Measurement Data for the LDRx Passive Preamps

0 Members and 1 Guest are viewing this topic. Read 3397 times.

tortugaranger

  • Facilitator
  • Posts: 1547
  • Handcrafted high performance audio
    • Tortuga Audio
Measurement Data for the LDRx Passive Preamps
« on: 1 Sep 2013, 12:54 pm »
Well, this leads to the next question then....when will you be posting measurements of your LDR?

The answer is as soon as I have them. I don't have the test equipment to do it properly at our shop and had planned way back when on having a third party testing lab handle it. The unvarnished truth is I put in on the back burner and it fell off the stove. Since you're not the first person to ask this question recently, it's now back on the front burner and will get some attention. So as soon as I get the results back, I'll post them.

But of course the LDRx passives will still sound great with or without measurements!   :green:

jtwrace

Re: Measurement Data for the LDRx Passive Preamps
« Reply #1 on: 1 Sep 2013, 01:12 pm »
The answer is as soon as I have them. I don't have the test equipment to do it properly at our shop and had planned way back when on having a third party testing lab handle it. The unvarnished truth is I put in on the back burner and it fell off the stove. Since you're not the first person to ask this question recently, it's now back on the front burner and will get some attention. So as soon as I get the results back, I'll post them.

But of course the LDRx passives will still sound great with or without measurements!   :green:
Any idea when you will have them? 

Yes, they will subjectively sound good.   :P  I no longer use anything that hasn't been measured...   :)

tortugaranger

  • Facilitator
  • Posts: 1547
  • Handcrafted high performance audio
    • Tortuga Audio
Re: Measurement Data for the LDRx Passive Preamps
« Reply #2 on: 1 Sep 2013, 02:39 pm »
Any idea when you will have them? 

Yes, they will subjectively sound good.   :P  I no longer use anything that hasn't been measured...   :)

I would expect to have it done and posted in September.

So just to be clear, as a matter of ...I guess you could call it audio principle.... no matter how subjectively good something sounded to your ears, would you reject it if it lacked sufficient accompanying objective test data to provide a rational basis for the subjective goodness? What I'd call a purist anti-snakeoil stance.  I'm an engineer/techie so I always look for an objective basis for understanding what I'm experiencing. I get that. I personally detest being spun with marketing B.S. that I can't make technical sense out of.  But I guess I fall somewhere in the middle where I'll also trust my subjective ears if it makes me happy despite the data (or lack thereof). I've had some fairly harsh critics take a crack at listening to our LDRx units and at least subjectively, the consensus is very positive. Who am I to argue? :lol:

Will post an update on the measurements and timing thereof as soon as I have something more definitive.

Davey

  • Full Member
  • Posts: 1458
Re: Measurement Data for the LDRx Passive Preamps
« Reply #3 on: 1 Sep 2013, 04:00 pm »
So you publish some measurement data.....then what?
Subjectivists don't care and measurement wonks will immediately start querying about the methodology and validity.
(It's a real double-edged sword.)  :)

The LDR-based attenuators have built up a following almost entirely on their subjective performance, so I don't see much point in entertaining too much technical discussion of.
And on AC you have a largely subjectivist crowd.  :)

Cheers,

Dave.

barrows

  • Industry Participant
  • Posts: 455
Re: Measurement Data for the LDRx Passive Preamps
« Reply #4 on: 1 Sep 2013, 04:37 pm »
Hmm.  I would certainly prefer to see measurements.  Including input and output impedance (very important for component matching with passive systems), distortion and noise at a number of different volume settings, and especially, I would like to know how precisely the LDRs end up matching at different levels: this is especially important for balanced circuits.

tortugaranger

  • Facilitator
  • Posts: 1547
  • Handcrafted high performance audio
    • Tortuga Audio
Re: Measurement Data for the LDRx Passive Preamps
« Reply #5 on: 1 Sep 2013, 07:42 pm »
....and especially, I would like to know how precisely the LDRs end up matching at different levels: this is especially important for balanced circuits.

On this question, which I agree is highly germane, we have abundant data.

Let's start with the performance of individual LDRs of the same make and model. The pic below shows the results of plotting R-resistance vs. I-current for approximately 150 LDRs between 100 and 100k ohms. The first pic shows this relationship with R plotted logarithmically.

The first thing you notice is there's considerable variability from one LDR to the next. This is challenge #1 with LDRs. Getting 2 or more that match.



To drive home the second challenge with LDRs, the pic below shows this same data but now the R is plotted linearly with I. The nonlinearity becomes very pronounced much below 10k ohms. 



So let's say we grab 4 of these LDRs that are most similar, put them into an LDRx preamp board. 2 LDRs are used in each stereo channel. One is in series with the audio signal and the second shunts the audio signal to ground. The output signal is taken between the series and the shunt. We run the whole board assembly through almost its maximum volume range which corresponds to internal IC/CPU steps from 0-255. I'm not going to go into any more detail on all that since its proprietary. But the results of this system test (measured data) are shown below for an assembled LDRx board with 4 matching LDRs.

Matching yes, kind of, but even so this is hardly evidence of perfectly matched LDRs. No doubt we could have tested another 50-100 individual LDRs and arrived at a more precisely matched set of four. But lets move on.



Now let's see what these series-shunt pairs of LDRs look like in terms of their attenuation characteristics. This next pic shows the resulting raw attenuation curve before we do any correcting or distill this down into the final volume step vs. dB audio taper. Not too bad really. But we can do better.



Next we apply a correction algorithm where we move the test results around to better match left and right channel attenuation. The resulting attenuation curves are now in lockstep over the full attenuation range (within fractions of a dB). Now we're rock'n!  Please note that this isn't just some visual gimmickry. We actually shifted the selected data points of one channel up or down a few steps so that it matched the other channel.



Lastly, we take the above corrected attenuation data and convert it into the 70 volume steps programmed into the final software. The 70 steps cover ~60dB. Notice we bend the audio taper over the last few dB. This helps to lessen the tendency to come on too strong with big sound pressure steps as the volume gets close to max (0 dB attenuation) level. 

The result is each LDRx passive preamp has its own custom corrected attenuation curve based on the performance test results of how the matched LDRs actually perform when installed in each build. 

We use the same methodology to arrive at perfectly matched attenuation curves using 8 LDRs in a balanced preamp configuration.  Without a software based design, this approach, and the positive audible results, would be impracticable if not impossible.

















cab

Re: Measurement Data for the LDRx Passive Preamps
« Reply #6 on: 1 Sep 2013, 09:49 pm »
What happens if the characteristics of the LDR's change with temp or time? Is your software adaptive?

barrows

  • Industry Participant
  • Posts: 455
Re: Measurement Data for the LDRx Passive Preamps
« Reply #7 on: 2 Sep 2013, 02:58 am »
thanks for the info.  I understood how you used SW correction to match the attenuation characteristics of the LDRs.  Would it be possible to give a +/- X dB for the resulting attenuation of each of the four phases in a typical master/slave balanced board set up?  Perhaps a spec from you guaranteeing matching of <+/-"x" dB for all four phases from 0 dB to -60 dB of attenuation would be the most relevant spec.

I also wonder what your experience is of drift in the LDRs over time?  Does anyone know this, or are they considered very stable?

Would the precision of the matching (generally) be comparable to, or better than, a switched resistor attenuator using resistors better than 0.1%?  Thanks for the info, I find this very interesting as a DIY product, thanks for going forward with this project and offering it to DIY enthusiasts.

tortugaranger

  • Facilitator
  • Posts: 1547
  • Handcrafted high performance audio
    • Tortuga Audio
Re: Measurement Data for the LDRx Passive Preamps
« Reply #8 on: 2 Sep 2013, 09:43 pm »
thanks for the info.  I understood how you used SW correction to match the attenuation characteristics of the LDRs.  Would it be possible to give a +/- X dB for the resulting attenuation of each of the four phases in a typical master/slave balanced board set up?  Perhaps a spec from you guaranteeing matching of <+/-"x" dB for all four phases from 0 dB to -60 dB of attenuation would be the most relevant spec.

I believe the answer is yes it's possible but I'd have to look further into the data we've generated to date to see how much variation there really is before considering a guarantee value. I grabbed data from a random LDRx unit and plotted it out to get a better feel for it. I've not seen this visually before. In this particular example, the channel db error is greater at very low volumes and then diminishes significantly up into normal listening range. Worst case was +/- 0.4 dB down around -50 dB. The average over the whole range is 0.0041 dB with a standard deviation of 0.12 dB.



Quote
I also wonder what your experience is of drift in the LDRs over time?  Does anyone know this, or are they considered very stable?

While largely anecdotal vs. hard test data, I've not been able to detect audible drift in the prototype unit that ran for about 4 years pretty much non-stop. Since each LDR contains an LED light source and all LEDs exhibit some drift over a finite life I would expect there to be some drift but I suspect they drift somewhat uniformly. We don't drive our LDRs hard against their current limit so I'd expect minimal drift and very decent longevity. Conventional LEDs have a general life expectancy of 50,000 hours. That's like listening to music 5 hours a day, 365 days a year for 27 years. Not too shabby.

It will be a while yet before we come out with our next generation LDR board but we are noodling with the idea of building in the ability to self-test and regenerate correction curves which would make drift, to the extent it might be happening slowly over time, totally moot. It would require more hardware on the board and considerable software modifications but probably doable without getting crazy expensive.

jtwrace

Re: Measurement Data for the LDRx Passive Preamps
« Reply #9 on: 19 Sep 2013, 02:45 pm »
I would expect to have it done and posted in September.
Any progress?

tortugaranger

  • Facilitator
  • Posts: 1547
  • Handcrafted high performance audio
    • Tortuga Audio
Re: Measurement Data for the LDRx Passive Preamps
« Reply #10 on: 23 Sep 2013, 02:00 pm »
Any progress?

Everything takes longer than expected but we've now engaged an independent testing service who will run one of our LDR3x units through it paces including THD, impedance, frequency response, S/N ratio, channel separation, sensitivity to power supply quality and maybe even sensitivity to small/large signals. Timing is such that the results probably won't be available until around mid October or so.

As an engineer, designer and audio enthusiast I've always been of 2 minds regarding this type of data. Performance data/specs tend to take on a life of its own separate and apart from the subjective experience. It comes down to subjective experience (do I like it?) vs. the objective data (does the data say it's good/better/best?). In an ideal world one compliments the other and but the real world is very often different. Personally, I rarely buy anything based purely on specs alone and will often discount specs (or totally ignore them) if I'm otherwise happy with the results (i.e I'm satisfied with how it sounds to me). There are others who engage in endless spec wars and tend to become detached from the purpose behind it all - does it sound good to you and make you happy? There's a natural tension between these 2  ways of knowing something and it's that tension that makes life interesting for audio enthusiasts and the people who design/manufacture audio equipment.

With that all said, I will publish the results once they're available. I will posit this: the results will NOT explain why these LDR based attenuators sound so damn good. And then hope I'm wrong about that.  :green:


rollo

  • Industry Participant
  • Posts: 4918
  • Rollo Audio Consulting -
Re: Measurement Data for the LDRx Passive Preamps
« Reply #11 on: 23 Sep 2013, 02:07 pm »
      Shoot man just hook the sucker up and listen to it already. 


charles

tortugaranger

  • Facilitator
  • Posts: 1547
  • Handcrafted high performance audio
    • Tortuga Audio
Re: Measurement Data for the LDRx Passive Preamps
« Reply #12 on: 23 Sep 2013, 02:29 pm »
      Shoot man just hook the sucker up and listen to it already. 
charles
+1 (exactly)