Can anyone shine a light on this ?

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

I am still playing with the planars. and i run into the following problem.


I have a small panel but at a certain volume i get some weird nasty distortion.

in the picture i measured one panel that has an openarea of around 10% single ended. plays fine. but when i add a backpanel so it works in push pull i gain 6 dB as expected. but this backpanel has lower open area around 4%. what i notice is that above 300hz all seems normal gain 6dB , but below 300 the distortion skyrocket upwards. particular 3rd harmonics. i did not change input volume, so it is not my amp. neither is it hitting the stator/magnets. you can also see at resonance it did not increase much is sensitivity.

when i put more power in the single ended version i run into the same problem be it slightly higher SPL then the push pull version

any one have an idea ? i would love to get that 6dB extra, but not when i get so much distortion for free.

few things i think it might be but i am not sure at all (hence the help needed)

- when in push pull there is not enough open area to let mylar move free and starts to self resonate?

- tension is so high on the foil and the magnets cant push it any further ?
(although i had panels with very low tension that did the same weird sounding distortion down low, they as well used the 4% open area)

hope someone experienced something like this ? if it is an open area problem this might occur in esl to.
It might save me allot of time.


RED - push pull (front plate 10%open area , backplate 4%open area)
BLUE - Single ended (front plate 10% open area, backside open)
 

Attachments

  • distortion planar low.jpg
    distortion planar low.jpg
    282.8 KB · Views: 303
Hi low mass, the panel is rather small 280 cm2. at wich volume i cant say my mic is not calibrated for SPL. but yeah. they could be to small (prety sure) but i wonder why it sounds so disgusting at a certain volume.

here i did some measurements






First Pic Metal VS non Magnetic panel. 4% open area vs 10%

i could not get the metal membrane to stretch as far as the non magnetic :( did it by hand. so there should be slightly less distortion but not going to resolve what we see here :).

Green is non magnetic HPL panel
Red is metal perforated sheet.

No the metal does give 1.5 dB more efficiency so it does increase the field. but at the same time it also screws up the distortion in the lower end.

PIC 2
When the 2 panels are combined we get a mix, distortion goes down everywhere, but there is still a peak from the metal variant.

RED Metal single stator
GREEN non magnetic single stator
BLUE Combined effort of both in series

SPL stays around the non magnetic only,

PIC 3
When i crank it up by 8 dB or something this is the result of HPL frame alone versus metal and hpl in series. at same output SPL. surface area wins, except for that nasty peak formerd by the metal version.

RED is single HPL
BLUE is HPL and METAL in series.
 

Attachments

  • metal vs HPL 5 percent 10 percent.jpg
    metal vs HPL 5 percent 10 percent.jpg
    187 KB · Views: 276
  • hpl metal and both.jpg
    hpl metal and both.jpg
    276.6 KB · Views: 276
  • dual compared to single hpl same spl.jpg
    dual compared to single hpl same spl.jpg
    430.4 KB · Views: 266
someone mentioned, Impedance radiation.

but i am not getting it , it can produce 150 hz but at a certain volume it starts to sound like ****. but there is excursion left in the panel. since hitting the stator sounds really different. this sound like a fart nasty. by adding panels in series helps it is also clear that the metal; perf variant has this problem more pronounced. but at the same time has lower distortion in the upper frequency.

Cant believe someone else never encountered this. or never tried :)
 
Administrator
Joined 2004
Paid Member
I think the hint is in the frequency response graphs? Below a certain frequency the air mass no longer loads the panel properly, excursion goes up (and is probably not uniform across the panel) and with it distortion.

This looks pretty good above a few hundred Hz, so why not high pass above 200Hz, or are you aiming for full range response from this panel?

I can't imagine linear excursion is much over a mm pp.. (Question, I don't know)

I lived with various Magnepans for almost a decade and remember excursion was extremely limited. (And they did not need to move much to couple either)
 
Thx Kevinkr some good explanation !


For this panel excursion limit was 1.6mm (one way) almost as much as a full blown SMGA. with 1,8mm (3.6 pp). just what i had around in terms of spacers

i intend to make big ones just like the SMGA , but i noticed this artifact every time for each panel i made. so i wondered what it was.

trick is go big and with allot of tension on the foil so it seems, the lower it is the more crap it looks.
 
I suspect you are seeing distortions related to strong resonant activity. Flat diaphragms do not behave all that well below about 700 hz and often between 200 and 400 can be the worst. This issue is even more of a problem with small diaphragms and usually to taim them you will see damping added by solid attachment of some kind of cloth to magnet structure to create a resistance to air flow. I believe bolender does this. As well the Wisdom panels do too. Every planer I have ever built needed this resistance to control the very high Q resonance issues below about 300 hz and to tune the main resonance at lower freqs.

Speakers like the Apogees and Magneplaners do the resistance by using no cloth but making the perforated steel and magnet structure closed up enough to give the effect.

I never measured the distortion related to these resonances so take above with a grain of salt, BUT I have done tests with identical diaphragms but with different magnet structures, some very closed up, others very open. Using a stroboscope you could see huge standing wave issues on the ones with very open magnet structure. The more closed in ones showed much less activity AND they sounded fast and clean by compare.
 
Last edited:
Pulling a SLooooowwwwmotion on the planar

playing sweep from 25 - 600 @ 240 frames a sec exported at 24

i wanted to see where it all goes wrong :) not sure if i managed but still is funny :)

resonance of the panel is at 45Hz distortion is High there and goes down to 200hz

on the left is the counter of frequency from the sweep i am not so sure how accurate that is. so i might do a manual once to see what really happens (poor neighbors)

https://youtu.be/_rb5mvXJVjg
 
I suspect you are seeing distortions related to strong resonant activity. Flat diaphragms do not behave all that well below about 700 hz and often between 200 and 400 can be the worst. This issue is even more of a problem with small diaphragms and usually to taim them you will see damping added by solid attachment of some kind of cloth to magnet structure to create a resistance to air flow. I believe bolender does this. As well the Wisdom panels do too. Every planer I have ever built needed this resistance to control the very high Q resonance issues below about 300 hz and to tune the main resonance at lower freqs.

Speakers like the Apogees and Magneplaners do the resistance by using no cloth but making the perforated steel and magnet structure closed up enough to give the effect.

I never measured the distortion related to these resonances so take above with a grain of salt, BUT I have done tests with identical diaphragms but with different magnet structures, some very closed up, others very open. Using a stroboscope you could see huge standing wave issues on the ones with very open magnet structure. The more closed in ones showed much less activity AND they sounded fast and clean by compare.

funy you posted this and i posted a slow motion :) not the best quality ill try to make a better one with more light and higher resolution.

but yeah you are right about the Q and not pistonic behaviour. but one weird thing is the closed one has higher distortion as the more open one. ill try to do this over somewhere this weak maybe we can see more stuff :)


Thanks for thinking with me Lowmass.
 
wen u say closed more distort than open...

How much more distortion and how "closed" is it really?

small diaphragms are much more suseptable to quality control issues like not perfect flatness, not perfect tension etc.

sooo some of the difference u see between your open and more closed units could be just diaphragm tension differences.

A big diaphragm has quite a bit more elasticity and can be tensioned more evenly.

Again just thinkin out loud so really not sure
 
wen u say closed more distort than open...

How much more distortion and how "closed" is it really?

small diaphragms are much more suseptable to quality control issues like not perfect flatness, not perfect tension etc.

sooo some of the difference u see between your open and more closed units could be just diaphragm tension differences.

A big diaphragm has quite a bit more elasticity and can be tensioned more evenly.

Again just thinkin out loud so really not sure



Ok its like this.

Panel 1
magnets are 9.5mm wide then a 1mm gap. of this one mm gap +-90% is open

so around 9% total open area, downside is the panel is 4mm thick and not made of metal compare to the 1 mm perforated sheet.


panel 2 only 40 percent of the 1mm gap is open so around 4% total open area. even lower since holes dont always line up from perf sheet.


both are not that much. a magnepan comes in at around 10% i believe of total open area.

Those maggies are not as simple as they look :)
 
WrineX , can u go into more detail on the design

1- exact length width of diaphragm( dont include frame)
2- material and thickness of film
3- thickness of foil
4- width of foil traces
5- width of undriven film between traces
6- tensioning method
7- type and approx thickness of adhesive


At the moment I am leaning a bit to the idea that its a small panel and thus will has some serious issues below about 700 hz unless it is very carfully tensioned and the diaphragm itself very carfully manufactured. Just a guess at this point however and need more info
 
WrineX , can u go into more detail on the design

1- exact length width of diaphragm( dont include frame)
2- material and thickness of film
3- thickness of foil
4- width of foil traces
5- width of undriven film between traces
6- tensioning method
7- type and approx thickness of adhesive


At the moment I am leaning a bit to the idea that its a small panel and thus will has some serious issues below about 700 hz unless it is very carfully tensioned and the diaphragm itself very carfully manufactured. Just a guess at this point however and need more info

I have to lookup the the size of the panel when im home. i can awnser a few already

2. Mylar 12 micron (chosen for low end duty)
3. I used wire on this panel. diameter 0.31mm aluminium magnet wire (i also used 40 micron 5mm alu tape)
4.if foil is used 5mm or 3. in these measurements only wire was used as used on magnepan.
5. undriven area approx 9.5 mm the width or the magnet (also like magnepan although they used 1/4 inch magnet)
6. by hand (not something i would use in the final models)
7.adhesive used for wires and mylar 3M 30nF contaxct adhesive as used by magnepan.


something i would like to try , but is allot of work but cutting my magnets down 3.8mm (i already did several meters i made a jig for it and works perfectly but lose some of the magnet in the process) and decrease undriven area and as well expand open area on the metal plate by adding more magnets i create more magnet gaps. for strength i might need to use push pull but we will see.




By the way in other posts i was shooting for the high frequency extend, this is another topic and i used other materials like etched foils and such, this topic was mainly about the distortion in the lower department.
 
Last edited:
hmmm well there are a bunch of things going on with those materials that would make a small panel very difficult IMO. 12 micron film is quite stiff and can be hard to tension evenly on a small diaphragm frame. Remember the Maggies may use similar stuff BUT there much longer dimension will likley give a better result. Also the undriven area and the mechanical properties of stiff riggid wire in a short distance(small diaphragm)

I would try some thinner film , 6 micron or less and foil with full width traces. When I was working small diaphragms I would actually put the foil traces for say a 6 inch by 6 inch diaphragm on a film that was about 18 inches square. I would tension on an 15 inch square frame by hand . Then just lay the whole thing down on the 6 by 6 frame. This gave even tension in that small area in center.
 
hmmm well there are a bunch of things going on with those materials that would make a small panel very difficult IMO. 12 micron film is quite stiff and can be hard to tension evenly on a small diaphragm frame. Remember the Maggies may use similar stuff BUT there much longer dimension will likley give a better result. Also the undriven area and the mechanical properties of stiff riggid wire in a short distance(small diaphragm)

I would try some thinner film , 6 micron or less and foil with full width traces. When I was working small diaphragms I would actually put the foil traces for say a 6 inch by 6 inch diaphragm on a film that was about 18 inches square. I would tension on an 15 inch square frame by hand . Then just lay the whole thing down on the 6 by 6 frame. This gave even tension in that small area in center.

Thanks ! lowmass. well i tried foil on smaller panels as well as etched and results are the same. one thing is clear the bigger you go the less pronounced the problem is :) overall. but this could be because of the reason of less excursion so it does not tell much. i really need to try the smaller magnets to see any difference , changing from 12 to 6 micron did not change anything. i tried this one :(
 
Hello everyone
Wrinex, you can use a stroboscopic method to determine if it is a membrane self-resonance problem. I use it when I have similar problems with my tapes and I can always see what's wrong, if the displacements are big enough to be visible. Then correcting the problem is another matter, but at least I know it depends
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.