Georges, my comments were a bit sharp. Panelhead's comments first appeared to me more dismissive than they now do. Accept my apologies.
My experience with tube heaters leads me to suppose the LED supply is likely audible. If this supposition is taken in line with my audio experience generally, the LED supply will be more audible than first suspected.
Cheers.
My experience with tube heaters leads me to suppose the LED supply is likely audible. If this supposition is taken in line with my audio experience generally, the LED supply will be more audible than first suspected.
Cheers.
My partial comparison
When I commented on tube heaters, I should have added indirectly heated tubes instead of just tube. I have experianced how much different heating methods effect direct heated triodes.
The two are completely different beasts. I think in my mind that the indirectly heated tubes are more effected by power supply than these optoislators. That is the real beauty of these, the resistive element is completely isolated from the power supply electrically.
One day, I hope to add the Lightspeed clone to a buffer. It may be even a little better. But for now, am completely satified with attenuators. No noise, plenty of gain, great tonal balance, and plenty of punch. Maybe I miss the added colorations from my previous linestages.
If someone finds batteries are better than the dual regs with wallwart I will give it a try. Currently using a +/- 12 volt battery supply for my amplifier input stage. It was am improvement over an active regulated supply.
But first measure either from input to ground as the pot is fuly rotated or use two meters to measure from input to output, and the other from output to ground. This will allow actually measuring the impedance of the pot as it is operated. Doing this and writting down the numbers with the two meter method helps in measuring the tracking of the attenutor also.
Do this with both the active and battery supply to make sure it is an apples to apples comparison. Changing the pot impedance will effect the sonics a bunch.
George
When I commented on tube heaters, I should have added indirectly heated tubes instead of just tube. I have experianced how much different heating methods effect direct heated triodes.
The two are completely different beasts. I think in my mind that the indirectly heated tubes are more effected by power supply than these optoislators. That is the real beauty of these, the resistive element is completely isolated from the power supply electrically.
One day, I hope to add the Lightspeed clone to a buffer. It may be even a little better. But for now, am completely satified with attenuators. No noise, plenty of gain, great tonal balance, and plenty of punch. Maybe I miss the added colorations from my previous linestages.
If someone finds batteries are better than the dual regs with wallwart I will give it a try. Currently using a +/- 12 volt battery supply for my amplifier input stage. It was am improvement over an active regulated supply.
But first measure either from input to ground as the pot is fuly rotated or use two meters to measure from input to output, and the other from output to ground. This will allow actually measuring the impedance of the pot as it is operated. Doing this and writting down the numbers with the two meter method helps in measuring the tracking of the attenutor also.
Do this with both the active and battery supply to make sure it is an apples to apples comparison. Changing the pot impedance will effect the sonics a bunch.
George
Hi guys!
It looks like my experiment has made this forum alive a bit
. I did a measurement after the regulator. The voltage is 5.08V. I'm not sure about the impedance. Anyway, I also did sumthing else. The cable from battery to pre is also shielded n with a ground out. Fyi, the sound with the cable grounded n not grounded also different. The grounded one will give a silencer background n better tonality. This is my finding using a shielded cable from the battery to pre. U guys can try it out. YMMV.
Cheers!
Eddy.
P/S : Also got few other tricks.
It looks like my experiment has made this forum alive a bit
. I did a measurement after the regulator. The voltage is 5.08V. I'm not sure about the impedance. Anyway, I also did sumthing else. The cable from battery to pre is also shielded n with a ground out. Fyi, the sound with the cable grounded n not grounded also different. The grounded one will give a silencer background n better tonality. This is my finding using a shielded cable from the battery to pre. U guys can try it out. YMMV.Cheers!
Eddy.
P/S : Also got few other tricks.
georgehifi said:
Hi George!
Actually I'm from Malaysia. The voltage here is like Australia. It is 240V n it can go as low as 235V in my area. Quite stable. Anyway, how's the listening test going on? Just tested the pre with the battery psu against reputable passive pre. The lightspeed is way ahead. It is so silence with no glare at all. My fren said this design will revolutinise the passive pre market. Good for the purist.
Cheers!
Eddy.
Good experiments, Eddy. I think you're onto something. My guess is the LED power supply *is* critical. When I said the optocoupler connects your audio signal to the wall, I meant that control of the amplitude of that signal is subjected to variations in wall voltage. IMO, those variations must transfer, to some degree, into the signal. Any such transfer would of course be nonlinear, if that's the right word, as the cell responds with a time lag. That lag is very likely frequency dependent, etc., etc. I suspect it probably changes the frequency of LED supply variations---perhaps in the manner of DA in capacitors. I seriously doubt it effectively cancels, by smoothing or otherwise, any such variations, regardless the frequency.
Of course, any AC transfer from LED supply to audio signal path raises the question of audibility. That's where listening becomes important, ey?
Of course, any AC transfer from LED supply to audio signal path raises the question of audibility. That's where listening becomes important, ey?
A couple of questions come to mind.
To achieve say 30dB attenuation, one need to switch the shunt element pretty hard on, with something like 20mA through the LED (I think even higher with Perkin Elmer). This means just the LED alone dissipates some 30mW. Even that the power in the resistive element is negligible, doesn't the package get warm ? How much does the resistance drift typically (combined effect of LED & LDR drifts) ? Any problems with thermal noise (from the LED) ?
How many LDR's would one have to order typically to get say 2 sets of 5 matched to better than 0.5dB ?
Many thanks for any hints.
Patrick
To achieve say 30dB attenuation, one need to switch the shunt element pretty hard on, with something like 20mA through the LED (I think even higher with Perkin Elmer). This means just the LED alone dissipates some 30mW. Even that the power in the resistive element is negligible, doesn't the package get warm ? How much does the resistance drift typically (combined effect of LED & LDR drifts) ? Any problems with thermal noise (from the LED) ?
How many LDR's would one have to order typically to get say 2 sets of 5 matched to better than 0.5dB ?
Many thanks for any hints.
Patrick
I don't know about the Perkin Elmer ones but the
Silonex NSL-32SR2S (sorted), the one I use in the Lightspeed, work well within their specified limit, I have had not one failure of the 100 odd in operation yet. My personal one has now been on for about 2.5 years 24/7 and still as good as the day it was made.
As for numbers you need to match for one stereo unit, you need 4 of the LDR's, to get this in the Silonex NSL-32SR2S (sorted) you'll need to get about 10 of them, as for the Perkin Elmer ones, they are not sorted, I would say about 20-30.
You see the the Silonex NSL-32SR2S (sorted) are alphabticaly graded from A to G and it would be wise to ask for all of the same letter to get an even better match.
Cheers George
Silonex NSL-32SR2S (sorted), the one I use in the Lightspeed, work well within their specified limit, I have had not one failure of the 100 odd in operation yet. My personal one has now been on for about 2.5 years 24/7 and still as good as the day it was made.
As for numbers you need to match for one stereo unit, you need 4 of the LDR's, to get this in the Silonex NSL-32SR2S (sorted) you'll need to get about 10 of them, as for the Perkin Elmer ones, they are not sorted, I would say about 20-30.
You see the the Silonex NSL-32SR2S (sorted) are alphabticaly graded from A to G and it would be wise to ask for all of the same letter to get an even better match.
Cheers George
> You see the the Silonex NSL-32SR2S (sorted) are alphabticaly graded from A to G and it would be wise to ask for all of the same letter to get an even better match.
If I understand correctly from their website / datasheet, the S version is only graded, but one cannot specify a specific grade or order a quantity of the same grade. See note 3 of the datasheet.
Or did you have a different experience ?
Did you order directly from the manufacturer ?
Most of their distributors (RS, Allied or Farnell) have stopped stocking the S (sorted) version and only sell the standard version now.
Also I wonder why you choose to use the 32R2, since they specify in their application note that 32R3 has the lowest distortion.
Patrick
If I understand correctly from their website / datasheet, the S version is only graded, but one cannot specify a specific grade or order a quantity of the same grade. See note 3 of the datasheet.
Or did you have a different experience ?
Did you order directly from the manufacturer ?
Most of their distributors (RS, Allied or Farnell) have stopped stocking the S (sorted) version and only sell the standard version now.
Also I wonder why you choose to use the 32R2, since they specify in their application note that 32R3 has the lowest distortion.
Patrick
I chose the NSL-32SR2S for it's on resistance and for it's dark resistance, but more also the fact they are sorted and I can get better quicker matching from them.
If you say they are no longer available, I will have to look elsewhere when my stocks run out, this is a pity but i have enough to last a year or so, by then they may start sorting them again, hopefully.
Cheers George
If you say they are no longer available, I will have to look elsewhere when my stocks run out, this is a pity but i have enough to last a year or so, by then they may start sorting them again, hopefully.
Cheers George
You can also get the NSL32SR3S as sorted, same as R2S. According to Silonex, it has the least distortion.
I am sure you can still get them sorted from Silonex, just no longer from Farnell or RS, which are convenient sources in Europe.
Maybe I should try to get them directly from Silonex.
Do they have minimum quantities ?
Maybe we should start a group buy here ......
Patrick
I am sure you can still get them sorted from Silonex, just no longer from Farnell or RS, which are convenient sources in Europe.
Maybe I should try to get them directly from Silonex.
Do they have minimum quantities ?
Maybe we should start a group buy here ......
Patrick
Count me in on a group buy if someone is willing to match them. Got a question. I have to much gain. 95db line sources. I am seeing about 1.45 volts after the 100K volume pot, with it all the way attenuated. Input voltage is 5.05 volts DC regulated. Would you try a resistor before the volume pot or after to attenuate the voltage?
EUVL said:
Just be careful with the NSL32SR3S I'd get samples before buying a bulk lot, as I don't think your off volume will be as low as the NSL32SR2S and the full up may not get up as high.
Also the input and output impedence will be different to the NSL32SR2S, it's around 7k input and output, if the output gets too high youwill start to need poweramps with input impedences much higher than when using the NSL32SR2S.
As for distortion it's is not a factor of the audio signal, the added distortion is zero by the ldr.
Cheers George
