Piezos can be great
Bitsmasher ,
Absolutely, i really admire his melded array concept!.....
What we did is a little different since it is driven by a high pass filter + step up transformer circuit using a cheap $5 70v transformer (re-purposed as a step-up) from Parts Express and a Poly cap from ApexJr , along with a few inexpensive resistors .... The Piezos are also Brian Steele style "Frankenpiezos" which are Goldwood 1016 horn bodies with GRS 1016 elements slapped onto the back ... We also added some damping material to the inside of the rear chamber (i used dryer lint) and some acetone treatment to the back of the horn in order to bond some caulking to the back of the horn body (i used PL premium) which deadens the resonance of the thin plastic that these are made out of ... There is quite a contrast between tapping your fingernail on a deadened horn body versus a stock body, the improvement is easy to hear using the tap test ....
Fun little project with high-fidelity results! One man's trash is another man's treasure if you don't mind the time it takes to polish these turds! The 1016s in their stock form sound like the notorious trademark piezo "nails-on-a-chalkboard" shriek, shout and spit but with some generous patience, love and proper treatment they become something surprisingly smooth and pleasant, even enjoyable to listen to 🙂
By ear i cannot even tell that these are piezos anymore, and i am not using any EQ whatsoever!😀
Bitsmasher ,
Absolutely, i really admire his melded array concept!.....
What we did is a little different since it is driven by a high pass filter + step up transformer circuit using a cheap $5 70v transformer (re-purposed as a step-up) from Parts Express and a Poly cap from ApexJr , along with a few inexpensive resistors .... The Piezos are also Brian Steele style "Frankenpiezos" which are Goldwood 1016 horn bodies with GRS 1016 elements slapped onto the back ... We also added some damping material to the inside of the rear chamber (i used dryer lint) and some acetone treatment to the back of the horn in order to bond some caulking to the back of the horn body (i used PL premium) which deadens the resonance of the thin plastic that these are made out of ... There is quite a contrast between tapping your fingernail on a deadened horn body versus a stock body, the improvement is easy to hear using the tap test ....
Fun little project with high-fidelity results! One man's trash is another man's treasure if you don't mind the time it takes to polish these turds! The 1016s in their stock form sound like the notorious trademark piezo "nails-on-a-chalkboard" shriek, shout and spit but with some generous patience, love and proper treatment they become something surprisingly smooth and pleasant, even enjoyable to listen to 🙂
By ear i cannot even tell that these are piezos anymore, and i am not using any EQ whatsoever!😀
Do you have photos of your Frankenpiezos? Are these the budget piezo horns on PE?
I assume the Frankenpiezo are done with 1016 clones - is there any difference in the the 1005 GRS and Goldwood horns?
hey MMJ - got any wood to do a PA310 Karlflex for fun?
hey MMJ - got any wood to do a PA310 Karlflex for fun?
More Franken Piezo information, how i did it
Xrk ,
Yessir , I purchased mine from PE , as Freddi said they are made using the 1016 clones, not sure about the differences in 1005s though, i haven't experimented with those ...
A Brian Steele Frankenpiezo is just a Goldwood 1016 body with a GRS element attached to the back with some extra fibrous material (i used dryer lint) inside and then caulking on the back of the horn body as damping methods ...
So you have to order both brand 1016s and you just end up with some GRS bodies and Goldwood elements left over in the end ..
I took a few pictures of the bodies, preparing them for the new elements , and i will attach a few other photos as well ...
You can see some extend down to 2k , other down to 2.5k and a few down to 3k but other than that the GRS elements are pretty consistent, much more consistent than the Goldwood elements in my experience but the GRS elements definitely play best in the Goldwood horn bodies ..
So this is best-of-both-worlds sort of combination, it just costs twice as much
..
The network i used is very similar to what is shown in the schematic , except i kept all four Piezos in parallel , and i used the "15w" (330 ohm) on the stepped up side ... I also used the 4ohm tap with 2uf with series cap on the low impedance side of the transformer, and with the massive "tilt" that the voltage step-up provides this allowed my piezo array to come in right around 3k which integrated with my mid-horns nicely ....... On the high-Z side i used a 400ohm swamping resistor and a 25ohm series resistor going to the array ....
At loud house party levels i never get more than a half of a volt going to the array , and this array can handle 25 volts! 😀 So this tweeter system is pretty much indestructible on my current 200w per channel amp that i use to power my mids and highs on this rig ...
Here is the audio transformer i used :70V 15W Line Matching Transformer
The "15w" 70v transformer from EDCOR might be a little better if someone wanted to spend the few extra bucks
Xrk ,
Yessir , I purchased mine from PE , as Freddi said they are made using the 1016 clones, not sure about the differences in 1005s though, i haven't experimented with those ...
A Brian Steele Frankenpiezo is just a Goldwood 1016 body with a GRS element attached to the back with some extra fibrous material (i used dryer lint) inside and then caulking on the back of the horn body as damping methods ...
So you have to order both brand 1016s and you just end up with some GRS bodies and Goldwood elements left over in the end ..
I took a few pictures of the bodies, preparing them for the new elements , and i will attach a few other photos as well ...
You can see some extend down to 2k , other down to 2.5k and a few down to 3k but other than that the GRS elements are pretty consistent, much more consistent than the Goldwood elements in my experience but the GRS elements definitely play best in the Goldwood horn bodies ..
So this is best-of-both-worlds sort of combination, it just costs twice as much
..The network i used is very similar to what is shown in the schematic , except i kept all four Piezos in parallel , and i used the "15w" (330 ohm) on the stepped up side ... I also used the 4ohm tap with 2uf with series cap on the low impedance side of the transformer, and with the massive "tilt" that the voltage step-up provides this allowed my piezo array to come in right around 3k which integrated with my mid-horns nicely ....... On the high-Z side i used a 400ohm swamping resistor and a 25ohm series resistor going to the array ....
At loud house party levels i never get more than a half of a volt going to the array , and this array can handle 25 volts! 😀 So this tweeter system is pretty much indestructible on my current 200w per channel amp that i use to power my mids and highs on this rig ...
Here is the audio transformer i used :70V 15W Line Matching Transformer
The "15w" 70v transformer from EDCOR might be a little better if someone wanted to spend the few extra bucks
Attachments
Last edited:
I have been thinking about exactly that Freddi, i already have two mid-high tops for a miniature PA-system that play down to 200hz, and i just have to figure what i want to build for mini-subs ...... They will probably end up being some sort of 50 or 60 liter Transflex/Karlflex sub design that reproduces well from 40hz up to at least 200hz .... Something with 10" or 12" woofers most likely ...
I called around to see if anyone had the SWE-10S4s left in stock and it looks like the supply has completely dried up (i have one more place to check on monday) but as an alternative the PA-310s might work in an offset design ... I have a couple of little 150w class-d amps that i will use on the mini-subs so the drivers wont need an outrageous amount of xmax
...
Still contemplating upon it ...
Last edited:
Here's a before and after response measurement of a GW 1016 with the back swapped with one from the GRS equivalent. Yep that's a +/- 2.5dB response from below 2kHz to above 15kHz. No further damping or other tweaking had been done to the piezo yet. the adding of damping in the chamber tends to smooth the response a little.
Attachments
The Good Doctor
Your new name will be Dr Brian Victor Piezenstein
Your mixed and motley amalgamated creations sound fully fantastic!
These particular parts are a match made in heaven.
Your new name will be Dr Brian Victor Piezenstein
Your mixed and motley amalgamated creations sound fully fantastic!
These particular parts are a match made in heaven.
Attachments
I might be building some DR280s in the near future, but I'm not sure I can justify buying twice the number of piezos.
Assuming you do both the damping inside the chamber and the rear of the horn as shown above, would you be ahead using just the GRS or Goldwood piezos?
I'm thinking the horns might not be so much important in an array, but the lower resonance of the GRS elements would still make a a difference.
Assuming you do both the damping inside the chamber and the rear of the horn as shown above, would you be ahead using just the GRS or Goldwood piezos?
I'm thinking the horns might not be so much important in an array, but the lower resonance of the GRS elements would still make a a difference.
Frankenlink
Zetta ,
The GRS definitely reaches lower than the Goldwood, and are generally smoother.. There is a little bit of a troublesome peak at 10k in the stock GRS , but thats not nearly as offensive as a big peak around 4k like you would get with a stock Goldwood...
To give you more examples of the different modified, unmodified and Franken responses take a quick look at this short discussion, it goes back to when Brian first discovered and measured these combinations :
GRS 2x5 tweeter mods...
Zetta ,
The GRS definitely reaches lower than the Goldwood, and are generally smoother.. There is a little bit of a troublesome peak at 10k in the stock GRS , but thats not nearly as offensive as a big peak around 4k like you would get with a stock Goldwood...
To give you more examples of the different modified, unmodified and Franken responses take a quick look at this short discussion, it goes back to when Brian first discovered and measured these combinations :
GRS 2x5 tweeter mods...
An idea based on personal exerience with these Frankens
Zetta ,
Here is an idea (since it is very worth it to put these GRS elements in the Goldwood horns, they reach lower and become more smooth this way) you can use half as many Piezos as you had originally planned and just do a straight array , but use one of these 5 dollar step-up transformers to boost the array's output ... The transformer also tends to smooth things out further if you can believe that! 😀
You can put all Piezos in parallel this way and the voltage going to the array is only 1 volt or two at most if using a large amplifier ... I am using a 200 watt per channel amp and I have to push the heck out of it (outrageously loud) just to get a half volt going to my 4 Piezos and they have no problem keeping up with my EV midhorn which has a response very similar to an OT12 .... It is extremely high fidelity , no need for EQing here which means maximum headroom ! 😀
How tall is the DR box you are using? You should be able to fit more piezos vertically than i was able to use in my Mid-horn ... 🙂
Due to the "tilt" that the step-up introduces i was able to use a 2uf series cap on the transformer's 4ohm tap , it all came together very nicely ..
Zetta ,
Here is an idea (since it is very worth it to put these GRS elements in the Goldwood horns, they reach lower and become more smooth this way) you can use half as many Piezos as you had originally planned and just do a straight array , but use one of these 5 dollar step-up transformers to boost the array's output ... The transformer also tends to smooth things out further if you can believe that! 😀
You can put all Piezos in parallel this way and the voltage going to the array is only 1 volt or two at most if using a large amplifier ... I am using a 200 watt per channel amp and I have to push the heck out of it (outrageously loud) just to get a half volt going to my 4 Piezos and they have no problem keeping up with my EV midhorn which has a response very similar to an OT12 .... It is extremely high fidelity , no need for EQing here which means maximum headroom ! 😀
How tall is the DR box you are using? You should be able to fit more piezos vertically than i was able to use in my Mid-horn ... 🙂
Due to the "tilt" that the step-up introduces i was able to use a 2uf series cap on the transformer's 4ohm tap , it all came together very nicely ..
Last edited:
what is the range of reflected impedance at the "primary" of your transformer? - IIRC, a piezo with characteristics of a 0.13uF capacitor will have about 56 ohms reactance at 22KHz. For one CTS 1016 piezo, I've used the 5 watt tap of a 25 volt transformer for around 103dB @ 2.83v into the primary and probably had about 56 ohms in series with that piezo. That Frankepiezo graph is great - are the GRS benders bimorph? at 103dB out and 5120Hz sine, 2nd harmonic was -53dB and 3rd harmonic ~ -68dB with a CTS1016A
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Well controlled impedance
Freddi ,
Looks like a good circuit! The 5w tap on a 25v transformer is closely equivalent to the 15w tap on a 70v transformer, they are both between 300 and 350 ohms ... A 1uf to 2uf series cap on the 8ohm tap is also very close to what i used ..... It appears to be the magik formula!
On the mid-high boxes that i will use for my mini-pa i used Frankenpiezos , an array of 4 on each which was total overkill in retrospect.. I wired them in series-parallel instead of fully parallel because i didn't need as much output for the smaller speakers (my big pa was fully parallel) (I also used the 8ohm tap on the small set instead of the 4ohm tap for the same reason, the 4ohm tap was louder than i needed it to be) .... I used the $6 CV 6" midwoofers from ApexJr in those boxes ....
When i put those little boxes on my handy dandy DATS the impedance graph was remarkably flat deviating only by a few ohms (from around 7 to 10 ohms) throughout the entire range of the tweeters ... This is of course nothing like the wild and wacky impedance of the Piezos themselves, so apparently the transformer does a great job of isolating impedances and also acting as a proper impedance interface ....🙂
Oh, and yes the GRS piezo benders look like bimorphs to me ..
Freddi ,
Looks like a good circuit! The 5w tap on a 25v transformer is closely equivalent to the 15w tap on a 70v transformer, they are both between 300 and 350 ohms ... A 1uf to 2uf series cap on the 8ohm tap is also very close to what i used ..... It appears to be the magik formula!
On the mid-high boxes that i will use for my mini-pa i used Frankenpiezos , an array of 4 on each which was total overkill in retrospect.. I wired them in series-parallel instead of fully parallel because i didn't need as much output for the smaller speakers (my big pa was fully parallel) (I also used the 8ohm tap on the small set instead of the 4ohm tap for the same reason, the 4ohm tap was louder than i needed it to be) .... I used the $6 CV 6" midwoofers from ApexJr in those boxes ....
When i put those little boxes on my handy dandy DATS the impedance graph was remarkably flat deviating only by a few ohms (from around 7 to 10 ohms) throughout the entire range of the tweeters ... This is of course nothing like the wild and wacky impedance of the Piezos themselves, so apparently the transformer does a great job of isolating impedances and also acting as a proper impedance interface ....🙂
Oh, and yes the GRS piezo benders look like bimorphs to me ..
Last edited:
hey MMJ - those CV must be great drivers - do you think an array of 4-1016 per channel would be good for home use? - or would timber thicken too much and combing be too apparent? that's a fantastic input impedance with the 70v xfomer.
Definitely
Freddi,
These are great for use at home, it is basically what my most recent arrays have been relegated to... They are mounted into the mid-horns that are part of our old bulky old PA system that is permanently sitting in our living room, we use it for movies and music ....
There isn't any noticeable combing (to my ears) but if a person wanted to you can trim the outer edge off of the 1016s in order to get a closer center-to-center spacing ...
These do sound best on vertical axis of course, but horizontal dispersion is fantastic ... I can walk around the living room with music playing and i hear nothing but extremely consistent coverage as if you had a pair of headphones on, it is almost eery ....
The sound that comes from these arrays is pristine and very enjoyable, but somehow warm and "diffused" is the best way i can describe it......... They have an immersive quality that is attractive and draws you in...
...................................... ... Metaphorically speaking imagine the difference between highly diffused light which is like being bathed in a pleasant glow of soft comforting incandescent illumination
(as opposed to an overly focused glare of a single sharp, nerve grating, headache inducing crummy compact fluorescent light source
) ... 😛
Hope that helps
Rock on
Freddi,
These are great for use at home, it is basically what my most recent arrays have been relegated to... They are mounted into the mid-horns that are part of our old bulky old PA system that is permanently sitting in our living room, we use it for movies and music ....
There isn't any noticeable combing (to my ears) but if a person wanted to you can trim the outer edge off of the 1016s in order to get a closer center-to-center spacing ...
These do sound best on vertical axis of course, but horizontal dispersion is fantastic ... I can walk around the living room with music playing and i hear nothing but extremely consistent coverage as if you had a pair of headphones on, it is almost eery ....
The sound that comes from these arrays is pristine and very enjoyable, but somehow warm and "diffused" is the best way i can describe it......... They have an immersive quality that is attractive and draws you in...
...................................... ... Metaphorically speaking imagine the difference between highly diffused light which is like being bathed in a pleasant glow of soft comforting incandescent illumination (as opposed to an overly focused glare of a single sharp, nerve grating, headache inducing crummy compact fluorescent light source) ... 😛Hope that helps
Rock on
Last edited:
CONE-OUT (DRIVER FACING FORWARD) KARLFLEX !
WHY?
Furthermore, you can choose your "offset" figure by mounting the woofer into the baffle lower down towards the bottom for less offset or higher up on the baffle for more offset .... With the sims you can determine what amount of S2 offset works best for your driver (adjust "L12" in HR's Loudspeaker Wizard, this sets the distance from the center of your driver to the closed end of the path) ...
AND , here is an additional fun and funky feature that is sure to be handy for those who seek sonic perfection: YOU CAN INVERT THE K-APERTURE !! Meaning you can turn it upside down to make the cabinet into more of a parallel tuned bandpass, as opposed to the orientation that you have in The Karlsonator or XKi which are series tuned ....... Some drivers prefer one alignment over the other so you have options to best compliment your driver , you can even make custom terminus/apertures of your own design , and since the front panel can be made conveniently removable for testing and refining you can actually tune in the cabinet's response until you are happy by only altering the front panel (aperture/terminus/mouth) as long as you are shooting for an FB in the range of 30hz to 40hz
One of the examples is the Dual 10" Fosgate drivers in the 60L 40hz tuned Karlflex, these drivers prefer some parallel loading to give them a lift in the midband response ..... The same is also true with the Lab12 and also the SWS12d , but these ones would need to be tuned to 35hz so you adjust "Z" til you get a FB of 35hz and then you can proceed to add the parallel loading (this is difficult to explain so i will just have to make a drawing of what i'm talking about here)
The difference between 40hz and 35hz is only about a dozen-ish centimeters worth of path length at "Z" .....
This cabinet is versatile enough to work well with lots of different drivers, even the cheap MCMs (like the 2981 and the 2992) work better in this cab now and they have a pretty high QTS , while lower QTS drivers like the PA-310 and Lab12 also work well ... This design seems to be somewhat forgiving..
Tomorrow i will try to scale one of these cabs up to 90 liters and try a few 15s in it , until then enjoy the screenshots , graphs , HR files and the sketch .... I plan on adding more detail to the sketch later but this will suffice for today🙂
P.S. Hey Freddi , check out that phase response!!!
P.PS. you may notice a few extra liters in some of these sims, it is from "VRC" used to emulate the inverted Aperture, it is redundant space ... These are still 60 liter cabs..
WHY?- "Cone out" Because not only does it look nice but the small front chamber allows more airspace to be dedicated over to the S1-S2 region, I see it as a more efficient use of space ....
- Cone facing forward with this small front chamber could potentially give you the extreme bandwidth that Karlsons are known for (with the right drivers)..
- Freddi wanted to see how a more traditional driver orientation might work with the Karflex design ..
- Palsa wanted something that performed better with the SWS-12D2.
- Mechanical noise from the driver's motor wont be as audible compared to the magnet-out style cab..
- Might be a good option for those desert dwelling folks like myself who might be concerned with dust or dirt getting sucked into the motor when using them outdoors at a desert party, backyard or park event, or at something like a burning man gathering or whatever where they have that dreadful sodium alkali dust ...
- We are back to all 90 degree angles and repeating measurements throughout, makes for an outrageously fast and simple build!
Furthermore, you can choose your "offset" figure by mounting the woofer into the baffle lower down towards the bottom for less offset or higher up on the baffle for more offset .... With the sims you can determine what amount of S2 offset works best for your driver (adjust "L12" in HR's Loudspeaker Wizard, this sets the distance from the center of your driver to the closed end of the path) ...
AND , here is an additional fun and funky feature that is sure to be handy for those who seek sonic perfection: YOU CAN INVERT THE K-APERTURE !! Meaning you can turn it upside down to make the cabinet into more of a parallel tuned bandpass, as opposed to the orientation that you have in The Karlsonator or XKi which are series tuned ....... Some drivers prefer one alignment over the other so you have options to best compliment your driver , you can even make custom terminus/apertures of your own design , and since the front panel can be made conveniently removable for testing and refining you can actually tune in the cabinet's response until you are happy by only altering the front panel (aperture/terminus/mouth) as long as you are shooting for an FB in the range of 30hz to 40hz
One of the examples is the Dual 10" Fosgate drivers in the 60L 40hz tuned Karlflex, these drivers prefer some parallel loading to give them a lift in the midband response ..... The same is also true with the Lab12 and also the SWS12d , but these ones would need to be tuned to 35hz so you adjust "Z" til you get a FB of 35hz and then you can proceed to add the parallel loading (this is difficult to explain so i will just have to make a drawing of what i'm talking about here)
The difference between 40hz and 35hz is only about a dozen-ish centimeters worth of path length at "Z" .....
This cabinet is versatile enough to work well with lots of different drivers, even the cheap MCMs (like the 2981 and the 2992) work better in this cab now and they have a pretty high QTS , while lower QTS drivers like the PA-310 and Lab12 also work well ... This design seems to be somewhat forgiving..
Tomorrow i will try to scale one of these cabs up to 90 liters and try a few 15s in it , until then enjoy the screenshots , graphs , HR files and the sketch .... I plan on adding more detail to the sketch later but this will suffice for today🙂
P.S. Hey Freddi , check out that phase response!!!
P.PS. you may notice a few extra liters in some of these sims, it is from "VRC" used to emulate the inverted Aperture, it is redundant space ... These are still 60 liter cabs..
Attachments
-
KFLX-12D2-FF.txt
-
KFLXPA310-FF.txt
-
KFLXRF10X2FF.txt
-
OFFSET ML-TRANSFLEX -35hz-60L-LAB12-OR-SWS-12D-24inch-150-cm 40hz or 160cm 35hz Path-Small-Front.jpg -
KARLFLEX-CONE-FORWARD-60L-40hz-Dayton PA-310 -29v-Halfspace.jpg -
KARLFLEX-CONE-FORWARD-60L-40hz-Surplus Fosgate 10x2 -20.5v-Halfspace.jpg -
KARLFLEX-CONE-FORWARD-60L-40hz-Dayton PA-310 -29v-Halfspace-Phase-Response.jpg -
KARLFLEX-CONE-FORWARD-60L-35hz-SWS-12D2 -54v-Halfspace.jpg
Last edited:
the PA310 version should make a good bass guitar cabinet which one could sit upon - (make a Velcro attached cushion) - what does Kappa12A do in the same Karlflex and same voltage? Kappa12 may have a more sharp/louder (?) midrange with tradeoff on the bottom from less xmax.
Last edited: