To infinity and beyond!
Although this isn’t specifically related to the S.E.X. amp i thought it worth sharing all the same. For those unfamiliar with LDR attenuators they use optical couplers (specifically light dependent optocouplers) as voltage controlled variable resistors to mimic the function of a potentiometer. In theory they are more transparent and have a lower noise floor than either a carbon based pot, or resistor based stepped attenuator. Curious yet? ...me too
The specific type of optocouplers used are Silonex NSL-32SR2 which have an unfortunate property in that they vary hugely in their response from sample to sample. They are so horribly matched you have to buy at least two dozen of them and painstakingly measure each one with a fixed current to find a couple of matched pairs. In an ideal world you want four matched, but you can make do with two sets of matched pairs. Most resistors have a tolerance of 1% to 5%, these have a tolerance more like 50%, but once you find a set that measure the same your set. They don’t have to be exactly the same as the final balance between them can be balanced out with trim pots, but the closer the match the better.
Parts required:
-x4 NSL-32SR2
-x4 100ohm resistors (current limiters to LDR’s)
-x2 1k trimpots (to balance left/right channels)
-x1 100k linear volume pot (adjusts voltage to LDR’s)
And that’s it. Only other catch is you need a quiet 5v DC supply for them, a simple 7805 or LM317 voltage regulator is perfectly adequate. The current draw of the entire circuit is very small, under 100ma so you can get away with tapping into a 6.3v ac winding and rectify into DC which will give you ~7-8v to feed into your voltage regulator for a clean 5V DC output.
Once you have recovered from spending hours slaving away at a breadboard with a multimeter matching sets you then have to assemble the circuit and do one final balance with the trim pots to get each channels response exactly the same. I almost forgot to mention the second catch, the resistance values vary wildly with temperature so they need to be running for about half an hour before they settle down. Just holding one of the LDR’s between your fingers will change its resistance enough for the volume change to be audible within a few seconds. Kind of amusing in some ways, but it does make their suitability for the inside a hot valve amp questionable.
The how does it sound question. In my test circuit i went from a regular 100k carbon pot so there was a noticeable improvement in clarity, better defined bottom end, and cleaner sounding treble. I then switched to an eBay ladder attenuator with Dale resistors and honestly, i couldn’t really tell any difference. Possible the LDR was a bit more detailed in the high end, but its vary hard to say when it takes 5mins with a soldering iron to switch between them so i’ll keep experimenting with it and see how it performs over time.
Overall conclusion. Huge variation between samples means you have to buy them by the dozen at a time. Hard to find, i could only find them stocked in the US so they cost me more in shipping than for the parts. Pain in the arse to match. Require a dedicated power supply to drive them. Huge drift with temperature. Pain in the arse to match.
On the plus side the entire board is compact, i fitted the entire circuit with AC to DC power supply on a 3†x 1†protoboard. The actual volume control action is nice as they react slowly to change so any volume adjustment has a fluid gradual change. A novelty when you’re used to stepped attenuators. Possibly sounds better than a ladder attenuator, jury is still out on that one. Unless i come to the conclusion there is a tangible improvement sonically over stepped attenuators i’m going to say that they are more trouble than they are worth. Nice idea, but a pain in the arse to implement.
