background noise

[greenbluegold]

so is ~80 dB signal to noise the norm with the audio inputs? Anybody getting much better?

[electrolover]

i've decided to use rme,at least for now,card itself grounded and all inputs very silent.i can sell scope.

[electrolover]

btw little ground loop isolator really works there is appr. 7db between direct connection with it. im really surprised.

[greenbluegold]

Well, at least I've got the peace of mind to now that I'm not going to be improving the S/N ratio much. Inside the computer needn't matter, the RME AEB-8i w/ Digi98 I still sometimes use does much better and that is installed right next to the Scope board. They're from the same era too, but are apples and oranges otherwise.

[Pete]

Hey greenbluegold, how should I test my card's analog s/n ratio, so I can compare to yours?

I'm not sure how to route the modules for it, and the first app I tried crashes.

[electrolover]

set everything off or disconnect from card (s). use digicheck, vumeter, rme/scope mixer to see level, also try to record silence and most important use your ear.

[Pete]

Digichck says this in the Readme: DIGICheck does not work with non-RME hardware!

[electrolover]

yes but you can try others for scope.at least if you have sound forge you can record silence and also you can check levels from scope's mixer but still everything illogical if you can hear noise(s) with you ears.

[valis]

If you have an RME card with an adat port, connect that to Scope & route the signal over to the RME...

If not there are other meters on the market that do RMS+peak with numeric readout.

[electrolover]

one of my friend told me that there is a regulator chip on scope and it could be produce noise if badly soldered or wrong chip used wrong country. it is on the card dsp's opposite side or i can say left bottom side. we searching on the internet but we could not find to chip for purchase. it is put and soldered later on the board. i mean its process is last process when producing the card and handmade soldering process.so a man who soldered the card if not in a mood he or she can fuck up the card.

[electrolover]

If you have an RME card with an adat port, connect that to Scope & route the signal over to the RME...

If not there are other meters on the market that do RMS+peak with numeric readout.

yep brilliant idea but signal could be filtered (+4/-10/lo gain).expect 3-6db clean signal.


i meant rme mixer for rme, scope mixer for scope earlier.

[valis]

yep brilliant idea but signal could be filtered (+4/-10/lo gain).expect 3-6db clean signal.

i meant rme mixer for rme, scope mixer for scope earlier.

If you're using adat there won't be any voltage/impedance mismatch.

Understood about the mixers too, just mentioning how to use Digicheck to monitor Scope (ie, to monitor Scope's inputs, route that to an adat out which is connected to an RME adat input, then set Digicheck to monitor that input).

[electrolover]

ok my bad sorry.

[greenbluegold]

Pete,

In a scope mixer route the audio in to a channel and then use the direct out to ASIO into whatever recording program you use -- I used wavelab. Set the meter at 0.0 on the scope mixer with the analog in. Record a clip. I then used Wavelab's analysis options to determine the peak level of the noise.

[greenbluegold]

I did some more measurements. For comparison, I get 85 dB S/N with an external preamp/converter combination (sound devices) running off batteries. So ~80 db for the scope card inside a computer isn't really too bad. I imagine if I had well designed wiring in my apartment it might even improve.

[greenbluegold]

I've done some more measurement comparisons to give the scope S/N ratio some context.

First, I followed the advice that I saw elsewhere (I think in another thread here at planetz) and ensured that my motherboard was properly screwed in to chassis. I'd been lazy with the screws and there had been several with brass risers that were for grounding that were not installed. Fixed that, and my noise levels improved by a few dB. Where I had been gettting -77.5 or so without any input into the scope analog ins, now I was below -80.

Here are the comparisons. All are done at 16 bit 48 kHz. The same loop was recorded from an external sequencer/module through a passive direct box (with the exception of the Emu -- I was too lazy to find my balanced XLR to TRS cables so skipped the transformers and put an unbalanced signal directly from the module to the balanced ins of the E4, this most likely reduced my S/N ratio a bit, I should redo that test with the proper cables -- gives me some motivation to get out my soldering iron and finally wire up my patchbays).

Reference: Scope Pro+ Analog In with no input: -80.77 dB noise floor

Scope+ Analog In with MP4x preamp, peak at -0.6: -74.75 dB noise floor
Scope+ Analog In with A&H MixWiz v3 Preamp, peak at -1.59: -72.25 -dB noise floor

Sound devices 702, battery powered, internal preamp, peak at -0.1: -78.27 dB noise floor

Emu E4XT with internal preamp, peak at -2.0 (unbalanced output to balanced input): -70.33 dB noise floor.


I think this goes to show that the analog inputs on the scope pro+ are decent enough.

Much of the noise is from a combination of your devices output and preamps.

If you measure your inputs and they are significantly noisier, check the grounding on the motherboard.

[valis]

You'd be amazed at how many 3rd party machines I find that are built with most of the motherboard standoffs missing.

The worst one I've seen was when I 'fixed' a friend's pc after discovering the only thing holding the motherboard down was the heatsink mounts (which don't act as ground) and the backplate (i/o ports) & graphics card exerting pressure due to being slotted. I could basically 'wiggle' the motherboard which meant it was not only lacking proper ground but it had been flexing during power on (heating) and power off (cooling) cycles. He'd been putting up with daily crashes and frequent disconnects of peripherals (keybd/mouse even) for about 2 years before he thought to ask me to look at it.

[electrolover]

hi,
well i 've just wanted to say something...
first of all my problem was the background noise (we call it deepnoise here) even if the card in null (no cable connected even its own cable)
then i 've learned scope and rme structure very different and rme 's already grounded but scope not (i 've used a lot of card but sope was the first one like this kind expensive and not grounded). anyway luckily i have rme and tried it. i've succeed. btw interesting thing about it; i have hum eliminator and i used it also i have simple a/d converter (but we can say analog rca to toslink converter) and analog connection better than digital that i made it with a/d converter. so stick with rme and analog input to card, this way better works for me for now...

then some day i've visited one of my friend which he has got same rme. i heard very disturb background noise from his krk. i realized background noise changes place to place, system to system.

thank you everybody.

[Pete]

I haven't looked, but I would think that motherboards may ground to the chassis via the connectors that either go to the power supply, or to the switches.

I always thought the mounting holes on a motherboard didn't have traces to anything.

Maybe there's an ieee standard or something that tells it.

[valis]

It is electrically grounded through the power supply's ground pin to household ground, but my experience has been that the standoffs make a difference as well. Note that I mentioned not only the possibility of cleaning up RF a bit (I'm no computer electrical engineer) but the flexing that a board undergoes when it's not completely mounted as well.