Hi,
I've learned more about the concept of "power" than I have in quite some time.
Anyway, re your tweeters and how you haven't blown any while at high volume levels; I used to blow tweeters regularly until I got a pair of "Tweeter Guards" from SpeakerLab in Seattle years ago. I appears that clipping was the cause of blown tweeters. You must have a fairly high-powered amp which gets loud enough for you without clipping.
Maybe the waveform, approaching a square wave, is too abrupt in its change in voltage?
Thanks for a very informative website!
Jon
I've learned more about the concept of "power" than I have in quite some time.
Anyway, re your tweeters and how you haven't blown any while at high volume levels; I used to blow tweeters regularly until I got a pair of "Tweeter Guards" from SpeakerLab in Seattle years ago. I appears that clipping was the cause of blown tweeters. You must have a fairly high-powered amp which gets loud enough for you without clipping.
Maybe the waveform, approaching a square wave, is too abrupt in its change in voltage?
Thanks for a very informative website!
Jon
Different purposes call for different measures, spectrum analysis for some information and just plain instantaneous voltage (like with the LED circuit above) for other purposes.
In so far as the instantaneous voltage gives you an indication of how the cone ought to be peaking (depending on frequency, of course), that may be what a "wattmeter" request is really looking for. Unless you are concerned about the amount of current heating the coil. Or where to put the crossover points....
Ben
In so far as the instantaneous voltage gives you an indication of how the cone ought to be peaking (depending on frequency, of course), that may be what a "wattmeter" request is really looking for. Unless you are concerned about the amount of current heating the coil. Or where to put the crossover points....
Ben
I actually got a patent on a digital wattmeter on an amp. Don't bother to look it up. its not significant. The dancing numbers will drive you nuts. Add some time constants and you can see what the peaks are. Flip the relationship and it will tell you what the instantaneous impedance is. Its still not that useful. A scope in X/Y will tell you a lot more about how close to voltage or current limits you are and how much the phase lag or lead is increasing the power dissipation in the amp. Wort case is zero volta and max current.
Indeed. I work for a company that does a lot of R&D and transducer prototyping and I'm trying to develop a way to effectively measure real power going through a transducer. This means RMS power calculations have to be made from 20 to 20kHz and must work with broad-spectrum noise instead of a single periodic signal. In my initial searching I have not found a product that can do that yet.
You would be better off getting one of these than building one: Clarke Hess #259 https://www.bmisurplus.com/products/48058-clarke-hess-model-259-digital-v-a-w-meter It has wide bandwidth, sensitivity and lots of engineering to make sure its accurate, even at large phase angles. This is the current version Wideband Power Analyzer | Clarke Hess lots more money. More http://rlab.org.uk/mediawiki/images/1/12/ClarkeHess_255.pdf
They are good for DC to 30 KHz. They were poplular for CCFL stuff for measuring power on the HF going to the bulb.
They are good for DC to 30 KHz. They were poplular for CCFL stuff for measuring power on the HF going to the bulb.
I use a Vallahala Scientific 2111A it is good to 50kHz. I use this and several other models to measure AC Power input and the Speaker power. I use 3 or 4 of them many time is the development and debug. They are typ $600-$1000 on Ebay now. New analyzers are about $2k. Be careful when choosing a model as not all have wide bandwidths.
Duke
Duke
I just stumbled upon a module in our Audio Precision 515x called "noise rms" and has a range from 20-20kHz. I'm going to have a word with the folks at AP and see if this will work with large signals. Seem more geared to measure noise floors but you can load custom stimulus files and receive a Vrms value from it.
The APX515 is one of several AUDIO ANALYZERS from Audio Precision. It will measure voltage up to 200vp. I needs a current to voltage device to measure current (Resistor, current transformer, other). It is a good tool for measuring signals from DC to 90kHz, Other models have other features. I worked for AP before moving back to LA.
Duke
Duke
RMS metering of noise- simple and not simple.
Simple, measure the added heating of a resistor by the noise (actually the hardest)
Hard, figure the necessary bandwidth, crest factor level of the noise. Make sure it matches the useable range of your meter. Typical crest factor of meters can be from 3 to 8 at full scale increasing as you go down scale. If the crest factor of your source exceeds that number your accuracy is gone. If the noise bandwidth exceeds that of your meter again your accuracy is gone.
If you are measuring power, particularly with noise, you need to be able to multiply the instantaneous V * A. The phase is critical and with noise and non-linear reactive loads you can't account for it any other way. I don't know if that's something the AP does.
The Valhalla is nice as well (I have one also). Doesn't have the low ranges of the Clark-Hess if they are needed.
Real randomness in noise actually is hard to get. probably not necessary for audio. Some tests call for clipped noise with a max crest factor of 2. GR used to make noise generators with this type of function. There is an EU headphone sensitivity test that uses bandlimited, crest factor limited noise for checking headphones. Its really hard to cheat that test. AP has the noise in downloadable files.
There are a number of standard driver stress tests that use noise. Those are probably what you are embarking on.
Simple, measure the added heating of a resistor by the noise (actually the hardest)
Hard, figure the necessary bandwidth, crest factor level of the noise. Make sure it matches the useable range of your meter. Typical crest factor of meters can be from 3 to 8 at full scale increasing as you go down scale. If the crest factor of your source exceeds that number your accuracy is gone. If the noise bandwidth exceeds that of your meter again your accuracy is gone.
If you are measuring power, particularly with noise, you need to be able to multiply the instantaneous V * A. The phase is critical and with noise and non-linear reactive loads you can't account for it any other way. I don't know if that's something the AP does.
The Valhalla is nice as well (I have one also). Doesn't have the low ranges of the Clark-Hess if they are needed.
Real randomness in noise actually is hard to get. probably not necessary for audio. Some tests call for clipped noise with a max crest factor of 2. GR used to make noise generators with this type of function. There is an EU headphone sensitivity test that uses bandlimited, crest factor limited noise for checking headphones. Its really hard to cheat that test. AP has the noise in downloadable files.
There are a number of standard driver stress tests that use noise. Those are probably what you are embarking on.
I need to sit down and look further into Noise RMS once things settle down. We're half way through our 120 degree mic array install
OK, may look at my Analyzer with PSP (Power Spectrum Bench) to get into Wattage.. Also do the calibration and finally add your load resistor for proper Watt calculations.
Keep in mind to measure noise levels using FFT is somewhat tricky.
use the following settings:
1. Do not use any FFT Window (= none = rectangle)
2. Use the Detector on Average mode (NOT PEAK MODE) ...
Otherwise the level displayed will not accurate, while
on FFT data to FFT display pixel reduction, the level would be using
the peak detector simple to high
Hp
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.