A precision LED/LDR-based Attenuator - Page 5 - diyAudio
Go Back   Home > Forums > Source & Line > Analog Line Level

Analog Line Level Preamplifiers , Passive Pre-amps, Crossovers, etc.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 15th August 2010, 02:01 PM   #41
diyAudio Member
 
Join Date: Nov 2008
Quote:
Originally Posted by jackinnj View Post
You mean the company whose IR spectrometer I used for qualitative organic chemistry made a typo?

VTL5C3 1.5 ohms at 40mA.
Yup, that sounds like the one where the graph shows 1.5K ohms instead of 1.5 ohms and 1.5K is correct. I actually got an order of parts before I discovered the error. Allied gave me a full refund and told me to throw the parts away.
  Reply With Quote
Old 15th August 2010, 07:29 PM   #42
gootee is offline gootee  United States
diyAudio Member
 
Join Date: Nov 2006
Location: Indiana
Blog Entries: 1
Quote:
Originally Posted by wapo54001 View Post
Well, the minimum resistance value is important in the max attenuation range. With 40 ohms, with a 5K Zo, the math says I only get 42dB attenuation maximum. At a lower Zo, a slightly greater attenuation is possible.

Is there a PerkinElmer device that reaches, or almost reaches, 40 ohms? I did not find one when I went looking . . .

At the minimum attenuation end, you only need a much greater number, so it's not a problem.

With reagard to the dual-LDR devices, the dual LDR shares a common wire (3-wire device) and that kinda limits its usefulness.
Well, I could have sworn that they had lots of them that went below 100 Ohms. But looking at the data plots again I see only two. One of those goes to 50 Ohms but it is undesirable otherwise. And I could have sworn that the VTL5C2 models that I measured went down to 40 Ohms or even less. But that was a long time ago and the Perkin Elmer data plot that I'm looking at right now shows between 100 and 200 Ohms minimum for the VTL5C2.

So I'm sorry. What I said was apparently INCORRECT.

But I wonder if there would be any advantage to putting them in parallel with fixed resistances, to get a lower minimum resistance. Their dynamic ranges do go from 65 dB to 112 dB. And they do seem to have a larger selection of other characteristics. You might also consider paralleling the Silonex devices with fixed resistors, if it happens to be advantageous.

Cheers,

Tom
  Reply With Quote
Old 15th August 2010, 08:51 PM   #43
diyAudio Member
 
Join Date: Nov 2008
Quote:
Originally Posted by gootee View Post

But I wonder if there would be any advantage to putting them in parallel with fixed resistances, to get a lower minimum resistance. Their dynamic ranges do go from 65 dB to 112 dB. And they do seem to have a larger selection of other characteristics. You might also consider paralleling the Silonex devices with fixed resistors, if it happens to be advantageous.

Cheers,

Tom
Well, I've thought of that, but first I want to try a 'purist' approach that puts LDRs only in the signal path. I've looked at wandering resistances at high values, I think it's partly due to actual variation in resistance, and partly due to the difficulty of measuring large resistance values accurately. I still think I can make a .5~42dB attenuator with a fairly constant Z of 5K, and that should be pretty good for many applications.

I've ordered some boards which will be here Friday. After I get one up and running and able to program it, we'll see how complicated the software has to get. It'll be fairly easy to set two diffierent current ranges for the series and shunt LDRs and have them respond in opposite directions as you change the input. Right off the bat, there will be advantages over the straight log pot technique due to being able to program two different resistance ranges to respond to a single input range; it remains to be seen how much math will eventually be required to make the native Silonex curves play correctly with smooth, consistent, attenuation and relatively constant Z.
  Reply With Quote
Old 16th August 2010, 06:01 AM   #44
gootee is offline gootee  United States
diyAudio Member
 
Join Date: Nov 2006
Location: Indiana
Blog Entries: 1
Quote:
Originally Posted by wapo54001 View Post
Well, I've thought of that, but first I want to try a 'purist' approach that puts LDRs only in the signal path. I've looked at wandering resistances at high values, I think it's partly due to actual variation in resistance, and partly due to the difficulty of measuring large resistance values accurately. I still think I can make a .5~42dB attenuator with a fairly constant Z of 5K, and that should be pretty good for many applications.

I've ordered some boards which will be here Friday. After I get one up and running and able to program it, we'll see how complicated the software has to get. It'll be fairly easy to set two diffierent current ranges for the series and shunt LDRs and have them respond in opposite directions as you change the input. Right off the bat, there will be advantages over the straight log pot technique due to being able to program two different resistance ranges to respond to a single input range; it remains to be seen how much math will eventually be required to make the native Silonex curves play correctly with smooth, consistent, attenuation and relatively constant Z.
Here are two other ideas (edit: the second one is probably more attractive):

1.) I wonder if you should try to use feedback control loops, so that basically "nothing would matter" regarding the actual LDR curves, and you could maybe greatly reduce the overall complexity of your scheme in that regard.

Essentially, if it were able to be implemented, your system could simply tell the LDR control systems what resistances were needed and the feedback systems would automagically adjust the LDRs' control currents to get them there.

To implement it, you would only need to be able to continuously measure the resistances, to provide the feedback signals. Then you just subtract the signals that correspond to the resistances from the command signals that correspond to the desired resistances to produce a signal that changes the current. When the resistances are what your system told them to be, the command signals (literally) automatically stop changing and the currents and resistances are exactly where you told them to be.

To do that, you would have to be able to sense current through and voltage across the LDRs' resistive elements, without the audio circuit being affected significantly (i.e. basically not at all). But that should be able to be done by using extremely high input impedance voltage sensors (maybe just the right type of FET-input opamps?).

Once you could sense the voltage across a very-low-value current-sense resistor and across the LDR's resistance, you'd almost have it.

The only other thing that might be a minor problem, at first, would probably be setting the time constant for your resulting command signals to be slow-enough for the LDRs to respond, so that you wouldn't over-correct at first, and also wouldn't potentially oscillate the (LDRs') control current.

-----

2.) An entirely different type of scheme that is also conceptually simple (and also requires no additional analog circuitry) would be to have a "calibration/learning" mode, in which the system would automatically test each LDR and store its entire resistance-vs-current curve in a table, so that from then on it could just do a table lookup to set the current to exactly what was needed for any desired resistance. And Voila!

(I would probably code this type of table-lookup routine so that it would always interpolate between data points, if necessary. Then, potentially, your software or firmware could piecewise-linearize each curve, within certain error margins you had set, when it built each table, possibly storing far fewer data points, but it would still all work within specs.)

Cheers,

Tom
  Reply With Quote
Old 16th August 2010, 02:04 PM   #45
diyAudio Member
 
Join Date: Nov 2008
Quote:
Originally Posted by gootee View Post
Here are two other ideas (edit: the second one is probably more attractive):

1.) I wonder if you should try to use feedback control loops, so that basically "nothing would matter" regarding the actual LDR curves, and you could maybe greatly reduce the overall complexity of your scheme in that regard.

-----

2.) An entirely different type of scheme that is also conceptually simple (and also requires no additional analog circuitry) would be to have a "calibration/learning" mode, in which the system would automatically test each LDR and store its entire resistance-vs-current curve in a table, so that from then on it could just do a table lookup to set the current to exactly what was needed for any desired resistance. And Voila!

Tom
For me, I would view the first as beyond my abilities to implement. It would introduce major technical challenges including a slew of additional analog parts and I can't imagine how you would extract resistance information from music. Not that it can't be done, I just don't know how to do it.

The second is what I am actually doing. I don't need any analog chips to make it happen -- aside from the PIC, just one capacitor, three resistors (one for current limiting), and one mosfet per LED. That's a pretty minimal parts count. Everything else on the board is for voltage regulation, volume control, and option selection.
  Reply With Quote
Old 16th August 2010, 02:40 PM   #46
diyAudio Member
 
jackinnj's Avatar
 
Join Date: Apr 2002
Location: Llanddewi Brefi, NJ
If I were going into the matched LDR biz, I would just use a programmable current source to establish the control law for each device, and bin them. (Now where did I put my HP6177C?)
  Reply With Quote
Old 16th August 2010, 03:11 PM   #47
diyAudio Member
 
Join Date: Nov 2008
Quote:
Originally Posted by jackinnj View Post
If I were going into the matched LDR biz, I would just use a programmable current source to establish the control law for each device, and bin them. (Now where did I put my HP6177C?)
The devil is in the details . . .

Have you found a programmable current source which, in a single chip, can
1) control current from 20ma to .0005ma?
2) control four (up to eight) different current loads simultaneously?
3) interpret two separate inputs and apply them differently to different outputs?
4) offer programmable options to offer flexible inputs?

If yes to all, please tell me and I'll have a look. If not, I'll stick with my single-chip PIC solution . . .

BTW The whole point in my project is to entirely do away with 'matched' LDRs -- I want to use LDRs off-the-shelf that require no matching other than to insure they meet minimum specifications.
  Reply With Quote
Old 16th August 2010, 04:27 PM   #48
diyAudio Member
 
jackinnj's Avatar
 
Join Date: Apr 2002
Location: Llanddewi Brefi, NJ
20mA to 500nA is a range of 92dB -- at 500nA you might find that the radiation from the PIC is going to modulate the diode current!

There's an interesting tale on the Analog Devices website, of a designer who was driving himself nuts with a strange modulation of a low noise circuit -- turns out that the radiation from a fluorescent bulb was impinging upon one of the diodes. It's not medical instrumentation with life-threatening applications we're discussing here, but interesting to note.
  Reply With Quote
Old 16th August 2010, 04:45 PM   #49
diyAudio Member
 
Join Date: Nov 2008
Quote:
Originally Posted by jackinnj View Post
20mA to 500nA is a range of 92dB -- at 500nA you might find that the radiation from the PIC is going to modulate the diode current!

There's an interesting tale on the Analog Devices website, of a designer who was driving himself nuts with a strange modulation of a low noise circuit -- turns out that the radiation from a fluorescent bulb was impinging upon one of the diodes. It's not medical instrumentation with life-threatening applications we're discussing here, but interesting to note.
If it does, it'll be at 32MHz!
  Reply With Quote
Old 16th August 2010, 05:39 PM   #50
diyAudio Member
 
Join Date: Nov 2008
Quote:
Originally Posted by jackinnj View Post
20mA to 500nA is a range of 92dB -- at 500nA you might find that the radiation from the PIC is going to modulate the diode current!
When I converted to new electronic ballast for the fluorescent fixtures in my workshop, my portable FM radio picked up nothing but intense static. I had to install an outside antenna to be able to listen to the radio again.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
LDR photoresistor attenuator preamp, in general romana Analog Line Level 54 14th December 2013 02:32 AM
TUBEpre and Lightspeed LDR attenuator tinitus Analog Line Level 5 31st January 2010 05:08 PM
LDR attenuator questions Carlp Analog Line Level 15 21st January 2010 01:01 AM
LED to LDR amplifier electro samurai Analogue Source 8 21st March 2004 05:07 PM
Spectrol, high precision, 10k attenuator. Peter Daniel Swap Meet 3 4th February 2004 05:15 PM


New To Site? Need Help?

All times are GMT. The time now is 03:49 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2