They finally got back to me, and the order is placed. Now, for a question: The product page says that 0, 50, and 75 ohm BNC terminators might be useful for certain types of measurements, but no further detail is given. Can someone be more explicit about what these might be? Excuse my ignorance -- I can see why you might want shorting terminators for SE measurements, but why the others?
I received my QA401 last week and have just started to dig into it. Preliminary impressions are pretty positive. So far I haven't seen any evidence of the dirty USB power issue that others have remarked on here, regardless of the sample rate. I have found however that I have to be a bit careful where I site the unit... it can't be anywhere near my (linear) Bench DC supply, or all kinds of crap gets coupled in.
One oddity so far is that running the 1.50 software, I can't seem to figure out how to access the new graphing feature for sweeps of THD versus frequency and amplitude, or frequency response at various amplitude. The 1.2 release of the manual suggests they should just pop up before being asked to save the data as a .CSV file.. but nothing ever pops up, it just prompts for a file name to save. The basic graphing tool they are using does seem to be functional, as the "View Residual" visualizer will open (although the link in the new manual regarding how to make use of this feature is already broken as a consequence of their website redesign). Anybody else have luck/insight with respect to the swept measurement graphing features? Maybe I'm missing something basic here, I dunno.
One oddity so far is that running the 1.50 software, I can't seem to figure out how to access the new graphing feature for sweeps of THD versus frequency and amplitude, or frequency response at various amplitude. The 1.2 release of the manual suggests they should just pop up before being asked to save the data as a .CSV file.. but nothing ever pops up, it just prompts for a file name to save. The basic graphing tool they are using does seem to be functional, as the "View Residual" visualizer will open (although the link in the new manual regarding how to make use of this feature is already broken as a consequence of their website redesign). Anybody else have luck/insight with respect to the swept measurement graphing features? Maybe I'm missing something basic here, I dunno.
I showed it here for the QA400, not difficult at all. Just go backwards with the calculation and if you don't use the Hann window you need to look up the equivalent noise BW of the window you use (Google).
http://www.diyaudio.com/forums/anal...-single-ended-phono-preamp-8.html#post5058502
Last edited:
Same for me, the 1.5 software release does not meet my expectations. I found another bug a while ago, but I can not remember...
Actually, I'm getting more and more used to the software application and must admit I like it! Today I did a short perfomance-check on an old Yamaha graphic equalizer and the QA401 was very useful.
Still, the stepped measurements fail. Currently with version 1.5 the stepped measurements loop infinitely and the results are also not displayed. If this gets fixed in a future update the QA401 will definately be a very valuable tool.
BTW today I also measured a TPA6120A2 headphone amplifier, which was built several years ago. At that time I had access to a R&S UPL16 Audio Analyzer. The results I receive today with the QA401 are quite similar to my initial measurements
Still, the stepped measurements fail. Currently with version 1.5 the stepped measurements loop infinitely and the results are also not displayed. If this gets fixed in a future update the QA401 will definately be a very valuable tool.
BTW today I also measured a TPA6120A2 headphone amplifier, which was built several years ago. At that time I had access to a R&S UPL16 Audio Analyzer. The results I receive today with the QA401 are quite similar to my initial measurements
V 1.51 out for the QA401
I just noticed that QA has a new V1.51 release of the QA401 software out:
https://quantasylum.com/pages/downloads-1
Looks like they have moved the downloads over to github:
https://github.com/QuantAsylum/QA401/releases
1.51 June 2017
I just noticed that QA has a new V1.51 release of the QA401 software out:
https://quantasylum.com/pages/downloads-1
Looks like they have moved the downloads over to github:
https://github.com/QuantAsylum/QA401/releases
1.51 June 2017
- Ensured that precise stop frequency will always be the final point in both freq and THD test plugins
- Fixed bug where sometimes the option for graph generation wasn't shown. This was due to install
problem where the old DLL wasn't overwritten. - Fixed bug where time domain display could be stuck in log mode. This was evident based on axis.
After a recent mishap with another item of USB test gear I bought a high speed USB isolator, specfically this one... https://hifimediy.com/high-speed-usb-isolator-480Mbps
I have tested the QA400 with this and it appears to work faultlessly.
I have tested the QA400 with this and it appears to work faultlessly.
With the latest Software release the graph display works fine... thumbs up!
Yesterday I trimmed my external filters for Common Mode Rejection. thereby, I started writing a set of python scripts for graph-display of the exported .csv and found another annoying bug (hope QA is following this thread):
Connecting both (hi and lo) inputs of the left channel directly to one unbalanced output with short unshielded cable
(four frequency sweeps of different amplitude ranging from -6dBV to 0dBV in 2dB steps, -20dB attenuator active) yields the following result:
Apparently, the sweep-amplitude info is stored in the first field of each measurement-vector (formatted in rows). But the the incrementation of the amplitude info is obviously wrong. I cross checked several types of measurements and this behaviour is consistent.
Writing my own scripts I was able to extract the step-incrememt and build a workaround but hopefully this issue will be fixed at some point.
Funnily enough, the legend in the new graph display window is correct.
Yesterday I trimmed my external filters for Common Mode Rejection. thereby, I started writing a set of python scripts for graph-display of the exported .csv and found another annoying bug (hope QA is following this thread):
Connecting both (hi and lo) inputs of the left channel directly to one unbalanced output with short unshielded cable
(four frequency sweeps of different amplitude ranging from -6dBV to 0dBV in 2dB steps, -20dB attenuator active) yields the following result:
Code:
Amplitude Data Below
, 100,000, 125,992, 158,740, 200,000, 251,984, 317,480, 400,000, 503,968, 634,960, 800,000
-6,00, -74,910497, -74,903333, -74,882056, -74,895487, -74,956802, -75,078654, -75,016886, -75,049466, -74,981070, -75,021936
-6,00, -74,946204, -72,754605, -72,699879, -72,713824, -72,755583, -72,705500, -72,721308, -72,805351, -72,739302, -72,778954
-4,00, -72,720040, -72,693076, -70,670730, -70,756776, -70,749314, -70,805396, -70,818693, -70,716160, -70,829948, -70,825450
-2,00, -70,763749, -70,826349, -70,763377, -68,928737, -68,989533, -68,999449, -69,037976, -69,021908, -69,083897, -69,004275
Phase Data Below
, 100,000, 125,992, 158,740, 200,000, 251,984, 317,480, 400,000, 503,968, 634,960, 800,000
-6,00, -1,390167, -2,756197, 2,006489, -2,342053, 1,879277, -2,991506, 2,668690, 2,218190, 1,085385, 1,711201
-6,00, 1,504559, -2,993846, 2,722176, 2,034556, -3,127304, 1,868848, 3,030987, 2,847412, 2,939334, 1,051388
-4,00, 2,270425, 1,611547, 3,119275, 2,684132, 2,034030, 2,817030, 1,659726, 2,530558, 2,361600, 2,265396
-2,00, 1,021925, 2,064429, 1,629389, -2,483875, 2,812901, 1,615485, -2,889849, 1,048358, 2,908780, 2,709984Apparently, the sweep-amplitude info is stored in the first field of each measurement-vector (formatted in rows). But the the incrementation of the amplitude info is obviously wrong. I cross checked several types of measurements and this behaviour is consistent.
Writing my own scripts I was able to extract the step-incrememt and build a workaround but hopefully this issue will be fixed at some point.
Funnily enough, the legend in the new graph display window is correct.
ASIO to be added in the upcoming QA401 V1.52
I just spotted a really interesting recent (7/12/17) blog post on the QA site about the QA401. Says they are adding ASIO - finally! - as an experimental option on the upcoming V1.52 firmware and even-later-down-the-road ASIO drivers:
https://quantasylum.com/blogs/news/asio-and-hypermiling-thd
Among other things the QA401 will apparently be able to be used as just a soundcard with ARTA. As he notes that makes for a wonderful double-check on the math in their QA401 software, comparing it vs. the results in ARTA.
He gets into running stats to find out which combination of inputs on the QA401 gives the best (lowest THD) results. On the box he was using it was R- out into R+ and he measures -120.1dB. I had a chuckle at this though: "...it can be easily replicated over and over with some tweaking as the measurement is sensitive to the position on the desk". The underline is my edit.
His is being realistic about that -120.1 dB though. This from the wrap-up at the bottom: "Like hypemiling, the utility of measurements at this level is marginal. You generally want to be a long ways away from the limits of your equipment when making measurements. For example, If you want to measure millivolts, your equipment needs to be able to resolve hundreds of microvolts." So this particular exercise is just to probe the absolute limits of the unit.
I just spotted a really interesting recent (7/12/17) blog post on the QA site about the QA401. Says they are adding ASIO - finally! - as an experimental option on the upcoming V1.52 firmware and even-later-down-the-road ASIO drivers:
https://quantasylum.com/blogs/news/asio-and-hypermiling-thd
Among other things the QA401 will apparently be able to be used as just a soundcard with ARTA. As he notes that makes for a wonderful double-check on the math in their QA401 software, comparing it vs. the results in ARTA.
He gets into running stats to find out which combination of inputs on the QA401 gives the best (lowest THD) results. On the box he was using it was R- out into R+ and he measures -120.1dB. I had a chuckle at this though: "...it can be easily replicated over and over with some tweaking as the measurement is sensitive to the position on the desk". The underline is my edit.
His is being realistic about that -120.1 dB though. This from the wrap-up at the bottom: "Like hypemiling, the utility of measurements at this level is marginal. You generally want to be a long ways away from the limits of your equipment when making measurements. For example, If you want to measure millivolts, your equipment needs to be able to resolve hundreds of microvolts." So this particular exercise is just to probe the absolute limits of the unit.
Last edited:
So in no particular order I have the following questions:
44Khz from Solid State florescent lamps, can we just put a .1uf/600v cap,
or .01uf/600V cap across the mains and ground to suppress the high frequencies?
Would the X or Y cap be suitable for this?
@agdr, pretty good stuff there, I always appreciate your insight.
Now I'm sounding like a Agilent/Keysight/HP publication.
Cheers,
44Khz from Solid State florescent lamps, can we just put a .1uf/600v cap,
or .01uf/600V cap across the mains and ground to suppress the high frequencies?
Would the X or Y cap be suitable for this?
@agdr, pretty good stuff there, I always appreciate your insight.
Now I'm sounding like a Agilent/Keysight/HP publication.
Cheers,
It's best to suppress impulse noise at the source before it's had a change to radiate.
Can you snub right at the lamp? How much of the radiation is RF with inter modulation products?
Mis-wrote that. I meant at the output of the lamp power supply feed.
That circuit has seven in parallel single florescent power supplys and
the 48 inch bulbs. With two five-ways to turn them on and off.
In that circuit, is also the bench circuit. whoop tee do.
Or course I have the cheesy circular florescent with integral magnifier.
I'm sure that one of the causes of HF hash that I see.
Another cause I believe that Demian discusses from time to time
is and out of phase drop. Somewhere else in the circuit.
Finally, there is an intermittent somewhere in the wiring that
no one has ever been able to find. It will rear it's ugly head
at various times and has been known to feed into the seven in series
lighting circuit, even after the breaker was in the OFF position.
I've had a few electricians try to figured it out, some brought their
special meters and sound generators and pick ups that can trace
through the walls etc.
The last time I tried I went so far to pull up the attic decking and
physically wrote numbers on the wiring to trace where they came
from and where they were going to.
I did hum "Paint Your Wagon's" theme song when going through it all.
...Where are we goin' I don't know,
When will I get there I ain't certain,
All I know is I'm on my way."
Cheers,
- required smiley...my kid said so!You will never sort out the problems as long as several are conflicting. Narrow them down to one.
Another issue I have encountered is two separate circuits getting connected at the far end. A real PITA to troubleshoot.
The radiated noise from Fluorescent needs to be addressed at each ballast. The traditional magnetic ballasts are not too bad and a few caps should do it. They will mechanically buzz. They are just large inductors in series and magnetostriction makes noise inevitable. Electronic ballasts are much better but suppressors at the power inlets help. LED replacements are the best of the lot and save money.
You have meters. Measure the voltage drop from no load to "full load" (10A, or 1 KW of load). Check at various points in the chain. If you see a big drop (more that 3%) you have connections that need repair/replace. Aluminium wire is not good and prone to starting fires. Tighten every connection with al wire as they loosen with temperature cycling. Check online for suggestions and replace where possible.
Another issue I have encountered is two separate circuits getting connected at the far end. A real PITA to troubleshoot.
The radiated noise from Fluorescent needs to be addressed at each ballast. The traditional magnetic ballasts are not too bad and a few caps should do it. They will mechanically buzz. They are just large inductors in series and magnetostriction makes noise inevitable. Electronic ballasts are much better but suppressors at the power inlets help. LED replacements are the best of the lot and save money.
You have meters. Measure the voltage drop from no load to "full load" (10A, or 1 KW of load). Check at various points in the chain. If you see a big drop (more that 3%) you have connections that need repair/replace. Aluminium wire is not good and prone to starting fires. Tighten every connection with al wire as they loosen with temperature cycling. Check online for suggestions and replace where possible.
Elliot of ESP did a write up on these new lamps. Seems they only have a power factor of 0.55. That means almost half the power is being pissed away. Not very efficient like they are said to be. I guess 50% on less power is better than nothing.
Additionally the power companies see the distortion rising from there use. But put up with it
due to the lower power consumption.
It's a strange world we live in.
Anyway, enough off topic.
Additionally the power companies see the distortion rising from there use. But put up with it
due to the lower power consumption.
It's a strange world we live in.
Anyway, enough off topic.