BTW, we also got a new amplifier, actually an AVR. It is one of the affordable Sony Atmos-compatible lines.
It sounds good, and we're watching GOT in 4K and Atmos with it but simultated Surround as I still have to finalise the real Atmos Surround setup.
Haven't measured it though, but paired with Sony towers it looks great, sounds great.
It apparently also does DSD but via HDMI and I don't think I can send DSD via HDMI from my PC Server, so this is as yet untested.
It sounds good, and we're watching GOT in 4K and Atmos with it but simultated Surround as I still have to finalise the real Atmos Surround setup.
Haven't measured it though, but paired with Sony towers it looks great, sounds great.
It apparently also does DSD but via HDMI and I don't think I can send DSD via HDMI from my PC Server, so this is as yet untested.
Hi all,
just an update. The technician of the school was able to get new CS4272 and a former colleague changed them. Now the UMC202 are working again. At two of them R105 (10 Ohm) was damaged (open), so only one leg of the input was working. I think I will change R105, R108, R101, R104 to a higher value, maybe 1 K. Hopefully this will protect the UMC from high input voltages. It will change the gain. 2/3 without pad, small for pad). I will also take out R106, R107, R68 and R70 which couple the legs if one is open. The TRS inputs are now in parallel to the XLR.
Alfred
Hi all,
I wrote an article about a Fliege Notch filter for ELEKTOR. (https://www.elektormagazine.com/magazine/elektor-272/60926)
You can download the layout files there.
I also have PCBs for the filter and the power supply.
Please find attached a list of filter PCB I did, in case somebody is interested.
Then I can provide detailed information.
Alfred
I wrote an article about a Fliege Notch filter for ELEKTOR. (https://www.elektormagazine.com/magazine/elektor-272/60926)
You can download the layout files there.
I also have PCBs for the filter and the power supply.
Please find attached a list of filter PCB I did, in case somebody is interested.
Then I can provide detailed information.
Alfred
Attachments
It's possible that 634R resistor might not be correctly connected in the original datasheet. Some info here:
https://electronics.stackexchange.com/a/146634
AN241 has it like the datasheet. But I did manage to find it connected the way that user proposed, in an Infineon AN
https://www.infineon.com/dgdl/Infin...aveo_II_Family-ApplicationNotes-v02_00-EN.pdf at page 27.
Sorry I didn't link it correctly:
https://www.infineon.com/dgdl/Infin...N.pdf?fileId=8ac78c8c7cdc391c017d0d3e9ceb67ee
You can search on google for "AN226043" if it still doesn't work.
https://www.infineon.com/dgdl/Infin...N.pdf?fileId=8ac78c8c7cdc391c017d0d3e9ceb67ee
You can search on google for "AN226043" if it still doesn't work.
Hi Alfred,
FYI, the AD8692 is in stock at Mouser. Personally, I would prefer the OPA1662 though. Slightly less noise and distortion, if that matters to you. The only minor drawback of the OPA1662 is its slightly smaller dynamic range, but it's still within the working range of the ADC/DAC. As you can read in my previous posts, I have used the OPA1664 successfully in the UMC202HD, replacing the original AD8694.
Btw, read your article in Elektor with interest! I was just hoping you would have compared the relatively expensive RME interface with the cheap UMC202HD.
I'm still working on my diy audio analyzer project. I had hoped I could have finished before summer but did manage to do that due to circumstances. Anyway, just finished a nice High-Speed USB isolator based on the TI ISOUSB211DPR chip. Works flawlessly and did help to lower the noise floor in some cases. I consider making the PCB available from PCBWay for anyone interested. Or sell as a complete product. Did not decide on that yet. Is anyone interested? Mind you, the smallest parts to assemble are 0402. Not everybody's cup of tea, I'm afraid. It must be externally powered, which I did deliberately. The interface itself is drawing quite some power and if I had included a 5V DC/DC converter, it could only deliver a few hundred mA to the client side and of questionable quality (noise). It could also have compromised common mode rejection from PC to the client side. Now you can power your audio interface, measuring device or whatever you connected as client using your own favorite ungrounded power supply. Furthermore, the design also includes ESD protection, Vbus filtering, status LEDs and jumpers to select cable compensation. See the ISOUSB211DPR datasheet for more information.
The last sub-project of my analyzer is the 4-channel voltmeter module using an Arduino. But my programming skills are not that good, unfortunately... And I have some mechanical issus to solve. So I don't expect this to be finished anytime soon. Once everything is finished, I will share the results and design files.
Greetz, Jan
FYI, the AD8692 is in stock at Mouser. Personally, I would prefer the OPA1662 though. Slightly less noise and distortion, if that matters to you. The only minor drawback of the OPA1662 is its slightly smaller dynamic range, but it's still within the working range of the ADC/DAC. As you can read in my previous posts, I have used the OPA1664 successfully in the UMC202HD, replacing the original AD8694.
Btw, read your article in Elektor with interest! I was just hoping you would have compared the relatively expensive RME interface with the cheap UMC202HD.
I'm still working on my diy audio analyzer project. I had hoped I could have finished before summer but did manage to do that due to circumstances. Anyway, just finished a nice High-Speed USB isolator based on the TI ISOUSB211DPR chip. Works flawlessly and did help to lower the noise floor in some cases. I consider making the PCB available from PCBWay for anyone interested. Or sell as a complete product. Did not decide on that yet. Is anyone interested? Mind you, the smallest parts to assemble are 0402. Not everybody's cup of tea, I'm afraid. It must be externally powered, which I did deliberately. The interface itself is drawing quite some power and if I had included a 5V DC/DC converter, it could only deliver a few hundred mA to the client side and of questionable quality (noise). It could also have compromised common mode rejection from PC to the client side. Now you can power your audio interface, measuring device or whatever you connected as client using your own favorite ungrounded power supply. Furthermore, the design also includes ESD protection, Vbus filtering, status LEDs and jumpers to select cable compensation. See the ISOUSB211DPR datasheet for more information.
The last sub-project of my analyzer is the 4-channel voltmeter module using an Arduino. But my programming skills are not that good, unfortunately... And I have some mechanical issus to solve. So I don't expect this to be finished anytime soon. Once everything is finished, I will share the results and design files.
Greetz, Jan
Hi Jan,
thanks for the hint to the OPA1662. I will get some and test them.
At the time I bought the RME and I was writing the article I was not aware of the UMC. As mentioned in the article I tested several other interfaces, but none was really good.
Would I buy the RME now? I don't know. It is better, no doubt. But expensive.
I also wrote an article about the UMC for Elektor, but it is not accepted so far. They don't like to have to many articles of the same author and the same subject in short time.
In the next issue I will describe active Filter design software.
Normally it takes up to one year before an article is published.
Alfred
thanks for the hint to the OPA1662. I will get some and test them.
At the time I bought the RME and I was writing the article I was not aware of the UMC. As mentioned in the article I tested several other interfaces, but none was really good.
Would I buy the RME now? I don't know. It is better, no doubt. But expensive.
I also wrote an article about the UMC for Elektor, but it is not accepted so far. They don't like to have to many articles of the same author and the same subject in short time.
In the next issue I will describe active Filter design software.
Normally it takes up to one year before an article is published.
Alfred
Hi all,
I found another small problem and a possible solution.
The signal Vcom_ADC at IC16:B Pin 7 is oscillating with a small amplitude of about 20 mVpp at around 140 kHz.
The frequency changes with temperature. Measuring with an RCL meter (power off) tells me that there might be a cap of 220 nF at this signal. Maybe directly at the output of the opamp, which is always a bad thing. Since I did not want to change the board, I just added two big caps of 47uF at the pots for the line and the headphone output to GND and the oscillation is gone.
You might need a filter in the signal path to see the 140 kHz signal.
Best regards
Alfred
I found another small problem and a possible solution.
The signal Vcom_ADC at IC16:B Pin 7 is oscillating with a small amplitude of about 20 mVpp at around 140 kHz.
The frequency changes with temperature. Measuring with an RCL meter (power off) tells me that there might be a cap of 220 nF at this signal. Maybe directly at the output of the opamp, which is always a bad thing. Since I did not want to change the board, I just added two big caps of 47uF at the pots for the line and the headphone output to GND and the oscillation is gone.
You might need a filter in the signal path to see the 140 kHz signal.
Best regards
Alfred
Hi Alfred,
I think it has to do with the CS4272 itself. IC16 works as a double DC buffer. Pin#7 gives buffered output of the Vcom of CS4272 for its input analog buffer only. Pin#1 gives output of 1/2*VA for all the rest circuitry using a voltage divider for reference. Thus, in my early efforts to tame unwanted spikes, I bypassed CS4272 Vcom and used the other half for everything. Later it was suggested that the spikes in question were introduced by the phantom power smps and members who tried that reported better results. Still, my conclusion is that the CS4272 Vcom is dirty. In my diy soundcard where the only common thing with the UMC 202HD is the CS4272, I can see a small bump starting from well within the audio range and progressively moving to higher frequencies to finally go away after a few minutes. Although I can't tell if its fundamental frequency is at 140kHz it seems to be temperature depended.
I think it has to do with the CS4272 itself. IC16 works as a double DC buffer. Pin#7 gives buffered output of the Vcom of CS4272 for its input analog buffer only. Pin#1 gives output of 1/2*VA for all the rest circuitry using a voltage divider for reference. Thus, in my early efforts to tame unwanted spikes, I bypassed CS4272 Vcom and used the other half for everything. Later it was suggested that the spikes in question were introduced by the phantom power smps and members who tried that reported better results. Still, my conclusion is that the CS4272 Vcom is dirty. In my diy soundcard where the only common thing with the UMC 202HD is the CS4272, I can see a small bump starting from well within the audio range and progressively moving to higher frequencies to finally go away after a few minutes. Although I can't tell if its fundamental frequency is at 140kHz it seems to be temperature depended.
Hi all folks here 
I posted this !here! ... but i have something on my mind, i better ask here that is.
Recently i bought this fine Topping E30II !Topping!, and ofcause i should try to measure it.
I'am wondering why i cannot reach more than MAX -105dB (THD+N / Sinad) with notch, while the Topping should go near -120dB
Someone in the other thread told me that it could be the limit of the codec (mention that i perhaps need ES9822PRO), but i think he could be wrong ? Another one that i can amplifi the signal directly after the passive notch too
@alfredr was kind and wrote a pedagogical explanation to me some 5 months ago, when i considered building one of his active notch designs.
But duo to component shortage and increased prices, i'am not going to build one at moment.
(Case is also that i can buy one of the Cosmos APU's for nearly the same price, but this is ofcause not that much fun! - and maybee it is, while they offer a kit to adjust the APU by hand by user)
Alfred told me, that by adding some gain in the end of an active notch (could be +20dB), would help bring the harmonics out of the noise. (I understand this)
The +20dB should the be subtracted from the result (this i'am sure could be implemented into the REW calibration file too) .. (This makes perfect sense too for me)
But how would that actually look like in the FFT ?, I mean how would the spectrum look like when adding gain at the end of a notch (active & passive)
I'am sure some of you guy's here can explain it to me ? - Did i get it right, is the modded Behringer up to the task @ > -120dB
Thank's in advance, hoping you all are doing well.
Jesper.
Direct with no passive notch. (I actually reached very near -100dB at some point)
998Hz passive notch with REW calibration file (notch / NOnotch)
(I actually reached -105dB at some point)
I posted the wrong notch screenshoot... here is the right one :
I posted this !here! ... but i have something on my mind, i better ask here that is.
Recently i bought this fine Topping E30II !Topping!, and ofcause i should try to measure it.
- Output from REW / Topping -3dB 998Hz sine. (Laptop battery powered) with and without notch.
- Voltage outputsving from Topping is 1395mV like this.
- Behringer is powered from laptop usb as usual, input into Magicbus's balanced input mod.
I'am wondering why i cannot reach more than MAX -105dB (THD+N / Sinad) with notch, while the Topping should go near -120dB
Someone in the other thread told me that it could be the limit of the codec (mention that i perhaps need ES9822PRO), but i think he could be wrong ? Another one that i can amplifi the signal directly after the passive notch too
@alfredr was kind and wrote a pedagogical explanation to me some 5 months ago, when i considered building one of his active notch designs.
But duo to component shortage and increased prices, i'am not going to build one at moment.
(Case is also that i can buy one of the Cosmos APU's for nearly the same price, but this is ofcause not that much fun! - and maybee it is, while they offer a kit to adjust the APU by hand by user)
Alfred told me, that by adding some gain in the end of an active notch (could be +20dB), would help bring the harmonics out of the noise. (I understand this)
The +20dB should the be subtracted from the result (this i'am sure could be implemented into the REW calibration file too) .. (This makes perfect sense too for me)
But how would that actually look like in the FFT ?, I mean how would the spectrum look like when adding gain at the end of a notch (active & passive)
I'am sure some of you guy's here can explain it to me ? - Did i get it right, is the modded Behringer up to the task @ > -120dB
Thank's in advance, hoping you all are doing well.
Jesper.
Direct with no passive notch. (I actually reached very near -100dB at some point)
998Hz passive notch with REW calibration file (notch / NOnotch)
(I actually reached -105dB at some point)
I posted the wrong notch screenshoot... here is the right one :
Last edited:
It looks like the 'N' signals that are contributing to a 'poor' THD+N level are actually harmonics above the 9th order, but still within the 20kHz bandwidth you have specified in the Distortion Settings box. You could confirm that by narrowing the bandwidth to say 9.5kHz.
The notched response is now more prone to the non-harmonic signals. Those extraneous signals may be internal to your equipment - have you tried different environments, and quality of interconnectors to clarify the circa 800Hz and 1600Hz signals, as well as different fundamental frequency ?
The notched response is now more prone to the non-harmonic signals. Those extraneous signals may be internal to your equipment - have you tried different environments, and quality of interconnectors to clarify the circa 800Hz and 1600Hz signals, as well as different fundamental frequency ?
I could of course be wrong but IME only ES9822PRO has low enough noise and high enough SNR/DR for measuring -120dB THD+N with a notch. Based on your measurements THD+N is dominated by noise.
Hi,
the measurement with a notch (passiv or active) does not help you at THD+N if noise is dominating.
Just at pure THD.
The function of the notch is to prevent the ADC to add its own harmonics since it does not see the signal itself in such a high level.
The noise is amplified by an active notch, since it has its own OpAmps which add their noise too.
In rare cases you might notice that one or the other harmonics is smaller when you measure without the notch.
So the phase of the incoming harmonic is 180° compared to the harmonic generated by the ADC and so the displayed harmonic is reduced.
Alfred
the measurement with a notch (passiv or active) does not help you at THD+N if noise is dominating.
Just at pure THD.
The function of the notch is to prevent the ADC to add its own harmonics since it does not see the signal itself in such a high level.
The noise is amplified by an active notch, since it has its own OpAmps which add their noise too.
In rare cases you might notice that one or the other harmonics is smaller when you measure without the notch.
So the phase of the incoming harmonic is 180° compared to the harmonic generated by the ADC and so the displayed harmonic is reduced.
Alfred