USB soundcard line-in overload

Discussion in 'General Electronics Chat' started by oneoldude, Aug 7, 2013.

  1. oneoldude

    Thread Starter New Member

    Aug 7, 2013
    12
    0
    Been looking all over and I can't find this answer.

    What is the typical USB soundcard line-in AC clipping voltage? I mean max voltage in with no headroom at all for a typical unbalanced line-in.

    I figure that since they are all powered with 5V it can't be more than 5V P-P (1.78V RMS, +7.22 dBU, +5 dBV). But this is a gross assumption on my part. My guess is that max in is less than that.

    Anybody have experience in this area?

    Thanks
     
  2. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    I can't believe a single 5V rail could work without boosting somehow. The "line" levels vary based on what country you're in and whether it's commercial or professional.

    An audio line has to process signals going both pos and neg, so I would think a a 5V rail alone would never be able to suffice since you would only be able to swing maybe 1.5V above and below midpoint.
     
  3. #12

    Expert

    Nov 30, 2010
    16,262
    6,768
    Line in, in my world, is about .775 volts RMS. Probably capacitively coupled to a mid-point voltage in the sound card.
     
  4. oneoldude

    Thread Starter New Member

    Aug 7, 2013
    12
    0
    You have a common misbelief due to using different standards, some of which date to the 1930's. I can tell you that USB only provides 5V DC. That is all a USB sound card has to play with and that is more than enough to cover all recording conventions re: amplitude I know of. If you could go rail to rail with that 5V you would would have the following max-in limits like I said before:

    5V P-P = +7.22 dBU = 1.78V RMS = +5 dBV

    As for recording amplitude standards, these are the nominal conventions I know of:

    Studio level international +4.00 dBU = 1.228V RMS = +1.78 dBV
    Standard 1 Volt +2.22 dBU = 1.000V RMS = 0.00 ref.
    Standard 0.775 Volt 0.00 ref. = 0.775V RMS = −2.22 dBV
    Domestic consumer level −7.78 dBU = 0.316V RMS = −10.0 dBV

    The important thing to note here is that under all the standards, 5V rail to rail has the capability to to handle all the nominal recording conventions. The toughest standard is the Studio level international and its RMS level is 1.288V RMS. A 5V rail to rail circuit would allow 1.78V RMS. But that is neither here nor there. The real question is whether anyone knows the typical line-in overload specs for USB sound cards. It could be as high as 1.78V, but the chip makers may have designed for a standard below that. Unfortunately, I can't find that info on the web. I need it to set the gain chain from mic capsule to ADC.

    Does anyone actually know?

    Thanks
     
  5. oneoldude

    Thread Starter New Member

    Aug 7, 2013
    12
    0
    My reply above relates to bountyhunter's post not #12's post.

    Thanks
     
  6. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    Really depends on the soundcard in question. Some can accept quite higher levels, some even have switchable attenuators to accomodate higher levels.
    Of course there is not much of a problem to get +/-15V rails from a 5V source, but swither supplies are noisy.
     
  7. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    But line levels do not drive speakers, they only have to drive high impedance inputs. That means extremely small currents. In such cases, you can make a charge pump to easily double or triple the voltage using small capacitors and a couple of transistors. That's what I would do if I was designing it.
     
  8. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    Here's what I know: 1.228 VRMS corresponds to ZERO dB in that standard. 1.228 VRMS is equal to 3.5V peak to peak. Even if the full 5V (p-p) was available for signal swing (which it could not possibly be) that only allows about a 4 dB headroom above 0 dB which is certainly not enough for most music.
     
  9. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    0.775 VRMS was the "standard" that Dolby set for the so-called "Dolby Tone" used to calibrate recording devices back in the 70's when we did their Dolby B chip, but it corresponded to +3 dB on the VU meter. There are other standards now, some higher and some lower than that value for the 0dB reference.
     
  10. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    The only way to actually know is feed a sine wave into the input and look at the output. Increase amplitude until it clips the top. That would give me the answer in about two minutes.
     
    GopherT likes this.
  11. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    Here is the spec of one very good sound card. No idea if yours is a s good:

    http://www.stereophile.com/content/digital-audio-labs-carddeluxe-pc-soundcard-specifications





    Professional audio +4 dBu Peak-to-Peak Amplitude, VPP 3.472V


    15 dB headroom above that is about 20Vp-p signal swing. You ain't getting that from a single 5V rail.
     
  12. oneoldude

    Thread Starter New Member

    Aug 7, 2013
    12
    0
    Yea, that is right. Apparently neither you nor anyone else has done that kind of test on line-in for simple sound cards. Everyone seems to rely on standards for analog technology that date as far back as far the '30s. Some designers even go so far as to include both pads and gain blocks on the analog front end for ADCs to allow the use of the ADC as though it was an analog circuit. Which it is not. How much headroom is there on an ADC above its 0 dB point? That's right, its zero. Anything above that is gone. So to fake it, the analog front ends are cobbled up to make them act like they are feeding an analog back end with headroom. All of that is way too complicated for a simple guy like me. I don't expect the situation to change anytime soon. Too many vested interests and too many people want to do things the same way grandpa did.

    So here is my solution.

    Instead of trying to come up with a general solution that would apply in many instances, I decided to attack one problem and deal with that.

    I bought a Behringer UCA222 that has a specified line-in max input of 1.25V RMS.

    That 1.25V RMS = 1.768V P = 3.536V P-P = +4.157 dBU = +1.938 dBV

    That will handle up to the Studio international level standard of +4 dBU on the USB 5V bus.

    Now I can design a mic-pre with gain to match my mic sensitivity against SPL input for the sound card overload point. I have already designed and built one with adjustable gain, just in case. But for my purpose, measuring loudspeaker drivers at about 1 Pa (94 dB SPL), a fixed gain should simplify matters greatly.

    Thanks
     
  13. oneoldude

    Thread Starter New Member

    Aug 7, 2013
    12
    0
    "Nominal analog input and output level: +4dBu, internally switchable to -10dBu. Nominal headroom: 15dB ref. +4dBu."

    Sorry,

    You posted while I was typing.

    Quickly scanned the review.

    Went to the web site. The product is no longer being sold. I could find no manual for it. Apparently not a successful product.

    It is a pci card. The pci bus is 5V at its best. This means that it should perform more or less like the cheapie UCA222. The pci card, especially for the price, should have better noise and distortion numbers depending on how much they paid for their chips. Both will handle +4 dBU max. But the UCA222 does not have any analog padding or gain. That is why I bought it.

    If, on the reviewed card, they put in a software selectable 15 dB pad they would get 15 dB headroom. Of course that is with a nominal -11 dBU input. The review did not discuss that issue from what I quickly scanned and the specs are open to marketing BS.

    Trust me, I tried to find out what the max input to line-in was for sound cards without padding or gain. I could find no info at all. That is why I originally posted here. Figured someone was inquisitive enough to have tested cards to find out. You know, its like the old time philosophers debating how many teeth a camel has. The debate went on for years. Finally a stranger suggested counting them. Yea, the beginning of real scientific knowledge.

    I haven't the equipment or sound cards to do the measurements. But I'll bet that mid grade and above sound cards will measure about what the UCA222 specs for line in-in overload if you account for analog padding and gain before the ADC.

    There is a lot of cool-aid out there. The amount a person drinks depends on what flavor he likes and the nature of his preconceptions, well founded or not.

    Me, I'm on my way with the UCA222, even though I find the color hideous.

    Thanks
     
Loading...