[carlito1]

Hi,

 like the title ,whats the maximum mismatch (in db) betwen channels that I can expect with this unit.

[tortugaranger]

Hi,

 like the title ,whats the maximum mismatch (in db) betwen channels that I can expect with this unit.

There are 70 attenuation steps in our LDR preamps. I haven't measured each of these with respect to each other over the full range although I did do a couple of spot checks a couple of years ago on a development board just to confirm we were in the ballpark of expectations.

The self-contained calibration system measures the resistance level of each step against a precise mathematically derived target dB level. This is done for each LDR for each of the 70 steps. The closure criteria for hitting the target dB level at each step during the calibration process is currently 0.2%.  That doesn't necessarily mean the actual channel-to-channel balance is within 0.2% but it's probably not way off either. Probably more like +/- 0.2% which for the 60 dB attenuation range would be around +/- 0.12 dB.  Again, this is not a specification or some guaranteed number. As a practical matter the channel balance has not been a problem with users.

[carlito1]

 the calibration is 0.2%...is that at 10k and 100k ? I know that higher the resistance less precision is achieved with LDR.

So is this 0.2% precision from 10 to 100k ?

 

[tortugaranger]

can I please have a reply on what can I expect as skew betwen the LR channel with the Tortuga LDR ? (db or %)

The LDR calibration algorithm is coded for a 0.2% accuracy of each resistance level of each LDR at each of the 70 attenuation steps which covers a resistance range from 100 ohms to 100,000 ohms.

The formula for dB = 20 * Log (Rshunt/(Rshunt + Rseries))

If both the shunt and series LDR are equally off by 0.2% then you still get the exact dB you're targeting because we're dealing with ratios. So let's instead assume that the Rshunt is off by +0.2% and the Rseries is off by -0.2%. If we crank that through the dB  formula you get 0.035 dB error from the target. Let's further assume that the other channel is just the opposite so you now have a 0.07 dB channel-to-channel dB error.

All fine and good but we all know the real world introduces additional errors. So let's assume the REAL error in the calibrated resistance levels is actually 5 times worse or 1%. Now we get a 0.174 dB error from the target dB schedule and up to 0.348 dB difference between left and right channel.

Reality is likely somewhere in between on average over the full attenuation range.

[carlito1]

Ok so the basis of my question is that I read on another forum that at 100k target  , the measured resistance of the photo resistor was off by 5%  while at 10k the measured resistance was off by 1.5%

 Basically I was trying to know if you actually measured the resistance at 100k not readings from soft.

[tortugaranger]

Ok so the basis of my question is that I read on another forum that at 100k target  , the measured resistance of the photo resistor was off by 5%  while at 10k the measured resistance was off by 1.5%

 Basically I was trying to know if you actually measured the resistance at 100k not readings from soft.

First, if someone wants to measure the actual resistance of the LDRs at a given volume level and use this to compare channel balance you would have to measure both the series and shunt LDRs on both sides and then compute the effective dB = 20 x log (Rshunt/(Rshunt+Rseries)) for both sides. Series resistance is measured between input and output. Shunt resistance is measured between output and ground. The fact that the absolute resistance levels may not be the same in both channels isnt all that important. What is important is the comparative ratios per the above dB calc.

Second, it wouldn't surprise me at all if the least accurate measurement turned out to be at step 1 which is supposed to be -60 dB which is barely audible. This is the only attenuation point where resistance goes all the way up to ~100k and it's only the series resistor that goes that high. Resistance drops down quickly such that by step 20 or so it's down to 20k or less (depending of course on impedance setting). Most listening will happen in the step 30-50 range. The reason step 1 is likely to be the least accurate is that controlling the resistance up around 100k is most difficult due to the really high control gain at this elevated level.

I measured the V2 controller's actual dB in each channel back when it was in development. Sufficient to know it was within expectations. Have not measured it since but would do so again if we made a material change to the calibration algorithm as a double check.

To my knowledge channel imbalance has not been a problem. I would usually hear about if from customers if it were.