Hi Lemon,
I built a resistor divider box. You will see it in the photo below. The red and black leads go to the amp and then I have a BNC and RCA post on the output. There is a switch to select divide by 20, 50 or 100. This accommodates most of what I have. The knob and potentiometer were for round one of the design. This allowed me to vary the division ratio. However, the pot creates too much noise...I removed the pot from the circuit and then built the resistor divider with the switch in the same enclosure.
I forget the resistor values at the moment; but I used ohms law to figure out the ratios. In loop back with the highest output of the EMU / ARTA combo, there is no noticeable noise or added distortion using the /20. For what I test, it works fine for me.
A Pete Millett interface would be nice, but I had these parts on hand and can use the second output connector to check my output voltage on a meter.
Let me know if you have any other questions.
I built a resistor divider box. You will see it in the photo below. The red and black leads go to the amp and then I have a BNC and RCA post on the output. There is a switch to select divide by 20, 50 or 100. This accommodates most of what I have. The knob and potentiometer were for round one of the design. This allowed me to vary the division ratio. However, the pot creates too much noise...I removed the pot from the circuit and then built the resistor divider with the switch in the same enclosure.
I forget the resistor values at the moment; but I used ohms law to figure out the ratios. In loop back with the highest output of the EMU / ARTA combo, there is no noticeable noise or added distortion using the /20. For what I test, it works fine for me.
A Pete Millett interface would be nice, but I had these parts on hand and can use the second output connector to check my output voltage on a meter.
Let me know if you have any other questions.
Thanks for the plots! I had to stare at the ARTA vs HP a bit to see how they related. at first I thought they didn't, but closer inspection shows they do. The 1% plots both show the high odd-order distortion from H5 and beyond.
The ARTA plot just reveals so much more noise. Is that an artifact or does the soundcard just see further down into the dirt?
The ARTA plot just reveals so much more noise. Is that an artifact or does the soundcard just see further down into the dirt?
Hmm... looking more closely the -90dB line on the HP would correspond to the -95dB level on the ARTA graph, right? If so, they look a lot alike.
The emu0204 is capable of measuring below 100nV and further if gain is added at the input So yes it can "just see further down into the dirt".
Ha! I push everything to the side for the photos!
Pano, The two images were acquired at the same time (ARTA and HP). I had the HP connected to the amp output along with the EMU. I set the peak of my ARTA at -5dBFS and the peak of the HP at 0dB. Yes, there is a 5dB difference in measurement scales. I need to purchase the license copy of ARTA to be able to clean up and change scaling in the software. Or, I could set the HP for -5. I would prefer 0dB for both.
The HP can see to -90 at its best. As Davada said, the EMU / ARTA combo can see much lower. The vertical scale on the HP is on the left and it is set for 1KHZ per division on the horizontal scale. So on the HP, you are seeing 1KHz signal and then harmonics up to 10KHz. I could have changed the ARTA scale to just see the same (1K to 10K), but I wanted the whole spectrum. And yes, it is a little 'grassy' on the higher frequencies.
Dave
Thanks David and Dave. 😀
I use an M-Audio USB sound card that can see pretty far down, too. I've just never compared it to a dedicated analyzer. Gary Pimm has a big old HP that runs up to 40MHz, I think. It can do 1Hz resolution. But it's sloooooooow cutting it that fine. Gary uses ARTA and his custom interface with an M-Audio Audiophile card. Works great.
You might look into Room EQ Wizard. I've just started using it and it does a lot. It's free. Might replace ARTA for you. ARTA is certainly worth the money, but there are other softwares out there.
I use an M-Audio USB sound card that can see pretty far down, too. I've just never compared it to a dedicated analyzer. Gary Pimm has a big old HP that runs up to 40MHz, I think. It can do 1Hz resolution. But it's sloooooooow cutting it that fine. Gary uses ARTA and his custom interface with an M-Audio Audiophile card. Works great.
You might look into Room EQ Wizard. I've just started using it and it does a lot. It's free. Might replace ARTA for you. ARTA is certainly worth the money, but there are other softwares out there.
When i started with inexpensive test systems using external ADC etc, I did some careful tests below -100dB and found they were not accurate below that level. But with a notch filter added to the front end, you could do much better... accurate for another 10-20dB. I eventually progressed to commercial/industrial instruments to get accurate measurements below -100dB.... first old Hewlett-Packard and modified them.... then to a couple Japanese instruments and to the Audio Precision 2722 system. You just can not get accuracy and precision below -100db with an inexpensive system. BUT, the good news is that if your DUT is at or below -100dB then you have a really excellent amplifier.
THx-RNMarsh
THx-RNMarsh
This is where some of the Analog Devices do so well, also.... lowest distortion at -60dB. Distortion compared at -40 and -60dB is a lot more important than at max/zero dB number.
Looking at your linearity plot at -100dB and below shows the non-linearity clearly.... and why it cannot be used as a test instrument below that level... even though FFT software would let you think you have ability much below that.
THx-RNMarsh
Last edited:
Hi Lemon,
I built a resistor divider box. You will see it in the photo below. The red and black leads go to the amp and then I have a BNC and RCA post on the output. There is a switch to select divide by 20, 50 or 100. This accommodates most of what I have. The knob and potentiometer were for round one of the design. This allowed me to vary the division ratio. However, the pot creates too much noise...I removed the pot from the circuit and then built the resistor divider with the switch in the same enclosure.
I forget the resistor values at the moment; but I used ohms law to figure out the ratios. In loop back with the highest output of the EMU / ARTA combo, there is no noticeable noise or added distortion using the /20. For what I test, it works fine for me.
A Pete Millett interface would be nice, but I had these parts on hand and can use the second output connector to check my output voltage on a meter.
Let me know if you have any other questions.
skidave thanks for the clarification.
I have a similar solution btw amp and emu 0404 input. See at the attachment photo at the upper(left) of devices.
It is a small attenuator box with a stereo 10K ladder pot for each channel. The inputs-outputs are xlr.
The attenuation is from 0...70 db almost.
My measuring setup follow the path: amp -> dummy loads -> fluke multimeters -> xlr attenuator -> emu inputs.
The emu calibration has done with 1Vrms signal. This level is FFT correlated at the 0dBV or ≈ -14dBFS.
I haven't any problem when I measure digital DUT, my problem is with amps, I have some spikes at the low frequencies (from 50Hz to 1K).
The lift emu gnd improves somewhat the problem but it doesn't disappear.
Attachments
Hi Lemon,
Grounding and noise are critical with the setup. 1audio (Demien) and several others have mentioned various ground configurations with the EMU and DUT. On my input and output cables I have the mini-grabber connected to my ground. This makes it easy to try different ground configurations or lift the ground completely.
You may want to try eliminating the voltmeters and other devices connected. I noticed when I had additional voltmeters connected on the output of the voltage divider, it caused noise and sometimes ground issues. Try an ARTA measurement without your meters connected and then try some ground configurations.
There are interfaces out there that have a built in meter and buffer circuits... I used to think I would need to measure the input to my soundcard; now I'm comfortable and I don't measure it anymore.
Hope that helps.
Grounding and noise are critical with the setup. 1audio (Demien) and several others have mentioned various ground configurations with the EMU and DUT. On my input and output cables I have the mini-grabber connected to my ground. This makes it easy to try different ground configurations or lift the ground completely.
You may want to try eliminating the voltmeters and other devices connected. I noticed when I had additional voltmeters connected on the output of the voltage divider, it caused noise and sometimes ground issues. Try an ARTA measurement without your meters connected and then try some ground configurations.
There are interfaces out there that have a built in meter and buffer circuits... I used to think I would need to measure the input to my soundcard; now I'm comfortable and I don't measure it anymore.
Hope that helps.
- Status
- Not open for further replies.