Single pot
Looking at it, looks like it should work fine. Guess I spent all those hours digging through surplus bins for dual linear 100K pots for nothing. Single pots are much easier to source.
Everyone but me seems to have used log pots. I find the taper works fine with linear. The reason is because the shunt and series move in opposite directions, so the result is a second order function. If a fixed shunt and variable series element was used the taper would require a log pot.
Just picked up new pots today, passed up a BUNCH of high quality cheap single pots. Guess it is back to surplus soon.
George
Looking at it, looks like it should work fine. Guess I spent all those hours digging through surplus bins for dual linear 100K pots for nothing. Single pots are much easier to source.
Everyone but me seems to have used log pots. I find the taper works fine with linear. The reason is because the shunt and series move in opposite directions, so the result is a second order function. If a fixed shunt and variable series element was used the taper would require a log pot.
Just picked up new pots today, passed up a BUNCH of high quality cheap single pots. Guess it is back to surplus soon.
George
This looks good, though I think you would have to use a 200k log pot as the entry of the 5vdc is now at the centre of the rseistance track.
Keep it up guys this is good thinking.
As for using a linear in my setup it comes on too quick, a log is just right, but with this single setup who knows you'll just have to suck it and see.
BTW. I think the sound will not change at all, but the feel of the volume control's advance will, and also it would make it a dollar or two cheaper to build.
Cheers George
Keep it up guys this is good thinking.
As for using a linear in my setup it comes on too quick, a log is just right, but with this single setup who knows you'll just have to suck it and see.
BTW. I think the sound will not change at all, but the feel of the volume control's advance will, and also it would make it a dollar or two cheaper to build.
Cheers George
Blues said:
The only thing is you have your 5vdc coming in on the wiper and one end of the resistance track on each section, where I only had it coming in on one end of the resistance track of each section and not the wiper as well. and you'll only need the 1k trimmer on the channel with the highest gain.
Cheers George (Nice drawing BTW).
Matching for balanced version
I'm going through 30 LDRs to find 2 sets of 4 matching ones for balanced. This is not working out very well-- I'll possibly need a total of 50 or more to find two sets of 4 that track +/- 5% within each set. If after matching within their personal inventory, would anybody be interested publishing a spreadsheet of LDR measurements at agreed-upon test values for control V & R, and possibly swap some LDRs around thru the mails? In any case this would help recover costs of the many rejects involved.
BTW if anybody can figure out how Melos dynamically tracked and converged LDRs in the Porzilli Pho-teniometer, this would solve the problem of matching for balanced, while adding the considerable refinement of precision resistors to the passive.
Dave
I'm going through 30 LDRs to find 2 sets of 4 matching ones for balanced. This is not working out very well-- I'll possibly need a total of 50 or more to find two sets of 4 that track +/- 5% within each set. If after matching within their personal inventory, would anybody be interested publishing a spreadsheet of LDR measurements at agreed-upon test values for control V & R, and possibly swap some LDRs around thru the mails? In any case this would help recover costs of the many rejects involved.
BTW if anybody can figure out how Melos dynamically tracked and converged LDRs in the Porzilli Pho-teniometer, this would solve the problem of matching for balanced, while adding the considerable refinement of precision resistors to the passive.
Dave
Graded LDR's
Dave,
I must have been lucky. Out of 30, I got three quads to match fairly close. This was only using three currents levels.
All mine were sorted, R3F. What was observed was a shallow exponetial decay when plotted out on linear grid. The slopes and offsets varied a bit.
The "sorted" varied this much, I would hate to match up the standards. But last week or so I looked at Allied and they had the sorted back in stock.
Somewhere in this thread the details of the Melos are listed. Seems it was a fixed series, variable shunt. Just like a shunt attenuator. The Lightspeed is much better, the input impedance on mine only vary about 20 -25% with full pot rotation. It is lowest in the middle ranges.
PM me, I might have something that you can use.
George
Dave,
I must have been lucky. Out of 30, I got three quads to match fairly close. This was only using three currents levels.
All mine were sorted, R3F. What was observed was a shallow exponetial decay when plotted out on linear grid. The slopes and offsets varied a bit.
The "sorted" varied this much, I would hate to match up the standards. But last week or so I looked at Allied and they had the sorted back in stock.
Somewhere in this thread the details of the Melos are listed. Seems it was a fixed series, variable shunt. Just like a shunt attenuator. The Lightspeed is much better, the input impedance on mine only vary about 20 -25% with full pot rotation. It is lowest in the middle ranges.
PM me, I might have something that you can use.
George
Sorting LDRs for Balanced
George,
Thanks. I bought my NSL-32SR2S (sorted) from Allied. On one side they are marked "466"; on the other side "R2D". I'm new to all this, but I can see that my desire to go balanced as well as high-impedance (as needed for a 100K volume control for my tube pre) has added complexity. I'm finding that not only are LDRs nonlinear, but across a wider range into the dark side of the LED (as needed to obtain resistance out to 100K) there are greater deviations in nonlinearity from piece to piece than are observed at lower R for the standalone passive. For example two LDRs that match closely out to 25K, may deviate in unpredicable ways 10% or more from 25K to 100K. I'm losing courage on whether this will make unacceptable imbalances between phases in a high-impedance balanced volume control.
Dave
George,
Thanks. I bought my NSL-32SR2S (sorted) from Allied. On one side they are marked "466"; on the other side "R2D". I'm new to all this, but I can see that my desire to go balanced as well as high-impedance (as needed for a 100K volume control for my tube pre) has added complexity. I'm finding that not only are LDRs nonlinear, but across a wider range into the dark side of the LED (as needed to obtain resistance out to 100K) there are greater deviations in nonlinearity from piece to piece than are observed at lower R for the standalone passive. For example two LDRs that match closely out to 25K, may deviate in unpredicable ways 10% or more from 25K to 100K. I'm losing courage on whether this will make unacceptable imbalances between phases in a high-impedance balanced volume control.
Dave
Blues said:I was hoping Bear would come back, start a thread and share his "secret circuit tricks" for his Super Phono Stage.
There's nothing spectacularly "secret" actually, just very clever implementations that are optimized both on the measurement/design side and the subjective/listening side so that the two converge... but that would be another thread.
Thanks for asking...
I downloaded the patented LDR circuit... haven't had the opportunity to read it yet. Looks on the surface like he solved the non-linearity of the LDRs with a third/fourth LDR going for feedback via a bridge... perhaps worthy of consideration here, although certainly increasing the component count and complexity.
If anyone does a computer soundcard based FFT of the LDRs, that would be cool to see here.
Otherwise, enjoy the experimentation!
_-_-bear
Re: Graded LDR's
BTW if anybody can figure out how Melos dynamically tracked and converged LDRs in the Porzilli Pho-teniometer, this would solve the problem of matching for balanced, while adding the considerable refinement of precision resistors to the passive.
Dave
From memory my last 50 NSL-32SR2S were all batch R2A and I had a fairly easy time matching these up, 10 sets of four out of the 50 for ten Lightspeeds.
I don't think that Melos or anyone ever published how they dynamicly tracked there LDR's for channel balance, but it was'nt very successfull as I believe they all were problematic, but then they did only use one shunted ldr and a series resistor, also an inferior ldr from what I've seen in pics it looked like a NORP 12 or similar mounted in a brass tube with a lamp or led mounted in the other.
The simple way I match mine up is at 3 volume settings, 9, 12 and 3 o'clock on the scope then use the trimmer for fine tuning, then have a final listen on the system.
I can imagine what you guys are going through to do a balanced Lightspeed version, it will be hell doing it, but you will get there.
Cheers George
Panelhead said:
BTW if anybody can figure out how Melos dynamically tracked and converged LDRs in the Porzilli Pho-teniometer, this would solve the problem of matching for balanced, while adding the considerable refinement of precision resistors to the passive.
Dave
From memory my last 50 NSL-32SR2S were all batch R2A and I had a fairly easy time matching these up, 10 sets of four out of the 50 for ten Lightspeeds.
I don't think that Melos or anyone ever published how they dynamicly tracked there LDR's for channel balance, but it was'nt very successfull as I believe they all were problematic, but then they did only use one shunted ldr and a series resistor, also an inferior ldr from what I've seen in pics it looked like a NORP 12 or similar mounted in a brass tube with a lamp or led mounted in the other.
The simple way I match mine up is at 3 volume settings, 9, 12 and 3 o'clock on the scope then use the trimmer for fine tuning, then have a final listen on the system.
I can imagine what you guys are going through to do a balanced Lightspeed version, it will be hell doing it, but you will get there.
Cheers George
Matching for balanced version
For now I'll retract some of what I posted concerning the difficultly of closely matching LDRs. Remeasuring today some of the pieces I tested yesterday, I'm seeing 10% differences that must be caused by temperature fluctuations in the house during winter! I'm working in a room where ambient fluctuates 5-10 degrees F with the tube stereo warming up or the house heater coming on. The process has been prolonged through temperature changes as I've been using 10 fixed resistors at values between 500R and 1M.
For now I'll retract some of what I posted concerning the difficultly of closely matching LDRs. Remeasuring today some of the pieces I tested yesterday, I'm seeing 10% differences that must be caused by temperature fluctuations in the house during winter! I'm working in a room where ambient fluctuates 5-10 degrees F with the tube stereo warming up or the house heater coming on. The process has been prolonged through temperature changes as I've been using 10 fixed resistors at values between 500R and 1M.