LMS pros, cons, and alternatives for measuring speakers

Status
Not open for further replies.
Jump to Latest
H

head_unit

Member
Joined 2004
#1
Inherited some LMS and LEAP setups, which I used some time ago.

LEAP I really liked as a simulator...LMS, less so. My application was measuring 2-way car speakers in a not-large space, and simulate crossovers. With MLSSA, I could get good matches between simulation and results. With LMS, not. Basically as I recall, the LMS does not truly measure phase but calculates it via Hilbert transform. That's fine if you have linear drivers and a good measurement...but if you measurement chops off below a few hundred hertz, the calculation freaks out. Also, if there are non-minimum-phase nonlinearities due to acoustic cancellation and such, then I think Hilbert is not valid (?).

Anyway, I would welcome any hints and comments as to:
- How to make good measurements with the LMS
- Alternative systems which make an actual phase measurement, thereby bypassing those problems, and their pros and cons. (I already know MLSSA well-works great, expensive. But what others still exist or have come out newly?)
- Any simpler/cheaper packages than LEAP which number crunch actual magnitude and phase data to simulate crossovers?


P.S. If some think this should be in the "software" forum, apologies, but when I looked there it was basically SPICE...plus I figured the users I want are here in this forum...
 
#2
I use LMS. Mine is the early DOS version (386 days). Works just fine even after all these years. Very happy with the performance. Extremely accurate.

LMS is not a simulator. It does frequency and impedance sweeps. A useful function is the Gated Sweep. With this feature, it is like measuring in an anechoic chamber.

By the way, LMS comes with a Full size AT card. You'll need an ISA slot (vintage motherboard). I think Linear-X has an external case for the newer LMS versions.
 
#3
What do you do with the LMS? (i.e. what are you measuring/designing? What is your application?)

What hints do you have for using this?

I want to import data into LEAP to use it for crossover work, but don't have a big measurement space.
 
#4
LMS basically measures a driver's response in real time. You can measure the frequency response of, say a tweeter with a crossover network in place. Likewise for the woofer. When the two of them are combined as in a 2-way speaker, you will be able to tell if the crossover network is correct.

For more information, click here for a project that uses LMS exclusively in designing a speaker. It is quite detailed, starting with Thiele/Small measurements.

Click here for newer speaker designs.
 
#5
Sorry, I was not clear. I'm familiar with LMS, just haven't used it in a long time.

I'm wondering, if I don't have a big measurement space available, if I can really use it for crossover work-or if the phase-deriving transform will freak out and yield garbage...
 
#6
All you need is 1 meter without any obstructions all round. With a Gated measurement, you should be able to measure down to about 700-800Hz. Below that, you'll need space. Either ground plane or pit.

I believe you can export LMS to LEAP and use the crossover design software there for simulations. I don't use LEAP, nor do crossover simulations, so not much help to you there.

The Minimum Phase feature in LMS allows you do certain things, like merging the port output of a Bass Reflex with a Near Field Response of the woofer. You can also generate the Group Delay of a Low Pass Filter.

Minimum phase plots can look very messy. It is fine if it is simply a phase plot of the impedance sweep.
 
#7
I know this thread is old, but, I stopped using LMS about 10 years ago. I find it isn't very flexible and behind the times. LMS doesn't really offer very much for the money and like has been stated relies on frequency data to produce phase data...also you cannot find true acoustic center of drivers with LMS. I use Praxis and before that CLIO. Praxis can do everything that LMS can do and more. There are other sound card based programs out there that are more advanced than LMS also if you don't mind doing a little work. Praxis isn't cheap either (around $1000), but still less than LMS or MLSSA and you don't need an ISA slot. I have put together a nice very portable set up behind Praxis that all fit into a 1x1x2' case. Everything I need is in that case besides a mic boom and laptop. If you are on a budget look into SoundEasy, but buy John K's guide for $11.
 
#8
Yeah, I like LEAP a lot but LMS not as much. It's nice to actually measure the amplitude instead of calculating, but I hate the assumption that phase is linear. And it's not always convenient to try and do all that splicing and so on.

Isn't there some cheap sound card based thing that will do a quick and dirty basic measurement by now? I thought there used to be something for like $190 but everything seems very expensive...
 
#10
LinearX is working to catch up with the immanent? release of the LX500.

As has previously been stated LMS is painful to use, it looks like the new platform will be a huge improvement.

The old ISA card has real trade in value, like $500 USD trade in value although part of the savings is in not buying the mic again as it is the same M31 used with LMS, but the LX500 is still expensive.
 
#12
I'm sure that Speaker Workshop also measures phase. It gives you the option of subtracting time-of flight. It is free.

Also, SoundEasy is a low-cost high performance software that does measurement, simulation and emulation. It is under continuous development, based on latest audio engineering learnings and user feedback. Link here: Bodzio Software (scroll down for SoundEasy). I highly recommend a read through the online user manuals and other articles available on the Bodzio web site.

Here are some links related to the topic of accuracy of driver measurements and phase:

Driver Model Accuracy and its Impact on Phase
Errors which can occur in the Hilbert-Bode Transform for phase in speakers
Finding Relative Acoustic Offsets Empirically
Crossover Offset Study
Gated response

About drivers not being linear phase: you will pick up in the articles that within their pistonic range drivers are considered to be minimum phase devices. One could use a phase model based on Hilbert transform (to derive minimum phase response) plus relative acoustic offset. It requires some pre-processing to determine an accurate phase model, and part of the process is to confirm that the phase model is accurate, by comparing the simulated response of the sum of two drivers' response to an actual measurement of the two drivers operating together.
 
#13
The other thing I never talked about is how rude Linear X was over the phone supporting their products. I always felt like such a big bother. That is why I didn't upgrade from LEAP 4.6. I know most DIYers look at LinearX like some sort of holy grail but the product line is honestly over priced, not cutting edge, and support treats you like a second class citizen.
 
#14
Well, at one time LinearX WAS the Holy Grail, as they were the only ones who could truly model a crossover from impedance and frequency/phase data.

Now many moons later, I should think much cheaper tools would be available to do the same. I guess it would mean a thread about speaker simulation software…but now I'm into different stuff, and my next speakers will probably be a predesigned-crossover kit type.
 
#16
Modern Crossover Shop and Speaker Shop LMS products are quite sophisticated and the diffraction modeling ability is impresive. I agree that the lack of phase info was a major strike against LMS. I also felt that LMS (Chris) was usually difficult to deal with, but would eventually answer your questions.

Holm is hard to beat for the price (free).

David
 
Status
Not open for further replies.
Top Bottom