Hi,
i also build Lightspeed attenuator.
LDRs are matched at 3 points -currents (0.15, 0.6 and 1.8mA).
It sounds better than relay volume attenuator with Vishay-dale resistors.
But soundstage is smaller, goes between speakers, it is much more deeper, instruments sounds more real - natural.
Maybe i missing little warmer sound. It is not so relax, maybe i need to try with better source. I will test with phono to be more objective.
I use battery 12V and 7805 regulator.
Test was with Hiraga Monstre amplifier like passive preamp.
Any suggestions? Better with buffer?
thanks,
bostjan
i also build Lightspeed attenuator.
LDRs are matched at 3 points -currents (0.15, 0.6 and 1.8mA).
It sounds better than relay volume attenuator with Vishay-dale resistors.
But soundstage is smaller, goes between speakers, it is much more deeper, instruments sounds more real - natural.
Maybe i missing little warmer sound. It is not so relax, maybe i need to try with better source. I will test with phono to be more objective.
I use battery 12V and 7805 regulator.
Test was with Hiraga Monstre amplifier like passive preamp.
Any suggestions? Better with buffer?
thanks,
bostjan
EUVL said:
@peufeu
No I have not tried the LDR with a output buffer. Should I? Did you do that before?
I have tried several preamps before but in my opinion passive it the best solution.
@EUVL
>>I suppose if you do all passive, probably not surprising.
>>I only intend to use between 50 ohm to 2.5k ohm.
My source has a 50 Ohm output, the input impendance of my amp is 100.000 Ohm. The cables are low capacitive. So 18.9 kOhm should not be a problem. But I will compare it with a LDR in shunt configuration with a lower over all impedance.
>>No, that is not what I meant. To do autocalibration ....
With a µC and a 10 bit ADC it is no problem to measure resistance over a wider range in good accuary of < 1% or less. For my setup I have used 6 different resistors which I can switch alternately in a voltage divider (the LDR is second R in the voltage divider). The measurements were done for each resistor several times to increase accuracy from 10 bit a little. After that the µC Software has to look in what resistance area is the LDR and use only the measurement from the R next to LDR resistance. I used mainly 1 % resistors for the voltage divider resistors, only with that the results in a range from 100 Ohm to 47 KOhm are quite good. But for use in a autocalibration this has to be optimized of course (0.01 % or less Rs). Alternatevely you could change the reference voltage for the ADC to have always best resolution or use a external ADC with better resolution.
For current control I used a DAC with 2048 (11 bit) steps from 0,01 to 20 mA. For the low range this is not enough, there is the need to use better resolution. Any ideas to realize a log progression with a voltage to current converter circuit?
As you mentioned best way would be to store the currrent/resistor results in EEprom and use this as a look up table. This is my plan also.
Arne
No I have not tried the LDR with a output buffer. Should I? Did you do that before?
I have tried several preamps before but in my opinion passive it the best solution.
@EUVL
>>I suppose if you do all passive, probably not surprising.
>>I only intend to use between 50 ohm to 2.5k ohm.
My source has a 50 Ohm output, the input impendance of my amp is 100.000 Ohm. The cables are low capacitive. So 18.9 kOhm should not be a problem. But I will compare it with a LDR in shunt configuration with a lower over all impedance.
>>No, that is not what I meant. To do autocalibration ....
With a µC and a 10 bit ADC it is no problem to measure resistance over a wider range in good accuary of < 1% or less. For my setup I have used 6 different resistors which I can switch alternately in a voltage divider (the LDR is second R in the voltage divider). The measurements were done for each resistor several times to increase accuracy from 10 bit a little. After that the µC Software has to look in what resistance area is the LDR and use only the measurement from the R next to LDR resistance. I used mainly 1 % resistors for the voltage divider resistors, only with that the results in a range from 100 Ohm to 47 KOhm are quite good. But for use in a autocalibration this has to be optimized of course (0.01 % or less Rs). Alternatevely you could change the reference voltage for the ADC to have always best resolution or use a external ADC with better resolution.
For current control I used a DAC with 2048 (11 bit) steps from 0,01 to 20 mA. For the low range this is not enough, there is the need to use better resolution. Any ideas to realize a log progression with a voltage to current converter circuit?
As you mentioned best way would be to store the currrent/resistor results in EEprom and use this as a look up table. This is my plan also.
Arne
JoshK said:
Josh is right on the money, just use the series shunt arrangement as I have outlined, do some matching then calibrate them together with the 1k multiturn trimpot in the + feed line to the series led and you have it matched.
I had to chuckle to myself on how complicated some of you guys have tried to make this relitively simple circuit, remember I have been doing the Lightspeed now for over 20 years, and it has come down to being this simple. KIS is the word
And try to use the NSL32SR2's as they are from 40ohm-1mohm where the NSL32SR3's are from 60ohm-25mohm, look at the impedence span the 3's are 25x greater, that is 25x less likely to get a close match, also the 2's will give you a lower minimum volume, better for dinner parties.
Cheers George
Arne, you are right in saying keep it as designed, as your setup with 50ohm c/d source and the 100k poweramp input is a perfect match for the Lightspeed as I have designed it.
If you or others start putting in buffers before and or after the Lightspeed all you are doing is to add colourations and adding thinness to the sound of what is great sounding stand alone product. Better to get the source impedence down and the poweramp input impedence up as you have.
Arne you have a perfect match with your 50ohm and 100kohm there will not be any gain to be had either in transients bass slam or attack by adding buffers, all you'll do is to butcher the richness of sound of the Lightspeed.
Cheers George
If you or others start putting in buffers before and or after the Lightspeed all you are doing is to add colourations and adding thinness to the sound of what is great sounding stand alone product. Better to get the source impedence down and the poweramp input impedence up as you have.
Arne you have a perfect match with your 50ohm and 100kohm there will not be any gain to be had either in transients bass slam or attack by adding buffers, all you'll do is to butcher the richness of sound of the Lightspeed.
Cheers George
Hi George,
You are right, KIS is the world but "Projects by the fanatics, for the fanatics" is the motto of this forum
The lightspeed in its original version is astounding simple and a great bargain. But why not use some state of the art technique to make it more handy to regulate the volume in a comfortable way and add remote control capability.
The world is waiting for the warp pre
Concerning the usage of buffers I agree that you will only put some coloration to the sonics that you can like more or not. Thats why I have been using passive attenuators for years.
Arne
You are right, KIS is the world but "Projects by the fanatics, for the fanatics" is the motto of this forum
The lightspeed in its original version is astounding simple and a great bargain. But why not use some state of the art technique to make it more handy to regulate the volume in a comfortable way and add remote control capability.
The world is waiting for the warp pre
Concerning the usage of buffers I agree that you will only put some coloration to the sonics that you can like more or not. Thats why I have been using passive attenuators for years.
Arne
FastEddy said:
As this is low production for me, by word of mouth, it is all I need for now. ( to buy it, email me through the forum)
Page 5 of this topic shows my finished product, and page 10 gives you the circuit if you want to diy.
My finished product is basic, one input one output, I have a prototype with mutiple inputs, but I swear it does not sound as good as the single even though it's all ofc cable gold plated Alps selector switch contacts, so this is the way it stays, if multiple inputs are needed, I say use a QED input switch box that has ofc cable and gold plated Alps switcher, that way you can still always go back to direct for the best possible sound.
As for remote motor driven pots, remotes and recievers this would change my price to almost double for the finished product, and this is not what (for me) Lighspeed Attenuator is all about.
My motto is KIS, if I could I would drive the speakers from the cd players output!
Cheers George
George,
My sincere apologies for appearing to "highjack" the Lightspeed thread.
LDRs are very interesting components and you brought us a lot of information and your practical experiences with commercial components, for which we are all very grateful.
But it appears a few of us might have other applications in mind for the device. I hope you could bear with us on that.
I am sure if anyone wants to build a pure passive preamp, there is no better approach than to follow your instructions.
Best regards,
Patrick
My sincere apologies for appearing to "highjack" the Lightspeed thread.
LDRs are very interesting components and you brought us a lot of information and your practical experiences with commercial components, for which we are all very grateful.
But it appears a few of us might have other applications in mind for the device. I hope you could bear with us on that.
I am sure if anyone wants to build a pure passive preamp, there is no better approach than to follow your instructions.
Best regards,
Patrick
No grumbling here, you guys are free to do as you please, I was just answering a couple of question for members who are taking the simple approach, and that have the perfect setup for input/output matching for that simple approach.
There are those of us who love a batttery of electronics between the source and the ears and those of us who want straight wire with no gain, gees I think I've just taken the straight wire with gain motto one step further.
Cheers George
There are those of us who love a batttery of electronics between the source and the ears and those of us who want straight wire with no gain, gees I think I've just taken the straight wire with gain motto one step further.
Cheers George
I was wondering if someone could explain something about this circuit to me (as pictured in post #92).
If the 5v source is going though a 100k pot, unless you are at the extremes of the pot's range, how do you get enough current to turn on the LEDs?
5v into 50k, for example, only gives you .1 mA, and that ignores the diode drop.
What am I missing?
If the 5v source is going though a 100k pot, unless you are at the extremes of the pot's range, how do you get enough current to turn on the LEDs?
5v into 50k, for example, only gives you .1 mA, and that ignores the diode drop.
What am I missing?
Hello preiter,
The pot works like a dimmer. At those very low currents through the diode it will be almost dark inside the LDR device, and the more you raise the current the brighter it gets.
The brighter the light, the more the resistive element of the LDR goes down towards below 50 ohms.
When the LED is completely dark, the resistance is up at a few megaohms.
The pot works like a dimmer. At those very low currents through the diode it will be almost dark inside the LDR device, and the more you raise the current the brighter it gets.
The brighter the light, the more the resistive element of the LDR goes down towards below 50 ohms.
When the LED is completely dark, the resistance is up at a few megaohms.
I get that's what it's supposed to do, and does do I'm sure.
But LEDs turn fully on at around 20mA. I don't know how much current is required to turn them on a little bit, but I would guess it was at least a few milliamp, or under a few kOhms (with a 5V source).
But the pot is 100kOhms, so at the very top end of the range, one LED would be on. At the very bottom end of the range, the other LED would be on. Through most of the range, both LEDs would be off.
At least that is the way it seems to me. Except that people have built this and obviously it works, so I must not be getting something.
But LEDs turn fully on at around 20mA. I don't know how much current is required to turn them on a little bit, but I would guess it was at least a few milliamp, or under a few kOhms (with a 5V source).
But the pot is 100kOhms, so at the very top end of the range, one LED would be on. At the very bottom end of the range, the other LED would be on. Through most of the range, both LEDs would be off.
At least that is the way it seems to me. Except that people have built this and obviously it works, so I must not be getting something.
Preiter: I haven't assembled a complete attenuator, but I made a little test to check what you were suggesting.
Hooked up a 9V battery and a 100k pot in series with an NSL32SR2 LDR, and measured steady changes in the "output resistance". It was around 1k. In this range each 5k drop in the pot resistance turned the LDR resistance down almost exactly 60 ohms.
9/100000 = 0.09mA, so I suppose it works just as well for the Lightspeed attenuator with its 5/100000 = 0.05mA range.
Hooked up a 9V battery and a 100k pot in series with an NSL32SR2 LDR, and measured steady changes in the "output resistance". It was around 1k. In this range each 5k drop in the pot resistance turned the LDR resistance down almost exactly 60 ohms.
9/100000 = 0.09mA, so I suppose it works just as well for the Lightspeed attenuator with its 5/100000 = 0.05mA range.