things can be the end goal for some. Roy does this modeling thing too but he also stops to LISTEN and verify the academic calculations with audio perception by well trained ear and arrives at great conclusions backed by physical examples. Which are often sold by the hundreds of thousands to happy campers.

 

So once again I want to know what you do with these copious amounts of data. If you use this data to make things with lets see the practical real world examples and not just another chart. I can understand being fascinated with the process of designing from the I want to know part of the mind but to me the more important part of the  equation is what can and is done with the data. Until you build it all you have is predictive data. In the right hands this is a really powerful tool which cuts down on the amount of physical builds to verify the predictive behavior.   Analytical data derived from a built sample leads to does this idea really work and by analytical I mean trained ears and electronics and real tangible put your hands on them goods.

 

 So what do you do with all this stuff anyway?

 

  As an aside here. Edgar builds a lot of stuff and has a long track record of design builds. I always look forward to his informative real world posts.

 

For my work I rely on simulations a lot before going into prototype phase. For me it starts with an idea and scribbles with pencils and paper, circuits and calculations. Next thing to do is simulation to check calculations and also fine tune the performance. The prototype phase helps us improve my simulation phase also as if something needs to be moved on the board I need to either input or change parasitic properties into the simulation.

 

Besides designing and building things to put food on the table I have been making radios, amplifiers, TV's, speakers, crossovers, and many other electronic devices for the past 50-60  years.

 

Of course the listening aspect of this hobby is an important one and specs often don't tell the complete story, they can tell you a lot but after making something it always comes down to the listening test, I think we can all agree on that. As for simulations, they are only as good as the models you are using. Use a wrong model you end up with the wrong results, so it's important to know what the equivalent circuit is in order to model accurately.

 

As for this specific scenario, I don't see the purpose of any listening test. We are all on the same page that the acoustic response is the end goal but I feel many have lost sight of how this discussion got started. My original response was to folks choosing to use a film capacitor in the woofer circuit. The original design uses a general purpose electrolytic which has a much higher loss angle and subsequent much higher ESR. This extra resistance in that specific part of the circuit flattens the resonance out and so when changing to a film capacitor one may have to add the extra resistance in themselves to keep the electrical properties flat and more so keep the acoustic properties of the speaker the same as intended. In cases such as yourself you actually preferred the difference in sound so you liked the change in electrical properties and subsequent acoustic properties.

 

The discussion evolved into someone questioning the accuracy of the simulation without the voice coil reactance. Good question, crunching numbers tells us the average voice coil inductance really isn't going to have much reactance at frequencies below 500Hz but people were still concerned so I put the voice coil inductance into the simulation and voila no discernible difference as expected. It is a very simple exercise, inductors and their properties are quite well known and simply calculated but people are visual learners and all of us in here are not electrical engineers so it makes sense to post the two simulations showing the change with and without the inductance. If you find anything wrong with the simulation I am happy to rectify it, if one thinks the T/S parameters will make a difference I can calculate them into an RLC network and add it to the simulation but it should have no bearing on the series inductance and it's effects on frequency response. The simulation is only there to work out the electrical properties of the crossover network, not to be the end game acoustic sound form the speaker as once the crossover is worked out then move onto the measuring acoustic properties of the speaker system as a whole and critique it.