By-passing speaker's passive crossover for DSP crossover - need to insert capacitor/s?

[sergeauckland]

When you have the amplifier and filter 'fixed' onto the rear of the speaker it won't be needed.
When there is the slightest chance you could make a mistake with wiring or filter settings I would use a series capacitor calculated to about 500Hz crossover with the tweeter. It will protect the tweeter from disasters in such case.

I left the capacitors in series with the midrange and tweeters not so much for crossover errors, although indeed that's of value, but in case the amplifier(s) become faulty and put 50+V DC across the driver. My amps are AC coupled to the 'speakers through the floating ground power supply, but with 13000uF on each rail even though eventually the DC will decay especially once the fuses blow, the pulse will most likely destroy a tweeter, won't do a lot of good to a mid-range unit although the bass driver should take it long enough.

S

 

[KaiserSoze]

Another silly question: the passive crossover had an old circuit board with the various components.

Now removing the passive crossover I will just have wire between drivers and binding posts.

If I'm inserting this protective capacitor in-line with tweeter, is it ok to just have it in between wire, just hanging this inside the speaker?

No risk of catching fire with the wool or wooden speaker cabinet?

Or is there a better way than just letting the capacitor hang there in between 2 wires inside the cabinet?

What is good practise for inserting this capacitor.

Thanks again

Just a couple comments.

The capacitance value doesn't have to be exact, because you're choosing a ballpark value (500 Hz) for the -3 dB of the high-pass filter (which is what you're doing, i.e., you're adding a high-pass filter, with a corner frequency well below the frequency of the eventual DSP-based filter).

The voltage rating of the capacitor is so that the capacitor can hold up under the voltage seen across the capacitor. It does not affect the filter behavior. 100 V is a common value for crossover capacitors. At very low frequency most all of the signal voltage will appear across the capacitor, and almost none of it will appear across the tweeter. (The sum of these two voltages will equal the signal voltage.) Since power is equal to V^2/R, V is equal to (PxR)^.5. The question to ask is what the voltage would be if power were 100 Watts and R were 10 Ohms (big to err on the safe side). This is about 30 V, which means that 60 V is probably good enough, however you want to be absolutely certain that the capacitor will withstand whatever voltage appears across it, so you use 100 V even though it is possibly overkill.

You can physically locate the capacitor at the crossover or at the tweeter or anywhere in between. However you want to gather up that bundle of wires and use some sort of wire tie or clip to run the bundle along the walls of the enclosure as much as possible. Since the purpose of this is to protect the tweeter, I would be inclined to put it physically at the tweeter. If the tweeter has a tab that you want to use with a mating connector that pushes on, you can solder the mating connector to the capacitor, using a short piece of wire if that seems appropriate. Or if you are going to solder a wire directly to the tweeter's tab anyway, you might want to just solder the capacitor directly to the tweeter tab. No matter which way you do it, you may want to use some heat shrink tubing to cover the leads on the capacitor. You do not want to leave the leads exposed if there is any possibility of any kind of short occurring with unintended contact. If you put the capacitor directly on the tweeter, you might be able to use a nylon wire tie to tie it to the tweeter.

 

[Music1969]

Just a couple comments.

The capacitance value doesn't have to be exact, because you're choosing a ballpark value (500 Hz) for the -3 dB of the high-pass filter (which is what you're doing, i.e., you're adding a high-pass filter, with a corner frequency well below the frequency of the eventual DSP-based filter).

The voltage rating of the capacitor is so that the capacitor can hold up under the voltage seen across the capacitor. It does not affect the filter behavior. 100 V is a common value for crossover capacitors. At very low frequency most all of the signal voltage will appear across the capacitor, and almost none of it will appear across the tweeter. (The sum of these two voltages will equal the signal voltage.) Since power is equal to V^2/R, V is equal to (PxR)^.5. The question to ask is what the voltage would be if power were 100 Watts and R were 10 Ohms (big to err on the safe side). This is about 30 V, which means that 60 V is probably good enough, however you want to be absolutely certain that the capacitor will withstand whatever voltage appears across it, so you use 100 V even though it is possibly overkill.

You can physically locate the capacitor at the crossover or at the tweeter or anywhere in between. However you want to gather up that bundle of wires and use some sort of wire tie or clip to run the bundle along the walls of the enclosure as much as possible. Since the purpose of this is to protect the tweeter, I would be inclined to put it physically at the tweeter. If the tweeter has a tab that you want to use with a mating connector that pushes on, you can solder the mating connector to the capacitor, using a short piece of wire if that seems appropriate. Or if you are going to solder a wire directly to the tweeter's tab anyway, you might want to just solder the capacitor directly to the tweeter tab. No matter which way you do it, you may want to use some heat shrink tubing to cover the leads on the capacitor. You do not want to leave the leads exposed if there is any possibility of any kind of short occurring with unintended contact. If you put the capacitor directly on the tweeter, you might be able to use a nylon wire tie to tie it to the tweeter.

Thanks heaps! Really appreciate everyone’s info

I’m now armed with enough knowledge to be (less?) dangerous

 

[Music1969]

No matter which way you do it, you may want to use some heat shrink tubing to cover the leads on the capacitor. You do not want to leave the leads exposed if there is any possibility of any kind of short occurring with unintended contact.

Is there a recommended method to apply heat shrink tubing to the capacitor leads?

Is there a chance of heat damaging the capacitor itself when applying heat?

I saw this thread but there are multiple opinions on how to do it and it left me confused:

https://audiokarma.org/forums/index...nk-tubing-for-electrolytic-capacitors.532021/

And kinds of tips or pointers would be greatly appreciated

 

[KaiserSoze]

Is there a recommended method to apply heat shrink tubing to the capacitor leads?

Is there a chance of heat damaging the capacitor itself when applying heat?

I saw this thread but there are multiple opinions on how to do it and it left me confused:

https://audiokarma.org/forums/index...nk-tubing-for-electrolytic-capacitors.532021/

And kinds of tips or pointers would be greatly appreciated

There are several good tips at that link, also an ad for liquid electrical tape, from Permatex, which I've never used, but which is probably good stuff. Don't try to use a regular soldering iron to apply heat shrink. The only way you'll get adequate temperature is by holding the iron so close that you might make contact and make a mess of it. A good heat gun is best. A torch type soldering iron might work but I've never tried that way of shrinking wrap. I have also used regular butane lighters which work fine but you have to watch what you're doing and keep the flame moving. I don't think a hair dryer will get hot enough. As long as you watch what you're doing there's little reason to be concerned about damaging the capacitor while applying heat shrink. I've never bothered with trying to wrap the whole capacitor because as long as you get the wrap right up against the body of the capacitor there's little opportunity for any unintentional contact. And you don't need to shrink the wrap all the way up to the body of the capacitor.

I've never had a problem with damaging any component while soldering, but it could happen if you were using a big iron with low temperature. When soldering you want the iron to be hot enough to melt the solder quickly. There's several good videos on YouTube that give good instructions for soldering. If you anticipate doing a lot of soldering in the future you might want to get one of those good irons with the digital temperature control, which keep the iron at the optimal temperature. In fact I need to get one of them myself. I like to use the fine silver rosin-core solder. Next thing you know you'll be building your own amps.

 

[restorer-john]

You can put 20Hz on a tweeter as long as the power remains (well) below the continuous rating of the tweeter.

Not a good idea, nor good advice. Tweeters continuous system ratings are only ever specified for the frequency range they are rated for. A 10W system rated tweeter with tear itself apart with even a few watts at 20Hz with no crossover. Very rapidly indeed.

 

[solderdude]

Do you think 0.4mW is really going to damage a tweeter ?
Arguably 20Hz is not going to give any output at all so might as well start from 200Hz .

 

[Music1969]

Arguably 20Hz is not going to give any output at all so might as well start from 200Hz .

Once I get the frequency response of both tweeter and woofer by measuring, then is it simple as knowing the overlap of flat FR region of woofer and tweeter, and then just picking the mid-point and choosing that as the DSP crossover?

Then running the full sweep with DSP crossover in action and hoping for the best !?

 

[Music1969]

I will share results of this experiment here on ASR.

 

[solderdude]

Then running the full sweep with DSP crossover in action and hoping for the best !?

You really have to be careful sweeping the whole speaker system as that requires much much much more power than the suggested very near-by tweeter sweep.

An idea is to use pink-noise and 'de-pink' the results. Sweeping tweeters at higher SPL is not really recommended as one easily reaches max. power ratings of a tweeter when the sweep is taking too long.
You can set the sweep length short in REW as well to get a feel of the frequency response.

 

[KaiserSoze]

Once I get the frequency response of both tweeter and woofer by measuring, then is it simple as knowing the overlap of flat FR region of woofer and tweeter, and then just picking the mid-point and choosing that as the DSP crossover?

Then running the full sweep with DSP crossover in action and hoping for the best !?

You can make it that simple but you want the crossover point to be high enough to avoid tweeter distortion and low enough so that the woofer will still have good dispersion. Subject to these considerations the crossover point could end up at either end of that overlap region. Also, unless you plan to use a notch filter at the tweeter's resonance (or use the DSP to deal with that), and depending on how steep the high-pass filter will be, you need for the crossover point to be well above the tweeter's resonance frequency.

 

[KaiserSoze]

You really have to be careful sweeping the whole speaker system as that requires much much much more power than the suggested very near-by tweeter sweep.

An idea is to use pink-noise and 'de-pink' the results. Sweeping tweeters at higher SPL is not really recommended as one easily reaches max. power ratings of a tweeter when the sweep is taking too long.
You can set the sweep length short in REW as well to get a feel of the frequency response.

Solderdude is saying that there's no reason to start the sweep that low and that if you do there is greater risk of damaging the tweeter, but maybe your meaning was to run the full sweep with both drivers and the DSP all in place. Not sure exactly what you meant.

 

[Music1969]

, but maybe your meaning was to run the full sweep with both drivers and the DSP all in place.

Yes once I've measured each driver and worked out an initial crossover, then the next step is to use DSP crossover and now check how everything looks with DSP crossover in place.

So I'm guessing at this point it's time for a full sweep 20-20kHz ?

Maybe start off at lower power to start with and maybe 1meter away, say 50dB? Just to make sure DSP crossover is working as intended.

And then gradually increase volume and re-measure?

 

[solderdude]

I was talking about the full sweep with the XO in place of the whole system.
You still need to be careful with slow sweeps and high SPL because of the relatively low continuous power ratings of tweeters.
Of course sweeping at 1m at 80dB SPL isn't really a problem as less than 1W is needed at any time.
Hearing 4kHz at 80dB SPL is not really that pleasant anyway.
I was merely commenting on the usage of sweeps at higher SPL measured further away from the entire speaker.

 

[Music1969]

Of course sweeping at 1m at 80dB SPL isn't really a problem as less than 1W is needed at any time.

Thanks! Ok so from 1m away, I'd start at 50dB with the 1st measurement with full sweep to see how it all looks the 1st time. And then increase to 80dB. Should be safe approach.

I guess your warning is most important when it's time to do DSP room EQ measurement sweeps, at 80dB from 5m away ?

 

[solderdude]

Yes, at larger distances and not using a dB meter (or calibrated REW).
When doing a slow sweep and trying to adjust a level by starting a slow sweep without a dB meter present one might turn up the level when listening to the first part of the sweep only. One might not get the right idea of the actual level being used.
No need to start with 50dB SPL though, can easily be 70 to 80dB. For a quick check a short sweep is better than a long one when you always want to remain on the safe side for the tweeters.

 

[Music1969]

You can make it that simple but you want the crossover point to be high enough to avoid tweeter distortion and low enough so that the woofer will still have good dispersion. Subject to these considerations the crossover point could end up at either end of that overlap region. Also, unless you plan to use a notch filter at the tweeter's resonance (or use the DSP to deal with that), and depending on how steep the high-pass filter will be, you need for the crossover point to be well above the tweeter's resonance frequency.

Would the tweeter resonance frequency be visible from full speaker sweep frequency response measurement, with DSP crossover in place ?

 

[KaiserSoze]

Would the tweeter resonance frequency be visible from full speaker sweep frequency response measurement, with DSP crossover in place ?

You might see a bump in the response, but it becomes most evident when there is another impedance (a simple resistor for example) in series with the tweeter. When a passive, series crossover is in place (your capacitor), the signal voltage is divided between series elements in the same proportion as the impedances. For example, if the impedance of the tweeter at a given frequency is 5 Ohms and if the impedance of a resistor in series with the tweeter is 2.5 Ohms, then since the impedance of the tweeter is 2/3 of the total series impedance (and twice the impedance of the resistor), 2/3 of the signal voltage will appear across the tweeter (and the voltage across the tweeter will be twice greater than the voltage across the resistor). The impedance of any driver increases sharply at the driver resonance. (Note that this is consistent with the fact that very little current is needed to excite the driver at the resonance point and that a small amount of current at a given voltage implies high impedance.) Since the impedance of tweeter will most likely increase sharply at the resonance point, the effect is that the proportion of signal voltage that appears across the tweeter at this particular frequency is a greater proportion than at frequencies both above and below the resonance. In other words the crossover doesn't do what it is supposed to do because it is designed with an assumption about the tweeter's impedance (the value you entered into the calculator to obtain the capacitor value), and this assumption isn't valid at the resonance.

If the frequency space between the resonance and the crossover point is adequately great, the impedance of the crossover element in series with the tweeter will be adequately high - it increases with decreasing frequency - such that the tweeter's impedance will be small (in comparison to the series filter's impedance) in spite of the fact that it increases sharply at that point. Many speaker builders nevertheless prefer to use a series notch filter even if the resonance point is two or more octaves below the crossover point. A series notch filter consists generally of a resistor, capacitor and inductor in series, connected in parallel with the tweeter, i.e., a shunt across the tweeter. It effectively lowers the impedance of the tweeter at the resonance, because it is sort of like a short circuit bypassing the tweeter but present only at the resonance frequency and not a complete short. This cancels the effect of the natural increase in the tweeter's impedance at the resonance, thereby allowing the high-pass filter in series with the tweeter to do what it is intended to do. (If there were no impedance in series with the tweeter, the notch filter would not have any voltage-lowering effect on the tweeter. In this case the effect would only be to increase the total amount of current at the resonance point.)

There are online calculators for determining the value of the resistor, the capacitor, and the inductor. The order in which you connect them does not matter, all that matters is that they are in series and connected across the tweeter terminals. The capacitor blocks frequencies below the resonance point, the inductor blocks frequencies above the resonance point, and the resistor establishes a floor of sorts for how low the shunt impedance will be at the resonance point. Trial and error may be needed to get the values optimized, especially if you don't know the actual impedance of the tweeter at its resonance. Clearly it is better to minimize the potential problem by insuring that the crossover point is at least two octaves above the resonance point, but this depends on how steep the crossover is. This is one way in which a steeper filter is better, because the steeper the high-pass filter is, the greater its impedance will be at the tweeter's resonance.

Since you will play the sweep tones with the simple high-pass filter in place, in series with the tweeter, you might be able to hear the resonance point, or see it in the recorded measurements. And if you decide that it will affect the sound quality, it may be possible to deal with using the DSP crossover you are using, depending on whether it will permit you to define a parametric EQ notch in addition to the high-pass filter. But if the DSP isn't able to do this for you, then depending on whether you think it is something that needs to be dealt with, you may decide that it is warranted to place a series notch filter in parallel with the tweeter, across the tweeter terminals.

 

[dualazmak]

Recently, in my multichannel multi-amplifier project post #258, I carefully tested ON or OFF of protection capacitors for Beryllium MID range squawkers, Beryllium HI range tweeters and Horn type super tweeters, 68 uF, 10 uF and 10 uF respectively.

Fortunately, I found little, almost no, inferior effect of these protection capacitors (Jantzen Audio Standard Z-Caps).

Since I really would like to protect my treasure Be-SQ and Be-TW, I decided to use these protection capacitors at least all the way throughout my coming tests with several amplifiers.

For further details, please visit my multichannel multi-amplifier thread.

 

[ROOSKIE]

Once I get the frequency response of both tweeter and woofer by measuring, then is it simple as knowing the overlap of flat FR region of woofer and tweeter, and then just picking the mid-point and choosing that as the DSP crossover?

Then running the full sweep with DSP crossover in action and hoping for the best !?

No, first see my comment to the second question. Additionally you need to match levels (of course you likely have that one), build in BSC and likely some other tweaks. Play around. Remember to gate your measurements - you want to simulate ancheotic response and aim for a flat line that you later adjust with PEQ once in room.

Would the tweeter resonance frequency be visible from full speaker sweep frequency response measurement, with DSP crossover in place ?

DO a distortion check. That will tell more than the resonance peak will. Some tweeters can be crossed closer to that peak than others, what really matters is distortion levels and frequency response. (for example one of my tweeters can handle something like 1200hrz 24db but because of the rising response below 1800 it crosses better at 2k)

Is there a recommended method to apply heat shrink tubing to the capacitor leads?

Is there a chance of heat damaging the capacitor itself when applying heat?

I saw this thread but there are multiple opinions on how to do it and it left me confused:

https://audiokarma.org/forums/index...nk-tubing-for-electrolytic-capacitors.532021/

And kinds of tips or pointers would be greatly appreciated

Just use some alligator clips or some screws sunk halfway into a board (wrap the cap leads and the wire around the screws) you are not going to accidentally cross wires as you don't need to strip one lead at all. This cap is for testing and can be out side of the box.
I realize you are using some drivers that may be hard to replace but if you are careful you can skip that cap. (especially if are ever using inexpensive tweeters, certainly don't skip it on big buck stuff) Just be super on point and if you fry a tweeter than all the more reason to explore again.