PC audio test set

[Tubetec]

I was able to improve the measured noise floor of my SSL2 interface by 13db with a few minor modifications and separating off analog and digital PSU rails .
I find the balanced Hi-Z input useful where you want to minimise any loading on the device on test .

REW doesnt know when I switch my interface from mic to line to hi-z ,
obviously the gain changes , that means the reference level needs to be reset or the numbers are off.

A midi controlled external attenuator , under the control of REW would be nice , so it automatically re-calculates .

 

[jp8]

In a similar way, I improved the performance of the UMC202HD that I use in my "Audio Analyzer" setup. The major improvement came from changing the internal Vdd level from 4.5V (!) to 4.9V. At the same time, I raised the USB voltage from the USB Isolator to 5.15V, btw. The Codec chip required at least 4.75V according to the datasheet, but Behringer decided it could also run from 4.5V... The low Vdd caused huge distortions above -6dBFs, but also at lower levels distortion can be improved with a higher Vdd. I also bypassed the input and output circuits, except the AD input filter section and the DA output buffer. This essentially leaves you with a clean Codec with USB interface, to which you can then connect your own front end.

In the tables below, you can see what improvements can be made and how the Analyzer with a modified UMC202HD rates against a cheap Mackie ProFX12 mixer.

A midi controlled external attenuator , under the control of REW would be nice , so it automatically re-calculates .

How about the other way around? I.e. when you design your own frontend, you know the input sensitivity. The most important thing for REW to know is the calibration level of each gain setting on your frontend. Is there any way to interface to REW and send calibration levels from an external program or through a file or whatever?

Jan

 

[abbey road d enfer]

For what it's worth, here's the schemo of my "front end".
Of course, if I had to do it again, I wouldn't repeat the mistakes I made at the time...

 

[jp8]

Would you mind elaborating on those mistakes...? And perhaps on what you do like about your design and what might be useful for us to know? As I said, I'm not going to do build a frontend again, but I'm shure there are forum members who like to learn from your design choices.

Jan

 

[abbey road d enfer]

Would you mind elaborating on those mistakes...?

I would duplicate the meter output on a buffered BNC (connected to one channel of oscilloscope) and lose the reconfigurable optional filter for a basic A-weighted. I would also lose the 10dB attenuator on the Gen input.

And perhaps on what you do like about your design

I like the 100kohm input Z and its large gain/attenuation range, the output Z management and the 22-22k filter, all of which are essential in a good test set IMO.

 

[blue_luke]

I would duplicate the meter output on a buffered BNC (connected to one channel of oscilloscope) and lose the reconfigurable optional filter for a basic A-weighted. I would also lose the 10dB attenuator on the Gen input.

I like the 100kohm input Z and its large gain/attenuation range, the output Z management and the 22-22k filter, all of which are essential in a good test set IMO.

Taking note !
Right now, even if I don't make much noise here, I read everyday comments and advices that people share..
I am not trying to make an Audio Precision killer here, just a "comfortable" interface to test pretty much any audio device in and out, within the limitations of a modern sound card, which are pretty remarquable devices nowadays, even the cheapest ones are quite sum'thin!
Remember, I started building audio stuff in the late 60's where a S/N ratio of 74db was deemed fantastique!
Walt Jung, John Roberts, Don Lancaster and a few others were considered gurus in the days of Audio Amateur, electronic, world, popular electronic and some other magazines!
I have learned so much from these!

 

[JohnRoberts]

Taking note !
Right now, even if I don't make much noise here, I read everyday comments and advices that people share..
I am not trying to make an Audio Precision killer here, just a "comfortable" interface to test pretty much any audio device in and out, within the limitations of a modern sound card, which are pretty remarquable devices nowadays, even the cheapest ones are quite sum'thin!
Remember, I started building audio stuff in the late 60's where a S/N ratio of 74db was deemed fantastique!
Walt Jung, John Roberts, Don Lancaster and a few others were considered gurus in the days of Audio Amateur, electronic, world, popular electronic and some other magazines!
I have learned so much from these!

Not a guru but back in the 80s I designed a piece of semi-pro test equipment (TS-1) and that was harder than I expected.

I have much respect for the guys who designed Audio Precision and lab grade audio test gear.

JR

 

[Matt Syson]

I find it curious how some excellent designs miss out on the 'bleeding obvious' such as my AP Portble One won't sit on it's back because the power cord (plug) is too long unless you find the right angle variant AND the BNC outputs cannot be used for the same reason. The headphone out should be easily configurable to balanced or unbalanced (headphones OR an external amplifier). I love my System One with Windows software but it was a pig in DOS. Neutrik mini (Minirator) boxes won't handle plus 21dBu input balanced or unbalanced, what were they thinking? The mini sig gen is OK considering it runs on batteries and is otherwise 'acceptable'. The Neutrik Brown test sets were reasonable but the oscillator relays weld together if you accidentally have phantom power on, or the output stage burns. I believe the predecessor to these was from Theatre Projects (??) in the UK which was good but did not have the autoranging meter (1980 era).