Go Back   Home > Forums > >

Planars & Exotics ESL's, planars, and alternative technologies

My newest pair of DIY electrostatic panels, advice needed
My newest pair of DIY electrostatic panels, advice needed
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Thread Tools
Old 3rd August 2021, 09:19 AM   #1
silvershadelynx is offline silvershadelynx  Netherlands
diyAudio Member
Join Date: Jul 2016
Default My newest pair of DIY electrostatic panels, advice needed

Hi, last week I finished my fifth pair of electrostatic loudspeakers. (see attachments #5 and #6) I think it's the best effort of my five attempts.

Sensitivity is around 86 dB / 1m / 1 VA / 1 kHz. Outer dimensions are 40 cm x 150 cm. Diaphragm (mylar) thickness is 6 micron and dimensions are 23 x 130 cm. Stator / membrane spacing is about 1,6 mm, a little less than I used in previous models. (about 2,4 mm). I'm aware this limits maximum excursion, but has the benefit of improved sensitivity. I use a custom made 1:125 full range quality setup transformer.

I use loudspeaker screen on the front because otherwise the panels don't fit esthetically with my room. This dampened the fundamental resonance frequency with 5 dB. After fitting a mesh damping screen with a rather low rayle number on the rear stator (on the inner side of the wire stator in order to damp resonances in mid range as well), I got around 2,5 dB more damping, resulting in 7 to 8 dB damping. Resonance frequencies of both panels are 62 and 76 Hz. I guess they will lower a bit after breaking in. I hope they will stabalize around 50 Hz.

I used the software "esl_seg_ui" by Edo Hulsebos to simulate / calculate the electrical segmentation. (see attachements #1)

In the simulation I included the 1 Ohms series resistor between amplifier and audio transformer.

I did extensive listening tests in the room where I built the speakers which is a smaller room than my living room where I put them after they were finished. Bass extension was ok in the smaller room, but in my larger (26 square meters) room at greater listening distance (5 meters from speakers) it feels a bit thin. Not too bad, but I would like to have a bit more lf extension / power.

For frequency and impedance measurements, see attachements #2 and #3.

Instead of building another, larger pair of speakers I was thinking about adding an extra panel per channel which serves as a bass panel to improve bass extension. By putting two panels per channel close to each other airload increases resulting in lowering of resonance frequency. So I'm considering to build another pair with same dimensions and use the same amount of wires (64 per panel) and connect them to the other, segmented panels by adding 1 resistor which makes a low-pass filter. The only possible disadvantage could be that this lf segment is not symetrically on both sides of the main panel, but on only 1 side. Would this be a problem? I made another simulation using esl_seg_ui by adding this new panel by adding one new wire group of 32 wires (2x 32 in this simulation equals 1 new seperate bass panel of 64 wires). See attachements #4.

Included attachements:

#1: segmentation simulation for current panel
#2: impedance and phase measurement in REW, using a custom 1:125 step-up transformer and a 1 Ohms series resistor
#3: frequency measurement in REW with microphone close to front stator (about 3 cm)
#4: segmentation simulation for current + new bass panel
#5: rear side of new panel
#6: front side of new panel
Attached Images
File Type: png esl_seg_ui 4_6_24_5m.png (467.0 KB, 371 views)
File Type: png impedance esl 5.1.png (50.0 KB, 379 views)
File Type: png fq esl 5.1 @3cm.png (40.5 KB, 373 views)
File Type: png esl_seg_ui 4_6_24_32_5m.png (476.3 KB, 359 views)
File Type: jpg 20210803_105054.jpg (160.0 KB, 383 views)
File Type: jpg 20210803_105029.jpg (129.5 KB, 201 views)

Last edited by silvershadelynx; 3rd August 2021 at 09:35 AM.
  Reply With Quote
Old 4th August 2021, 08:12 PM   #2
silvershadelynx is offline silvershadelynx  Netherlands
diyAudio Member
Join Date: Jul 2016
In addition to my previous post, I found an very useful excel sheet by Bolserst to calculate far / near field, and intermediate field. Although I was aware of this sheet, I never seriously applied it. See post below:

Segmented Wire Stator ESL simulator (esl_seg_ui)

Tonight I filled in the dimensions of my latest stats (I used the outer dimensions of the panel, not the diaphragm dimensions), see attachement #1. I used a listening position of 5 meters as this reflects my current situation. In this case the intermediate field ranges from 756 Hz to 10625 Hz. So that would result in a -3 dB/oct slope below 756 Hz, as I used electrical segmentation calculated / simulated by esl_seg_ui. Resonance frequency at around 60 Hz corrects this dip a little bit. Above 10 kHz there would be a +3 dB slope, but LC resonance of audio transformer + panel, and weight of the 6 micron mylar will compensate for this raise in output at higher frequencies.

In order to extend low frequency range, (it's not a big deal, but I would think this would make an improvement because of the increased listening position) I'm wondering what would be the best option, if I want to build a second panel per channel: put it next to the current panel as described in my first post, and use this panel as a bass (only) panel, or just copy the current panel (and it's segmentation) and put it on top of the current panel? In the last case, this would get closer to a line-source.

Widening the panel would result in a graph in attachement #2: the near field (yellow dashed line) moves to the left, resulting in a field transition point at 2656 Hz. So, assumed I apply segmentation as suggested by esl_seg_ui, as proposed in my previous post, I would get a +3 dB/oct slope from 2656 Hz upwards. The -3 dB/oct slope in the far field region doesn't change. To me, this seems not answering my wish to extend low frequency range, but rather boost mid / high frequencies.

Doubling height of the panel, and keeping width the same would result in a graph in attachement #3: this lowers the field transition point (red dashed line) down to 189 Hz and keeping the other transition pont (yellow dashed line) at 10625 Hz. This would increase low frequency range with approximately +6 dB. (two ocatves @ 3 dB/oct: 756/2/2=189Hz)

I'm not sure if I made any errors or forgot things (like core saturation, etc.), but at the moment I would conclude that stacking both panels vertically would be preferable as this would extend low frequency range and would come closer to a true line source?
Attached Images
File Type: png current_situation.png (173.0 KB, 127 views)
File Type: png widening_stator.png (172.4 KB, 49 views)
File Type: png heighten_stator.png (172.8 KB, 44 views)

Last edited by silvershadelynx; 4th August 2021 at 08:21 PM.
  Reply With Quote
Old 6th September 2021, 07:27 PM   #3
bolserst is offline bolserst  United States
diyAudio Member
bolserst's Avatar
Join Date: Jun 2005
Location: Fort Worth, Texas
My newest pair of DIY electrostatic panels, advice needed
Since the spreadsheet is calculating field transition break-points for an unsegmented ESL far away from any room boundary, it can be difficult to draw conclusions for segmented ESLs in a room. To get a better idea of the difference between adding width and adding height, we can compare a theoretical infinite line with a finite ESL line source, as well as a finite line sitting on the floor receiving contribution from the floor mirror image source. I believe the 3rd case is closest to what you have.

Attachment #1: This set of plots shows response for the 3 line source models (ideal, finite, and finite-on-floor) for your original panel, and then with doubling width, or height. For the double width, I used a lower ladder resistance value that you had so as to lift the LF response for the finite-on-floor. For the double height, I hooked the segments together, so same segment widths but twice the height requiring the ladder resistors to be halved. Both doubling width and height can provide improved LF response with little effect on the HF polar response.

Attachment #2: One other thing to look at when evaluating adding height or width, is how the response changes with distance from the ESL. Adding height reduces the variability of the LF response(<200Hz) with varying distance. This is one of the reasons I prefer to add height rather than width. It also happens to take up less floor space. But some people feel the added height makes them dominate a room's decor, and so prefer a shorter, wider panel from an aesthetic standpoint.
Attached Images
File Type: png ESL_add_height_width_01.png (847.4 KB, 41 views)
File Type: png ESL_add_height_width_02.png (489.9 KB, 44 views)
  Reply With Quote
Old 7th September 2021, 10:38 AM   #4
esl 63 is offline esl 63  Sweden
diyAudio Member
Join Date: Aug 2010
Location: Hisingisland
If you attach "wings" to each side of the panel pointing for example 45deg backwards, and they have an extension length same as the 1/2 with of the panel you should get +6dB in the low end. I have experimented with this on esl 63s and the extra bass you get is distortion free, flat, clean punchy extra bass, "for free", extending the further gives some more dB. If you look at bolserst graph at 10 Hz you clearly see which one to go for. If you want clean, good bass response.
  Reply With Quote
Old 14th September 2021, 08:12 AM   #5
silvershadelynx is offline silvershadelynx  Netherlands
diyAudio Member
Join Date: Jul 2016
@esl 63: thanks for your wings suggestion, I will try / experiment by adding some wings to the current stats.

@Bolsterst: thanks again for your help and simulations. Can your excel sheets for these simulations be found on this forum?

I agree on your conclusion that making a panel taller seems preferable when looking at your simulations, although it differs "only" a few dB's.

The only thing I'm wondering is when making the current design taller, let's say about 2.5 meters tall rather than the current 1.5 meters (as 3 meters is a little too tall maybe), capacitance will raise and LC resonance of panel and transformer will lower in frequency. I tested this by connecting 2 panels in parallel to 1 transformer and measured 5 cm in front of one stator and I can see a reduced high frequency roll-off.

In order to keep high frequency range about the same, would proportionally increasing d/s distance compensate for this? (as panel capacitance will be the same as the smaller panel)?
  Reply With Quote
Old 15th September 2021, 03:10 AM   #6
bolserst is offline bolserst  United States
diyAudio Member
bolserst's Avatar
Join Date: Jun 2005
Location: Fort Worth, Texas
My newest pair of DIY electrostatic panels, advice needed
Yes, if capacitance from 2.5m tall ESL is an issue for your transformers, you could reduce capacitance by increasing d/s distance.
Alternatively, you could reduce the segments widths. Or, if you already have DSP in the signal chain, you could use it to flatten the top end.
Not much concern about boosting highs as there is little musical content up there.

Sorry, the simulation spreadsheet has not been posted yet. I have been working on an expanded capability segmented ESL spreadsheet that includes polar maps and everything from transformer parasitic, to room boundary reflections, stator transparency, and damping mesh. But, it is not near user friendly enough to post yet. Your question did remind me though, that I had performed some measurements of a finite line source ESL suspended in a gymnasium to confirm I had calculations working properly.

Attachment #1 is a comparison of theory and measurement for a 64” x12” line source measured at 3ft, 6ft, and 9ft.

While looking for that measurement, I found another one where I had measured the effects of adding wings to an ESL panel.
Attachment #2 shows measurements and calculated increment for adding 6” and 12” wings to a 12” wide ESL.
So you get a nice LF boost, but can start impacting the midrange if taken too far. eAudiostatic baffle step filter
I thought you had tried wings previously and didn’t like the results…but I could easily be confusing you with another diyAudio member.

Attachment #3 shows room average response with and without wings from a while ago.
The LF boost is as expected, and the midrange notch shown in previous windowed measurements is not quite as prominent since its frequency and depth changes with off-axis angle and smooths out in a time averaged room response.

Some years ago, Capaciti posted a similar measurement for his Element 160 with Plexiglas wings. Some questions on ESL dimensions
Although he didn’t post a picture of the ESL, I had saved an image from his website which I included here as Attachment #4 .
  Reply to this post


My newest pair of DIY electrostatic panels, advice neededHide this!Advertise here!
Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Electrostatic speaker panels and supplies in Indianapolis mbcouple Swap Meet 1 23rd October 2010 04:16 PM
Newbie question on suitability of A40 with Shackman electrostatic HF panels njgale Pass Labs 6 1st July 2010 08:20 PM
electrostatic accoustat panels boxcustom Swap Meet 0 13th July 2009 03:37 AM
FS: Jantzen Electrostatic panels $20.00 Zero Cool Swap Meet 7 1st October 2006 08:22 PM
electrostatic panels for sale sasha Swap Meet 8 18th September 2003 08:05 PM

New To Site? Need Help?

All times are GMT. The time now is 08:15 AM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2021 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2021 DragonByte Technologies Ltd.
Copyright ©1999-2021 diyAudio