HELP required

Hi there All

was looking to find a kindly soul to guide me thro a tutorial on using speaker workshop with the wallin jig to measure and design crossovers
I think I have the necessary tools anyway but am having trouble adapting to this software
Am in leeds UK and willing to travel a reasonable distance but would like to see how it is done and have the opportunity to ask questions as I go


reply via PM if not comfortable posting

TA
 
I can almost help you. I am writing a quick start guide but for a slightly different tool set: SimpeS, Holm, LIMP,EDGE, and PSD-Lite. More of a beginners guide. I started doping this from my frustration learning SoundEasy. There was no clear outline. Tools like HOLM have lots of theory, and no how to. Some of the new ones have no docs yet at all.

As Dave is progressing on WinPCD which may be easier to start on. Jeff's EXCEL version is a bit intimidating until you work your way through it a few times.

My manual, ( 20 pages in WORD) is not ready to post time as a couple of the tools are still getting their kinks worked out.

I guess I could post pieces of it as I go along with the understanding it will change. I don;t really know how or where I will post it when I am done.
 

wintermute

Administrator
Paid Member
2003-08-03 11:43 am
Sydney
I'd recommend as a bare minimum reading through the Speaker Workshop manual http://home.exetel.com.au/wintermute/diyAudio/sw_manual.zip Claudio's web site also has a good tutorial. Claudio Negro's home page, how to project and test a loudspeaker using Speaker Workshop

I've stopped using SW for acoustic and impedance measurements. But I still use the crossover modelling functionality. I use Holmimpulse for doing the FR measurements and REW for impedance export and import into SW. I was having difficulty getting consistent results with SW.

Certainly not going to be practical for any in person coaching with me being in Sydney ;)

Tony.
 

wintermute

Administrator
Paid Member
2003-08-03 11:43 am
Sydney
room eq wizard (but it is much more than that) REW - Room EQ Wizard Home Page

I found that the latest woofers I bought the resonant freq was way off (something like 80hz instead of 60Hz)... figured it might be my measurements and decided to try out something other than SW to do the measurement, I tried ARTA, audioTester and REW. Arta didn't like my sound card much, AdioTester and REW were both good, REW was free so it won :) I got resonant freq of 60Hz with both REW and AudioTester.

Interestingly I then measured the impedance of my 10" Vifa's in their B/R enclosures, and the result I got from SW and the other two was consistent. So it is something about the 4" SB drivers that SW doesn't like. I thought it might be the very light cone, as I got similar high (compared to manufacturers spec) resonant freq on my tweeters.

Tony.
 
Last edited:
Thanks. I''ll check it out. As I have a Woofer Tester II, I use that, but I a looking for free impedance tools ( LIMP is great but expensive) to write up in my cheapskates beginners guide, which is just what the OP needs. Once you get through the process once, then they make a lot more sense. I am trying to help the gentleman doing SimpeS as it too is free.

We have identified several issues with accuracy. I was using a 10 Ohm load resistor as that's what SoundEasy suggested. I found the measurements way off. I now use 200. You can also have issues with the values if the resistors in the probes. There is a bit more on this over on the SimpeS thread I started.
 
What frustration. I actually have a Verizon FIOS account I have never used, but after 5 technical support people, they can't reset my password unless I know what date my service was installed. YEARS ago, how would I know? They have the computer records.

I may have a dormant Gmail account, but so far, all I get is automated replies saying the bogus data I had to enter to get the form to submit was bogus.

These folks need to have some lessons in customer support.
 
Hi!
I have been using many programs with great success for measuring speakers and and audio equpiment.

This is the list of ones I use,

REW v5 Beta 10 as mentioned.
I am not sure if this beta 10 version can still be downloaded as it does allow you to get THD measurments.

HOLMSimpulse,
Great piece of software.

Audacity,
This is the very first software that I started using for frequency response measurement as it does have a decent FFT function built in to it.
It does full duplex as does most of thes programs.
You just create a chirp track and play it back while it is recording the sample from your microphone or out of your preamp,filter or poweramp.
Providing you use a buffer and proper voltage divider so that you don't blow the inputs to your sound card.
This is necessary for all of these programs to work.

Wave Spectra,
This is a simple and very fast Realtime FFT analyser and it is great for analyzing spots in your room realtime using noise signals.
It also makes very accurate FR graphs using the hold function and sweep tones.
It displays realtime THD as well and this program was the very first one I used to measure THD of my speaker drivers.
Note the input level must be as close to 0db you can get in order for the THD data to be accurate.

Visual Analyzer,
This is a super piece of software!
It has everything you would need wrapped up in one package.
Once it is learned it is very easy to use.
The VRLC function is the most important one that I just learned to use.

However there are several versions of this program and some seem glitchty and erratic and some don't, sometimes it will just work perfectly.
It seems to favor WinXP as far as stability.
Version VA2011xe Beta 0.3.2 is the one that I am using in Win7 and appears to be working good.

It does support 192Khz but only in 16bit mode with my hardware.
Occasionally I can get it to work in 24bit mode but it surely crashes every time.
This is confusing as some of the other programs work fine in 24bit/192Khz just fine on my machine.
Also this program is very CPU heavy when it is in VRLC mode and is very sluggish on some slower older machines as it is not multi-threaded and hogs everything that one core can do.
It is still a great program if you can tolerate any sluggishness and make it run properly.
Again this is only while using the VRLC function.

LMSBridge,
This is a RLC metering program and work great ever time I fire it up.
It doesn't have as much of a value resolution the VA offers but it does return the same results as VA.
It is a very straight forward program I can just start it up and go.
The only glitch I have found is that sometime I have found that using it after VA was used my results were off a great deal.
I think this has something to do with the clock of the sound system getting changed and doing a reboot solves this everytime.

SimpleS
This is a super Impedance plotting program I have just done some extensive testing of this program and it works very good.
However there seems to be issues with the newest versions and I am sure that are being dealt with.
I was using the very first version and the only issue I had was with not choosing the proper reference resistor.
If to low of a resistance is used then in loads down the output of the sound card and changes the reference signals level through out the sweep as the impedance of the part being tested changes.
A buffer can help take care of this issue
And if to high of a resistance is used it throws the sampled signal out of the sound cards linear dynamic range and makes it too low to be accurately sampled.

There are many other scope and signal generator programs out there but these are my favorite ones to use and seem to work the best.
I have a scope so this function is nill to me.
It is the FFT function that I have sought after.

Wave Spectra has a standalone signal generator that is very nice.
I also like the one that is on the Tolvan Data website as well as many more.
That would have to be a whole another list.
Lots of times I will just use a chirp track from Audacity or the signal generator that is built into Winlsd.
Depending on what I am doing and then again you can't help twisting your arm off on the dial of a standard standalone generator as well.

The key to making these programs work is using a properly buffered interface especially on the inputs to the sound card.
I have found that the impedance of the input of most sound cards change with frequency and this will throw your measured data way off.
Some of these programs may not even work at all without it being buffered as I have found out the hard way.

Here are some tips of how I built mine here,

A TEST JIG FOR FINDING ESL STEP-UP TRANSFORMER PARAMETERS

And for measurements of higher voltages such as poweramp outputs and stepup transformers and such here,

A TEST JIG FOR FINDING ESL STEP-UP TRANSFORMER PARAMETERS


All of these programs are free and work great!!!
The links to them can be found in this forum and I think I have reposted all of them recently in this forum in various threads and in the ESLDIY forum as well as they are all the center of my testing arsenal right now.

Enjoy !!!!

jer :)
 
Yes, this is something that has crossed my mind several times.
In the past I have found some that perform better than others and I was quite surprised by this.

I have seen some with great UI's and performed terribly and some that produced a great signal with very little to offer in the form of control over the signal.

I don't have a standalone THD analyzer to do a proper comparison.
Although I do have three good quality sound cards to work with and every one of them produces a better signal than my standalone signal generator.
Except of course when it comes to square waves.
Nothing beats a good ole discrete switching circuit when it comes to those.

I was just thinking last night how I have been wanting and really need to build a signal generator for audio that will give a stable programable frequency resolution of at least .1 hz and even .01 hz.

I know that there are some software versions that do or may allow this but typically I have found that they are usually off by a certain small percentage or don't allow for a full small step such as 1/10th or 1/100th of a hertz as this has to do with being a fraction of what the sample rate would allow.

I will consider this study sometime in the future as right now I have my head buried in coming up with a good DIY ESL step transformer.


Cheers !!!

jer :)
 
Yes, this is something that has crossed my mind several times.
In the past I have found some that perform better than others and I was quite surprised by this.

I have seen some with great UI's and performed terribly and some that produced a great signal with very little to offer in the form of control over the signal.

I don't have a standalone THD analyzer to do a proper comparison.
Although I do have three good quality sound cards to work with and every one of them produces a better signal than my standalone signal generator.
Except of course when it comes to square waves.
Nothing beats a good ole discrete switching circuit when it comes to those.

I was just thinking last night how I have been wanting and really need to build a signal generator for audio that will give a stable programable frequency resolution of at least .1 hz and even .01 hz.

I know that there are some software versions that do or may allow this but typically I have found that they are usually off by a certain small percentage or don't allow for a full small step such as 1/10th or 1/100th of a hertz as this has to do with being a fraction of what the sample rate would allow.

I will consider this study sometime in the future as right now I have my head buried in coming up with a good DIY ESL step transformer.


Cheers !!!

jer :)

Just watch e-bay for an HP or Wavetech. You should find something for under ten grand. :D
 
A software generator is only as good as the systems clock at best. Pro studio stuff you can sync with a rubidium clock. They are only about $1600.

.01Hz. Is that at 1Hz, so 1% accurate or at 20Khz? You are not going to do better than a couple of Hz at 20K unless you greatly up the sample rate. You could do this with an external high speed clocked D to A. Make it USB as you are only sending control signals to it.

In reality, how good does it need to be for audio work? You may have a idea I have not thought about, but I can't imagine what you would need resolution below 1% in frequency. Low distortion and a constant amplitude are probably more important, and harder. Using a full 24 bits and 96K would help. Most freeware tools are 16 bit and 48K. 192? right. Of course, unless you have an analog preamp to deal with the amplitude, the 24 bit goes away real fast right along with the low distortion. Maybe a pure sine wave would integrate OK at lower resolution if your sound card has teflon caps in it.

What might be cool is a way to calibrate it. Say put out a 10K tone and hook up a counter. Enter the reading and let the software figure out the adjustment. Do it again at 100Hz to be sure it scaled. Same thing on the amplitude. Output at 60 Hz ( so a DVM can measure it) and enter the actual at-load. A sound card has a pretty high output impedance though, not like a good bench generator.
 
Most likely a resolution to 1Hz would be good enough for most work.
But to at least to 1/10 would be nice.

You can enter in values to the 1/100 in VA but I have no way of verifying its accuracy as my frequency counter is only good to 1Hz resolution.

Wave spectra's signal generator and Tolvan data's has a very fine adjustment of the frequency and I like this and I can see the results using my scope.

But if try to use VA to verify this the best can get it to work is about 1.5Hz resolution or so and the display gets stuck while the frequency is some where in between.
The only other options are some odd numbers as .73, .37, .18 or .092 hertz increments and rarely can I get one of these ranges to work.

I guess it is just that my Cheapy meter has gotten me so mad that I realize that it is not that hard to build your own test equipment that has a decent amount of precision to it for the cost of a real good and decent meter.
Especially if you have the parts needed lying around in your parts bin for many years as I do.

If I can come up with a variable 0V to 13.8KV bias supply with control over the voltage of better than 1volt of resolution in one continuous range it can't be that hard to do.

In fact I just found my old XR2206 signal generator board that I built ages ago and all I need to do is add a PLL and some divider's.
I even found my ICL8038 chip should I want triangle waves with a variable duty cycle.
Even these are good but it is hard pressed to beat the low THD of a 24bit 192Khz sound card.

When you using the stuff to measure the values of components the resolution is a big deal especially at the lower frequency's.

At 100hz if your off by 5Hz are calculated values will be off by 5% plus what ever the tolerance of the part may be.
At 50Hz this would mean 10% and this could be a big deal depending on what you are working on or trying to do.

For instance the other day I was measuring some core saturation points on a transformer and when I got to 330hz everything was looking okay and by the time I got to 327hz the core was in complete saturation.
This is a only difference of .9% shift in the frequency.

VA has been a god send when it comes to measuring voltages at different frequency's as well.
But you must use an input buffer as the input impedance to the sound card changes with frequency and loads the circuit you are trying to measure differently giving you the same false reading as your typical cheapy DVM will do.

This is not good when you are trying to work in audio and when your meter says 1volt at 60hz is coming out of your amplifier and then .5V when you shift up to 1Khz or 10Khz when your scope clearly shows that the level has not changed and is in fact still 1Volt!

Depending on your scope you may not be able to get a voltage measurement accuracy of better than 10% and thus the need of some type of reliable measuring tool or system.
So far the software tools are working very good for me in this aspect.

For D.C measurement I am going to have to build an A/D interface eventually.
My recent attempts to bypass the input capacitor on the sound card has failed.
The input pins have there own D.C bias on them and the inputs will only respond to A.C. levels.
Even if I shift the DC bias with another voltage or by dragging it down with a resistor it self centers on the VA display while the input DC voltage doesn't change and still only responds to an AC signal.
However I have been able to get to about 1hz or less but it always self centers when the level is not changing.
I had got some extremely clean 10Hz and lower square waves into the input of the PCM chip by using a larger input capacitor but it was still only AC.

jer :)
 
Last edited:
"Most likely a resolution to 1Hz would be good enough for most work.
But to at least to 1/10 would be nice."

1Hz at 1Hz is 100%, at 20KHz it is .005%, which you are not going to do with a 48K sample rate let alone clock jitter.

Generating a signal is not the same as measuring it and displaying a value. Remember in digital signals processing, there is nothing between the points. THat requires the integrator and anti-aliasing filters.