Tony_J

Measuring T/S parameters using REW

Recommended Posts

Moderator

I've been attempting to use REW to measure Thiele-Small parameters of a driver, with no success so far - entirely failing to get the impedance measurement feature to work, which is a necessary pre-cursor. I'm using the setup described here:

https://www.avsforum.com/forum/155-diy-speakers-subs/1340740-lilmikea-s-diy-impedance-measurement-jig.html

with a slight variation - I had a two-gang switch so used one gang to short out the sensing resistor and the other to disconnect the speaker leads for calibration. So should work exactly as described in the article and in the REW help, but it doesn't - the calibration scan gives a suspiciously good 100% reading, and the impedance scan suggests an impedance in the meg-Ohms range, which is clearly wrong. I am using the current version of REW - V5.20 Beta 8 - so the software should be current. I have tried it using the Behringer interface suggested in the article, and also using a different USB sound interface - same results.

Anyone out there that has attempted to do this, and succeeded? 

Share this post


Link to post
Share on other sites

Tony, I have built one of those jigs to measure impedance vs frequency of drive units in some DIY passive speakers in have built.

To test it, I have measured the impedance of some passive components across the frequency range and the measurements were close to the theoretical values (allowing for component tolerances). Having based passive crossover designs on the drive-unit measurements made with the jig and found the resulting measured frequency response very close to that predicted by XSim, I concluded that the jig was sufficiently accurate for it's intended purpose. I use a Behringer UCA222 USB interface.

I take it you have tested the connectivity with the multi-meter and found it satisfactory, so it's probably something in the REW settings?

I think the article was written with an earlier version of REW. When I used v5.X with it recently, I think I had to play with the left / right source settings in the measurement dialog box, which I think have changed since I first used REW. I can check which ones work for me but  might not have a chance to dig all the gear out until tomorrow night.

  • Upvote 1

Share this post


Link to post
Share on other sites
Super Wammer

I tried using a similar jig and REW, gave up and bought a Dayton DATS, its excellent for measuring T/S parameters. I still use REW for everything else.

  • Like 1
  • Upvote 1

Share this post


Link to post
Share on other sites
Moderator

Thanks Guys. I'm pretty sure that my wiring is correct, so more than likely it is down to the REW settings, but for the life of me I can't see it after several hours of fiddling. Having seen the DATS spec I am tempted to give that a go if I can't persuade REW to work.

Share this post


Link to post
Share on other sites
Moderator
Posted (edited)

Finally figured out what was going wrong. When you plug the Behringer UCA202 into a Win 10 machine, the default driver that it loads doesn't work properly - it doesn't allow you to select the line inputs properly. You can download an ASIO driver from here:

http://www.musictribe.com/Categories/Behringer/Computer-Audio/Interfaces/UCA202/p/P0484/downloads?active=Downloads

Slightly non-intuitive to set it up - in the REW Preferences page you select "ASIO" in the Drivers box, and use the ASIO Control Panel button and the Reload button to select the USB device, then you get to be able to select the right inputs. I find that REW complains of a timing difference between channels, so I have had to swap the inputs to make it go away, which is a pain, but at least it is producing impedance graphs that look sensible and makes a stab at generating a set of T/S parameters. A UK 1p piece makes a useful added weight - 3.5 grammes.

I will probably go for the DATS kit anyway, because the REW setup seems a tad clunky, to say the least!

Edited by Tony_J

Share this post


Link to post
Share on other sites

Glad you have it working, Tony. I was using a Mac so didn't have the same issue.  I suspect the REW documentation might be a bit out of date now as it seemed more straight forward a few years ago. If you get the DATS I would be interested to see some comparisons of results.

Share this post


Link to post
Share on other sites
Moderator

Thanks Robert. I will certainly do that for my own interest, so will post the results here.

Share this post


Link to post
Share on other sites
Moderator

A slightly better driver seems to be found in BEHRINGER_2902_X64_2.8.40.zip, available from various places on Tinternet if you Google for it. 

Share this post


Link to post
Share on other sites
Moderator
On 11/04/2019 at 13:09, bobovox said:

Glad you have it working, Tony. I was using a Mac so didn't have the same issue.  I suspect the REW documentation might be a bit out of date now as it seemed more straight forward a few years ago. If you get the DATS I would be interested to see some comparisons of results.

OK...the DATS gizmo arrived in the post today, and I have been repeating the set of measurements that I took using the REW rig. First impressions - the DATS software/hardware is REALLY easy to use - took me about half an hour to measure 8 drivers, including the time taken to remove 4 of them (2 DC units) from their cabinets. The displays are good too.

I've done a quick-and-dirty spreadsheet to show the comparison. The first four drivers are Tannoy 8" bass units - two pairs, each pair coming from a different Tannoy speaker, but all 4 nominally the same Tannoy driver model number. The 5th and 6th are the bass and treble sections, respectively of a Tannoy 8" DC drive unit, ditto the 7th and 8th. I had not recorded anything other than f(s) for the treble units on REW, and of course Vas isn't measured for them. 

Tannoy 8 inch - comparison DATS vs REW.pdf

Looking at the data, I get the impression that the DATS measurements are more consistent than the REW ones. Whether that is a fair reflection on the accuracy of the two systems or not, the ease of use of the DATS system makes it a clear winner as far as I am concerned!

  • Upvote 2

Share this post


Link to post
Share on other sites

Thanks for posting the comparison, Tony. The Qts figure for Bass 1 with REW looks like a rogue result and I have to agree the DATS results look more consistent.

The fs & Re figures for the two methods are not too far apart. The Vas figures are quite alarmingly different between the two methods and also within the sample. If you do a box design using alignment tables the required box volume is proportional to the Vas figure for any given Qts. One might expect quite audible differences in the bass response from the sample in a given enclosure. Do Tanny publish any info about these drive units?

I think I might consider the DATS for future projects.

  • Upvote 1

Share this post


Link to post
Share on other sites
Moderator

I'd agree with your assessment Robert - the REW data doesn't look good to me either.

I haven't looked up the bass units online yet, but I couldn't find any published data for the DC units last time I looked.

An interesting difference between the two systems was that when doing the second pass to determine Vas, DATS seemed to require sufficient added mass to shift f(s) by at least 25%, so where I was using 3.5g (a 1p piece) added to the cone in REW, the DATS needed 19g (two pound coins) added before it was happy. That may account for some, but not all, of the differences in measurements.

Anyhoo...time to use the data and design some cabs!

Share this post


Link to post
Share on other sites

As I remember, @orangeart organised a seminar with  Dr. Scott Lindgren and he explained that T/S parameters are not constant with respect to signal level, perhaps this may explain some of the discrepancy?

I look forward to reading about the new project.

Share this post


Link to post
Share on other sites
Moderator
Posted (edited)
41 minutes ago, bobovox said:

As I remember, @orangeart organised a seminar with  Dr. Scott Lindgren and he explained that T/S parameters are not constant with respect to signal level, perhaps this may explain some of the discrepancy?

I look forward to reading about the new project.

My impression from the sound level produced during the sweeps is that both systems are using a similar signal level. Certainly, the TS parameters are "small signal" parameters - i.e., intended to be measured within the linear operating region of the drivers.

I will let you know how I get on. The intent is to use the DC drivers (which I used in one of my Kegworth systems) along with either 2 or all 4 of the bass drivers in a 3-way - the DC doing mid bass/mid and treble. Not yet sure whether to go for a conventional ported or sealed box for the bass, or whether to go a bit off piste with a band-pass design, or even an isobaric... A bit of modelling is going to be needed!

One thing that seems clear having measured the tweeters is that my Kegworth system was using a crossover frequency that was pretty much the same as the tweeter's resonant frequency, which wasn't too clever - received wisdom seems to be that it should be at least 2* f(s). So that needs some playing with too.

Edited by Tony_J

Share this post


Link to post
Share on other sites
Super Wammer
Posted (edited)

A bit late to this, sorry.  I spend ages trying to get REW to work before discovering that my laptop at that time, like many others, does not have a stereo mic input.  A dual-mono mic is useless for REW as its trying to measure a different between right and left mic channels.

Edited by sunbeamgls
  • Upvote 2

Share this post


Link to post
Share on other sites
Moderator

Had to redo my measurements because I was using new-style pound coins as known weights & the website I got the weights from was wrong...so the added weight figure I was using wasn't correct :doh:

So, the detailed values generated by the DATS system have changed slightly, but still looks to be producing more believable numbers than I was getting with the REW setup.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.