LDR sources

Mike McLane

Active member
I was on the waiting list for GL5516's forever. Now I see they're not in the offering of components on PPCB. I wanted to build another General Tso, but have been frustrated by the difficulty in sourcing them (or equivalent) elsewhere as I'm not a REAL electronics guy able to fully evaluate the specs. There are Ebay or Amazon sources that leave me leery as to quality and you gotta buy in "bulk". Any suggestions?
 
  • Like
Reactions: fig
Ebay has GL5516 for cheap. Sure you have to buy 50 or 100 but they're like $5 for the bulk package. Buying bulk is an advantage as there is so much variance between units, it's good to get several so they can be matched up.
Check out this link for some suggestions on how to test:

 
Thanks for the feedback. Matched up, huh? How so??? I understand about matching up JFET's and vacuum tubes and proper biasing, etc. But LDR's? Are you saying they should be tested to exhibit the same resistance given the same light intensity? Will minor variance mess up the performance of something like the General Tso?
 
Just curious. . . did you order a "generic" green 5mm LED (I think around 20 mcf +/-) or did you get something with a higher mcf? I put the generics in mine and could have sworn that the compression effect was unacceptably low. I acquired LED's with 40 mcf, but managed to trash the PCB while trying to replace them. However, Chuck Bones cautioned me that the Tso is a subtle beast and not the boa constrictor most of us have come to know vis-a-vis the Ross/Dynacomp clones. That's to say I probably should have spent some quality time with it before setting it in the corner and waiting on the replacement LED's. Curious as to your experience.
 
Just curious. . . did you order a "generic" green 5mm LED (I think around 20 mcf +/-) or did you get something with a higher mcf? I put the generics in mine and could have sworn that the compression effect was unacceptably low. I acquired LED's with 40 mcf, but managed to trash the PCB while trying to replace them. However, Chuck Bones cautioned me that the Tso is a subtle beast and not the boa constrictor most of us have come to know vis-a-vis the Ross/Dynacomp clones. That's to say I probably should have spent some quality time with it before setting it in the corner and waiting on the replacement LED's. Curious as to your experience.
In the ones I built I used all green run of the mill LEDs. I love my Tso, it’s stayed on my board this entire time. I sold one to a friend who has a Cali 76 and he said the Tso was as good the Cali, which is a pretty high remark, and made me fee pretty confident in the build.
 
Thanks for the feedback. Matched up, huh? How so??? I understand about matching up JFET's and vacuum tubes and proper biasing, etc. But LDR's? Are you saying they should be tested to exhibit the same resistance given the same light intensity? Will minor variance mess up the performance of something like the General Tso?
Most builds called for a specific dark or off resistance (Roff), which you can measure with a multimeter. You need to block any stray light while testing. Try putting the LDR in a closed drawer or drilling 2 holes in a film canister so just the leads poke out and then measure the resistance.
 
Don't overthink the LED. Whatever LED you use will make more than enough light. A high brightness LED will overdrive the LDR, causing slow response.

As for LDRs, beside min ON resistance and max OFF resistance, the other importance performance parameter is response time. You need a scope and sig gen to measure that, but believe me there is a huge variation in response time for various LDRs.
 
As for LDRs, beside min ON resistance and max OFF resistance, the other importance performance parameter is response time. You need a scope and sig gen to measure that, but believe me there is a huge variation in response time for various LDRs.
IMO, this is the biggest difference between real vactrols & homebrew LED/LDR versions. For effects like optical compressors and envelope filters, the response (rise & fall) time is a critical parameter - with rise time often in the sub-10millisecond range. Unlike LDR resistance, which can often be modified by changing LED drive current, the response time can't be easily adjusted. There are cases where response time doesn't matter that much, like when the control signal is quite slow or when the LDR is part of an LFO.

This may be a result of different LED types (Red/Green/Yellow/UV/IR/etc.) as well as a different LDR material. You’ll often see CdS LDRs given a material “Type” from 0-7 on the data sheet. Each of these will bring their own characteristic response time, min/max resistance, spectral response, and even different “light memory” effects.

All that’s pretty damn cool, but is also part of the reason why some builds essentially require an actual vactrol and may not work well with a DIY one. Here are some things I've learned that may be helpful:
  1. Builds that require VTL5C2, VTL5C3 or NSL-32 vactrols can often be replaced with a higher off resistance LDR + a red or yellow LED - Use a 3mm LEDs for small form factor. Place LED & LDR face-to-face and wrap with 2-3 layers of black heatshrink. Use pliers when heatshrink is hot to clamp ends closed for better light isolation. Don't use superbright LEDs, they often have higher Vf, a smaller viewing angle, and a steeper/high-current I/V curve (we're not driving anywhere near 50mA into an LED): High-Efficiency Red 3mm LED EDIT CROP.png
  2. If the build really requires a really fast response vactrol (like VTL5C1 or VTL5C10), then you’re shit out of luck, and only the real thing will do. Their characteristic speed & slope just isn’t feasible with a homebrew vactrol.

If you want a good PDF reference for all things vactrols & LDRs, this Perkin-Elmer VLT5Cx-series datasheet is a great resource on the basics, and has a great table for easily comparing the different vactrol specs:

Vactrol Comparison Chart.png
 
Good stuff, B&E!

The above table is useful, but it's qualitative. It doesn't define where the dividing line is between Very Fast, Fast, Average, etc., or the test conditions. The slowest portion of the response curve is at the high resistance end and occurs when going dark. The way we use LDRs in pedals, we don't much care about resistance over about 1MΩ; in some pedals, even less. Examples:
  • Fat General LDRs are in parallel with a 100K pot. Once the LDRs are each over 200K, their effect on gain diminishes significantly. If they take a long time to get to 1MΩ or 10MΩ, we don't care.
  • Warden LDR is in parallel with 2.2M. Once the LDR gets to 2.2MΩ, any further increase in resistance affects the gain by 6dB or less.
  • Mutron III LDRs are in parallel with 220K resistors.
  • Bi-Phase LDRs are in parallel with 220K resistors.
  • Univibe derivatives bias the LED or bulb so that is never goes completely dark.
In optical compressors, we need the LDR resistance to drop quickly because that is when the comp is responding to the note attack. It's ok if the resistance increase is slower because the guitar's decay is much slower than the attack. There is such a thing as too slow and that is the "breathing" we hear when the compressor can't increase the gain quickly enough and then finally catches up. If we want very fast and accurate automatic gain control, then we don't use LDRs. We use OTAs or digital compressors. All have their advantages & disadvantages.

Phasers, Choruses & Vibes run the modulation slowly, so a little LDR lag is unnoticeable. The light bulb itself is slow and its response curve is part of the characteristic sound of the sweep.
 
The above table is useful, but it's qualitative. It doesn't define where the dividing line is between Very Fast, Fast, Average, etc., or the test conditions. The slowest portion of the response curve is at the high resistance end and occurs when going dark. The way we use LDRs in pedals, we don't much care about resistance over about 1MΩ; in some pedals, even less.
Ahh, Chuck, always bringing "reality" into the equation!

Too true on the relative response times, I think they're more interested in displaying the relationship between material type (0-7) and response time anyways. The more helpful (but still not THAT helpful) is a plot from the VTL5C3 datasheet, which illustrates your point on the disparity between high & low resistance, as well as turn-on vs. turn-off behavior:
VTL5C3 response plot datasheet.png

I love the sound of builds with vactrols, but the in-circuit behaviors are SO hard to intuit because of all the compounding non-linearities. It doesn't help that the datasheets tend to only have data for much higher drive currents than what we usually work with.
The way we use LDRs in pedals, we don't much care about resistance over about 1MΩ; in some pedals, even less.
This isn't something I had really thought about, but it does explain the indifference of many builds to the specific LDR used.
 
Back
Top