This 29v pk-pk sound card range is the essence of my concern about sound cards. The gain must not be constant.
Sound cards use A/D converter. The two best audio A/D converts I know of (and there may be others) are the Texas Instruments PCM1084 and the Wolfson WM8785 or WM876. The links for the data sheets can be found at: http://focus.ti.com/lit/ds/symlink/pcm1804.pdf http://www.wolfson.co.uk/products/digital_audio/adcs/WM8785/ A sound card can be no better than it's A/D converter. Both of The above converters are very similar. The TI part, as specified on page 7, has an input range of +/-2.5v, centered around 2.5v or 5v peak to peak. This is the maximum input voltage to this ADC. Distortion + noise is rated at -101 db with an input at 0.5 db. Thus, the noise floor is about -100 db. Since the +/-2.5v input maximum is a level higher than a 0 db level, the dynamic range of the A/D converter is specified on page 7 as being 112 db (A weighted). Lets look at the Wolfson part. On page 6, the input rage of this part is specified to be 2v RMS which can also be expressed as +/-2.8v or 5.6v pk-pk. This seems a bit odd because the part has a supply voltage of only 5v, and I would not expect the input to be larger than the supply range. The distortion is -102 db (I am assuming this is distortion + noise), similar to the TI part, and the dynamic range is 111 db A weighted and 108 db non-weighted. The A weighted vs. non-weighted difference is probably significant since we probably use these in an unweighted manner. Thus the TI part is probably really a 109 db dynamic range part. Ideally, you should be able to get 6 db of range for every bit in the A/D converter. Thus, it would seem that 24 bit A/D converters would be capable of 144 db of dynamic range (there might be a n-1 factor in here, 138 db, I do not remember). This is obviously not the case with real converters. I have seen real converters approach this only when they are running very slow, such as a 10 Hz sampling rate. Since we want 48+ KHz sampling rates, we get less conversion accuracy and a smaller dynamic range. Thus, the range on either of these parts from the weak to the strongest signal than can be handled is 108 to 109 db. Thus, it is not possible for this A/D converter to have a blocking dynamic range greater than 112 db. Since the current SDR1000 implementation is a wide band receiver, meaning the A/D converter has no hardware filtering protection from large signals far away, the burden of receiver performance falls squarely on the A/D converter. If the A/D converter has a smallest signal to largest signal range of 108 to 112 db, then the receiver can do no better than this. Thus my comment on why it would be nice to have a sound card with 145 db of dynamic range. They do not exist because the very best A/D converters that the sound card uses can do no better than 108 to 109 db. A sound card is capable of reducing its gain when it sees a signal of 29v pk-pk, and increasing it gain when the signal is much smaller. However, the dynamic range, the ability to hear a weak signal in the presence of a large signal, does not change. It can be no better than 108 to 109 db in a wide band receiver such as the SDR1000. If the detector is good for 5v peak to peak, and the A/D converter is good for +/-2.5v signals, the best large signal performance would be to run with no gain at all. The problem would be an MDS of only -100 dbm, as the spec sheet indicates. A gain of at least 40 db would be required to lower the MDS down to -140 dbm. However, not matter what you do, the A/D converter still has only ~110 db of dynamic range. Thus, with an MDS of -140, I would expect blocking to occur with a signal any where with in +/- 100 KHz with a level of -30 dbm or more (110 db of dynamic range). Although I have found -140 dbm sensitivity to be useful on 20m on at least one occasion (I was comparing one of my homebrew DC rigs to a K2 capable of -136 dbm), I understand that the lower Bands get progressively more noisy. For example, I understand that 40m rarely has a noise floor of less than -116 dbm (500 Hz BW). Thus the SDR approach of applying variable gain is indeed useful, given the 110 db dynamic signal range the best A/D converters are capable of producing. The good news is that to the extent A/D converters improve (more dynamic range), simply buying a better sound card should be all that is needed to upgrade the performance of the receiver. - Dan, N7VE -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Ahti Aintila Sent: Saturday, September 24, 2005 7:47 AM To: FlexRadio@flex-radio.biz Subject: Re: [Flexradio] Audiocards, USB, etc. Thank you Dan and Jim for the good comments. Sure I noticed how difficult it is to measure the sound cards without proper instruments. The clipping (or compression) levels are easy, but the noise in the present computer environment and with signals approaching the thermal noise levels are challenging. Instead of measuring the audio card only I decided to continue with the whole SDR-1000 system. I recorded 1) the noise floor (dBm/500 Hz) with audio card input cable input connected to the radio and the antenna connector terminated to 50 ohm and then 2) with a signal to radio until clipping or compression was indicated at the line-in connector of the radio or at the SDR-1000 own measurement systems. The results were: Preamp Setting HIGH, -140 dBm/500 Hz, -26 dBm, INA163 out 25 Vpp Preamp Setting MED, -130 dBm/500 Hz, -16 dBm, INA163 out 25 Vpp Preamp Setting LOW, -130 dBm/500 Hz, -13 dBm, INA168 out 4.8 Vpp (1.4 dB compressed) My conclusion is that the QSD can take about 4 Vpp until it starts to saturate and my sound card can take 29 Vpp, so the amplifier after the QSD could have 17 dB voltage gain for optimal results. The front end gain need to be adjusted accordingly. Dan mentioned:"... ideally 130 to 145 db to match the blocking performance of other rigs...". This should be our target and to achieve that we need audio cards handle signal from tens of nanovolts to tens of volts. I estimate, the accuracy of the above measurements is about 1 dB. The measurements were made with PowerSDR 1.4.5 console with unmodified RFE. These figures serve as the reference when comparing the results of the ECO-25 modifications. 73, Ahti OH2RZ ----- Original Message ----- From: "Jim Lux" <[EMAIL PROTECTED]> To: "Tayloe Dan-P26412" <[EMAIL PROTECTED]>; "Tayloe Dan-P26412" <[EMAIL PROTECTED]>; "Ahti Aintila" <[EMAIL PROTECTED]>; <FlexRadio@flex-radio.biz> Sent: Saturday, September 24, 2005 6:20 AM Subject: Re: [Flexradio] Audiocards, USB, etc. > At 05:40 PM 9/23/2005, Tayloe Dan-P26412 wrote: >>I would think that on top of the audio blocking test we also want to >>run audio IP3 tests and audio IP2 dynamic range tests as well to make >>sure that the distortion characteristics are at least as good as the >>SDR1000 front end. >> >>- Dan, N7VE > > > Such tests would be useful, but are quite challenging to make for high > performance systems. You can take two general approaches: > 1) Obssessively account for all the error sources, use very clean sources, > etc. so you can truly know that what you've measured is just the unit > under test, and not quirks of the experimental setup. > > 2) Measure it in a "typical setup", in which case the performance will > certainly measure out worse, but at least it's representative of what > you'll really get. > > > Consider that if you're looking for 140 dB relative levels, you're looking > for signals of a 0.1 microvolt on a 1 volt signal. I've done DC > measurements to 6 digits, and it's, frankly, an ordeal. > > You'd also need sources that are that good, which is no easy matter, > especially if you want to cover the full frequency span of the device > (several decades). For instance, the SRS DS360 claims <-100dBc distortion > from 10mHz to 20 kHz. It's not too pricey at about $3000. > > > > _______________________________________________ FlexRadio mailing list FlexRadio@flex-radio.biz http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz
