OK, more precisely, why do "home builders have a line that says "occupation" so they can charge an extra 50K for anyone that is a lawyer".
Probably a joke, but I just don't get it...
Probably a joke, but I just don't get it...
CMRR
- Thread starter Gus
- Start date
Help Support GroupDIY Audio Forum:
Because lawyers sue everybody all the time for almost anything....
a crass stereotype but sadly with more than a little truth.
JR
a crass stereotype but sadly with more than a little truth.
JR
OK, thanks.
Newmarket
Well-known member
In the context of a book on VLSI it seems reasonable to focus on hf aspects.
I'm far from being opposed to OT, but this thread is in the Microphones section...
We could switch back to Bills original CMRR thread , seems like the obvious place to continue . I dont mind either way .
Abbey,
In relation to the cosmos tweak app when we vary the proportions of 2nd and 3rd D. , I think what were actually doing is compensating for the effects of imbalance at the input to the ADC , both in the cables and any mismatch in the output drive circuitry of the device on test The first logical test is to make sure the patch cables arent contributing to that unbalance .
Lets say we have two xlr-xlr patch leads on bench , there different values ,all over the place both conductor to conductor and capacitance on each leg to ground , if we inadvertantly swap an input for an output cable its going to put a scew in our results both at the input and output at both ends , when looking at noise floor or very tiny harmonics over a very wide range these things seem to seem to matter more . Were caught in a funk until we know to a high degree of certainty our probe or patch cables for what they do contribute are identical .
IMPORTANT:
Ivan highly reccomends screen shot-ing the original calibration set up of the ADC on the Tweek app as he doesnt store the calibration data info for each device. You need access to an AP if you loose the calibration .
Incidently including a calibration chart with the device would be a nice touch , and maybe cut down on guys crying over spilt milk when they've broken it .
Heres some new data about the 9822 ADC . the newer documents relating to the 9038 are also available ,also worth a look . Not sure are these releases are official or not but at least we dont need to sign the non dis anymore , F&^%$rs
Abbey,
In relation to the cosmos tweak app when we vary the proportions of 2nd and 3rd D. , I think what were actually doing is compensating for the effects of imbalance at the input to the ADC , both in the cables and any mismatch in the output drive circuitry of the device on test The first logical test is to make sure the patch cables arent contributing to that unbalance .
Lets say we have two xlr-xlr patch leads on bench , there different values ,all over the place both conductor to conductor and capacitance on each leg to ground , if we inadvertantly swap an input for an output cable its going to put a scew in our results both at the input and output at both ends , when looking at noise floor or very tiny harmonics over a very wide range these things seem to seem to matter more . Were caught in a funk until we know to a high degree of certainty our probe or patch cables for what they do contribute are identical .
IMPORTANT:
Ivan highly reccomends screen shot-ing the original calibration set up of the ADC on the Tweek app as he doesnt store the calibration data info for each device. You need access to an AP if you loose the calibration .
Incidently including a calibration chart with the device would be a nice touch , and maybe cut down on guys crying over spilt milk when they've broken it .
Heres some new data about the 9822 ADC . the newer documents relating to the 9038 are also available ,also worth a look . Not sure are these releases are official or not but at least we dont need to sign the non dis anymore , F&^%$rs
Last edited:
It's not my understanding. According to the website:
" Why you may need that app? In case, if you have a precise sine-wave source with surely ultra-low distortions(AP analyzer, custom -150db clean 1kHz oscillator, or Cosmos DAC), you can re-calibrate Cosmos ADC distortions. For example, AP analyzers require regular annual re-calibration, hence, such a precise ADC as Cosmos needs to be re-calibrated for THD as well."
I read somewhere that some use it to compensate for THD in the DAC. I certainly don't want to do that.
Thanks for these documents.
However, they don't give more pertinent info regarding frequency response at 384k SR, since they are simulated, not measured, and there is no guarantee that the responses scale linearly with the sampling rate. Actually not; if it was the case, the FR of the "dongle DAC's", that use the same chip, would be close to 180kHz.
Tweaking a bench test rig for lowest possible measurement floor is nice but academic unless you designing paths that clean, few are.
JR
JR
Newmarket
Well-known member
Just connect screen at both ends tbh.
Also bear in mind that for passive Hi-Z pickups you will probably want to use 'low noise ' cable ie with a semiconducting layer between the central conductor and the screen. Reduces 'cable noise' from cable movement.
There are, of course, options to transform the Hi-Z output to Lo-Z and balance. And also to go back to Hi-Z at the guitar amp end if that is important for you.
Last edited:
Simulation seems to indicate that connecting the shield at one end and using the conductors as signal and ground could result in better rejection of parasitics magnetically induced into the cable, for frequencies in the audio range, without any difference at RF.
Since electric guitars are usually floating, the noise mechanism is not the same as connections between pieces of mains-powered gear.
Noise levels generated there are dwarfed by noise directly induced into magnetic p/u's, both magnetically and electrostatically, though.
A similar simulation applied to non-floating unbalanced equipment shows that using 2cond+shield cable with the the shield connected at the source results in better RF noise rejection.
The problem with this arrangement is that active electronics are needed, close to the instrument. It can't be passive, because even with the best available xfmrs, it's a tone-killer.
AFAIK, there are (were?) a few commercially available systems. Some relied on phantom powering, others on battery operation. Some used balanced connections, others are unbalanced . Actually, a standard DI box is an adequate solution in most cases, except that they lack the interaction between the cable capacitance and the pick-up's inductance. Very few DI's include a "cable compensator".
At the other end, it is essential, in order to realize the interaction with the amp's input, to convert the low-Z signal to a signal comparable to that of the instrument, i.e. with adequate level, but also presenting the correct source impedance. The actual "re-amp" systems can be used for that task; however, many do not model satisfyingly the source impedance.
Newmarket
Well-known member
Yes. My real point was that 'standard' balanced mic cable, StarQuad or not, isn't really suitable for use with passive Hi-Z pickups. And I'm not aware of any balanced cable that incorporates the 'semiconducting' layer. So converting to/from Lo-Z offers a route to incorporate StarQuad. I'm not really saying it's generally worth it btw - depends on cable length and sources of interference of course. And as I think you said, a load of interference is due to pickup from the err...Pickups and, relative to a Lo-Z transducer, a Hi-Z pickup is inherently noisy in "Hiss" terms with a resistive impedance of say typically 10K.
Agree that the level at the other end needs to be in the right ballpark for practical reasons at least. Whether the impedance is essential really depends on if you do regard it as essential. If you want close to exactly same interaction as between the pickup/cable/input then yes of course. But in many cases there won't be a great deal of difference eg if your amp front end is a simple Hi-Z input around a TL07x or similar.
One of the ways I see it is that a guitar signal chain will often have the guitar going directly into some fx pedal - maybe a chain of them - then into an amp.
So the impedance of the signal at the amp input is almost certainly far from that of the guitar itself. But maybe that does miss some of the 'magic'.
Confession ! - I'm mainly a bass player myself and it's often about keeping a good clean signal as the first option rather than losing it in distortion (we'll do that later maybe
On the "re-amp" configuration - yes it has occurred to me that the impedance may not be well replicated esp the several Henries of inductance ?
Can't find the one I was thinking of but Radial have something as you might expect>
SGI - Radial Engineering
Last edited:
Thats the old guitar cable trick, shield at the amp input and run hot and ground as the "balanced" wires. It really makes a difference.
I was digging around and found an old style in line hi to low z transformer ,
the large barrel jack contains the transformer which plugs directly into the instrument ,
seeing as the output is balanced cable shouldnt be much of a worry , handy thing is we have no unbalanced cable to eat away at our high end and allow extraneous handling noise in ,it all kept to a minimum . Even if we took a 25db loss in level in the step down we could still have around 100db S/N at the ADC input , yeah we'd be recording at well below maximum 0dbfs and have a little bit more distortion as a result but the noise of the pickups themselves is always going to be the limiting factor in terms of S/N . Once the track is layed down we simply normalise it upto what ever max level we choose . We could add gain and attenuate like Abbey suggests , its not going to change the basic S/N of the pickups even if we might get a slightly more faithful recording at a higher level .
the large barrel jack contains the transformer which plugs directly into the instrument ,
seeing as the output is balanced cable shouldnt be much of a worry , handy thing is we have no unbalanced cable to eat away at our high end and allow extraneous handling noise in ,it all kept to a minimum . Even if we took a 25db loss in level in the step down we could still have around 100db S/N at the ADC input , yeah we'd be recording at well below maximum 0dbfs and have a little bit more distortion as a result but the noise of the pickups themselves is always going to be the limiting factor in terms of S/N . Once the track is layed down we simply normalise it upto what ever max level we choose . We could add gain and attenuate like Abbey suggests , its not going to change the basic S/N of the pickups even if we might get a slightly more faithful recording at a higher level .
The load presented by the loaded transformer will take care of it. A passive guitar is supposed to see a minimum load of 250k, preferrably 500k or 1Meg.
I just gave the Hi to Lo adapter a quick measure with the LCR and multimeter
The hi-z winding reads about 545H and 2 meg at 100hz and around 2.8 meg at 1khz , I read about 4pf in parallel , 3.3kohms dc resistance . Lo-z winding measures just under 20 ohms dc with around 470pf from each leg to ground including 1.5 meters of twin core screened and the XLR.
This thing is old enough that good quality hi-z transformers were still being made , my guess is its 50s or early 60's .
I have other nice input transformers with big mumetal sheilds , 50 ohm (ac)centre tapped primary and large inductance on the Hi-z side , more than 2000H ,they date back to the late 40's , something like 1:50 ratio , that equates to around 34db attenuation if used from hi to lo . So both those transformers should give a reasonable load to a passive pick up, depends on pickup output if there usable direct to the ADC or not .Humbuckers stand a chance , passive single coils maybe not so good .
The hi-z winding reads about 545H and 2 meg at 100hz and around 2.8 meg at 1khz , I read about 4pf in parallel , 3.3kohms dc resistance . Lo-z winding measures just under 20 ohms dc with around 470pf from each leg to ground including 1.5 meters of twin core screened and the XLR.
This thing is old enough that good quality hi-z transformers were still being made , my guess is its 50s or early 60's .
I have other nice input transformers with big mumetal sheilds , 50 ohm (ac)centre tapped primary and large inductance on the Hi-z side , more than 2000H ,they date back to the late 40's , something like 1:50 ratio , that equates to around 34db attenuation if used from hi to lo . So both those transformers should give a reasonable load to a passive pick up, depends on pickup output if there usable direct to the ADC or not .Humbuckers stand a chance , passive single coils maybe not so good .
Last edited:
