New sub design? Constricted Transflex, simple build (series tuned 6th order)

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
By the way , who was asking about refining some square piezo horns a while back? Was it GM or someone else? This conversation has grown so long now ..... I suppose i could scroll through the pages to find the exchange .... I have something that works VERY well as a piezo horn body with square dimensions ..... I was pleasantly surprised to find out that these measure better than the Goldwood 1016 horn body! :eek:

Hi MMJ,
glad you survived the holiday madness :D

IIRC it was @freddi who started the piezo discussion. I'm really looking forward to your new findings.

Regards
Gerald
 
Santa operates a sweatshop

Hi MMJ,
glad you survived the holiday madness :D

We are survivors! :fight:.... Hehe

I hope all of the DIYaudio folks had a great holiday season!!

Gerald, this time of year must be gorgeous in Austria ......

I have a question for you: Do you ever order from Thomann.de?

Thomann is carrying a driver that i thought i should mention, The 15LB100-8W, it was first brought to my attention by my friend Sebastian in Denmark, he owns a set of them ......

The 15LB100 has a respectable amount of motor strength for a budget PA 15 , enough motor force in fact to make it perfect for a KARLFLEX 15! :nod: (and also great for a few other 6th order designs i have been working on)...... The driver may not have seemed entirely attractive to some if they saw the same spec sheet that i was initially shown which stated a humble linear excursion rating of "5.5mm", but i noticed that the 5.5mm figure is based upon a very conservative formula ..... After looking at the gap depth and voicecoil length (which they also make available) and applying those measurements to the same formula that B&C, 18sound, JBL, and RCF uses to calculate their "Xmax" i end up with a much higher value for the 15LB100...... (Hvc-Hg)/2 + Hg/4 = 8mm+ of one-way XMAX!!

That formula above doesn't necessarily tell us where the 10% THD mark is, at least not precisely, it is only a rough approximation but still indicates that this driver is very conservatively rated making it an extraordinary performance value for that price! :note::happy2::$:


https://www.thomann.de/intl/the_box_15lb1008wersatzspeaker_15.htm

I have more driver options to share but i figured this would be a good place to start ...
 
Last edited:
We are survivors! :fight:.... Hehe

I hope all of the DIYaudio folks had a great holiday season!!

Gerald, this time of year must be gorgeous in Austria ......

I have a question for you: Do you ever order from Thomann.de?

Thomann is carrying a driver that i thought i should mention, The 15LB100-8W, it was first brought to my attention by my friend Sebastian in Denmark, he owns a set of them ......

The 15LB100 has a respectable amount of motor strength for a budget PA 15 , enough motor force in fact to make it perfect for a KARLFLEX 15! :nod: (and also great for a few other 6th order designs i have been working on)...... The driver may not have seemed entirely attractive to some if they saw the same spec sheet that i was initially shown which stated a humble linear excursion rating of "5.5mm", but i noticed that the 5.5mm figure is based upon a very conservative formula ..... After looking at the gap depth and voicecoil length (which they also make available) and applying those measurements to the same formula that B&C, 18sound, JBL, and RCF uses to calculate their "Xmax" i end up with a much higher value for the 15LB100...... (Hvc-Hg)/2 + Hg/4 = 8mm+ of one-way XMAX!!

That formula above doesn't necessarily tell us where the 10% THD mark is, at least not precisely, it is only a rough approximation but still indicates that this driver is very conservatively rated making it an extraordinary performance value for that price! :note::happy2::$:


https://www.thomann.de/intl/the_box_15lb1008wersatzspeaker_15.htm

I have more driver options to share but i figured this would be a good place to start ...

Hi MMJ,

weather is sunny but rather cold. Fortunately no snow here in the city ;)

Sure I know Thomann, have also seen this particular driver before, but didn't think of it... Thanks for the hint. Now I just have to redraw the plans for the enclosure.

Best wishes
Gerald
 
15LB100 Karlflex chat

Hi MMJ,

weather is sunny but rather cold. Fortunately no snow here in the city ;)

Sure I know Thomann, have also seen this particular driver before, but didn't think of it... Thanks for the hint. Now I just have to redraw the plans for the enclosure.

Best wishes
Gerald

We ended up with a lot of rain, the change in weather is nice though as it is normally dry in Arizona :) Seems like it has been warmer than a typical winter here in the low desert so i just cleaned up my outdoor shop area, i think i might attempt to get a cabinet building project underway ...

Gerald, I hope your shop is indoors, i am sure it gets much colder at your latitude than it does here ....

That 15LB100 driver appears to be a good one for the price, my friend who owns the set doesn't have a DATS, WT3 (or similar device) so we were unable to verify T/S parameters but the driver performs as expected and matches simulations closely enough so that is a very good sign ....... If you like i can put you into contact with Sebastian and he can share his impression of those drivers with you ..... He has had the chance to compare them with various other drivers because he is always working on experimental DIY speaker projects over at his place :smash:....... The Danish DIY audio community is incredibly active from what they tell me...



As with most drivers they will sound best once the suspension is broken-in (if you want to do it the fast way you can just drive the woofer with 30hz to 40hz sine waves overnight in free-air) ..



Gerald, what frequency do you plan on using these Karlflex 15s up to?

Being a 15" driver it should be good out to around 1khz before it's midrange dispersion starts getting too narrow as a general rule .... So 1khz is a good (maximum) crossover point limit unless you expect for your listening to always be straight out in front of the cabinet (on-axis) ......



NOTE: In theory there is a workaround to this "beaming" challenge using pattern controlling schemes.... One method is a slot shape aperture (obscuring the sides of the driver somewhat, or at least the upper portion of the sides) which acts like an acoustic lens of a sort (as i have heard it described by Mr XRK971) causing the upper-midrange energy to become more evenly dispersed in the horizontal plane ............... I tried the slot idea with the Karlflex and indeed it really does work , verified by polar measurements, but (for me)this effect was at the expense of knocking down the on-axis upper midrange output by some decibels and the PA310 really didn't have enough extra decibels to spare up in that range in my opinion so I decided to go without the K-slot on mine for the time being .....

The 15LB100 does have somewhat of a rising upper mid response according to it's published curves so it may do well with this sort of pattern control extending it's usefulness out to what looks like about 1.75khz ......... If someone really wanted to experiment with this it would be easy enough to do by first building the Karlflex cabinet according to the plans and then making easy modifications to the front panel (and taking measurements , refining, etc):magnify: ........

I personally plan to revisit the slot idea, because i now have some new tricks up my sleeve that i would like to try.:bulb:
 
Last edited:
Gerald, I hope your shop is indoors, i am sure it gets much colder at your latitude than it does here ....

My shop is definitely indoors, being either kitchen or linig room - and we have a temperature drop down to -20°C forecasted... :mad:

Gerald, what frequency do you plan on using these Karlflex 15s up to?

Being a 15" driver it should be good out to around 1khz before it's midrange dispersion starts getting too narrow as a general rule .... So 1khz is a good (maximum) crossover point limit unless you expect for your listening to always be straight out in front of the cabinet (on-axis) ......



NOTE: In theory there is a workaround to this "beaming" challenge using pattern controlling schemes.... One method is a slot shape aperture (obscuring the sides of the driver somewhat, or at least the upper portion of the sides) which acts like an acoustic lens of a sort (as i have heard it described by Mr XRK971) causing the upper-midrange energy to become more evenly dispersed in the horizontal plane ............... I tried the slot idea with the Karlflex and indeed it really does work , verified by polar measurements, but (for me)this effect was at the expense of knocking down the on-axis upper midrange output by some decibels and the PA310 really didn't have enough extra decibels to spare up in that range in my opinion so I decided to go without the K-slot on mine for the time being .....

The 15LB100 does have somewhat of a rising upper mid response according to it's published curves so it may do well with this sort of pattern control extending it's usefulness out to what looks like about 1.75khz ......... If someone really wanted to experiment with this it would be easy enough to do by first building the Karlflex cabinet according to the plans and then making easy modifications to the front panel (and taking measurements , refining, etc):magnify: ........

I personally plan to revisit the slot idea, because i now have some new tricks up my sleeve that i would like to try.:bulb:

I have planned on using the K15 up to approx 250Hz, for the upper range I'll build some Mini-Karlsonators, 0,53 scale, with 2x TC9FD.

I'm really curious what you're gonna come up next!

Best wishes
Gerald
 
Modern K-variants on top of modern K-Variants!

My shop is definitely indoors, being either kitchen or linig room - and we have a temperature drop down to -20°C forecasted... :mad:

-20 c ! BRUTAL! :cold: Glad you have a way to do this indoors ..


I have planned on using the K15 up to approx 250Hz, for the upper range I'll build some Mini-Karlsonators, 0,53 scale, with 2x TC9FD.

Gerald,
I love the sound of your plan! Be sure to get plenty of pictures to show us during the build and after you are finished building :D .....Would be very cool to see a system made entirely out of modern Karlson inspired speaker designs stacked! The Karlflex and Karlsonator are both modern K-Variants and at the same time are also both series tuned 6th order designs .....

The designs should play well together crossing over at 250hz but don't be afraid to move the crossover point up and down during listening tests to see what filter point sounds best to your ears (200hz , 250hz, 300hz etc) ...

If you are bi-amping then it should be very easy to get a pleasing tonal balance and will surely sound great! :cloud9:
 
-20 c ! BRUTAL! :cold: Glad you have a way to do this indoors ..




Gerald,
I love the sound of your plan! Be sure to get plenty of pictures to show us during the build and after you are finished building :D .....Would be very cool to see a system made entirely out of modern Karlson inspired speaker designs stacked! The Karlflex and Karlsonator are both modern K-Variants and at the same time are also both series tuned 6th order designs .....

The designs should play well together crossing over at 250hz but don't be afraid to move the crossover point up and down during listening tests to see what filter point sounds best to your ears (200hz , 250hz, 300hz etc) ...

If you are bi-amping then it should be very easy to get a pleasing tonal balance and will surely sound great! :cloud9:

As soon as the sawdust has settled, I'll start making fotos :-D Glad you approve of my idea!

One more question:
If I have to mount the K'nator 0.53x flush under the ceiling (not at home, another project), do I have to turn it upside down?
 
There is no shame in the ole Trial & Error

If I have to mount the K'nator 0.53x flush under the ceiling (not at home, another project), do I have to turn it upside down?

I would test both ways to see which yields better results, but yes upside down would probably be the first thing i would try... :usd:

After putting it up there you could play some music over the K'nator and walk around the room while listening to confirm that there is sufficient coverage .......

You can always tilt the cabinet down by several degrees+ if that is what it takes to optimize...
 
Thomann will also ship to the other side of the planet

NOTE:
If any folks in the United States are curious about that 15LB100-8W driver from Thomann you guys should know that Thomann has a very reasonable shipping rate to the United States, as long as the order is under a certain weight limit (at least that was the deal when i checked a few weeks ago) ...
 
Brands unwilling to hype up their calculation

As i mentioned earlier: one of the things that makes the 15LB100 a good value is the fact that it's Xmax figure was calculated using a conservative formula, and when you apply the less conservative formula that many popular brands use currently (like B&C , 18sound , RCF and others) you find that the Xmax figure you end up with for the 15LB100 is very competitive, especially at that price!

There are other brands that also rate conservatively and likewise also offer competitive products that people may have passed up because their figures aren't as hyped.... But you don't have to take my word for it, recalculate and see for yourself! :happy2:


  • P-Audio - (some models)


  • BMS


  • CELESTION
 
Last edited:
OTHER DRIVERS WORTH MENTIONING (bandwidth limited)

There are a few other drivers worth mentioning which i have managed to stumble across in web searches, these were the ones that looked like they had useful parameters so i plugged them into Hornresp for a pleasant surprise:bigeyes: ...


Not all of these drivers are necessarily Karlflex related but all of them model extremely well in some easy to build 6th order and 4th order quarterwave resonant designs and produce great output for the money spent ...:happy1:

The Pro guys may scoff at this because these bass only sub style drivers i will mention here in this post are not PA style drivers :nownow:...... In my defense the only reason i chose to take these particular models seriously is due to the fact that they get great reviews and have the right parameters (according to what is published) .... The level of maximum output in the right cabinet exceeds similarly priced Pro driver solutions in similar sized cabinets by such a large margin that i do believe these are worth taking a look at for DIY sub duty (i will list a few later in this post) .......

Perspective ..

Didn't Danley use MTX subs in some of his early Tapped Horn projects?? How about his Matterhorn eh? :smirk: .. . Danley's work is generally considered to be excellent engineering is it not? ......So why not use decent quality Car Audio style high excursion drivers for reproducing low frequencies? After all it is precisely what they were designed to do, and some do it very well under the right circumstances ....There is no shame in it.... :rolleyes:

What prompted this search?

The Eminence Lab15-4 (special run) driver was built much like an autosound sub and is a solid performer in the 35hz tuned 150L ML-Transflex cabinet that i came up with a while back, UNFORTUNATELY the Lab 15-4 is no longer available , so i went on a quest to find whatever the heck else could be obtained for $180 , lo and behold i found a few options that promise to exceed the performance of the Lab!


So what were my requirements?

I was looking for drivers with an Fs that wasn't too low , and a price that wasn't too high .... Great Xmax, Power handling and Motor Force, along with great reputations ..... T/S Parameters must also be optimal for loading into high performance quarterwave resonant cabinets ...

Enough ranting already, What did i find???


  • American Bass XFL1522
    Can still be picked up from Jet dot com or Ebay for around $180 shipped .... Put one of these into a Tapped Horn simulation or one of my compound loaded quarterwave models and Holy Goats! High SPL subbass potential here ... Great motor force and excursion capability .... Car Audio guys like this brand, good reviews .... I know a fellow who owns this model driver and he can attest to the robust motor on this thing.. He says that he pushes the heck out of it abusively and claims the motor doesn't ever even get warm for him, he hasn't been able to kill it yet despite his persistent efforts :tongue:.... No mercy, and this XFL sub just refuses to die ...... "Tough as an old boot" as they say ..

  • Skar Audio DDX15
    Can be obtained for even less than the XFL driver with similar power handling specs and great motor force + great Xmax yet with less moving mass and a Free-Air resonance (Fb) that is up around 40hz, making this one an option for higher cabinet tuning and extraordinary output :D ...... Keep in mind that these are NOT high compliance drivers (Vas is relatively low) so tuning well below Fb is something i do not recommend with these, even in a quarterwave resonant cabinet design it would be counterproductive (i learned that lesson the hard way last year with my custom Definimax 12s) ....

    NOTE: As with any driver it is advisable to get it's T/S parameters measured (after break-in ideally) to verify figures before designing an optimized cabinet for them...




  • Skar Audio IVX-12v2 D2 (in multiples)
    Can be had for a modest $69 per.... Put two of these in an optimized quarterwave cabinet tuned to 38hz and expect it to outperform any budget class PA 18 out there in terms of subbass output ......

    With the combined displacement (Sd*Xmax*2) and motor force (BL^2/RE*2) it is easy to see why a pair do well for $140 :nod: The current parameters from Skar's website indicate moving mass is relatively light for a high excursion car audio sub, and the stated Free-Air resonance (Fb) is 38hz, making it very useful in my opinion for the practical music range tunings that i personally prefer ........

    My only gripe with Skar is the fact that i am unable to find a 1khz Inductance (Le) figure for these subs, i can find the figure for an old version but not the current revision... I have emailed Skar about the Le but still have no response from them which was a little disconcerting to be honest ... ....

    Hopefully i will hear back from Skar sometime, or maybe i will just have an opportunity to put one of these on my DATS ....


    NOTE:
    Coincidentally a single one of these drivers does model very well in a 60 liter KARLFLEX :D
    For bass only of course ..




  • JBL GT5-10 (used in multiples)
    For those who are weary of the above brands because they aren't familiar with them, well, here is an option from a familiar brand :) JBL is loved, trusted and respected to say the least....

    I must say that the GT5-10 on sale is one hell of a deal guys:checked:... When i checked just the other say you could still get these for $39 a piece on Ebay from Sonic Electronics! That is with free shipping for buyers in the United States ...

    Some friends and i did a lot of modeling with this driver and due to it's characteristics (Fs / Vas etc) i would recommend tunings in the range of 25hz to 35hz-ish (though one could go a little higher if you really wanted) .......

    NOTE: This JBL driver has an Re (DCR) of 4 ohms so is basically a 6ohm driver by PA standards and can be treated as such in modeling ..

    These little guys have a surprising amount of motor strength and excursion for their price..... So if you buy 4 of these you will be getting an outlandish amount of combined displacement (SD*Xmax*4) and combined motor force (BL^2/RE*4) for no more than $160 which is really not much at all considering that this amount of displacement and motor force surpasses that of even some expensive 18" drivers!

    This means more decibels of subbass output per dollar spent .... It means good value .. :)

    Just do the math and compare to 18" drivers that cost twice as much as this set and you will see what i mean ....... Then run the simulations and you will observe with your own eyes that a set of 4 in a quarterwave design tuned in the low to mid 30s delivers the decibels in abundance! :whip:
Lets do a comparison:

  • A set of 4 JBL GT5-10s for $160 total: 1000w rms combined power handling and more combined cone area than any 18" woofer.

    MOTOR FORCE (BL^2/RE*# of drivers) works out as follows:
    15.29*15.29=233.7841 divided by 4 ohms = 58.446 multiplied by 4 =
    233.784

    DISPLACEMENT
    (Cone Area*Xmax*# of drivers) works out as follows:
    SD is 363cm sq multiplied by 14.2mm Xmax= 5154.6 multiplied by 4=20,618


  • Eminence Definimax 4018LF for $250: 1200w rms thermal power handling.


    MOTOR FORCE 26.98*26.98= 727.92 divided by 6.1= 119.33

    DISPLACEMENT 1188 multiplied by 8.57 = 10,181


Interesting right? As you can see above the 4 JBLs have twice as much motor strength and twice as much displacement as the Eminence:bigeyes:, and the JBLs cost less by $90! :bigeyes:

So you may ask: "But Matthew what about efficiency?!?!"

I should point out that the 1w/1m SPL figure that is listed on a PA driver's spec sheet has little (or nothing) to do with performance under 80hz ......... That figure used to be based on a measurement at 1khz if i understand it correctly, though i think some manufacturers are using a different standard now, more like a midrange average which still does not pertain to subbass ......

Ok, and what does Hornresp have to say about this?

An externally hosted image should be here but it was not working when we last tested it.



Hornresp says the JBLs have quite an advantage! :headshot:


One could make the argument here that the JBLs will be suffering from more thermal power compression but will they really? We are talking about 4 voicecoils instead of one here ...........

I could make the argument that the Definimax will begin suffering from dynamic compression at this drive level because it is being pushed to it's linear limits (8.57mm) and the JBLs on the other hand have a few more millimeters to go at this drive level before reaching their Xmax of 14.2mm .. .

NEVERTHELESS to be fair either way if you are pushing the drivers this hard (with either option) on a regular basis you are probably working in a larger room and really are just asking too much of one cabinet and need to build another matching cab for +6db output .....

In a domestic scenario just one of these cabinets with 4 JBL GT5-10s in a typical sized living room will be plenty and will generate 132db (simulated) at 30hz in quarterspace (1pi) or up to 137db (sim) @ 30hz cornerloaded (.5 pi)... That level should surely give most sane folks more than enough intense subbass satisfaction on demand to keep them happy, unless you really are hellbent on violently vibrating the sheetrock loose from your walls within an even shorter amount of time;)...
 
Last edited:
Supplement

Just a few more to add to the list of drivers in the last post .....

Seemingly useful drivers which appear to be great decibel-per-dollar (bass-only) values based upon the PUBLISHED specs:rolleyes:.... So just take it all with a grain of salt until someone actually measures these drivers to verify their t/s parameters .. .:cool:


Skar Audio VVX15v2 - $129
Makes a fantastic replacement for the LAB15-4 (special run) driver in the 35hz tuned ML-TP (ML-Transflex) V3.1 box featured here on DIYaudio earlier in this discussion (from a while back) .....

This driver also works very well in some simpler designs ..



Skar Audio DDX15 - $169
High power handling and strong motor (according to what is published), an Fs value and other parameters that are useful in compound loaded quarterwave designs, Tapped Horns, and even just a simple Transflex, or a TL (QWP) ............ Produces more output (35hz to 80hz range) than Pro style options in this price class (loaded into cabinets with the same internal volume and tuning) .....

Used in relatively small cabinets
(170 liters internal volume or less) with fundamental tunings of 40hz these drivers produce impressive SPL ..



Skar Audio VD12 -- $99

Also seems to be somewhat useful.





NOTE: If anyone would like to see comparisons in Hornresp just let me know and i will gladly provide the screenshots ....:happy1:



REQUEST: If anyone out there already has measured parameters for any of these drivers please do share :yes:
 
Last edited:
Moving along now ... Some updates on Piezos

A while back (in this discussion) someone was asking about square piezo tweeter horns wanting to know if they could be made to work just as well as the GRS piezo elements plugged onto the back of the rectangular Goldwood 1016 horns (Brian Steele style Frankenpiezo.... i had been working on a variant to be driven by a step-up transformer & filter circuit) .... This person had purchased a GRS square piezo, i think it may have been the 1005, and it was terrible .... They measured it and it was really just awful, typical crummy piezo ....

I have since found a square piezo that is definitely worth mentioning!

Ok , so fast forward to just a few months ago (Nov 2016) , i had a bunch of these old Pyle brand PSN1167 piezo horns that previously were part of a crossfiring array attached to the front of a Bill Fitzmaurice designed DR250 horn ....

These piezos had been abused with lots of power :fight: as they were used out at desert parties here in Central Arizona, covered in dust and dirt, many were burnt out:skull: and some were weak:faint: .... There were only a few that were still good, and they measured surprisingly good, VERY good in fact!
:bigeyes: This intrigued me!


So i decided to plug a decent Goldwood element onto the back of one of these PSN1167s and WOW! Great result! Hmmm.. Ok ...... Then i tried plugging a decent GRS element onto the back of the Pyle horn and WOW again! :happy2: These might be better than the Goldwood 1016 horns (but perhaps not quite as much horizontal dispersion, I would have to measure the polar responses to be sure of that statement) ..

Unfortunately, i am not sure how widely available these are anymore... The PSN1167 horns that i have are the old style with the long silver phase plug... The new ones look different and lack the long phase plug (as far as i can tell from the photos) ...

(continued in the next post along with some photos and snippets copied from my project log on facebook)
 
Last edited:
Pyle PSN1167 piezo experiment log exerpts

Here is what my humble but effective piezo testing station looks like when it is set up....

An externally hosted image should be here but it was not working when we last tested it.


"Here (below) are the type of PSN1167, Pyle makes a few different styles of these but the ones i have are the type with the long silver phase plug ..... They seem to have (possibly by chance) done an exceptionally good job at shaping the throat & phase plug and have achieved something wonderful (i dare say even a little better than the Goldwood 1016 horn body)"

An externally hosted image should be here but it was not working when we last tested it.


"Here (below) is what the original Pyle piezo bender element looks like, quite large, even larger in diameter than the modern GRS elements, this was a surprise ...... Between each of these elements their performance and response varied but the good ones seemed to be solid down to around 2.2khz to 2.5khz and were quite hot but there is usually a dip in the 5khz to 8khz range with them ..."

An externally hosted image should be here but it was not working when we last tested it.


"Here (below) is the part that blows my mind: I can take a Goldwood element (or a GRS element) and plug it onto the back of one of these Pyle horns and get response that is even better than what i could get with the Goldwood horn body!! .....

In this graph we have a decent Goldwood element which i have plugged onto a Pyle horn body (black trace) versus the Goldwood horn body (green)... The Pyle body gives us better low end extension , a bit more output between 7k-10k and better top end
1f603.png
=D

I couldn't believe what i was seeing here because i thought the Goldwood horn bodies were the best, so i had to try this again with other elements , and had similar results!"

An externally hosted image should be here but it was not working when we last tested it.



"Here is what i saw when putting a decent GRS element into a Pyle horn body (black trace) versus the GRS element in a Goldwood 1016 horn body ... Once again we have better extension on the bottom , hotter on top and some small amount of midband improvement..."

An externally hosted image should be here but it was not working when we last tested it.



HYPOTHESIS:
"The increases on the top end could partially be due to the profile and dispersion pattern of the Goldwood 1016 ..... The Pyle's pattern is likely just 40 degrees to 45 degrees conical but the Goldwood's pattern seems to be wider in the horizontal plane according to some simple tests i performed (therefore the GW 1016 is spreading the high frequency energy horizontally, and thus a little less output on-axis is to be expected) ..."

However upon more thought i am not sure if the above hypothesis applies to the highest frequencies because these piezo tweeters (like most tweeters and compression drivers on typical horns) are likely to become "beamy" with increasingly narrow dispersion above 10khz or 12khz ....................... More testing is required ...
 
Speaking of

When it comes to these Piezo horn tweeters being driven by step-up transformers (repurposed line-matching transformers of the type typically used for distributed speaker systems) there is something to say about their radically downward sloping impedance response of the piezo which can have an impedance of several hundred ohms to a thousand or more at 1khz and by the time we reach 20khz the impedance has dipped to well below 100 ohms ! .... This is due to the Piezo bender looking (electrically) more or less like a capacitive load ...... Largely capacitive reactance (with some resonant ripple) .... In other words it looks basically like a capacitor to your source which would normally be an amplifier, or in this case our audio transformer ...

An externally hosted image should be here but it was not working when we last tested it.


Someone had once asked me if there was an issue with "reflected power" (i think that was the term used) meaning an impedance mismatch between the transformer and the load (the Piezos, especially an array of them wired as parallel multiples) ..... The simple answer is YES , but with how many piezos and at what frequency?

An externally hosted image should be here but it was not working when we last tested it.



Here is the network that i use basically, this is an old file (I need to clean this up or recreate this thing , my image editing skills have improved a lot since i made this) .... Notice that i made it flexible to accommodate various setups ....

An externally hosted image should be here but it was not working when we last tested it.


With three of the Frankenpiezos paralleled up on the 15w winding on the transformer (333 ohms) and 20 ohms of series resistance added after a 400ohms swamping resistor (the xfmr is being fed with a 2.2uf series cap on the 8ohm winding) i observed that the transformer did saturate somewhat above 15khz , and was noticeable as a nonlinearity when i took measurements at both high volume levels and then low volume levels then adjusted the curves in HolmImpulse to overlap them .... Everything matched up perfectly except for the response above 15khz which fell behind by a few decibels when the system driven to very loud levels ........ So some saturation does occur but really not bad considering the amount of impedance mismatch, the transformer is surprisingly well behaved .................................. Some distortion is surely being created but i imagine it must be falling well above the limits of my hearing, my cats might notice though :tongue: ..

My hearing is not very good above 15khz so i don't mind so much really but ultimately when i permanently install this piezo array into my Electrovoice midrange horn cabs i plan to use a transformer with lower impedance or perhaps i will just parallel up two of these budget 70v 15w transformers to cut the impedance down by half (which should reduce the problems with saturation considerably by providing an improved impedance match) ..................... In that case other components like the series cap and swamping resistor values will need to be adjusted accordingly to get the response flat again ...


....

NOTE:

I noticed the other day that Parts Express is selling a high powered audio transformer for a VERY low price! This is a very good value :) (link below) I haven't tried one of these yet so i cannot say for sure if they are great or not , but i am tempted to buy a few to experiment with ....

https://www.parts-express.com/70v-100w-high-power-speaker-line-matching-transformer--300-225
 
Xfmr , Piezo stuff and talk of updates

Hello Freddi! So good to hear from you!! :cheerful:


What do you figure would be a good piezo load and array for that 100 watt transformer?


That is a good question, I haven't yet bought the 100 watt transformer but it is really tempting being such a great deal ....

With more piezos they could be arranged as a line-array, or as a "splayed" horizontal array ... Maybe a BFM style cross-firing or melded array line-array sort of thing .... I suppose it depends on what the application is, or what dispersion pattern or aesthetic a person is after......

How many parallel piezos can be supported by this 70v 100w xfmr?? I cannot speak directly about the 100w transformer but I can take a shot at it and attempt a guess by extrapolating from what we already know ..... Maybe upwards of a dozen GRS or Goldwood piezo elements in Parallel (depending on how tolerant this transformer is), and of course many more piezos could be employed if using series-parallel wiring .......


If anyone wants to know about my reasoning in detail look below to the text in red ...


When it comes to the 70v 15w transformers i am driving the piezos using the 15w tap which is around 333ohms impedance and driving a single piezo or even two of them in parallel seems to work well enough, but three piezos in parallel (when driven harder) start to show an unacceptable amount of dynamic compression above 10khz and i am assuming this is due to the transformer saturating early which is caused by the piezo's low impedance load seen at high frequencies ........ The piezo presents a capacitive reactance to the xfmr (like a capacitor in parallel with the transformer winding) ...........I swamp some of that reactance out using parallel swamping resistors but that doesn't do much to help the falling impedance above 10khz.... Some small amount of series resistance can be used but excess series resistance only creates a lowpass filter effect due to the nature of the piezo.........

Tricky business right???

Ok now, the 100w 70v xfmr is around 50ohms on the 100w tap, which is about ONE-SEVENTH the amount of impedance of our 15w tap on the other xfmr, so going on our observations mentioned above i would imagine we should certainly be able to drive at least 10 parallel piezo elements of the GRS or Goldwood variety with no problem whatsoever, and if the xfmr is tolerant then perhaps as many as 14 in parallel, or a few more if we are lucky ........ Hard to say for sure until we try it and measure ....

___________________________________________________________


Have you found any current square piezo bodies and elements to match up pretty good ?


Freddi , I haven't yet found any other square piezo body that can outperform these old Pyle PSN1167 horn bodies with the long silver phase plug ....These things really are marvelous (with modern GRS and Goldwood elements loaded in them even) .................... Not sure if the new Pyle 1167 horns are any better or worse , i haven't had the chance to try any yet .... From the photos i can see that the phase plug is entirely different though ... .

I did recently try some of the GRS brand PZ1165 square piezo horns and they are a total disaster .... Just terrible :yuck:............ I understand the GRS 1005 (PZ1005) is also a trainwreck :Ohno: ... ... The GRS 1016 is better but not as good as the Goldwood 1016 ...

I have a bunch of the dual piezo 1177 types as well but i have not been at all impressed with them, they are loud but the response curve is a mess .... I may try damping the heck out of these by stuffing their rear chamber with fine fiber to see if they improve at all but i don't have much hope for them since they are so far off-the-mark to begin with .. . ...


_______________________________________________________

Freddi, i have other updates to post here ....... Good things , i just need to get it all caught up ..... I had a very successful experiment recently with a High-Q low pass filter which provided a nice boost to my upper-mids followed by a sharp cutoff , something that i think might be really useful with K-slotted cabs (in fact i think i will next try this with the PA-310s) .. .:nod: .... Also seems to be highly useful for a more natural sounding integration with piezos!

I also have a couple of exciting low-cost yet high performance driver finds complete with DATS measurements for verified T/S parameters !! :D

A few other odds and ends as well .. . Some photos of successful new designs that i came up with and had built by a talented craftsmen back east , high efficiency designs that compete well as alternatives to FLH cabs in PA work .....
 
how much of an impedance drop re you getting with the upper midrange boost circuit? how far up might it work with average pa woofers?- that could be useful in some K-slotted applications. I've got a big pile of original 1016 and may see if can get an array and step-up to keep up with my midbass horn. Paul Eizik suggested a curved array like Keele - that might be interesting.
 
A High-Q low pass filter on my mid allowed for better integration with piezos

how much of an impedance drop are you getting with the upper midrange boost circuit?

Freddi, the impedance was originally up in the 15 to 20 ohms range around 2khz (without modification) and then after I added the High-Q Low Pass Filter the impedance was pulled down to about 5.5ohms centered around 1.9khz , this was measured with my Dayton DATS unit ..... This was actually an easier real-world load than i thought it would be :D ... The simulation estimated that the impedance would dip down as low as 4 ohms .... I used XSIM software for designing this High-Q low pass filter , and had to create my own FRD and ZMA files for the driver so i could model it in XSIM........

This was for an Electrovoice DL10X driver in a midrange horn (FLH) ... The horn is the SH-1810ER model and it has never had enough presence around 2khz but it did have an inconveniently placed and annoying bump in response around 3khz, so this High-Q filter was designed to fix both of those problems.. Ideally i wanted it to bring vocal clarity forward in the mix while reducing listener fatigue, and it worked :happy2: (some of which may have to do with the reduced overlap & reduced interaction with the piezo array due to the very sharp filter curve above 2khz .... The upper-mids just sound much more smooth and natural now even though the measured amplitude/frequency response of the system doesn't appear to be drastically altered when i compare the before & after curves .. )

I measured the phase response with Holmimpulse using a Dayton microphone and despite our High-Q filter the phase response is still solid, and doesn't swing radically in the range where the boost is occurring .. (this was measured with the EV woofer active only, the Piezo array was disconnected) ...

I have all of the measurement images and XSIM screenshots for this project saved on another computer, i will post them all here tomorrow ...


how far up might it work with average pa woofers?


I guess it depends on the situation .... With the simple filter (only using a series coil and parallel cap) if the designer moves the High-Q filter much higher up in frequency it can create a peak with a sagging response below the peak .........
:magnify:
Perhaps using more components would allow a broadened effect, pulling the impedance down over a wider range .... Someone would need to take the time to model it and work out the values ..



- that could be useful in some K-slotted applications.

My thoughts exactly :D

I was thinking about next trying this idea on the PA310 loaded in the Karlflex, with the filter tuned the same as i did with the EV set (1.9khz-ish) , I can then form the Karlflex aperture into some sort of cool K-slot :cool:

The extra output between 1khz and 2khz would hopefully offset the on-axis reduction seen in the upper-mids when using a slot to control dispersion ..


I've got a big pile of original 1016 and may see if can get an array and step-up to keep up with my midbass horn. Paul Eizik suggested a curved array like Keele - that might be interesting.

A curved array of 1016s driven by a step-up network sounds like a great idea
:nod:
 
Last edited:
that must be a cool cabinet with a 10" driving a horn for midrange - how does your new filter differ from the original which appears to be ~60uFseries with a 6.5mH shunt inductor (that can be switched in/out) ?

would horresp's filter wizard produce ballpark results with XSIM? - or is your filter something that hornresp can't handle - ?

I will probably need some "boost" with a midbass horn - If I can get much as there's a lot of wire from one room to the room with the speaker.

NG3jtqE.jpg
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.