5th element said:
Thanks 😉
I'm currently in the process of expanding the A1 section of the website, so if you fancied sending me some comments about the pre-amp replacement, I'd be happy to put them on the page. I've had quite a bit of positive feedback about this simple mod...
I use a Wavetek 162 as the signal source. It has a precisely defined 50 ohms output impedance, so I sometime use that for quick and simple measurements (like FS). For more precise work, I use a power amp (my gainclone) and a 7.5 ohm series resistor - this particular resistor was selected because it's spot-on its value, and meets the recommendation in the instructions on the spreadsheet that I linked to a few posts back.richie00boy said:
To measure the voicecoil or resistor voltage, I use a scope in X-Y mode - that's more precise than looking at sine waves because all you need to do is measure the length of a vertical line. When finding peaks or nulls (for resonance), I expand the trace as much as possible and shift it down the screen so I can see the top - so the real length of the line might be approaching 20 divisions, giving much more resolution. As long as you don't clip the Y-amps, this works well. My Wavetek has a "vernier" frequency adjust control, which gives around +/-1Hz of adjustment. By zooming in this much on the 'scope trace, you can find the peak or null to within a few degrees of rotation of this control, so it really is very accurate.
Finally, having found resonance, I read off the frequency on my Marconi frequency counter. It takes 10 seconds to perform the reading at low frequencies! Should more resolution or precision be required, then I can perform a reciprocal measurement - this measures very low frequencies very well, but I have to do the reciprocal sum myself 😉
Yes, it's easier to find peaks and nulls with higher value resistors. But, the scope method outlined above is much easier and more precise than using a DVM. It also avoids any issues with RMS conversion that cheaper DVMs suffer from.
Good idea to calibrate - makes scope readings much quicker. I normally set the signal source to be around 3V - this avoids dissipating significant power in the resistor or voicecoil, as heating will change the value of these.
Another thing I have to pay attention to when using the scope - be consistent with the use of RMS, Peak and Peak-Peak. I enter RMS values into the spreadsheet, but the scope measurements are all Pk-Pk, of course. So you have to convert when making the Q measurements
Yes, that's why good test equipment is essential - my function generator is flat from DC to 30MHz. The 'scope is flat from DC to 18Mhz. The -3dB points of the amp are 3.4Hz and 100kHz. I wonder how PC soundcards compare to this? I haven't tried using PC's for this sort of work yet, so as they're an "unknown" to me, I'm inherently suspicious of them
Anyway, that's the basic method that I've used over the years - if there's anything I'm doing wrong, or anything I could be better, I'd appreciate any comments...
Cheers,
Mark 😉
It desn't sound like your actual kit is can be doubted and I'm glad you are aware of the RMS conversion etc. errors 🙂 I have used the scope XY method myself as well.
However, I do have concern over one aspect of your technique, and that is the value of resistance you insert after your amp to facilitate the potential divider. The potential divider method is intended to approximate a constant-current source, which has a very high impedance. 50 ohms is not particularly high and 7.5 ohms is a disaster. Let me show you why.
If you are calibrating the setup so that a 10 ohm speaker will give 100mV across it (as an example), then the voltage you set your power amp to output can be calculated by:
I=V/R = 100mV / 10 = 10mA
V=IR = 10mA * (7.5 + 10) = 175mV
Now, if your speaker impedance rises to 20 ohms:
I=V/R = 175mV / (7.5 + 20) = 6.363mA
V=IR = 6.363mA * 20 = 127mV
Yet we were expecting to see 200mV as we had calibrated to 10mV per ohm.
Even if we use your output voltage of 3V we get errors.
I=V/R = 3 / (7.5 + 10) = 171mA
V=IR = 171mA * 10 = 1.71V
Again, speaker impedance doubles to 20 ohms:
I=V/R = 3 / (7.5 + 20) = 109mA
V=IR = 109mA * 20 = 2.18V
Yet we were expecting 2*1.71=3.42V
I'll let you do the calcs for when the amp resistor is much higher, and see how the results get much more accurate 🙂 Remeber that the current ideally should be constant so that when resistance changes the voltage across it is directly proportional.
Of course, if you are not using this method, and are measuring current through the 7.5 ohm resistor, the results are still bad, as then we are looking for the voltage to be constant, which it will not be. In fact at a speaker impedance of 7.5 ohms it will be halved.
EDIT: just looked at the subwoofers.org site and it seems that the method is OK, just a different way of approaching it
However, I do have concern over one aspect of your technique, and that is the value of resistance you insert after your amp to facilitate the potential divider. The potential divider method is intended to approximate a constant-current source, which has a very high impedance. 50 ohms is not particularly high and 7.5 ohms is a disaster. Let me show you why.
If you are calibrating the setup so that a 10 ohm speaker will give 100mV across it (as an example), then the voltage you set your power amp to output can be calculated by:
I=V/R = 100mV / 10 = 10mA
V=IR = 10mA * (7.5 + 10) = 175mV
Now, if your speaker impedance rises to 20 ohms:
I=V/R = 175mV / (7.5 + 20) = 6.363mA
V=IR = 6.363mA * 20 = 127mV
Yet we were expecting to see 200mV as we had calibrated to 10mV per ohm.
Even if we use your output voltage of 3V we get errors.
I=V/R = 3 / (7.5 + 10) = 171mA
V=IR = 171mA * 10 = 1.71V
Again, speaker impedance doubles to 20 ohms:
I=V/R = 3 / (7.5 + 20) = 109mA
V=IR = 109mA * 20 = 2.18V
Yet we were expecting 2*1.71=3.42V
I'll let you do the calcs for when the amp resistor is much higher, and see how the results get much more accurate 🙂 Remeber that the current ideally should be constant so that when resistance changes the voltage across it is directly proportional.
Of course, if you are not using this method, and are measuring current through the 7.5 ohm resistor, the results are still bad, as then we are looking for the voltage to be constant, which it will not be. In fact at a speaker impedance of 7.5 ohms it will be halved.
EDIT: just looked at the subwoofers.org site and it seems that the method is OK, just a different way of approaching it
richie00boy said:
Hi Richie,
Thanks for this. I understand all of your working, but what I don't understand is the premise that you require current drive. The maths I used to use, and the spreadsheet (as far as I know), does not rely on this. Perhaps when you say "potential divider method", you're refering to something different. For reference, the method I use is described here: http://www.diysubwoofers.org/measure.htm and here: http://sound.westhost.com/tsp.htm
Ok - just seen your edit 😉
So, could you point towards more detail about your potential divider method? I'm happy to try another approach...
Cheers,
Mark 😉
You could make the enclosure deeper and close off the upper (tweeter end) of the enclosure to get you the 6 litres and enough depth for a port, or just bend the port around a corner, if overall size if paramount.mhennessy said:
I find it difficult to judge port air speed because it depends on how hard you push your drivers. For example, at 2w input xmax is exceeded below ~42Hz. Therefore if you look at port air speed at 2w, or even 5w input for that matter, port air speed is acceptable (peaks at 21 m/s). This is with a 150mm wide, 5mm high slot port (120mm long).
It would most definitely be hard to tune after it is built! 😉
By potential divider method, I mean a calibrated divider so that you can read the impedance figure directly from your meter without any calculations or conversion. It's handy for impedance plots. If you read the first part of my post again you should be able to work it out, it's extremely simple 🙂
Basically my amp output resistor is 2.2k ohms and I calibrate so that my meter reads 100mV across a 10 ohm cal resistor in place of the drive unit. Then, if current is constant, a 200mV reading indicates that the impedance is 20 ohms.
There will be no real difference between this method and the one you used in terms of accuracy, and you need an amp with a pretty high voltage swing to be able to push 10mA through 2.2k ohms (22V RMS -- 31V pk). Or you can build a little constant-current amp like I'm doing 🙂 Then you just connect the speaker directly to the amp and measure the voltage across it.
The only benefit of this method is the much increased ease of finding the points of interest because you can work to a much larger scale.
Basically my amp output resistor is 2.2k ohms and I calibrate so that my meter reads 100mV across a 10 ohm cal resistor in place of the drive unit. Then, if current is constant, a 200mV reading indicates that the impedance is 20 ohms.
There will be no real difference between this method and the one you used in terms of accuracy, and you need an amp with a pretty high voltage swing to be able to push 10mA through 2.2k ohms (22V RMS -- 31V pk). Or you can build a little constant-current amp like I'm doing 🙂 Then you just connect the speaker directly to the amp and measure the voltage across it.
The only benefit of this method is the much increased ease of finding the points of interest because you can work to a much larger scale.
Yes, bending a port using standard plumbing fittings is definitely an option. I must admit that creating dead air in a box seems slightly odd, but I suppose the overall depth of a centre speaker is less important because it will be placed in a rack, under the tv...Vikash said:
As I understand it (I might be wrong!), air speed is calculated at the resonant frequency of the port, when driver displacement is at a minimum anyway...
Indeed 😉 I've been wondering about a sliding panel arrangement as the section forming the slot in conjunction with the base would sit in slots routed in the side panels. I don't know how practical this would be, and what effect, if any, the external section of the panel would have. It would be annoying to determine the required panel size, only to cut it down to size and discover that the tuning has changed!
Thanks,
Mark 😉
richie00boy said:
Right - this makes sense. I can see how it could make life easier - in the past I used to calculate everything the hard way - I even wrote a BBC BASIC program to do the maths and plot a graph… Reading it directly from a scope screen or DVM definitely has some appeal 😉
Thanks, by the way, to everyone for responding. I'm learning a lot here 😎
Mark 😉
I would still use a slot around a corner 😉. Not using all the air volume in an enclosure is a perfectly valid approach 😉 I think Matt will be doing something like this to make his floorstanders with only 12 litres internal volume. I would fill the empty cavity with something, probably sand.mhennessy said:
Umm..I don't follow? I was saying that if you listen to the speakers at a level that that outputs minimal distortion (in my modelled case this is ~2w before xmax is exceeded), then I only need to consider port speed at this level (+ a little extra). I use Unibox and the graphs show how port air speed increases as I change the input power.mhennessy said:
Just out of interest, why do you want to tune after building the enclosure? You can always keep the port external to the enclosure. 🙂mhennessy said:
Vikash said:
I hadn't considered floorstanders - yes, in that instance, you're right. I'm getting too focused on my projects (compact passive 2-ways and an active 2.5 way centre speaker)
Can Unibox plot port air speed versus frequency? If you do this, you should find that the air speed in the port is max at the resonant frequency of the port. And, simultaneously, the excursion of the driver is at a minimum (this is the fundamental operating principle of a ported design). I suspect at this particular frequency (~49Hz), you'll be able to get a lot more than 2W into the driver before exceeding Xmax 🙂
Certainly I can put 10-20W into my sealed boxes with no obvious signs of strain… And I guess you'll want to put more than 2W into your speakers because they won't be terribly efficient, even though you're using a pair per box. If 2W is sufficient for your needs, then your neighbours are very lucky 🙂
Go below the port frequency and driver excursion increases massively because the port looks like an "open circuit" to the air in the box - hence only ~2W. Can you get Unibox to plot woofer excursion verses frequency? That might help to prove some of this… It would be interesting to compare ported to sealed...
Someone please correct me if I'm wrong 😉
Because invariably, whenever I've built a vented box and measured the tuning frequency of the port, it's wrong! It's normal (IME) to need to make a fine adjustments to the length of the port to get the frequency just right. Easy with pipes, not so easy with a slot - hence my idea of a sliding panel to enable the exact internal slot length to be determined once the box had been built.
Hope this helps, and again, someone please correct me if I'm wrong...
Mark 😉
You are of course correct with your ported theory, but my point is that you can't only consider one freq. point as music isn't a single frequency (in most cases 😀). It's fine saying that port speed is acceptable and xmax is not exceeded at some input at port tuning (eg 20 watts), but one octave below excursion is ~9.5mm, and an octave above it's ~5mm (xmax is 3mm). So worrying about port speed at a power level that won't be used doesn't seem right to me. Maybe my logic is flawed, or perhaps we're just talking about different thingsmhennessy said:
Sorry, saying 2W was a bad example. More like 6w, and a filter to control excursion when it begins unloading.
In case I'm overlooking something, here's some plots. What do you think?
An externally hosted image should be here but it was not working when we last tested it.
Vikash said:
That's a relief! Actually, I've just downloaded Unibox and done some tests to convince myself - I didn't know about it until you mentioned it, so thanks for the pointer 🙂
but my point is that you can't only consider one freq. point as music isn't a single frequency (in most cases 😀). It's fine saying that port speed is acceptable and xmax is not exceeded at some input at port tuning (eg 20 watts), but one octave below excursion is ~9.5mm, and an octave above it's ~5mm (xmax is 3mm). So worrying about port speed at a power level that won't be used doesn't seem right to me. Maybe my logic is flawed, or perhaps we're just talking about different things
Your logic is working just fine 🙂
Just to clarify, I was responding to your comment, made a few posts back:
I was just trying to explain that port speed is at it's greatest at the tuning frequency of the port, so that's normally the frequency at which you perform the "mach" calculations on the port. If the port diameter is too small in order to keep the length down (my original problem), then whenever a (say) 49Hz note is played, the port will make "chuffing" noises because of the high air speed in the port. I realise you know all of this 😉
I wasn't really factoring in Xmax at this point, but you're right, and your plots show it perfectly. Below port tuning, driver unloads. Above port tuning, it's possible to exceed Xmax because the port isn't helping to reduce excursion. Choosing 12W prevents the latter happening, and again, you're right - this will result in a lower mach at port resonance, so you can get away with a smaller diameter port than one based on 3mm Xmax at "F port"
With hindsite, I wonder if this is what you originally meant? Re-reading the posts, I think your 2W limit of power-input was based on the below-resonance <40Hz Xmax limit - your 12W (above-resonance 70-80Hz Xmax limit) is the right way to go 😉
I don't think you're missing anything with your design, I just trying to help you clear up that initial question/point, and at the same time, confirming with everyone that my grasp of the theory is sound. As I might have said before, it's been a few years since I played with ported designs...
So, I don't think I've said anything that you didn't already know. But hopefully it's been useful to have the conversation, because it's good to double-check and confirm these things. It's certainly been useful revision for me, and it could help others who are trying to learn.
The extra bass possible from a slightly bigger vented box is tempting, although the phase plots from WinISD are enough to worry me
Cheers,
Mark 😉
Guys,
i am currently looking into purchasing the crossover components as per Zaphs post,
can some one explain the differencies between air cored and ferrite cored (not the construction, that part is clear 😉 ) inductors.
The ferrit cores seem to be nearer the resistive value quoted for each inductor but i am assuming the air cored to be of a superior audio quality, but, could this be off set if the only air cored available advertise a higher resistance value due to the limited conductor size?
What should i do??
M
i am currently looking into purchasing the crossover components as per Zaphs post,
can some one explain the differencies between air cored and ferrite cored (not the construction, that part is clear 😉 ) inductors.
The ferrit cores seem to be nearer the resistive value quoted for each inductor but i am assuming the air cored to be of a superior audio quality, but, could this be off set if the only air cored available advertise a higher resistance value due to the limited conductor size?
What should i do??
M
By my calculations using LspCAD pro to simulate with a box of 18 litres tuned to 49hz. You can get away with a port area of 16cm squared. This produces a max air velocity of 15.6ms which is fine.
This takes 15 watts to reach. 15 watts into an 8 ohm load which is the same as 10.952 volts if you want to convert it into a 3ohm load. To see how many watts youll need into 3ohms.id imagine its something like 40.
So in other words if you have a slot length of 16cm it needs to be a cm tall. However with a circular port you require a diameter of 4.5cm and a length of 7.1cm which to me would be much easier to do then a slot. If you happy with an air speed of 20ms then a port of 4cm diameter is fine haha. Not really much difference.
This takes 15 watts to reach. 15 watts into an 8 ohm load which is the same as 10.952 volts if you want to convert it into a 3ohm load. To see how many watts youll need into 3ohms.id imagine its something like 40.
So in other words if you have a slot length of 16cm it needs to be a cm tall. However with a circular port you require a diameter of 4.5cm and a length of 7.1cm which to me would be much easier to do then a slot. If you happy with an air speed of 20ms then a port of 4cm diameter is fine haha. Not really much difference.
Inductors
It depends on whether the inductors are for the bass or treble.
If for the bass, ideally you should use the lowest possible DC resistance ones as the DCR will affect the Q of the box at resonance, and also rob a little power.
The tweeter is not so critical as you will not be using it significantly at resonance. Also, you often need a bit of resistance there to reduce the level as tweeters are usually more sensitive.
Another thing to consider is the fact that an air cored inductor has a rather large stray field, so placement is quite critical. Most of the field is contained within the core on a cored inductor, so it's less of a problem.
A good cored inductor is my choice everytime unless only a very small inductance is needed. A good part will use a linear core that is oversized so does not suffer from saturation and/or hysterisis problems usually associated with cheaper cored types. It will also be easier to locate and have much lower DCR.
But there will always be some that insist on only air cored inductors
It depends on whether the inductors are for the bass or treble.
If for the bass, ideally you should use the lowest possible DC resistance ones as the DCR will affect the Q of the box at resonance, and also rob a little power.
The tweeter is not so critical as you will not be using it significantly at resonance. Also, you often need a bit of resistance there to reduce the level as tweeters are usually more sensitive.
Another thing to consider is the fact that an air cored inductor has a rather large stray field, so placement is quite critical. Most of the field is contained within the core on a cored inductor, so it's less of a problem.
A good cored inductor is my choice everytime unless only a very small inductance is needed. A good part will use a linear core that is oversized so does not suffer from saturation and/or hysterisis problems usually associated with cheaper cored types. It will also be easier to locate and have much lower DCR.
But there will always be some that insist on only air cored inductors
Re: Inductors
Someone over at madisound did tests on magnetic coupling between inductors. They ran a voltage (larger then most speaker applications) thru one inductor and saw if it had any effect on another inductor that was located within very close proximaty to the first. Net effect hardly anything, showing that magnetic coupling doesnt really do much.
richie00boy said:
Someone over at madisound did tests on magnetic coupling between inductors. They ran a voltage (larger then most speaker applications) thru one inductor and saw if it had any effect on another inductor that was located within very close proximaty to the first. Net effect hardly anything, showing that magnetic coupling doesnt really do much.
D1GGY said:
Hello,
First, let me mention that I dropped all the information on my design onto a web page. This is going to be there a while so you don't have to worry about me deleting it. (the other images I linked to have been relocated to the web page)
Audax Mini
I've added a bit more information about the design and cleaned up the schematic into something more readable.
You probably don't have to worry much about the inductors, just get the right values. and keep the 1.5 mH DCR somewhere below .3 ohms
I agree with richie00boy's comments about the inductors, he covers your question well. In general, iron, ferrite or steel laminate core inductors saturate sooner than air core. But also in general air core has higher resistance.
We're not exactly going to be pumping megawatts into the little speakers, so cored inductors all around would be fine.
Other qualities about better performance of air core are highly debatable.
Actually a cored inductor probably would be better in a low power situation. Less wire needed less signal degredation etc.
AP100Z0 Fs
I've been trying to measure an AP100Z0 this evening. It's straight out of the box. I measured Fs at 96Hz.
I'm having problems with the Fl and Fh for calculating Q though -- the sanity check yields Fs of 90Hz. I'm using the Brian Steele method, but without any series resistor -- just putting my meter in series and reading off current. Should be the same thing
The meter is staying on the same range so it's resistance is constant, so that rules that out.
Must try and get my constant-current circuit sorted and see if that gives any better results.
I've been trying to measure an AP100Z0 this evening. It's straight out of the box. I measured Fs at 96Hz.
I'm having problems with the Fl and Fh for calculating Q though -- the sanity check yields Fs of 90Hz. I'm using the Brian Steele method, but without any series resistor -- just putting my meter in series and reading off current. Should be the same thing
The meter is staying on the same range so it's resistance is constant, so that rules that out.Must try and get my constant-current circuit sorted and see if that gives any better results.
- Status
- Not open for further replies.