FWIW: My experience confirms what Mike said:
Years ago SB mag had an article about the loose tolerance on drivers.
I was luckier w a MCM 8 - I only bought one but, indeed it was not exact. for instance Post burn-in Fs is still over 10% above the advertised 25Hz.
And in addition to what Mike said: sometimes there are bad runs or defects with otherwise good products.
Admittedly in a simple sealed cab, a driver can be out of spec ( to some extent ) and the actual results won't be too far from predicted.
If the design utilizes tighter tolerance and tuning ( B/R, Bandpass, T/L, horns ) then it would stand to reason because everything is based upon an assumption of T/S values vs actual values, the effects of cumulative error will be greater.
Before I had the means to run impedance measurements, I built a bandpass box, it needed tuning and was much more difficult to tune ( by ear ), than simpler designs.
Given a couple of old McCauley drivers to evaluate for possible "Hornifcation" - They had very close serial #'s One was closer to manufacturer spec the other was way off and would have been unusable.
I have measured over 15% variance.
Years ago SB mag had an article about the loose tolerance on drivers.
I was luckier w a MCM 8 - I only bought one but, indeed it was not exact. for instance Post burn-in Fs is still over 10% above the advertised 25Hz.
And in addition to what Mike said: sometimes there are bad runs or defects with otherwise good products.
Admittedly in a simple sealed cab, a driver can be out of spec ( to some extent ) and the actual results won't be too far from predicted.
If the design utilizes tighter tolerance and tuning ( B/R, Bandpass, T/L, horns ) then it would stand to reason because everything is based upon an assumption of T/S values vs actual values, the effects of cumulative error will be greater.
Before I had the means to run impedance measurements, I built a bandpass box, it needed tuning and was much more difficult to tune ( by ear ), than simpler designs.
Given a couple of old McCauley drivers to evaluate for possible "Hornifcation" - They had very close serial #'s One was closer to manufacturer spec the other was way off and would have been unusable.
I have measured over 15% variance.
Being able to measure things is really cool. It is even cooler when what you measure corresponds to what you hear. If I had to pick (and I am glad I don't) - I'd give up SPL measurements to have impedance and TS testing, 'cause I can't measure impedance or calculate TS parameters with my ears (or an SPL meter).
Really - if you have a computer (and you most probably do if you're reading this), an investment of about $200 can open up a whole new world of measurement capabilities. I have a little over that tied up in all my stuff (SPL from 10 Hz-20KHz +/- 2 dB, Impedance, and TS parameters) and there are others who have done more for even less.
Measurements trump specs, period. The ability to run an impedance sweep of a box really helps find leaks and other problems (you listening Jim????
) as well as compare what you built with what your software predicted you'd get.
I do preliminary designs and buy drivers based on specs. Once I get the drivers, I break them in and test them. The specs are usually consistent between multiple drivers in the same order, and usually not equal to what is published. Measurements and testing right out of the box also help identify bad drivers right away, as well as match drivers for use in multi-driver cabinets.
Usually the drivers I order work fine for my intended purpose with only minor tweaking of the initial design. Generally speaking, the more complicated the enclosure gets, the more likely the mismatch will get ugly.
Really - if you have a computer (and you most probably do if you're reading this), an investment of about $200 can open up a whole new world of measurement capabilities. I have a little over that tied up in all my stuff (SPL from 10 Hz-20KHz +/- 2 dB, Impedance, and TS parameters) and there are others who have done more for even less.
Measurements trump specs, period. The ability to run an impedance sweep of a box really helps find leaks and other problems (you listening Jim????
) as well as compare what you built with what your software predicted you'd get.I do preliminary designs and buy drivers based on specs. Once I get the drivers, I break them in and test them. The specs are usually consistent between multiple drivers in the same order, and usually not equal to what is published. Measurements and testing right out of the box also help identify bad drivers right away, as well as match drivers for use in multi-driver cabinets.
Usually the drivers I order work fine for my intended purpose with only minor tweaking of the initial design. Generally speaking, the more complicated the enclosure gets, the more likely the mismatch will get ugly.
4 cabinets was the point.
Though as I said I did not dwell on it anymore than that as I had no intention of building one. It was an exercise into the abilities of the driver.
The #'s are playable so make it what you want.
The driver seems to be very well suited for MAX SPL. Even within a small enclosure.
It is like the long standing debate in the car community.
Of "no replacement for displacement". It is always easier to get more from something bigger but what is the fun in that?
Though as I said I did not dwell on it anymore than that as I had no intention of building one. It was an exercise into the abilities of the driver.
The #'s are playable so make it what you want.
The driver seems to be very well suited for MAX SPL. Even within a small enclosure.
It is like the long standing debate in the car community.
Of "no replacement for displacement". It is always easier to get more from something bigger but what is the fun in that?
littlemike said:The ability to run an impedance sweep of a box really helps find leaks and other problems (you listening Jim????
I hear you loud and clear. been doing some of that lately, with tops and subs just for the fun of it.
Here's a low tech approach that I was just testing... just to see what I could get.
Attachments
littlemike said:Being able to measure things is really cool. It is even cooler when what you measure corresponds to what you hear. If I had to pick (and I am glad I don't) - I'd give up SPL measurements to have impedance and TS testing, 'cause I can't measure impedance or calculate TS parameters with my ears (or an SPL meter).
Really - if you have a computer (and you most probably do if you're reading this), an investment of about $200 can open up a whole new world of measurement capabilities. I have a little over that tied up in all my stuff (SPL from 10 Hz-20KHz +/- 2 dB, Impedance, and TS parameters) and there are others who have done more for even less.
Measurements trump specs, period. The ability to run an impedance sweep of a box really helps find leaks and other problems (you listening Jim????
I do preliminary designs and buy drivers based on specs. Once I get the drivers, I break them in and test them. The specs are usually consistent between multiple drivers in the same order, and usually not equal to what is published. Measurements and testing right out of the box also help identify bad drivers right away, as well as match drivers for use in multi-driver cabinets.
Usually the drivers I order work fine for my intended purpose with only minor tweaking of the initial design. Generally speaking, the more complicated the enclosure gets, the more likely the mismatch will get ugly.
THIS
All you guys building tapped horns need to read this. If you're trying to do it without measuring your drivers and the impedance curve of the finished product, you're wasting your time.
but you need to start SOMEWHERE, and hornresp and published specs are that somewhere.but yeah you guys have a point, it seems that my big fancy loud PA, is turning more and more into a fairy tale daily.
however, when it comes down to it, as long as the performance beats a woofer in a BR box, im realy not complaining, cause it seems like im the only one in toronto (apart from a couple clubs) who've ever heard of a horn, let alone a tapped one.
littlemike said:Being able to measure things is really cool. It is even cooler when what you measure corresponds to what you hear. If I had to pick (and I am glad I don't) - I'd give up SPL measurements to have impedance and TS testing, 'cause I can't measure impedance or calculate TS parameters with my ears (or an SPL meter).
Really - if you have a computer (and you most probably do if you're reading this), an investment of about $200 can open up a whole new world of measurement capabilities. I have a little over that tied up in all my stuff (SPL from 10 Hz-20KHz +/- 2 dB, Impedance, and TS parameters) and there are others who have done more for even less.
Measurements trump specs, period. The ability to run an impedance sweep of a box really helps find leaks and other problems (you listening Jim????
I do preliminary designs and buy drivers based on specs. Once I get the drivers, I break them in and test them. The specs are usually consistent between multiple drivers in the same order, and usually not equal to what is published. Measurements and testing right out of the box also help identify bad drivers right away, as well as match drivers for use in multi-driver cabinets.
Usually the drivers I order work fine for my intended purpose with only minor tweaking of the initial design. Generally speaking, the more complicated the enclosure gets, the more likely the mismatch will get ugly.
What $200 product are you talking about? Just curious what other are using to make their measurements.
Rgs, JLH
It is a DIY system I have pieced together over a few years, not any one item or any one piece of software.
For T/S parameters and Impedance, I use Smith & Larson's Woofer Tester 2. This is NOT the same as the P/E WT3, but is the same as the Tenma/MCM Speaker Measurement Interface. The MCM version is currently $145, but is occasionally available for ~$125 when purchased with a 15% off coupon. The WT2 is $159.99 direct from Smith & Larson. I purchased mine from Smith & Larson at a discounted price a while back as an upgrade to my old WT1, which did not work well with my newer laptop.
For relative SPL measurements, I use a Radio Shack 33-2050 Analog SPL meter modified per Eric Wallin's website. His instructions are exceptional, but you have to know which end of the soldering iron to hold. It also helps if you are at least comfortable with electronics, but the modifications were not that hard to accomplish. I picked my meter up off Ebay for $15, then spent another $25 or so @ digikey for the capacitors and the panasonic mic capsules I needed for the modifications.
With a windows-based computer that has a decent (16-bit, full duplex) sound card, these two pieces and a few patch cables are really all you need to get started. This is all you need for impedance, T/S parameters, and relative SPL from 10 Hz to 20 KHz. When comparing multiple speakers, I measure and set voltages with my digital multi-meter. It is not true RMS, but works well enough for my needs. Absolute SPL requires calibration with a known standard, something I have not worried about yet. If you are less concerned about the frequency response of the mic preamp, you can even run the panasonic capsule off of the mic-in port on a quality sound card and get good results.
All the software I use is freely available or bundled with the WT2. I use freqresplot (free download and free to use) and roomeq wizard (free to use, requires a forum registration for download) for swept sine bass measurements, and free versions of Praxis, SynRTA, or ARTA for other measurements. I understand that ARTA and Praxis will also measure impedance and T/S parameters, which sort of negates the need for the WT2, but did not discover this until after I purchased the WT2.
In my case, I also elected to purchase a sound card that was better than the one in my laptop so I could have a portable measurement system that provided meaningful results below 100 Hz. After looking at several options, I purchased an E-mu 0202 USB sound card. This added ~$75 to my total, but may or may not be required based on the PC that will be used.
Further discussions on this may warrant a separate thread (so that others can find it).
For T/S parameters and Impedance, I use Smith & Larson's Woofer Tester 2. This is NOT the same as the P/E WT3, but is the same as the Tenma/MCM Speaker Measurement Interface. The MCM version is currently $145, but is occasionally available for ~$125 when purchased with a 15% off coupon. The WT2 is $159.99 direct from Smith & Larson. I purchased mine from Smith & Larson at a discounted price a while back as an upgrade to my old WT1, which did not work well with my newer laptop.
For relative SPL measurements, I use a Radio Shack 33-2050 Analog SPL meter modified per Eric Wallin's website. His instructions are exceptional, but you have to know which end of the soldering iron to hold. It also helps if you are at least comfortable with electronics, but the modifications were not that hard to accomplish. I picked my meter up off Ebay for $15, then spent another $25 or so @ digikey for the capacitors and the panasonic mic capsules I needed for the modifications.
With a windows-based computer that has a decent (16-bit, full duplex) sound card, these two pieces and a few patch cables are really all you need to get started. This is all you need for impedance, T/S parameters, and relative SPL from 10 Hz to 20 KHz. When comparing multiple speakers, I measure and set voltages with my digital multi-meter. It is not true RMS, but works well enough for my needs. Absolute SPL requires calibration with a known standard, something I have not worried about yet. If you are less concerned about the frequency response of the mic preamp, you can even run the panasonic capsule off of the mic-in port on a quality sound card and get good results.
All the software I use is freely available or bundled with the WT2. I use freqresplot (free download and free to use) and roomeq wizard (free to use, requires a forum registration for download) for swept sine bass measurements, and free versions of Praxis, SynRTA, or ARTA for other measurements. I understand that ARTA and Praxis will also measure impedance and T/S parameters, which sort of negates the need for the WT2, but did not discover this until after I purchased the WT2.
In my case, I also elected to purchase a sound card that was better than the one in my laptop so I could have a portable measurement system that provided meaningful results below 100 Hz. After looking at several options, I purchased an E-mu 0202 USB sound card. This added ~$75 to my total, but may or may not be required based on the PC that will be used.
Further discussions on this may warrant a separate thread (so that others can find it).
mike:
The thought on the cabinet I posted was, how can I make the simplest/cheapest cabinet possible that's usable for PA? I'm on a kick to try to put together PA/band in a box, that's simple/cheap, fits in an average sized car, and beats the average plasti-box system.
The smaart measurement was taken against (2) mcm55-1740's, (4) 1025 piezo's, (5) 12x24 and (2) 24x24 pieces of plywood, and a little fiberglass..
What was of interest to me, that required me to make sawdust, was my ports inside the horn. Ports inside dropped the overall response down lower than predicted. (because hornresp does not allow ports inside the horn) I had solid output to 60hz. So, I need to readjust the ports to raise tuning a little, and get rid of my small dip at 100hz. With 60hz output... wow, usable (not great, but usable) without subs!
Anyway, with the 1k dip and the little 100hz caveat's... not bad. I was expecting 1k to be really bad, due to the angle of the 55-1740's... but it really wasn't bad... It was really listenable, which surprised me. Once the top is acceptable, then I'll turn my attention to a sub... any good ideas for a <$100 PA sub (total) that'll do 120db 40-100hz in 2pi?
To your point -- measuring makes all the difference in the world. When you invest time and effort, that makes things 'sound' better... regardless if it really does or not...
The thought on the cabinet I posted was, how can I make the simplest/cheapest cabinet possible that's usable for PA? I'm on a kick to try to put together PA/band in a box, that's simple/cheap, fits in an average sized car, and beats the average plasti-box system.
The smaart measurement was taken against (2) mcm55-1740's, (4) 1025 piezo's, (5) 12x24 and (2) 24x24 pieces of plywood, and a little fiberglass..
What was of interest to me, that required me to make sawdust, was my ports inside the horn. Ports inside dropped the overall response down lower than predicted. (because hornresp does not allow ports inside the horn) I had solid output to 60hz. So, I need to readjust the ports to raise tuning a little, and get rid of my small dip at 100hz. With 60hz output... wow, usable (not great, but usable) without subs!
Anyway, with the 1k dip and the little 100hz caveat's... not bad. I was expecting 1k to be really bad, due to the angle of the 55-1740's... but it really wasn't bad... It was really listenable, which surprised me. Once the top is acceptable, then I'll turn my attention to a sub... any good ideas for a <$100 PA sub (total) that'll do 120db 40-100hz in 2pi?
To your point -- measuring makes all the difference in the world. When you invest time and effort, that makes things 'sound' better... regardless if it really does or not...
Something like this perhaps?
Two of your favorite $25 dollar MCM's in a pretty simple 250L TH.
Not a tough build, pretty flat response, but nothing at all below 40 Hz.
Don't run much more than 100 watts per driver to it either.
Achieves > 120 dB 40-100 in 2pi without exceeding 8 mm excursion above 40 Hz. Not a small box, but not that huge either. I know it would not fit in my car easily.
If you take most of the taper out, it gets as small as 170 liters. Still a bit big for two tens, but it actually flattens the response and excursion is similar, and it only loses a few dB.
I know it is not optimized - just a concept here....
Two of your favorite $25 dollar MCM's in a pretty simple 250L TH.
Not a tough build, pretty flat response, but nothing at all below 40 Hz.
Don't run much more than 100 watts per driver to it either.
Achieves > 120 dB 40-100 in 2pi without exceeding 8 mm excursion above 40 Hz. Not a small box, but not that huge either. I know it would not fit in my car easily.
If you take most of the taper out, it gets as small as 170 liters. Still a bit big for two tens, but it actually flattens the response and excursion is similar, and it only loses a few dB.
I know it is not optimized - just a concept here....
Attachments
I've actually already tried some variations on that.. TH is what I thought I would want for this project... The 55-1740 looks great on paper.. but falls short on actually delivering... (back to that measuring thing??) Here's one of my attempts at a 55-1740 TH, stole one of the drivers for my latest top experiments.....
I know this sounds crazy, but a PE 295-470 In a FLH, 200-2500, 240 long with a 20liter rear chamber looks promising... Could do it up and back, 48" tall with driver in the bottom of the cabinet with the 45' reflector as the back of the chamber. that would be easy. (bass stick anyone??)
It has the advantage of not needing a HPF, and a simple 12db/oct cross would be sufficient, and it's 4ohm.
hmmm.. still thinking.
I know this sounds crazy, but a PE 295-470 In a FLH, 200-2500, 240 long with a 20liter rear chamber looks promising... Could do it up and back, 48" tall with driver in the bottom of the cabinet with the 45' reflector as the back of the chamber. that would be easy. (bass stick anyone??)
It has the advantage of not needing a HPF, and a simple 12db/oct cross would be sufficient, and it's 4ohm.
hmmm.. still thinking.
Attachments
j bell, i have to say im really interested in your current project, seems exactly what i want, around 100$ a sub, that will beat the average woofer in a BR run of the mill PA sub. depending on how it turns out i may forgo designing my own, as i feel a little outclassed by what you guys are designing, and the depth at which you do it.
also, your driver selection brings up an interesting point, dayton usually has pretty trustworthy specs, and i can get the DVC15 for quite cheap here in canada(158$ a driver), would it be safe to say that a TH designed in horn resp with one of those drivers would perform close enough to the simulation?
im also starting to notice that modelling says its easier to get higher SPLs out of conventional horn loaded systems. but i believe it is accepted that models dont show real life results.
my head is swimming with all these new factors to take into account.
also, your driver selection brings up an interesting point, dayton usually has pretty trustworthy specs, and i can get the DVC15 for quite cheap here in canada(158$ a driver), would it be safe to say that a TH designed in horn resp with one of those drivers would perform close enough to the simulation?
im also starting to notice that modelling says its easier to get higher SPLs out of conventional horn loaded systems. but i believe it is accepted that models dont show real life results.
my head is swimming with all these new factors to take into account.
xstephanx said:
xstephanx:
hmm....
There are some drivers that work well in TH, and some that work well in FLH, BLH, BP, BR, TL, etc... not every driver works well in all designs, and picking a random "good quality" driver and choosing a style of cabinet for it without finding out if it's a good fit... is asking for trouble in my experience. I haven't modeled the dayton DVC in all those different designs, so I don't know for sure what cabinet style it's best suited for.
Yes for this project, I am primarily interested in designing a TH, due to my past success with TH. However, finding a suitable driver in this price range that will give the required output... has been challenging. (anyone have a good suggestion?)
On FLH, usually they have a peak that's higher in output than other designs, but once you eq them flat to 40hz... things change dramatically. Also, with FLH, I see that hornresp is again, usually spot on, when looking at FLH as well as TH, and other designs, and I have great faith in the math behind david's GENEROUS work for the community.
I typically take issue to undersized 1/4 wave FLH, and their ragged response, and especially their rising response as they approach 1/2 wave. Having a sub that's 13db down in response at 40hz vs 100hz is not my idea of a good sub. The $80 295-470 driver happens to look better in a FLH, than a TH.. The FLH also has the advantage of a small rear chamber, which does not require a HP filter that the TH does, when playing recorded music with studio stuff added in below 40hz...
So for my 120db goal from 40hz -100hz, at $100 or less, well the FLH dayton is close, but not quite, by the time you add plywood, finish, jacks and corners. However, for your needs, if you're willing to go a little over $100 for a sub, it might be worthwhile, so I'm attaching the hornresp sim.
Attachments
im starting to wonder how important it is to stay within the xmax of the driver(to a certain extent). eminence gamma15a has 3mm of Xmax, but 11.4mm of excursion before it starts to become damaged.
people seem to say that horns sound very clean up until you blow the driver, so you should be cautious when powering them. this leads me to believe that maybe horn loading helps eliminate distortion caused by nonlinierities of the driver?
if you put the gamma in a 400 litre tapped horn you can get 120db with 100 watts in 2 pi from 40 to 120hz, if you highpass it at 40hz, the most it will ever see is 6 or 7mm excursion, how bad would distortion be at this level? it is still far from the stated mechanical and thermal limits of the driver, so presumeably it could achieve this all night, even though its not always staying within Xmax.
ive also noticed that hornresp predicts that excursion lessens as you decrease the radiating angle, would this same effect be observed in subs whose mouths are coupled?
try this out
s1=303
s4=2800
L12= 39
L23=252
L34=20.2
the driver will hit mechanical limits roughly at the drivers 300w rms rating.
just throwing some ideas out.
people seem to say that horns sound very clean up until you blow the driver, so you should be cautious when powering them. this leads me to believe that maybe horn loading helps eliminate distortion caused by nonlinierities of the driver?
if you put the gamma in a 400 litre tapped horn you can get 120db with 100 watts in 2 pi from 40 to 120hz, if you highpass it at 40hz, the most it will ever see is 6 or 7mm excursion, how bad would distortion be at this level? it is still far from the stated mechanical and thermal limits of the driver, so presumeably it could achieve this all night, even though its not always staying within Xmax.
ive also noticed that hornresp predicts that excursion lessens as you decrease the radiating angle, would this same effect be observed in subs whose mouths are coupled?
try this out
s1=303
s4=2800
L12= 39
L23=252
L34=20.2
the driver will hit mechanical limits roughly at the drivers 300w rms rating.
just throwing some ideas out.
How about a 2:1 compression ratio with an inexpensive car stereo woofer?
(JBL GTO1514, script is marked CSS SDX15a)
Sim is in 1Pi, so this represents a stacked pair in 2Pi.
The CSS SDX15 is a better driver, twice the price, but limited availability. I made the JBL GTO1514 fit the design for someone that had picked up a pair on the cheap, and wondered what could be done with them.
(JBL GTO1514, script is marked CSS SDX15a)
Sim is in 1Pi, so this represents a stacked pair in 2Pi.
The CSS SDX15 is a better driver, twice the price, but limited availability. I made the JBL GTO1514 fit the design for someone that had picked up a pair on the cheap, and wondered what could be done with them.
"im starting to wonder how important it is to stay within the xmax of the driver(to a certain extent). eminence gamma15a has 3mm of Xmax, but 11.4mm of excursion before it starts to become damaged"
On this model the coil is very short, and the top plate is very thin. The coil will be driven out of the gap before it reaches X-mech, and may hang up (I've had this happen). The driver bursts into flame shortly thereafter (had to use a towel to beat it out).
On this model the coil is very short, and the top plate is very thin. The coil will be driven out of the gap before it reaches X-mech, and may hang up (I've had this happen). The driver bursts into flame shortly thereafter (had to use a towel to beat it out).