Lightspeed Attenuator a new passive preamp

A 50K volume control with a fixed series resistor and a variable shunt resistor that attenuates -10db to -63db, requires LDR shunt resistance up to around 25K. This is inside the range of reasonably close tracking for matched LDRs.
what about a switchable series resistor?
100k for the low volume setting and 100k//22k for the higher volume setting.
If this used a two pole normally open relay then the attenuator could sit at the lower volume setting at power on and revert to the lower volume setting at power off. That way you need to action the higher volume setting.

Because you are using just a shunt LDR you can very easily add a balance trim without upsetting the LDR matching.
Using the series/shunt LDRs makes balance trim much more difficult.
 
Sepolansky, The Atma-Sphere MP-1 is an all-tube balanced integrated phono/line stage. It would be possible to squeeze a SS buffer in there. However I'm inclined toward toob minimalism. The output impedance of the stock 12AT7 phono section was something like 10K. I've brought this down to around 2K by subsituting ECC99 after a rework of CCS.

AndrewT, Thanks, interesting idea.
 
Or you could use LDRs as series resistors and use them just as you were going to use the other resistors. Set them for a specific resistance and leave them there. Use 2 in parallel with each other, series with signal. Switch THEM on and off as Andrew suggested but without a relay in the signal. You dont have to have matched LDRs this way as you can 'set em and forget em.'
 
Or you could use LDRs as series resistors and use them just as you were going to use the other resistors. Set them for a specific resistance and leave them there. Use 2 in parallel with each other, series with signal. Switch THEM on and off as Andrew suggested but without a relay in the signal. You dont have to have matched LDRs this way as you can 'set em and forget em.'
I doubt very much that you can maintain tolerance matching for Balanced Impedance connection.
0.1% match is suggested as the minimum standard for maintaining the rejection capability of the balanced link.
0.01% will give better rejection.
 
Sepolansky, The Atma-Sphere MP-1 is an all-tube balanced integrated phono/line stage. It would be possible to squeeze a SS buffer in there. However I'm inclined toward toob minimalism. The output impedance of the stock 12AT7 phono section was something like 10K. I've brought this down to around 2K by subsituting ECC99 after a rework of CCS.

David,

Outboard of the phono stage, add a cathode follower, 6SN7 or 6CG7/6FQ7. (of course a pair for balanced operation, one tube per channel).

Couple the input with an 0.01uF V-Cap/10Meg grid leak.

Remove the much larger output coupling caps from your existing preamp (or bypass, add a second set of jacks, etc.).

Under the cathode follower goes a servo, sampling the output DC voltage and comparing it to 1/2 the negative power supply voltage. This allows DC coupling of the output. The servo op-amp controls a MOSFET in the cathode circuit of the CF. The (-) return is to a -12 to -24V supply.

Stuart Yaniger and Gordon Rankin have variations on this theme. SY's applies the corrective signal to the grid.

I've heard and used both Yaniger's and this version (modified Broskie circuit) and they both are superb and transparent.

The signal path is therefore only modified by replacement of a cap with one probably 1/100 (or even smaller) the size, the addition of one triode section of a nicely linear tube and a small value (47R-100R) Mills NI wirewound resistor.

Economizing, even an IRC resistor and good quality PP cap can sound very nice.

Ideally they are used as an integral part of a source, but they can be used as stand alone. The modified Broskie version I use can be used without an input coupling cap if you choose not to modify your existing preamp. However, the elimination of a big coupling cap at the output of the preamp is a nice sonic improvement.

Stuart
 
Andrew/David
I think you can do it if you set the LDR at, for instance, 50k and leave it there. I dont mean that you would ever change the value after you set it. I am sure .01 is not possible. Certainly I think its possible to get 1%. .1% probably not. They will drift but if you do preliminary matching at about 50k then you should be able to get them to drift alike. So if one goes to 52k I would expect the other to also go to 52k. Sticking them near tubes or anything with a large voltage drop wouldnt help things, I admit, but I think its worth trying. Its as simple as a bit of regulation and a few caps followed by another trimmer to get you down to 50k. Whole additional circuit would cost less than the resistors considered for use right now.
I wont keep chasing after you about it, I just think that if you need a resistor for series that is not going to change value... that its possible down at 50k and perhaps a bit higher. If you want to vary them then I dont think it is. I suppose if .1% is the max then this is not going to work well. 1% I think its worth trying.
Stuart, the piece you sent me a few months back is working great as I type. Would like to try those other options you mentioned if you could pm me some part numbers.
 
Hi Andrew
REF7. I had to know if it could be better than REF5. Its got better depth. Things sound a little more real. However there was already great depth and realism. Gushing about it is not what I am about to do yet. The REF5 was incredible and this seems on par and a tiny bit better but I have done no side by sides yet.
Its massive. At least 2.5 times the size of the REF5. Ridiculously large but packed with power supplies.
Would I buy the REF7 again over the REF5. Probably not. The REF5 is as good in most cases if my ear's memory serves me correctly. I will have to do side by side but maybe thats over the weekend.
Uriah
 
I finally managed to make one lightspeed attenuator. After testing it seems ok.
I have now - in testing configuration bypassed the TKD pot in my somewhat modified Audio Note preamp.

It sounds very good...

Vel it looks like a mess but sounds good... :)

I am using a 50k pot for volume but it is exstremely sensitive. I think I have to use a 100k pot or maybe 250k? Is that smart?

attachment.php
 

Attachments

  • IMG_0080.JPG
    IMG_0080.JPG
    64.5 KB · Views: 530
Last edited:
Sometimes, when they are so sensitive you may consider finding a way to keep the shunt LDRs at a lower max resistance or a way to keep series at a higher min resistance. This way you do not get incredibly loud volume on accident with touchy LDRs. It may be your pot and it may be the taper of your pot as well. Since all LDRs have a different resistance taper to them you may have a pair that increase in resistance faster than most in shunt or the opposite in series.
Anyway.
Here is the things to try
100k pot dual gang log. This will help for sure.
10k-30k resistors in place of the 100R resistors that are feeding your Series LDRs. You will NOT get max volume this way and your resistance curve will change a lot, but it will help with your immediate problem.
5k-20k resistors in parallel with the resistor side of your Shunt LDRs. This will keep shunt from increasing in value to fast and will also have negative side effects as mentioned above with series fix.

Probably your only problem is that you are not using 100k dual log pot. However if you have high gain amp and you have super efficient speakers sometimes the normal LS setup can be tuned to work better with that sort of system in these ways. There are other options like different voltage to drive them or series 10R pot before 100k pot.
Try 100k first and let us know.
Uriah
 
All LDRs that are operating in any circuit are operating within the current tolerance of the LDR. Any that arent.. dont for long.
Dead silence is achieved by pumping up the resistance of the series LDRs to extremes which emulates very high resistance pots. The higher the resistance the higher the Johnson noise and the higher the resistance the more roll off you get of high frequencies.
Its a solution. Just as imperfect as others.
The best solution for low volume with LDRs is to parallel shunts to achieve low to zero volume with lower series resistances that allow a better sound. I suppose that sound is better in my opinion.
Uriah