Am enjoying the 1.16 release so far, particularly having the dbr track, track 
and the change of the color of the lettering to orange. And the Harmonics summary display 
It is great how the software continually is improving. I do have a request for the AMP THD vs Frequency automated test- it would nice to have an option for specified input levels (say 5 max) for the test rather than a start level, stop level and increment. That way it is easier to desired output powers when testing an amplifierā¦This option may be useful for other tests but I mainly use the AMP THD vs Frequency automated test.
Hi @VAR, so youād like to be able to specify āIf the amp under test has an output above 5 dBV, then stopā is that right? What would be the purpose of this?
If you try the PWR THD v Power test, that will let you abort a test early IF the THD exceeds a threshold AND you are above a certain power level. Is it similar to that?
1 Like
Not Exactly. I am looking at a āvintageā Carver power amp right now. I am driving an 8ohm load, but am tapped off it so I am 2.6dB down I know that at -13.5dBv I am getting 5w (all at 1kHz). With .6dBv I get 100w and at 2.4dBv I get 200w. Realizing that the amp is not flat across the band edges, for the most part my traces can be re-named to give the output power level rather than the input signal level, which is more meaningful in this case. It would be nice in this case just to say change the input to -13.5, 2.4 and .6dBv via a listā¦ I looked at the PWR THD v Power test, but I would have to take data at several frequencies and combine the results best I can tell. Maybe it could be automated with TRACTOR, but I have not programmed with itā¦ Thanks for checking into this.
Hi @VAR, OK, so, youād like THD versus Freq at a constant power, where the level into the amp would be adjusted to account for roll-off?
Matt- that would be great if you were able to do that- somehow run a Freq Chirp and use the data from that to offset the power over freq for the thd vs freq test- but that was not what I was asking for. I just would like to have a ālook up tableā for the specific input power level(s) that were specified. That input level equates to an output power for me- usually measured at 1khz in my case- but I am assuming the frequency response is flat for the ampā¦I will try to post a little video later that may explain it betterā¦
Here is a short video segment that explains better what I am doingā¦?
āIt would be niceā¦ā if the name of the settings file that you loaded was displayed on the panel somewhere on the screen, maybe at the bottom where the firmware is listed or the top where the software rev is listedā¦?
Just in time for Christmas Thanks!
Hi Matt,
Thanks for all the nice updates!
I know itās hard with supply issues, Iām dealing with this as well.
One possible update would be option for a Square wave output to look at peaks and ringing. Currently I just change to my regular function generator.
Many thanks and have a good New Years!
Hi @Botte, the square wave output probably needs dedicated hardware. The reason is that the DAC runs through a digital (and analog) filter, and that filter will limit the slew rate of the edges AND it has its own ringing associated with the DAC filter FIR response. So, a squarewave generated from the DAC would have a slew rate far below that of even a lousy opamp.
Is it slew rate you are wanting to look at?
Hi Matt,
Yes I look at it for slew rate and watch the front edge for ringing and overshoot.
Not a problem I can keep using my regular Function generator.
Thanks for the reply.
Dear Matt:
I have a feature request and it is to have the 2nd User Weighting functionality up and runnnig.
This feature was written about about a year or two ago but it has not been implemented yet.
I need this feature to add both a RIAA equalisation file and a QA402 correction file for frequencies up to 90kHz.
Any idea when the 2nd User weighting will be added?
Hi @SigurdR, I agree as I have run into the same thing where Iād like to apply user weighting (for notch) AND also A weighting at the same time. We will try to speed up.
I would love to see a windowing function built into the AMP Frequency Response for doing acoustic measurements. My specific application would be for testing hydrophones, but this of course could be useful for testing both drivers and microphones in air as well. We would just need to enter delay and gate times. I donāt know the physics of it, but I would assume that, if the program were really smart, zero-crossing calculations could be built in to minimize windowing errors. It would be nice to be able to use the right channel for a reference as well. I donāt have the skills to program this. Anybody else ever consider it?
This has been on my wish list for a long time (along with simply being able to set overlays on the main plot window). Recent posts have got me thinking that it doesnāt hurt to ask. Thanks for any consideration of the idea.
Hi @RobbN, how do you normally test a hydrophone? I assume under water. But is it in a large tank, or the ocean from a boat? Iād assume the surface of the water will also reflect sound.
In the frequency response context menu (right click on FREQ button) in GENERATOR group, you can see the following settings for a chirp. This will let you window and also let you use the right channel as a reference.
But the hard part is dealing with reflections. Iāll bet a lot of folks, like me, also donāt quite understand what is involved with a hydrophone measurement and would love a short intro! I see the speed of sound in water is about 4.3X faster than airā¦wowza.
Hi, Matt (and All). Thanks for the interest.
Testing hydrophones is certainly tricky. We contract with the US Navy Underwater Sound Reference Division for calibration. As you might expect, they have several options for testing hydrophones at different bandwidths, but the Open Tank Facility is used most, at the bargain price of $1100/hour. In those tests, they use a tank that is approximately 10 x 10 x 20 meters (if Iām remembering correctly) and use a pulsed constant wave as the excitation source for the DUTs. That is what Iām asking for with this request of putting gating into the AMP Frequency Response automated test. They are basically doing what is included in that test, but are using a custom program that was created in MatLab to account for propagation time and settling time and zero crossing and stopping the acquisition before reflections arrive. Even with a tank of that size, that test is only good down to 1KHz. As youāve noted, the higher speed of sound in water makes those reflections (surface or any acoustic impedance boundary included) come back to DUT pretty quickly.
Currently, I test hydrophones in several ways. We have a 5000-liter tank, wherein Iāll place a reference hydrophone and the DUT on a turntable that is attached to a forked holding fixture that extends down into the water. The two hydrophones are then stimulated with white noise and rotated slowly through a 360-degree arc (to simulate random positioning in the tankābut equal to both hydrophones), while taking continuous averages. This technique does a pretty decent job of averaging out the constructive and deconstructive interference from reflections, but it is not perfect and doesnāt allow for polar response measurements. I will occasionally test from a boat or a pier, but weather and other environmental issues make this challenging. It can be useful to just go to an industrial waterway where there is lots of ambient noise and record the sound in stereo with a reference on one channel and the DUT on the other, then look at the response difference. We can get pretty decent low-frequency response measurements this way. Iāve been trying to figure out ways to test hydrophones for 25 years now and this would get way too long if I elaborate too much. We do test in air for some QC work. When doing so, weāll calibrate a reference hydrophone of the exact same build in water, assuming that response deviations from the acoustic impedance mismatch between air and water will act the same on both hydrophones.
I have known of others that use a log-sweep method in a swimming pool to test response. I tried it myself using ARTA software years ago, but couldnāt get good results. I donāt know why. The pulsed constant wave technique seems to be the gold standard. Using the right channel as a reference is obviously very beneficial for minimizing errors related to projector non-linearity and from ambient noise in the test environment.
1 Like
Thanks very much for this, itās very helpful.
I think the expo sweep is what you want. The first link below shows how to use the plugin āMIC compare to referenceā used to compare two micsāone being a āgoldenā reference mic. This hasnāt been ported to the QA40x software yet, but it would be pretty quick to do. For thee post below, two closely matched Earthworks mics were used.
The math here is pretty straightforward, and there are some pictures in the second link. Letās ballpark and say sound in air travels about 1 foot per millisecond. If your stimulus (speaker) and two mics (reference and DUT) are 3 feet part, and the nearest wall in is 5 feet away, then a wall reflection would take 5 mS to hit the wall, and then another 5 mS to reflect back to the mics. So, 10 mS. But in that 10 mS, the expo chirp has moved on in frequency, and so itās immune to the reflection.
And so, you could pick a window that was, say 7 mS (which is less than 10 mS). That 7mS window gives you a lower bound on the frequency you can measure. Itās an inverse relationship. So, the 1/7mS = 150 Hz.
Now, if you are in water, and everything is moving 4X faster, then the 150 Hz becomes 600 Hz.
So, if you had a setup where the transducer and mic were 3 feet apart, and everything else was much further away (5 feet or more, including the surface of the water), then I think you could get a reasonable measurement down to 600 Hz or so. If you are in a smaller tank, you could start comparing two mics (golden and DUT) at higher frequencies and build confidence there. And once you are convinced on the accuracy, then move to a large tank ($) and lower your frequency.
The big benefit of a chirp is it is really, really fast.
For the low frequencies, why not using the near field measurement as is done for speakers? Then you can splice togethere the NF and FF curves.