John,
Those are the right ones. However, be warned that Allied can be VERY expensive on the shipping costs. Ask them to commit to a range so that you have an idea.
In my case, they delayed with the order, eventually (about 3 weeks after I placed the order) saying that they did not have stock. When I queried this with Silonex (Mary), Allied changed their tune and shipped the stock, but, given the delays they had incurred in dealing with my order, chose the fastest way to get them to me (without my prior consent) and I ended up paying Euro60 for 10! I guess the cost was exacerbated by the fact that I live in South Africa.
Ryan
Those are the right ones. However, be warned that Allied can be VERY expensive on the shipping costs. Ask them to commit to a range so that you have an idea.
In my case, they delayed with the order, eventually (about 3 weeks after I placed the order) saying that they did not have stock. When I queried this with Silonex (Mary), Allied changed their tune and shipped the stock, but, given the delays they had incurred in dealing with my order, chose the fastest way to get them to me (without my prior consent) and I ended up paying Euro60 for 10! I guess the cost was exacerbated by the fact that I live in South Africa.
Ryan
jkeny said:
I always get a relative in the US to do digikey/allied/mouser/allied purchases for me (usually lots of things in one go) and then forward them in the regular post. Saves a fortune on shipping/customs and I've never lost anything. You must have a good friend/relative stateside -- we (the irish) are everywhere...
I think the cap across the LED mod will always improve the sound regardless of how good the regulators are.
This is based on the theory that the LEDs themselves are inherently noisy. Even if the regulators a totally silent, if the load ( the LED ) is noisy, you will get noise. ( unless you redesign the cct and have a pair of variable regs directly across the led's ) With the cct as it is, the resistance of the pots will always give the LED's the freedom to express there inherent noisiness
However, I still think that it will be best to use the quietest regulators available because the there are plenty of other noise sources - mains, diodes etc - that can mess up the sound and with these noise sources reduced to a minimum the LED caps ( mostly ) only have to cope with the LED and pot noise.
mike
This is based on the theory that the LEDs themselves are inherently noisy. Even if the regulators a totally silent, if the load ( the LED ) is noisy, you will get noise. ( unless you redesign the cct and have a pair of variable regs directly across the led's ) With the cct as it is, the resistance of the pots will always give the LED's the freedom to express there inherent noisiness
However, I still think that it will be best to use the quietest regulators available because the there are plenty of other noise sources - mains, diodes etc - that can mess up the sound and with these noise sources reduced to a minimum the LED caps ( mostly ) only have to cope with the LED and pot noise.
mike
Series Chokes and Cpas across LED's
I went all out on both of mine. Dual regs to start with. Then series resistors inline on the + line and ground feeding the first reg.
Next was a series 1 mH choke on the +5 and ground going to the pots. Finally slapped some Silmic II caps across the LED's.
Got to think all this helped the sonics some. But it was very small in my system. The Lightspeed was very good with just the regs and no chokes or caps. Guess I could remove them to see, but am too happy to fiddle with it.
Overall I think it is audionervoso, the effect is small if any. But maybe my chokes are too small, or the caps too high ESR to address the noise. Or it could be my 51 year old ears are shot. Or the rest of my system is so crap that it is not evident.
What did make a big difference was playing around with the value of the pots. The pot value is what really sets the impedance of the LDR circuit. Initially it was set at about 22K. Then subbed some lower valued pots to get it down to 9 - 11K.
The lower impedance is a big improvement. But I think the Lightspeed commercail units are about 10K to start with. If you have a diy job, check the impedance to see what it is as the pot is rotated.
I think there is posibility that George might add bypass caps to the commercial Lightspeed attenuators. If the buyers realy feel they need them, it is simple to add while building up the boards. The addded cost should not be more than 10% or so at most.
Either way, I removed the PEC carbon pots from both of my linestages after listening to the LDR circuit. It is that good, bypassed with caps or not.
And I do not think it is the contacts. Most systems have lots of contacts. Maybe it is the disimiliar materials between the wiper and resistive track.
George
I went all out on both of mine. Dual regs to start with. Then series resistors inline on the + line and ground feeding the first reg.
Next was a series 1 mH choke on the +5 and ground going to the pots. Finally slapped some Silmic II caps across the LED's.
Got to think all this helped the sonics some. But it was very small in my system. The Lightspeed was very good with just the regs and no chokes or caps. Guess I could remove them to see, but am too happy to fiddle with it.
Overall I think it is audionervoso, the effect is small if any. But maybe my chokes are too small, or the caps too high ESR to address the noise. Or it could be my 51 year old ears are shot. Or the rest of my system is so crap that it is not evident.
What did make a big difference was playing around with the value of the pots. The pot value is what really sets the impedance of the LDR circuit. Initially it was set at about 22K. Then subbed some lower valued pots to get it down to 9 - 11K.
The lower impedance is a big improvement. But I think the Lightspeed commercail units are about 10K to start with. If you have a diy job, check the impedance to see what it is as the pot is rotated.
I think there is posibility that George might add bypass caps to the commercial Lightspeed attenuators. If the buyers realy feel they need them, it is simple to add while building up the boards. The addded cost should not be more than 10% or so at most.
Either way, I removed the PEC carbon pots from both of my linestages after listening to the LDR circuit. It is that good, bypassed with caps or not.
And I do not think it is the contacts. Most systems have lots of contacts. Maybe it is the disimiliar materials between the wiper and resistive track.
George
Re: Pot value
A 100k pot would have an output impedance of 50k at half level, that into say a standard power amps input impedance of 50k is a 1 to 1 ratio, a far cry of my minimum ratio of 1 to 10 or better, for your pot to have the minimum ratio of 1 to 10, the poweramp would have to be 500k input impedance, virtually impossible to reach except on a tube amps.
My Lightpeed Attenuator led control pot is 100k log using the
NSL-32sr2s, this has the best feel, smooth gradual increases from 7am to around 2pm then progressivly stronger from then on 5pm
Cheers George
Are you using the 100k dact as a stand alone passive volume control? if you are, then YesDr.H said:
A 100k pot would have an output impedance of 50k at half level, that into say a standard power amps input impedance of 50k is a 1 to 1 ratio, a far cry of my minimum ratio of 1 to 10 or better, for your pot to have the minimum ratio of 1 to 10, the poweramp would have to be 500k input impedance, virtually impossible to reach except on a tube amps.
My Lightpeed Attenuator led control pot is 100k log using the
NSL-32sr2s, this has the best feel, smooth gradual increases from 7am to around 2pm then progressivly stronger from then on 5pm
Cheers George
Impedance across LDR
Terry,
Measure from input to ground. This will measure the series and shunt leg in series. In the passive unit, 10K sounded faster, punchier, more dynamic.
Somehow I ordered the wrong parts from Allied. Mine are R3, not R2. This means I get different impedance than most. A 100K pot ended up as 22 - 24k impedance as the pot is rotated. 50K pot measues 9 - 11K for the Lightspeed. But most have said the R2 give 10K with a 100K.
Since you are using the LDR's as a volume control in front of an active preamp this may not be as critical. With a passive line stage this is much more critical. Driving the cables and input of the amp is demanding. With your 27K input amp, the lower value would be more important.
The amps I use have 100K input impedance. And the cables from amp to Lightspeed clone are very short. The passive works great with them. In fact better than my line with an AD815 and LDR volume controls.
George
Terry,
Measure from input to ground. This will measure the series and shunt leg in series. In the passive unit, 10K sounded faster, punchier, more dynamic.
Somehow I ordered the wrong parts from Allied. Mine are R3, not R2. This means I get different impedance than most. A 100K pot ended up as 22 - 24k impedance as the pot is rotated. 50K pot measues 9 - 11K for the Lightspeed. But most have said the R2 give 10K with a 100K.
Since you are using the LDR's as a volume control in front of an active preamp this may not be as critical. With a passive line stage this is much more critical. Driving the cables and input of the amp is demanding. With your 27K input amp, the lower value would be more important.
The amps I use have 100K input impedance. And the cables from amp to Lightspeed clone are very short. The passive works great with them. In fact better than my line with an AD815 and LDR volume controls.
George
Depends
Ryan,
It depends on your sources. If your sources have strong putput stages that either are dc coupled or have large value coupling caps the lower value might be better in your case. If you have vacuum tube output stages, the 22K might be better.
Measure what it actually is first. This is fairly easy, even with an active stage after the LDR's.
Select an input, then measure from that input to ground. This willl include the series LDR and the shunt LDR together. This is paralleled with the input of the active. I think your active is a 1K series and 100K to ground. So whatever the input of the LDR's is, it is paralleled with 101K. This will lower the measured impedance of the LDR only slightly. I think you will measure about 10K.
Mine was 22K due to accidently ordering the wrong LDR's. The R2 LDR's are lower resistance than the R3 I have.Other were measuring 9K or so, and I was getting 20 -24K.
I seem to remember georgehifi stating that the commercial Lightspeed attenuations are a little lower than 10K.
If you want to lower the impedance of the attenuator, try a 50K pot. This lower lower it to less than half of what is is with a 100K.
Good luck, if you try a different value pot, let us know how it works out.
George
Ryan,
It depends on your sources. If your sources have strong putput stages that either are dc coupled or have large value coupling caps the lower value might be better in your case. If you have vacuum tube output stages, the 22K might be better.
Measure what it actually is first. This is fairly easy, even with an active stage after the LDR's.
Select an input, then measure from that input to ground. This willl include the series LDR and the shunt LDR together. This is paralleled with the input of the active. I think your active is a 1K series and 100K to ground. So whatever the input of the LDR's is, it is paralleled with 101K. This will lower the measured impedance of the LDR only slightly. I think you will measure about 10K.
Mine was 22K due to accidently ordering the wrong LDR's. The R2 LDR's are lower resistance than the R3 I have.Other were measuring 9K or so, and I was getting 20 -24K.
I seem to remember georgehifi stating that the commercial Lightspeed attenuations are a little lower than 10K.
If you want to lower the impedance of the attenuator, try a 50K pot. This lower lower it to less than half of what is is with a 100K.
Good luck, if you try a different value pot, let us know how it works out.
George
high source impedance
Probably been asked before, but what if your source has an output impedance of, say, 2.3 kOhm?
I'm talking about the 47 Labs Shigaraki CDP. I have one. And I want to go passive after hearing the result with an DIY pot/resistor based model from a friend.
Then someone had to spoil it by saying my source output impedance is too high for the Lightspeed...
Any workarounds?
The Shigaraki doesn't seem to mind the mismatch in the current passive set up. Even dynamics are almost good. And tonally it's heaven.
Sorry to just pop in, but this seems to be the thread for Lightspeed questions.
Tunes10
Probably been asked before, but what if your source has an output impedance of, say, 2.3 kOhm?
I'm talking about the 47 Labs Shigaraki CDP. I have one. And I want to go passive after hearing the result with an DIY pot/resistor based model from a friend.
Then someone had to spoil it by saying my source output impedance is too high for the Lightspeed...
Any workarounds?
The Shigaraki doesn't seem to mind the mismatch in the current passive set up. Even dynamics are almost good. And tonally it's heaven.
Sorry to just pop in, but this seems to be the thread for Lightspeed questions.
Tunes10
Thanks, George.
My friend is checking with the DIY builder, I'll get back on it.
This impedance matching is starting to confuse me.
Everyone agrees that mistmatches cause flawed dynamics with traditional resistor based passives.
TVC's are supposed to solve this because source and power amp don't see eachothers impedances. Still, TVC-makers advise low source imputs, <1 kOhm. Than I asked one maker about my high source imp and he answered: no problem, I have clients with 2.2kOhm source impedances and that works fine.
Either the theory is relative or the sales opportunity absolute.
I don't know. But I think the first assumption might be true. I have tried direct injects with players with built in digital and analoque attenuators. These players had lower output impedances than the Shigaraki. But they had all the dynamic problemes that the Shigaraki doesn't. Go figure.
My friend is checking with the DIY builder, I'll get back on it.
This impedance matching is starting to confuse me.
Everyone agrees that mistmatches cause flawed dynamics with traditional resistor based passives.
TVC's are supposed to solve this because source and power amp don't see eachothers impedances. Still, TVC-makers advise low source imputs, <1 kOhm. Than I asked one maker about my high source imp and he answered: no problem, I have clients with 2.2kOhm source impedances and that works fine.
Either the theory is relative or the sales opportunity absolute.
I don't know. But I think the first assumption might be true. I have tried direct injects with players with built in digital and analoque attenuators. These players had lower output impedances than the Shigaraki. But they had all the dynamic problemes that the Shigaraki doesn't. Go figure.
Hi Tunes,
it seems complicated due to a number of different effects all working together to try to reduce the performance of your system.
There is usually a DC blocking capacitor at the output of a source and again at the input of the next stage (amplifier).
This DC blocking cap will act as a high pass filter when feeding into a resistive load (the amplifier).
Two caps in series will reduce the bandwidth more than a single DC blocking capacitor.
Removing the DC block can affect the way the amplifier sets it's output offset. Don't interfere with this unless you know how to measure and correct the output offset.
All cables and inputs have some capacitance.
Any source feeding this capacitance will act as a low pass filter.
Changing the resistances and capacitances will change the effective bandwidth of the system. Some amplifiers rely on accurate bandwidth setting of this low pass filter to ensure stability into any complex load. Again interfering with the low pass filter is only for the experienced DIYer.
To help alleviate these filter problems, a number of rules have developed over the decades.
One of these is that Rs << Rload, at least a 1:5 ratio and preferably 1:10 Rs:Rload is recommended.
If a passive pre (volume control) is in place, then the 1:5 ratio should apply on the source side and again should also apply on the receive side.
Lets look at some real numbers by way of explanation.
Rs=100r, DC block=1uF.
Rload=Zin=50k, Dc block=4.7uF, Cin 1nF, Rin=680r.
Cable capacitance 70pF/m and 3m long (=210pF).
Low pass filters: 1uF & 50K + 1uF & 50k.
combine the two series caps to give 0.83uF.
F-3db=3.8Hz, F-1db~7.6Hz. F=1/2/Pi/R/C (in Farads and ohms).
High pass filters:100r & 1nF + 100r & 210pF + 680r & 1nF,
combine the capacitance to give an effective value of 1.21nF.
F-3db=7.6MHz (100r & 210pF)
F-3db=169kHz (100r+680r & 210p+1nF). F-1db~85kHz.
These two are far enough apart that they do not combine 9in the audio sense) to any significant effect.
The bandwidth for the interconnection is thus 3.8Hz to 169kHz. This will work, but some builders will look for slight alterations both up and down to suit their preference.
Put in some values for high Rs or low Zin or low DCblock and see how narrow the bandwidth can become.
Insert a passive pre into the system and do the numbers before the pre and again after the pre and you will begin to see how complicated it becomes. For even more complication adjust the volume and examine the effect on the low pass filter (yes, frequency moves around with volume).
No wonder there is conflicting advice, particularly from salesmen who care more about their bonus than finding out the audible effect of changes in the interconnects. Fortunately some are very diligent and honest, find them and use them.
it seems complicated due to a number of different effects all working together to try to reduce the performance of your system.
There is usually a DC blocking capacitor at the output of a source and again at the input of the next stage (amplifier).
This DC blocking cap will act as a high pass filter when feeding into a resistive load (the amplifier).
Two caps in series will reduce the bandwidth more than a single DC blocking capacitor.
Removing the DC block can affect the way the amplifier sets it's output offset. Don't interfere with this unless you know how to measure and correct the output offset.
All cables and inputs have some capacitance.
Any source feeding this capacitance will act as a low pass filter.
Changing the resistances and capacitances will change the effective bandwidth of the system. Some amplifiers rely on accurate bandwidth setting of this low pass filter to ensure stability into any complex load. Again interfering with the low pass filter is only for the experienced DIYer.
To help alleviate these filter problems, a number of rules have developed over the decades.
One of these is that Rs << Rload, at least a 1:5 ratio and preferably 1:10 Rs:Rload is recommended.
If a passive pre (volume control) is in place, then the 1:5 ratio should apply on the source side and again should also apply on the receive side.
Lets look at some real numbers by way of explanation.
Rs=100r, DC block=1uF.
Rload=Zin=50k, Dc block=4.7uF, Cin 1nF, Rin=680r.
Cable capacitance 70pF/m and 3m long (=210pF).
Low pass filters: 1uF & 50K + 1uF & 50k.
combine the two series caps to give 0.83uF.
F-3db=3.8Hz, F-1db~7.6Hz. F=1/2/Pi/R/C (in Farads and ohms).
High pass filters:100r & 1nF + 100r & 210pF + 680r & 1nF,
combine the capacitance to give an effective value of 1.21nF.
F-3db=7.6MHz (100r & 210pF)
F-3db=169kHz (100r+680r & 210p+1nF). F-1db~85kHz.
These two are far enough apart that they do not combine 9in the audio sense) to any significant effect.
The bandwidth for the interconnection is thus 3.8Hz to 169kHz. This will work, but some builders will look for slight alterations both up and down to suit their preference.
Put in some values for high Rs or low Zin or low DCblock and see how narrow the bandwidth can become.
Insert a passive pre into the system and do the numbers before the pre and again after the pre and you will begin to see how complicated it becomes. For even more complication adjust the volume and examine the effect on the low pass filter (yes, frequency moves around with volume).
No wonder there is conflicting advice, particularly from salesmen who care more about their bonus than finding out the audible effect of changes in the interconnects. Fortunately some are very diligent and honest, find them and use them.
tunes10 said:
As you have already discovered it's all rather system specific. The impedance issue is determined by the power amp and so far i haven't noticed you mentioning yours. IME most solid state power amps require an active pre for happiness while tube amps do not.
I've used a high inductance (S&B) TVC for several years and you may drive from a 2.3k source only if you particularly dislike listening to the frequency extremes.
Even if the impedance matching looks great on paper a passive pre may still sound transparent but anaemic.
analog_sa said:
Yes it is the most transparent of all, but your way off with the anaemic statment. This will only happen if you have loaded down the source or the output of the passive too much.
If not loaded down by making sure your source is low <100ohms and your amp input load (impedance) is high >60kohms, there is no active pre that can match the true dynamics of a passive.
Note: I said ( true dynamics ), not the squeezed out shoot you in the face dynamics that active pre's give, that is very impressive for 5 mins then wears you out very quickly, but the big, rich, tight full bodied dynamics that you can listen to cd after cd I hear with a correctly impemented Lightspeed passive.
Cheers George
georgehifi said:
Hi George
If you read what i wrote you would notice a "may" qualifier. Excellent dynamics with a passive stage are certainly possible but by far not guaranteed. Good bass is similarly tricky
And impedance matching, although essential is not necessarily sufficient. This is a general observation as i have no experience with Lightspeed.<100ohms source and >60k input impedance will certainly make things much easier, especially if the passive controller is integrated in the power amp. Sadly, a lot of commercial units do not fit this criteria.