It's been hard to allready understand as much as i do now, using the notch, and i'am not there yet.
You're doing good, no worries.
I will post again on other thread, when i got the notch adjusted to excatly the ~1000HZ.
IMHO, don't post before you read at least a handful of pages before your own post - this has already been described there. The rest is in the R.E.W. help file and once you understand that, together with MagicBus's description above, you're more or less done.
I'm not worried, you'll be there in no time. In fact, you're already ahead of me
Done properly, it does exactly that.Shouldn't the calibrated file restore the "non notch filter" graph as close as possible?
Okay...It looks to me you're confusing cal. files. The difference is explained in the R.E.W. help. Once you solve that, you're all set I think and will have plenty of enjoyment with your new setup.
I did read all the stuff say 25times... I'am not fully understanding it... remember i'am not native English
You are very kind if you could be explaining me what i'am missing?.
EDIT :: Posting between post's... thank's @YashN
Jesper.
Hi.
I tried different, to have the 1KHZ notched closer to the 1KHZ NONnotched graph.
I can't really tell, why it's not more equal. I will do some more testing when i got my notch adjusted to the 998Hz, which my sine delivers.
I hope to have some more capacitors by end of the week to match.
I was reading some more stuff today, and something i cannot completly understand is why we are interested in getting as close to 0dB as possible.
Is it because we get more information closer to 0db bitwise?
Another thing regarding bits.
Is it correct that our soundcard can get a noisefloor of ~96dB with 16 bits, and around -144dB with the 24 bits our Behringer got. (without dither)
Jesper.
I tried different, to have the 1KHZ notched closer to the 1KHZ NONnotched graph.
I can't really tell, why it's not more equal. I will do some more testing when i got my notch adjusted to the 998Hz, which my sine delivers.
I hope to have some more capacitors by end of the week to match.
I was reading some more stuff today, and something i cannot completly understand is why we are interested in getting as close to 0dB as possible.
Is it because we get more information closer to 0db bitwise?
Another thing regarding bits.
Is it correct that our soundcard can get a noisefloor of ~96dB with 16 bits, and around -144dB with the 24 bits our Behringer got. (without dither)
Jesper.
All this is obscure to me too... Sure, it is explained thoroughly by experts but for non EE like us it's hard to follow when it gets too technical. OTOH, I don't see why not have a conversation in a more popular language running in the backround. Offering some amusement is not bad either! Your numbers refer to ideal situation, I think. It's the best you could expect. The CS4272 datasheet says -114dB dynamic range and -100dB THD+N. Not sure how the distortion analyzer software handles that either. In my diy soundcard, without input signal, the noise floor stays below -160dB but with full scale signal it hardly goes to -135dB with harmonics staying above -120dB so, should we say this is the -114dB dynamic range?
Regarding the notch filter, in the calibration procedure, I think you should try to get this blue and green traces to the same level. Perhaps you could edit the text file manually if its not too big and not before you save a copy.
Regarding the notch filter, in the calibration procedure, I think you should try to get this blue and green traces to the same level. Perhaps you could edit the text file manually if its not too big and not before you save a copy.
I fully agree...
And yes, we should have an "loose" conversation regarding this!
I also am aware that i reached the limit of this little critter
i found this somewhere on the www. link:
Our Behringer can handle much more (16777215), proberly giving better resolution.
EDIT::
Jesper.
And yes, we should have an "loose" conversation regarding this!
I also am aware that i reached the limit of this little critter
i found this somewhere on the www. link:
Any A-D/D-A converter has a maximum input voltage it can process. Staying with 16bit, the converter assigns a value of 32767 (or -32767) to this maximum voltage. This voltage is the converter's 0dBFS reference. Different makes of converters process different ranges of voltages, so 0dBFS is just a statement about the maximum voltage a particular converter can process.
If an A-D converter is provided with a higher input voltage than the maximum, it can still only assign a value of 32767, which alters (clips) the signal until it drops down to a workable voltage again.
Our Behringer can handle much more (16777215), proberly giving better resolution.
EDIT::
The blue and green is the same "notch" sweep, i just did manuel adjust the one at 0dB, just to try it out.Regarding the notch filter, in the calibration procedure, I think you should try to get this blue and green traces to the same level. Perhaps you could edit the text file manually if its not too big and not before you save a copy.
Jesper.
Hi all,
the UMC202HD does not have differential outputs. You can add an external single ended to differential amplifier (I have PCBs for that).
The AudioTester SW I am using allows you to use the same signal at both outputs, but with 180° phase offset. It is not a "native" differential signal coming out of a fully differential amp, but it looks like one. Does have ARTA and REW the same feature?
the UMC202HD does not have differential outputs. You can add an external single ended to differential amplifier (I have PCBs for that).
The AudioTester SW I am using allows you to use the same signal at both outputs, but with 180° phase offset. It is not a "native" differential signal coming out of a fully differential amp, but it looks like one. Does have ARTA and REW the same feature?
Hi all.
I did some research regarding FullScale.
I found this : Link!
It's telling me this :
* I also read somewhere that using -3dB prevents "peak" clipping, therefore i feel comfortable measuring around this area (-3dB ---> -6dB)
Another Link! with some good info.
Regarding my Notchfilter i asked in other threads, and the owner of REW answered my "tricky" quistion regarding the [FS Sine Vrms] dialogbox in REW.
Link! - Sry. that i could not spell the word understand correctly that day (Notch filter with REW, cannot undertand the use of [FS sine Vrms])
EDIT : Can someone explain this from the CS datasheet ?
Jesper.
EDIT EDIT : Getting wiser; could be i should stop making cranes, and do my living as prof. sound measuring expert... That is, if i wanna go to bed hungry every night
I did some research regarding FullScale.
I found this : Link!
It's telling me this :
So it must be correct that it's bitwise resolution, e.g as close to 0dB as possible gives us the most precise information about the DUT.0dBFS (Full Scale) is the clipping point for a signal in a digital audio product. Rather than measuring from the noise floor up, digital signals are measured (or referenced) from the clipping point, or full scale, down. A 0dBFS (Full Scale) signal contains the maximum amount of digital information that can be used to represent the signal being defined.
* I also read somewhere that using -3dB prevents "peak" clipping, therefore i feel comfortable measuring around this area (-3dB ---> -6dB)
Another Link! with some good info.
Regarding my Notchfilter i asked in other threads, and the owner of REW answered my "tricky" quistion regarding the [FS Sine Vrms] dialogbox in REW.
Link! - Sry. that i could not spell the word understand correctly that day (Notch filter with REW, cannot undertand the use of [FS sine Vrms])
EDIT : Can someone explain this from the CS datasheet ?
Jesper.
EDIT EDIT : Getting wiser; could be i should stop making cranes, and do my living as prof. sound measuring expert... That is, if i wanna go to bed hungry every night
Last edited:
Perhaps making sound systems for cranes?
Say analog power supply VA=5V, then the ADC would need 1.07*5 - 1.19*5V peak to peak input signal to reach full scale. That is differential. See datasheet page #32. As it's written -and empirically verified with this soundcard- it works for VA<5V too.EDIT : Can someone explain this from the CS datasheet ?
Maybe its also practical to find a sweet spot input level for lowest THD spikes and choose dBc scale to display? That way always refers to zero and has small sticking spikes but the dBFS way goes deeper THD+N. Depends on what aspect you want to focus for a DUT each time I guess.
Attachments
@MagicBus ...
I got it now with the VA - More visdom to me
@Salas ...
That's a nice interface the Motu, and thank's for the tip, with the dBc.
I did not try it out yet, but having an extra long weekend i will for sure try it.
Not for making this offtopic, an for hijacking this thread, i have a quistion regarding rectifier noise (100Hz / EU)
Some of my "test" amp's, including my amanero usb + cheap dac on my lab, all suffer with 100Hz spikes from the powersupply.
I read about snubber's but it's not clear if and how this should be tried and / or implemented afterwards.
* Would be fun to do some testing, just need a little kickstart.
Jesper.
I got it now with the VA - More visdom to me
@Salas ...
That's a nice interface the Motu, and thank's for the tip, with the dBc.
I did not try it out yet, but having an extra long weekend i will for sure try it.
Not for making this offtopic, an for hijacking this thread, i have a quistion regarding rectifier noise (100Hz / EU)
Some of my "test" amp's, including my amanero usb + cheap dac on my lab, all suffer with 100Hz spikes from the powersupply.
I read about snubber's but it's not clear if and how this should be tried and / or implemented afterwards.
* Would be fun to do some testing, just need a little kickstart.
Jesper.
In REW its readily available and easy to use. Stay around your interface's least visible THD spikes input level. In ARTA dBc is not in the scaling menu, it takes to fake dBFS level calibration for that sweet spot and keep measuring with an input level around there.thank's for the tip, with the dBc.
- Home
- Design & Build
- Equipment & Tools
- Behringer UMC 202HD for measurements