Speakerplans.com Homepage
Forum Home Forum Home > General > General Forum
  New Posts New Posts RSS Feed - Skram vs. C2E - Subwoofers info research
  FAQ FAQ  Forum Search   Events   Register Register  Login Login

Skram vs. C2E - Subwoofers info research

 Post Reply Post Reply Page  <1 4567>
Author
Message Reverse Sort Order
T Willy View Drop Down
Registered User
Registered User
Avatar

Joined: 07 February 2022
Location: Wisconsin
Status: Offline
Points: 43
Post Options Post Options   Thanks (0) Thanks(0)   Quote T Willy Quote  Post ReplyReply Direct Link To This Post Posted: 07 February 2022 at 10:51pm
The reasoning behind the lack of data is due to the DIY nature of it. Some don't have the means or proper knowledge to take measurements. Some don't have the space to do it. Some don't have a working sample built to test. Everyone is at varying stages of knowledge, capabilities, and ability.
Granted there is always someone who has more knowledge and i would be glad to test anything with some guidance of the guru's.
For now, attached is the REW response taken at 1m/2.83v of the Paraflex C-2E Silver Formula loaded with SB Audience Nero 18SW1100d driver. UMIK1 was used as measurement mic. Overlay is the Hornresp model for same enclosure/driver combo using actual T/S parameters pulled from one of my nero drivers using DATS.
Take this data as you wish. Is it perfect or "uniform" test? No, but some of us are working on a knowledge and ability limited circumstance here.

https://www.facebook.com/photo.php?fbid=10220100161723166&set=p.10220100161723166&=3

If the link does not work, please reply. I can find another way to satisfy the archaic means of photo sharing here.


Edited by T Willy - 07 February 2022 at 10:55pm
Back to Top
bee View Drop Down
Old Croc
Old Croc
Avatar

Joined: 14 June 2011
Location: Middlesex
Status: Offline
Points: 4553
Post Options Post Options   Thanks (0) Thanks(0)   Quote bee Quote  Post ReplyReply Direct Link To This Post Posted: 07 February 2022 at 6:38pm
Wow some very good detail in that response. The driver tested was the 18 sound 21 9601c (carbon). The rear projection in all sims is on par with your findings, but in real would we are getting a massive 6 to 9 db drop off, behind, I would love to post why, in truth I have no idea.. a positive gain yes but why... I see your point re 10m distance readings. This can be done. Soon as weather gets a bit nicer, I'll get them back outside and measureLile alot of things, some of the designs are or where done by a few people all over the world, with different measuring kit ECT. The c2a golden was designed and prototyped in the UK, by my self and Mathew's assistance via video links ECT. The plotts need to be updated on the plan, as I did do a revised set, this included a few updates to the tuning of the front chamber. Plans show updates. I could possibly arrange to get some boxes to you, if you would like a play. 
https://www.elements-audio.com
Back to Top
toastyghost View Drop Down
The 10,000 Points Club
The 10,000 Points Club
Avatar

Joined: 09 January 2007
Location: Manchester
Status: Offline
Points: 10919
Post Options Post Options   Thanks (0) Thanks(0)   Quote toastyghost Quote  Post ReplyReply Direct Link To This Post Posted: 02 February 2022 at 9:48pm
Originally posted by snowflake snowflake wrote:

any measurements or model of second and third harmonic distortion in paraflex? this could account for perceived loudness and directivity...


Unfortunately, if the measurement process isn’t correct for magnitude, the harmonic data produced from the Farina sweep method is also questionable. Especially at low drive levels.

In modeling terms, it is possible to derive an estimation of the lower order linear harmonic distortion components from FEA or BEM models, but you need to integrate over a lot of finely spaced data points along a line from the driver to the far field.

There are also time-dependent solvers which can be used to simulate nonlinear acoustics in the newer versions of COMSOL, but I tried to run the example model of a simple, axisymmetric horn (https://www.comsol.com/blogs/simulating-nonlinear-sound-propagation-in-an-acoustic-horn/) and I don't think my desktop PC has forgiven me since.

It is a shame there's not more quantified, repeatable data shared publicly in the pro audio world. My cynical side says that's at least partly intentional, but then I also see the variety of standards and understand a little why practitioners are less likely to pick up a book or paper, let alone a microphone. The folk in the actual engineering jobs seem to understand this stuff pretty well, though.

The upcoming AES standardisation of the M-Noise process should go some way to knocking s few of those barriers down, at least in practical terms.
Back to Top
Ricci View Drop Down
Registered User
Registered User


Joined: 13 March 2017
Location: Louisville, KY
Status: Offline
Points: 69
Post Options Post Options   Thanks (0) Thanks(0)   Quote Ricci Quote  Post ReplyReply Direct Link To This Post Posted: 02 February 2022 at 8:49pm
It would be great to have a few more outlets of in depth performance evaluation style testing of "pro"  subs and speakers, but I can understand why there are not. It is tedious, time consuming and requires a healthy dose of grunt work. It's really only the last 15 years that has seen this type of measurements enter into the mainstream. I find it slightly amusing that the home audio segment appears to be adopting objective data like spinorama and CEA-2010 testing more rapidly than the professional market. 

Back to Top
snowflake View Drop Down
Old Croc
Old Croc


Joined: 29 December 2004
Location: Bristol
Status: Offline
Points: 3122
Post Options Post Options   Thanks (0) Thanks(0)   Quote snowflake Quote  Post ReplyReply Direct Link To This Post Posted: 02 February 2022 at 5:03pm
any measurements or model of second and third harmonic distortion in paraflex? this could account for perceived loudness and directivity...

Edited by snowflake - 02 February 2022 at 5:03pm
Back to Top
toastyghost View Drop Down
The 10,000 Points Club
The 10,000 Points Club
Avatar

Joined: 09 January 2007
Location: Manchester
Status: Offline
Points: 10919
Post Options Post Options   Thanks (1) Thanks(1)   Quote toastyghost Quote  Post ReplyReply Direct Link To This Post Posted: 02 February 2022 at 3:07pm
I actually wanted to do one of the C2E since the size is way more practical for most people, but I was quite put off by the fact that the pinned post on the group had 3 ‘drawings’, all with different dimensions, and the lack of any apparent measurements for comparison or validation.

One thing to consider for anyone interested in measuring subwoofers is that you can actually get useful data even if you’re stuck indoors. Using the principle of superposition, if the sub’s mouth is placed at the very centre of a room, then putting the mic’s capsule right into one of the corners where the walls meet the floor gives you the fewest reflections possible in that space.

It does of course produce a massive boost to the overall response since you’re in 1/8th space, so you have to manually adjust the measured response down by approximately 9 dB for that, as well as adjusting for the distance offset.

The room still has to be quite large, and ideally square, and have solid, dense walls on all sides to ensure that there’s an even distribution of reflections and a relatively easy calculation of the modal response. However, if you can get a distance of >8 metres from the sub’s mouth to the nearest boundary, then you should push the reflection down to well below 30 Hz. Any peaks in the data can be compared to the calculated modes, to see if they can be ignored.

You can also make a simple sealed subwoofer, and measure this to create an inverse calibration curve including the modal response. That can then be applied to any future measurements in that space.

I'm pretty sure most folk have access to a reasonably sized warehouse or barn. You could also try this outdoors, if your nearest car park or yard isn't massive...

The same principle applies when looking for the best subwoofer placement in your house or a venue too. Stick the speaker where you'd normally have your ears, and move the mic to the various positions where you could reasonably put the speaker. Far less effort than shifting the heavy box around!

It is however essential that the charts are labelled properly, and are accompanied by photos or text explaining what you did, what kit was used, how cold and humid it was, etc etc. The boring part where you document the method is as important as the actual process of gathering the data for any acoustical measurement - and a foundation of scientific process, regardless of field.

Funnily enough, this sort of thing is documented as part of the CEA 2010 process

Edited by toastyghost - 02 February 2022 at 3:11pm
Back to Top
sushi View Drop Down
Registered User
Registered User


Joined: 27 March 2012
Location: Milan
Status: Offline
Points: 90
Post Options Post Options   Thanks (0) Thanks(0)   Quote sushi Quote  Post ReplyReply Direct Link To This Post Posted: 02 February 2022 at 2:26pm
Gentlemen, thank you all for you answers and for the data provided. I may not perfectly understand some of those, being me a newbie for what regards audio measurements, but this is for me a big motivation to learn more and achieve a deeper understanding of this matter.
Ricci, thanks for showing up, as well for sharing your good work! I'm currently re-reading the Skram topic on data-bass, couldn't see pics the first time i read it.. that measurement taken by Jay might not be good for a proper and complete review of your sub, but it shows a smooth and even response which is really promising.. i definitely could be building a pair of Skrams in the future, as they seem to well fit my requirements for now.
As regards the paraflex subs, i really like they are "free" plans and get constantly improved through shared knowledge, but i find nowadays some people are kinda taking them for "the best subs ever - period" and contribute to fuel a hype that could illude some non-expert users (like me) that they are just as good as one could dream a sub to sound. A badass look makes things worse..
Pay attention! I'm not saying C2E are not good! I believe they must sound loud and clear since lots of people are happy with them, and i'm really looking forward to hear one personally. But if the outdoor measurements are correct, then 30hz 15db down from 60hz seems no such dream performance for a sub, also that big.. it has a powerful overall bass response indeed, but the graph also suggests the low end is not so loud. It could maybe get better with some eq lowering that large peak around 65hz but, again.. we don't know cause no detailed eq setting are posted with the plans on the fb page, neither real life measuraments, sometimes not even simulations.. I understand many just want to build and play, but that doesn't work for me. I think it would work better if common half-space test setting standards had been given to the builders, and the resultant measurements and eq tips had been posted with the designs. This way it would really be society work, as people could give something back while sharing their own experience. That said, i reckon admins and designers are quite helpful when asked.

Edited by sushi - 02 February 2022 at 2:32pm
Back to Top
kipman725 View Drop Down
Registered User
Registered User


Joined: 02 September 2020
Location: Warrington
Status: Offline
Points: 231
Post Options Post Options   Thanks (0) Thanks(0)   Quote kipman725 Quote  Post ReplyReply Direct Link To This Post Posted: 02 February 2022 at 2:10pm
Thanks for doing all that sim work Kyle, I had suspected all the talk about rear rejection was incorrect as I couldn't think of a physical mechanism beyond the frontal area.  The HOQWs in general is lacking in test data, its disappointing that ASFIK the only people doing proper testing are data-bass and production partner.
Back to Top
citizensc View Drop Down
Young Croc
Young Croc
Avatar

Joined: 16 October 2015
Location: Perth,Australia
Status: Offline
Points: 532
Post Options Post Options   Thanks (0) Thanks(0)   Quote citizensc Quote  Post ReplyReply Direct Link To This Post Posted: 02 February 2022 at 6:37am
Originally posted by Ricci Ricci wrote:

I'd be very interested to see a fair comparison with 2 Skrams since the dual cabs take up the same volume as the 21" paraflex. 


A 900L volume and 800 euro driver budget opens up a lot of options. To be honest, I would like to see it compared to a couple of simple 21 inch reflex with something like the 21PW1400Fe.
https://www.facebook.com/voyager.system

@voyager_soundsystem
Back to Top
Ricci View Drop Down
Registered User
Registered User


Joined: 13 March 2017
Location: Louisville, KY
Status: Offline
Points: 69
Post Options Post Options   Thanks (1) Thanks(1)   Quote Ricci Quote  Post ReplyReply Direct Link To This Post Posted: 01 February 2022 at 9:04pm
Toasty,

Hell of a post you made there. Thanks for the Ted Talk. LOL
I can tell you put a fair amount of time into looking at that. 

A couple short observations if I may...
You are dead on that the low end response will droop by a couple of dB when measuring a front firing sub of this size at 1m vs 10m. These are big subs with a very large frontal area. At minimum a 4m ground-plane measurement would be advisable if 10m isn't possible. Or use the 10m measurement for a cal file on the shorter distance. 

I suspect that the response differences between your sim and the provided measurement are primarily due to a different driver being used in the measurement. Perhaps a 21NTLW5000. I could be wrong but that'd be my quick guess. Also some of the extra upper end response in the measurement would be from cab diffraction. That's a good sized frontal area. You may have taken this into account already and I missed it, 3dB extra gain by 100Hz already wouldn't be out of the question. Combine those 2 possibilities and it may close the gap between the measurement and sim considerably. 


Edit to add a Skram FR measurement. This was taken by a user Jay Michael outdoors at 1m GP in his backyard. Driver is the B&C 21SW152-4. This is the most trustworthy FR measurement of one I've seen yet. I don't trust the SPL calibration of the measurement but the overall FR shape should be close. 





Edited by Ricci - 01 February 2022 at 9:18pm
Back to Top
Ricci View Drop Down
Registered User
Registered User


Joined: 13 March 2017
Location: Louisville, KY
Status: Offline
Points: 69
Post Options Post Options   Thanks (1) Thanks(1)   Quote Ricci Quote  Post ReplyReply Direct Link To This Post Posted: 01 February 2022 at 8:43pm
I was informed of this GTG at Motion Labs a few weeks before it took place. I was told there was going to be a comparison of a few subs with measurements and someone would be building a Skram and was asked if the 21NTLW5000 was a match. It's not ideal for the Skram due to slightly lower motor force than the cab was designed to work with (NSW6021, etc...) and I said as much, but told them it would be serviceable. As far as I was aware all of the other subs were the paraflex type that suddenly seems to be the popular flavor in DIY subs. Unfortunately I never saw any good measurements or data come from this meeting and the only thing I was told is that the Skram was placed differently, built very quickly and barely listened to at all at the gathering because it wasn't as "exciting" as the much larger, I believe pairs of paraflex subs. Not very confidence inspiring for a fair or scientific comparison. I'm not sure what driver, amp, settings were used for the subs and I don't think it was level matched. 

The Skram is a 30Hz tuned 450L net cabinet with a lot of effort put into balancing: Smooth linear response, output capability, low distortion, tuning flexibility to work in multiple applications and manageable size / weight. Scaling the system with multiple cabs to fulfill the output requirements of the app is intended. I intentionally choose not to tune any higher than about 30Hz on any of my subs. I know this gives up some loudness to achieve this type of extension. It'd be easy to tune it to 35 or even 40Hz and gain some sensitivity and loudness in the 40-100Hz bandwidth but it's a tradeoff that I choose to make for being able to really hit that 25-30Hz content when it does show up. I design for the top tier drivers because the extra xmax and power handling can help make up some of the deficit from tuning lower and/or being smaller than a lot of other subs. It's not a philosophy that fits or is cost effective for everyone and that's ok. There are a lot of ways to skin the cat. 

The 21" paraflex appears to be 2X the size at 900L net and with a higher 35Hz corner. 
900L is a huge cabinet. My Skhorn 2x 21" is quite a bit smaller at 680L net.  As we all know bass efficiency is all about air volume / size balanced against extension. Is a 21" paraflex going to be more sensitive and louder than a Skram that is half the size and tuned lower? Yes.
I'd be very interested to see a fair comparison with 2 Skrams since the dual cabs take up the same volume as the 21" paraflex. With equally good amplification, proper level matching, processing, etc... 

+1 on pretty much everything TG has posted in here. We need data!

I'm a hard data kind of guy. Hence my website...
It's crucial to understanding the behavior of a system and understanding the causes of the difference in sound between designs. Part of the qualification / improvement process if you will. I don't put much value on subjective reviews/opinion especially from people who have built the cabinet they are evaluating. 

 I'll eventually get around to testing Skram and Ckram subs. I should've done it long ago but I took a hiatus from DB testing. I'd really like to see an equivalent ground-plane measurement workup on a few of the paraflex designs considered to be the most refined. I'd like to see some popular scoops done as well. Impedance and 1w sensitivity measurements are a start but the interesting bits are what happens with compression, thermal shifting of output and distortion at war volume. CEA style short term SPL capability. Etc. Most designs look great at a few volts. What happens at 50 volts? 100? I want to see where things start to degrade and fall apart. 
Back to Top
toastyghost View Drop Down
The 10,000 Points Club
The 10,000 Points Club
Avatar

Joined: 09 January 2007
Location: Manchester
Status: Offline
Points: 10919
Post Options Post Options   Thanks (0) Thanks(0)   Quote toastyghost Quote  Post ReplyReply Direct Link To This Post Posted: 01 February 2022 at 5:03pm
Pictures tell a thousand words and all that... plus this has been floating about on my hard drive for a couple of months, so might as well share.



This is a coupled lumped element and boundary element model of the Paraflex C2A Golden 21" cabinet in AKABAK3. This is the largest and most performant design as I understand it, a no-compromise version of the loading with outer dimensions of 1.2 metres by 1.2 metres by 0.6 metres.

The PDF plans don't indicate what driver was used, but I believe it was the Eighteen Sound NLW9601-4, so my model uses that as well. The shared traces from CLIO Pocket on the first page are sadly super compressed as part of a single image. I did the best I could with curve tracing tools, to get comparison data.

The base CAD model was drawn by myself from the plans shared on the HOQWS group, using quarter-symmetry on the XY plane. Eight subdomains are used, with the interfaces between each subdomain being meshed at twice the element count of the subdomains on either side. The CAD model has been manually meshed using the Frontal-Delauney algorithm in Gmsh, using frequency-swept element length parameters based on the internal geometry - in particular, channel dimensions were considered versus the element size, to ensure there was no overlap between integration points across the boundaries.

Stuffing was modelled in the corner pieces of the driver's rear chamber, using a measured absorption coefficient curve for a densely compressed piece of dacron.

There are 4,319 elements per quadrant, and an Infinite Baffle plane is used to create half-space (2pi) radiation conditions.

The recommended 21NLW9601-4 driver was modelled via the AKABAK voice coil model of complex, frequency-dependent inductance, using the driver-identification tool in the accompanying VACS software. For those who aren't familiar with this model, it uses exponential functions for resistance and reactance:
Zvc = Re(f) + j·Xe(f)
where
Re(f) = Re·(1 + f/fre)ExpoRe
and
Xe = (ω·Le)r
with
r = (1 + ExpoLe·q2)/(1 + q2)
q = ω·Le/Re

It is still a linear model overall, but this method produces a non-linear behaviour of Z(Le) as frequency increases.

Frequencies from 20 Hz to 250 Hz at 24 points per octave were calculated, for a total of 111 data points on each complex curve. Additionally, a 3-metre x 3-metre field across the ground plane was generated, with a resolution of 4 elements every 0.07 metres.

2V RMS driving level was used, based on the nominal impedance in pass-band being approximately 4 Ohm. This is equivalent to 1 Watt @ 1 metre distance, and matches the reported method on the plans.

Of interest to most people is the axial SPL, and electrical impedance, which I've shown compared to the measured data for qualification.


The electrical impedance tracks well, although it appears there is a slight shift down in tuning. Also, the upper response is reduced compared to measurement - this is likely caused by the lack of absorption from stuffing in the initial chambers of the model, and also the linear modelling method not including higher-order harmonic distortion components.

Another possible cause is that AKABAK3 is a BEM model in the physical domain, so does include effects of diffraction and wavefront distortion around bends and the like.

There's also a chance that the CLIO Pocket mic used for the original measurements was distorting, with the 3 dB crest factor of the log chirp (swept sine wave) stimulus and the CLIO Pocket mic's stated max peak SPL capability of 130 dB linear SPL at 1 kHz. The small plastic electret condensers typically perform worse at low frequencies, and I recently saw some strange behaviour at high signal levels when comparing two 'full-fat' CLIO mics to my own mics in a hemi-anechoic chamber.

If this wasn't just a 'bit of fun', and I had the cabinets to hand to experiment, I'd investigate further to determine what causes this discrepancy. As it is, close enough...


You can see that the electrical impedance dips to just under 3 Ohms at a couple of points between 30 Hz and 120 Hz, even in the measured data. As these are both small-signal data, it indicates that a 2 Ohm capable amplifier with one cabinet per channel would be a good idea for this box.

More importantly, you can see the problem of measuring at 1 metre. The light blue SPL trace tracks the 1-metre measurement very well, but the near-field condition provides an apparent +3 dB boost below 50 Hz when compared to a measurement taken in the far-field (green trace) and scaled back to 1 metre (purple trace). I don't know about you, but on my gigs and installs, only a tiny percentage of the punters are anywhere close to 1 metre away from the mouth of the boxes.


While we're here, one thing that is modelled well regardless of SPL discrepancies is cabinet diffraction. That is the primary effect on directivity, for low frequency enclosures. Since I've heard it said that Paraflex are supposedly very directional, that's worth a look at:


This is a broadband sum of all frequencies from 25 Hz to 125 Hz, looking from above the box at the 'ground'. The same pattern can be observed in the normalised horizontal polar plot - this is normalised relative to the axial response, so removes the SPL discrepancy between model and measurement from the equation:



It looks like there's little to no rear rejection - at least for a single cabinet - but there is a significant edge diffraction toward the lateral directions of +/-90 degrees, toward the rear cabinet corners.

One final bit of fun; stepping through the pressure response pattern from 10 Hz to 250 Hz, across that 3m x 3m plane. The meshing resolution of the observation field isn't high enough to show the entire internal path, so the 'gaps' are just where the cabinet walls are, but you can see the development of the pressure wavefront. That includes frequencies where the rearward radiation from the diaphragm is not in phase with that coming from the series resonant chambers.





Thanks for coming to my TED Talk
Back to Top
 Post Reply Post Reply Page  <1 4567>

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 12.06
Copyright ©2001-2023 Web Wiz Ltd.

This page was generated in 0.109 seconds.