I have a little Transistor tester I bought from M3Electronics. Pretty cool thing. I measured voltage drop but didnt see a correlation to resistance. Will be interesting if you can find a correlation and we could match now with voltage drop rather than just a set voltage/resistor/LDRresistance.
Uriah
Uriah
I had two objectives in mind:
First, to become acquainted with these devices, particularly the relationship between the 20ma and 40 ohm specifications, which I view as critical.
Second, to verify the correct current limiting resistor value, which is critical to the long-term health of the devices.
For myself, I confirmed that 20ma would result in 40 ohms or less in LDR resistance, and that is a good starting point for me because I can then adjust individual current limits for each device to deliver a certain resistance, maybe 40 ohms, maybe a little more.
I confirmed my earlier opinion that 100 ohm current limit resistors will result in over-driving the LEDs. I originally thought that 125 ohms would be about right (based on the Silonex documentation indicating a 2.5V drop across the LED). However, it turns out the drop is more like 1.9V, and thus the correct limit resistor is 150 ohms for a 5V supply voltage. This is important. If people are out there with their pots set to minimum or maximum with a 100 ohm resistor, they are driving 30ma through their LED, and this is more than they should by 50%.
As for my system, curves will not be important because I will drive each of the three device sets in the T attenuator to the precise current required for a given attenuation level. So, rather than matching curves, I will drive individual units to the correct resistance for a given attenuation, no curve matching required.
For a fixed Zo of 5K and attenuation of up to 54dB, the resistors across the top of the T will vary between 40 and 4980 ohms, and the resistance to ground will vary from 310K to 20 ohms (two devices in parallel). The minimum attenuation with 40 ohm resistors is .14dB, and requires 310K to ground, .5dB requires 86K, and 1dB requires 43K, so only two or three levels will require very large resistance values, and below that the values become very reasonable.
For the 40 ohm settings, I will start with a current limit via 150 ohm resistor to 20ma, then individually adjust the current further so that each device sits at exactly 40 ohms, which will be a lower current than the 20ma which generates resistances in the thirties and some twenties ohmic values.
As of now, I plan my first board to be selectable to stereo (1 volume, 1 balance) or dual mono (two volume) controls. If my first impression of the Silonex devices proves correct, the control pots will be linear and the Silonex devices will generate an audio taper output with some interim adjustment by the processor.
Oops. NOT 50%, I don't know where that came from, sorry. Anyway, 10ma, which is still way too much.
This Member has his head screwed on correctly.
He measures and checks.
He allows the arithmetic to remove the need for guessing.
whether you are using the LED/LDR for volume control or for T attenuator then this method of trimming the lowest usable/reliable LDR resistance is perfect.
The volume controller then uses the 1k0 trim pot to balance the gains in the the most used region of the volume control.
1. If you test the voltage drop across LEDs as I did and can confirm values for each device, you can set the exact resistance required to control current to within a few tenths of a milliamp, and a higher voltage to smooth out the inequalities between devices becomes moot. A preponderance of the devices I tested can run from a single value resistor because the voltage drop is consistent. Having now tested this batch of twenty or so, I am satisfied that you can safely run the devices with a 150R for current limiting without sacrificing performance. Even if you get slightly less than 20ma with that value, you probably will still get a 40 ohm result because most of these devices deliver better than 40 ohms at 20ma unless you get one like the one I rejected that has a minimum resistance of 58 ohms.
2. With a microprocessor managing current to the devices, I don't want any voltage on the board higher than the supply of the micro. Any voltage even a fraction of a volt above the micro's supply voltage can be deadly to the micro's health and well-being. Thus, I will run my devices at 5.00V, exactly the same as my micro.
Higher voltage is still desirable in a system where minimum limit resistance is not individually checked and you want a system that is safer forout-of-tolerance devices because the higher voltage allows a higher value limit resistor, which minimizes maximum current differences between devices with different voltage drops.
I like around 150R as well. In the little board I sell that is basically the Lightspeed backend I use a 200R pot for the Shunt LDRs. They reference ground through it with the wiper going to ground and each shunt tied to one of the other legs. Then between the outer legs and the LDR is a series 68R resistor. So approx 168R for each shunt. Works well and no burnt LDRs.
Uriah
Uriah
Uriah, could you explain the purpose of this setup -- is it intended as a control to balance the two channels? What about the series device?
Yep, its for balance. Series you need two pots. This way only one. Also when I match I cant match under about 200-250R so the 200R pot allows matching that last little bit without effecting the matching at higher resistances. If you match several hundred of them at high mA then you end up drawing a serious amount of current and not having a super accurate power supply. If I match at that level then I dont have to draw a huge amount of current and I have a super precise supply.
Uriah
Uriah
Varying Ground to each LDR for balance.
Here is a drawing further illustrating my grounding technique that allows balance with one pot. This is what's on the tiny board I sell on my site. Anyone can feel free to use it for their own uses as long as youre not selling my exact representation of it. Meaning, go build your own boards if you like.
Here is a drawing further illustrating my grounding technique that allows balance with one pot. This is what's on the tiny board I sell on my site. Anyone can feel free to use it for their own uses as long as youre not selling my exact representation of it. Meaning, go build your own boards if you like.
Attachments
wapo54001 has started his own thread
http://www.diyaudio.com/forums/analog-line-level/170381-precision-led-ldr-based-attenuator.html
http://www.diyaudio.com/forums/analog-line-level/170381-precision-led-ldr-based-attenuator.html
That board I just showed the schematic for is part of whats available on my site. There is a sale on a small kit right now. Includes 4 matched LDRs, the board, all parts that are mounted on the board plus the 5k pot. You will need to source a dual log 100k pot and a 5V supply. One member is trying Teddy Pardo's regulator and I think this is a fantastic idea that will provide excellent sound. For a while I will have the board/parts/LDRs for sale for $30. I think thats a fine deal. I hope you guys will write about the sound you hear and help get others interested in LDRs.
http://www.buildanamp.com/LDRs-for-Lightspeed-Board-and-Parts-Included-LDRsBOARDKIT.htm
Uriah
http://www.buildanamp.com/LDRs-for-Lightspeed-Board-and-Parts-Included-LDRsBOARDKIT.htm
Uriah
Last edited:
Feedback from a new constructor
I hope you guys will write about the sound you hear and help get others interested in LDRs.
Happy to. I recently finished building UD's version of the basic Lightspeed circuit using a decent PSU and have been listening to it now for a day or two.
First, thanks to Uriah for making matched LDRs and a PCB available at very fair prices. They make what is otherwise an esoteric project pretty easy.
Second, very hearty thanks indeed to George for making his design available to the audio DIY community and for his on-going help to constructors. His generosity is unusual and should be acknowledged more than perhaps it is. He and UD have the patience of a brace of saints. (Yup, I think I've read every page of this looooong thread . . .)
Third, sound quality. I replaced a pretty good stepped attenuator which was itself a significant step up from an elderly Alps pot. Simply put, the Lightspeed circuit blows it away.
My experience is that almost everything the designers and others claim for LDR-based attenuation seems to be true, especially the razor-sharp detail, well-defined bass and almost tactile vocals. The effect is particularly marked on complex orchestral music where it is now much easier to follow what is going on. That alone makes it a worthwhile upgrade.
On the technical side, my source is a highly-resolving audio PC (cMP^2), the Lightspeed's PSU is a step up from run-of-the-mill LM317 circuits but nothing overly exotic, the driving DAC's Z(out) is 90R and the power amp's Z(in) 20K. This last is of course significantly lower than George and others recommend but it's working just fine. (I'd anticipated having to modify the amp's input or even to buffer it but it really seems unnecessary.)
In short, a super upgrade and extraordinary value for money (as BTW I assume George's production unit is even if it's out of my reach). Anyone dithering about whether to proceed should just go for it.
In a world where folk happily pay $1,000 for a piece of wire, they'd be daft not to.
I hope you guys will write about the sound you hear and help get others interested in LDRs.
Happy to. I recently finished building UD's version of the basic Lightspeed circuit using a decent PSU and have been listening to it now for a day or two.
First, thanks to Uriah for making matched LDRs and a PCB available at very fair prices. They make what is otherwise an esoteric project pretty easy.
Second, very hearty thanks indeed to George for making his design available to the audio DIY community and for his on-going help to constructors. His generosity is unusual and should be acknowledged more than perhaps it is. He and UD have the patience of a brace of saints. (Yup, I think I've read every page of this looooong thread . . .)
Third, sound quality. I replaced a pretty good stepped attenuator which was itself a significant step up from an elderly Alps pot. Simply put, the Lightspeed circuit blows it away.
My experience is that almost everything the designers and others claim for LDR-based attenuation seems to be true, especially the razor-sharp detail, well-defined bass and almost tactile vocals. The effect is particularly marked on complex orchestral music where it is now much easier to follow what is going on. That alone makes it a worthwhile upgrade.
On the technical side, my source is a highly-resolving audio PC (cMP^2), the Lightspeed's PSU is a step up from run-of-the-mill LM317 circuits but nothing overly exotic, the driving DAC's Z(out) is 90R and the power amp's Z(in) 20K. This last is of course significantly lower than George and others recommend but it's working just fine. (I'd anticipated having to modify the amp's input or even to buffer it but it really seems unnecessary.)
In short, a super upgrade and extraordinary value for money (as BTW I assume George's production unit is even if it's out of my reach). Anyone dithering about whether to proceed should just go for it.
In a world where folk happily pay $1,000 for a piece of wire, they'd be daft not to.
Thanks for the pat on the back Ryelands, makes me feel good people appreciate it.
Just so people don't think the production Lightspeed Attenuator is the same price as the above piece of wire, it's not, it's $450 posted worldwide.
Cheers George
Hi
I've recently completed Uriah's LDR mount & board. I'm finding I have to rotate the 5K series balance pot fully clockwise to achieve a near central image. If I start to rotate counter clockwise it starts to shift to the right speaker and the volume increases. I thought I had calibrated the shunt LDRs correctly prior to testing...maybe not?
I've recently completed Uriah's LDR mount & board. I'm finding I have to rotate the 5K series balance pot fully clockwise to achieve a near central image. If I start to rotate counter clockwise it starts to shift to the right speaker and the volume increases. I thought I had calibrated the shunt LDRs correctly prior to testing...maybe not?