I put a 125H inductor (from a tube amp, cough) between the 12V and the 5V supplies, put 1uF capacitors from rail to ground before and after the inductor, and placed 1uF capacitors across the LEDs. Values were derived from looking in my parts bin to see what I had.
Mike, my source is my computer feeding wav files to an Empirical Audio I2S DAC. Handily beat my Emm Labs SACD setup, which I promptly sold.
Mike, my source is my computer feeding wav files to an Empirical Audio I2S DAC. Handily beat my Emm Labs SACD setup, which I promptly sold.
A little slower
I have built two units from scratch. Both use mono controls to get away from the balancing circuit and to allow for playing one channel louder than the other.
My chokes are 1 mH 250 ma type. I have wired them up two ways.
After the snubbed 5 volt reg, one choke in the positive line to each of the pots. This way in my mind it prevents crosstalk from one pot to the other.
In the other method, one choke goes to both pots carring the 5 volts. The other pot is installed in the ground leg of the regulator to the LED's.
The second method is how both of mine are wired now. This way it isolates any hash on the ground plane from getting to the LED's.
I thought about using a film cap across the LED's. My reason for not using one is the noise is self generated. Not sure if a cap will help to lower the noise of the LED's.
T obe honest, I could not tell if hooking up the chokes either way made any difference. It just felt better with them hooked up as they are now.
One thing I have done is to add a cunstant current load to the regs also. I am powering the blue LED on the front of the unit from the 5 volt reg also.
Moved the blue LED from the reg input to output. Reason is mine now sport switched volume pots. The power supply runs all the time, the optoisilators get turned off when not used. The diode keeps a couple ma running on the output of the reg. Could have used a 2 - 3k resistor to do the same, but running the diode lets me know the power supply is cranking along before powering up the optoisolators. I hope this will triple the lifespan of the ones used. Maybe 10 years.
I like seeing how others do theirs. Learn something all the time. May have to try a more elaborate power supply down the road. The 3 pin regs are not very trendy, a Borbely or Jung reg might feel better, even if it sounds the same.
OT, Tom you feel your computer system is better than the EMM you had? I must think about going in that direction one day. Sure would make some room if all the CD storage cabinets were gone.
George
I have built two units from scratch. Both use mono controls to get away from the balancing circuit and to allow for playing one channel louder than the other.
My chokes are 1 mH 250 ma type. I have wired them up two ways.
After the snubbed 5 volt reg, one choke in the positive line to each of the pots. This way in my mind it prevents crosstalk from one pot to the other.
In the other method, one choke goes to both pots carring the 5 volts. The other pot is installed in the ground leg of the regulator to the LED's.
The second method is how both of mine are wired now. This way it isolates any hash on the ground plane from getting to the LED's.
I thought about using a film cap across the LED's. My reason for not using one is the noise is self generated. Not sure if a cap will help to lower the noise of the LED's.
T obe honest, I could not tell if hooking up the chokes either way made any difference. It just felt better with them hooked up as they are now.
One thing I have done is to add a cunstant current load to the regs also. I am powering the blue LED on the front of the unit from the 5 volt reg also.
Moved the blue LED from the reg input to output. Reason is mine now sport switched volume pots. The power supply runs all the time, the optoisilators get turned off when not used. The diode keeps a couple ma running on the output of the reg. Could have used a 2 - 3k resistor to do the same, but running the diode lets me know the power supply is cranking along before powering up the optoisolators. I hope this will triple the lifespan of the ones used. Maybe 10 years.
I like seeing how others do theirs. Learn something all the time. May have to try a more elaborate power supply down the road. The 3 pin regs are not very trendy, a Borbely or Jung reg might feel better, even if it sounds the same.
OT, Tom you feel your computer system is better than the EMM you had? I must think about going in that direction one day. Sure would make some room if all the CD storage cabinets were gone.
George
I have done some searches here and at the Audio Asylum on LED noise. There was a thread on this that recommended at least a 10uf cap across the LED and also suggested a resistor. I also found a thread here on LED noise.
http://www.diyaudio.com/forums/showthread.php?postid=417008#post417008
http://www.diyaudio.com/forums/showthread.php?postid=417008#post417008
Self generated
I think the bypasses are used to prevent LED noise from polluting a supply shared by other circuits. This is debatable, some very good current sources use LED 's as a voltage reference with no noise suppresion.
Here the LED is the only circuit. I have never seen anything that indicates LED electrical noise causes light intensity fluctuations. With the time constant of the resistive material high frequency variations in light output should be smoothed out.
That does not explain why some hear improved smoothness and resolution after upgrading the supply. maybe there are two time constants. A very short one the can respond to small, fast light intensity variations, and a longer higher swing one that takes a little while to stabilize.
All my testing was done while matching these up. The supply voltage used was very stable. The resistance would take a while to peak out in response to LED current change. Usually I would give it a minute or more before recording the value.
George
I think the bypasses are used to prevent LED noise from polluting a supply shared by other circuits. This is debatable, some very good current sources use LED 's as a voltage reference with no noise suppresion.
Here the LED is the only circuit. I have never seen anything that indicates LED electrical noise causes light intensity fluctuations. With the time constant of the resistive material high frequency variations in light output should be smoothed out.
That does not explain why some hear improved smoothness and resolution after upgrading the supply. maybe there are two time constants. A very short one the can respond to small, fast light intensity variations, and a longer higher swing one that takes a little while to stabilize.
All my testing was done while matching these up. The supply voltage used was very stable. The resistance would take a while to peak out in response to LED current change. Usually I would give it a minute or more before recording the value.
George
George, it seems to me these LDR units are not well investigated. Your speculation about time constants seems reasonable. Here's a question: LED voltage goes up from X then returns to X in one-half the time of the LDR's time constant (Tc). Is there zero effect on the LDR? I think there must be some effect, as how does one otherwise derive Tc? Something must be happening to the LDR from Tzero to Tc.
Assuming an audible dynamic range of 70db and situation with a max signal of 0.1V the minimum audible signal is about 3uV.
3uV is a very small signal.
5 x TC = time for about 90% of change... Right ?
but most change happens in initial period
so we can assume that even with noise frequency periods of less than TC there is still some change - and as we think we can hear about 70 db range that "some change" may well be audible.
I'm guessing broadband noise on LEDs and my guess is the troublesome frequencies will be in low / mid audio frequencies.
With this premise I just put 10V 100uF OSCON caps across my LED's - and listened
They are staying in !
mike
3uV is a very small signal.
5 x TC = time for about 90% of change... Right ?
but most change happens in initial period
so we can assume that even with noise frequency periods of less than TC there is still some change - and as we think we can hear about 70 db range that "some change" may well be audible.
I'm guessing broadband noise on LEDs and my guess is the troublesome frequencies will be in low / mid audio frequencies.
With this premise I just put 10V 100uF OSCON caps across my LED's - and listened
They are staying in !
mike
Where?
Mike,
Did you put them right on the leads to the LDR? I had considered a 1 ufd stacked film cap, or maybe a FM cap.
My power supply is too far away from the LED, guess I need four small caps strapped across the LDRs to see.
You guys may be completely correct on this. But mine will have to wait till next week. Also guess the extreme low level of the LED noise means that no series sub ohm resistor is required.
On the time constants, the large one seen that takes minutes is temperature related. The LDR MAY have a very fast response, just the tempco effect takes a long time on top of actual LED current to resistive material function.
George
Mike,
Did you put them right on the leads to the LDR? I had considered a 1 ufd stacked film cap, or maybe a FM cap.
My power supply is too far away from the LED, guess I need four small caps strapped across the LDRs to see.
You guys may be completely correct on this. But mine will have to wait till next week. Also guess the extreme low level of the LED noise means that no series sub ohm resistor is required.
On the time constants, the large one seen that takes minutes is temperature related. The LDR MAY have a very fast response, just the tempco effect takes a long time on top of actual LED current to resistive material function.
George
Re: Where?
Yes right on the LEDs - ( with polarity observed )
The difference between 1uF in this position & 100uF is major !
I mean M A J OR ! ! ! like yesterday I thought I needed a new CD player . . .
1uF sounded like it might need snubber resistors. This is another reason I thought about a larger value cap. Snubbers are not needed for higher values.
I suggest you just put ur favourite 100uF cap in and see what you think.
Panelhead said:
Yes right on the LEDs - ( with polarity observed )
The difference between 1uF in this position & 100uF is major !
I mean M A J OR ! ! ! like yesterday I thought I needed a new CD player . . .
1uF sounded like it might need snubber resistors. This is another reason I thought about a larger value cap. Snubbers are not needed for higher values.
I suggest you just put ur favourite 100uF cap in and see what you think.
Re: Re: Where?
Hi Mike!
I have been thinking of putting the caps across the leds quite sometime ago to reduce the noise. In this case it will be a 2X blackgate NX Hi-Q 47uf in an anti parallel configuration across each diode since I have few units left . Since u've found out that it will make a different, It is making me itchy again . Coz in my cases, I found that by improving the psu, the sound improves n I'm using a battery supply now.
Cheers!
Eddy.
mikelm said:
Yes right on the LEDs - ( with polarity observed )
The difference between 1uF in this position & 100uF is major !
I mean M A J OR ! ! ! like yesterday I thought I needed a new CD player . . .
1uF sounded like it might need snubber resistors. This is another reason I thought about a larger value cap. Snubbers are not needed for higher values.
I suggest you just put ur favourite 100uF cap in and see what you think.
Hi Mike!
I have been thinking of putting the caps across the leds quite sometime ago to reduce the noise. In this case it will be a 2X blackgate NX Hi-Q 47uf in an anti parallel configuration across each diode since I have few units left . Since u've found out that it will make a different, It is making me itchy again
. Coz in my cases, I found that by improving the psu, the sound improves n I'm using a battery supply now. Cheers!
Eddy.
Re: Re: Where?
Mike, you're confirming my speculation in an early post in this thread. My intuition leans to thinking PSU noise *has to* affect the LDR.
mikelm said:
Yes right on the LEDs - ( with polarity observed )
The difference between 1uF in this position & 100uF is major !
I mean M A J OR ! ! ! like yesterday I thought I needed a new CD player . . .
1uF sounded like it might need snubber resistors. This is another reason I thought about a larger value cap. Snubbers are not needed for higher values.
I suggest you just put ur favourite 100uF cap in and see what you think.
Mike, you're confirming my speculation in an early post in this thread. My intuition leans to thinking PSU noise *has to* affect the LDR.
Another Clue ...
.
I was turning my pot today and I heard a crackle ! ! !
...so I would agree with you tom but I also suspect that some of the noise contribution is from the LED itself, and that it is some variety of broadband noise. So a bigger cap across the LED may address noise issues that caps & chokes in other positions cannot address.
The difference I noticed with the 100uF cap was across the bandwidth with a fuller bodied, more etched out, 3D kind of a sound. Not 3D imaging as such but rather each instrument have depth & occupying a 3D space.
I would be interested to hear what others notice with this mod if you try it out.
mike
ps tom did u get the email I sent ?
.
I was turning my pot today and I heard a crackle ! ! !
...so I would agree with you tom but I also suspect that some of the noise contribution is from the LED itself, and that it is some variety of broadband noise. So a bigger cap across the LED may address noise issues that caps & chokes in other positions cannot address.
The difference I noticed with the 100uF cap was across the bandwidth with a fuller bodied, more etched out, 3D kind of a sound. Not 3D imaging as such but rather each instrument have depth & occupying a 3D space.
I would be interested to hear what others notice with this mod if you try it out.
mike
ps tom did u get the email I sent ?
Crackle
Mike,
That crackle may be your pots. Most pots do not like dc across them for long periods.
I have tried several values and tapers so far. All the 2 watt pots have been quiet. The only 1/2 watt pots tried also made noise when adjusting volume.
This is another reason I like the switched pots now installed. Keeps power off the pot unless actually listening.
Gotta install those electros across the leds soon. If it gets much better I will not be able to stand it.
And, thanks again to Georgehifi for sharing his circuit and this wonderfull method of attenuating input signals.
George
Mike,
That crackle may be your pots. Most pots do not like dc across them for long periods.
I have tried several values and tapers so far. All the 2 watt pots have been quiet. The only 1/2 watt pots tried also made noise when adjusting volume.
This is another reason I like the switched pots now installed. Keeps power off the pot unless actually listening.
Gotta install those electros across the leds soon. If it gets much better I will not be able to stand it.
And, thanks again to Georgehifi for sharing his circuit and this wonderfull method of attenuating input signals.
George
Makes sense that the LED light output stability is important. After all, it is multiplying the audio signal to control the volume, so if it varies according to the audio signal or other noise, you'll hear that...
Has anyone tried to feed the LEDs with a constant current source ? (keeping the // cap) ?
Has anyone tried to feed the LEDs with a constant current source ? (keeping the // cap) ?
peufeu said:
Mmmm ?
A constant current source feeding 4 potentially noisy LED's - isn't that a recipe for interaction ?
...and it would also be cross channel
Perhaps that's why you suggest keeping the caps in the circuit ?
I think it would need 1 CCS per channel - it adds complications, what advantages would you predict ?
mike
quick question
Hi guys,
Have not looked at this thread for a long time (other distractions). The good news is that I got the LDR pre to work (I had connected the thing incorrectly the first time...). I am impressed by the mid-range clarity as well as the clean treble. But, the pre-amp is just to loud, even at the lowest setting.
My pre consists of balanced inputs which run into THAT1200 balanced to single-ended converters. This then runs into the LDR volume pot and this then runs into an AD815 chip which runs to single ended outputs
Balanced input-THAT1200-LDR volume pot-AD815-single ended output
Should I increase the size of the 100 ohm resistors used in series with the LDR's or rather adjsut the gain on the AD815?
Thanks
Ryan
Hi guys,
Have not looked at this thread for a long time (other distractions). The good news is that I got the LDR pre to work (I had connected the thing incorrectly the first time...). I am impressed by the mid-range clarity as well as the clean treble. But, the pre-amp is just to loud, even at the lowest setting.
My pre consists of balanced inputs which run into THAT1200 balanced to single-ended converters. This then runs into the LDR volume pot and this then runs into an AD815 chip which runs to single ended outputs
Balanced input-THAT1200-LDR volume pot-AD815-single ended output
Should I increase the size of the 100 ohm resistors used in series with the LDR's or rather adjsut the gain on the AD815?
Thanks
Ryan
Current limiting resistors
Ryan,
I think that lowering the value of the 100 ohm resistors will increase the maximum attenuation. This will lower the shunt resistance at maximum attenuation.
I have one built with an AD815 also. Great combo!
My best recommendation is to make the gain of the AD815 2. Sounds better to me than it did as a buffer, unity gain. Depends on system gain, I do not need any gain, all sources are plenty loud with unity gain or passive.
Try checking how the impedance varies. Measure from inut to ground and rotate pot to see how constant the impedance is. Then mesure from output to grpound to see how low the shunt resistance will drop.
George
Ryan,
I think that lowering the value of the 100 ohm resistors will increase the maximum attenuation. This will lower the shunt resistance at maximum attenuation.
I have one built with an AD815 also. Great combo!
My best recommendation is to make the gain of the AD815 2. Sounds better to me than it did as a buffer, unity gain. Depends on system gain, I do not need any gain, all sources are plenty loud with unity gain or passive.
Try checking how the impedance varies. Measure from inut to ground and rotate pot to see how constant the impedance is. Then mesure from output to grpound to see how low the shunt resistance will drop.
George
Thanks George.
Does your passive use the AD815 as a buffer and do you prefer it that way or with the AD815 set to provide gain?
Have you tried the pot without the AD815?
I think I'll try without the pot this evening.
My power amp has an input impedance of 27k. My source is a Theta Gen Va which I believe has amazing driving ability.
Ryan
Does your passive use the AD815 as a buffer and do you prefer it that way or with the AD815 set to provide gain?
Have you tried the pot without the AD815?
I think I'll try without the pot this evening.
My power amp has an input impedance of 27k. My source is a Theta Gen Va which I believe has amazing driving ability.
Ryan