LED reverse breakdown in VTL5C4

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

bcarso

Well-known member
Joined
Feb 20, 2005
Messages
4,055
Location
San Fernando Valley, Los Angeles
To All: If anyone has some of these parts could you find out where the Vbr of the LEDs are? The experiment would place a 100k or so R in series with a reasonably high voltage d.c. supply and observe what the voltage across the LED is with a decent high-Z (say a typical 10M input R) voltmeter. If you read about 1.5V then reverse the connections to the LED ;-).

For loose LEDs of various types I've been measuring, the Vbr is anywhere from 25 to 40V. I sent Perkin Elmer an email but not surprisingly I've gotten no response. I may try to get them on the phone, but they may not know the answer themselves and might well be unwilling to divulge who their LED supplier(s) is. All any vendor will say is don't hit the part with more than 3V reverse, or in some cases as high as 5V. Of course they are assuming a voltage source, with no intrinsic current limiting to protect the part from excessive power dissipation if the Vbr is exceeded.

I've also been unable to find any data on potential degradation of performance due to reverse current---the only parameter cited for loss of light output over time is die temperature.

In any event, this is a crucial parameter when using a circuit such as presented recently, in which the LED is driven solely by a series capacitor from a current-limited but large voltage swing source (such as the 12AU7 cathode follower with 22k cathode load R). As outlined by PRR such a circuit functions by routinely reverse-biasing the LED into breakdown to discharge the capacitor. It will affect threshold but even more will tend to make the LED drive circuit more of a peak detector when the breakdown voltage is large relative to the nominal forward voltage. This will have a very significant effect on the audio performance.
 
I have a bunch of them, almost 20. And it's a significantly smooth audio performance, no complaints from my clients.
 
OK---I managed to remember where I saw one and retrieve it. It has a 338 date code (33rd week of '98 or 38th week of '03 or '93 or....). It is a single channel part. Forward voltage at a milliamp is 1.51V, suggesting that it is probably a GaAsP "standard red" part.

With ~0.9mA of reverse current (100V behind 100k) the voltage is 11.2V.

I would like a larger sample size, but this is a start.
 
Wouldn't it make sense to shunt the LED with a 10V Zener + some diode in order to present the reverse-voltage with a discharging path other than the poor little opto?

Selecting the right zener voltage could probably get us quite close to the working of the original design - but without stressing (and thus - over time - degrading) the opto?

Jakob E.
 
Jakob, for me the issue is the consistency of Vbr. With no spec it's a gamble, one which may be appropriate for DIY and even low volume manufacturing. As mentioned the spec on Vreverse max is 3.0V, which is surely a CYA spec. But as I mentioned I have some other GaAsP parts that have much higher Vbr.

If we were certain that Vbr was always comfortably more than 10V then a zener + diode as you suggest would be fine. It would remove any concerns (which may be groundless) about long-term degradation from reverse current. Perhaps it's time for a reverse current vs light output experiment as well. Clearly it can't be a huge effect as analag has had these running for a while.

The design as such has the appeal of low parts count. But if I were doing it I think I would do something where the peak detection was separated from the LED drive, and would make the latter unipolar. I might also want to consider making the peak detection bipolar at least as an option--- a matter of taste as some may prefer the predominance of 2nd that occurs now, especially at lower frequencies.

This could be done hollow-state as PRR calls it, although a little judicious sand would make life easier.
 
In my first incarnation of this project I did use an opamp to drive the opto an op275 to be specific. Strangely the tube seem to drive it way harder and faster, but only in the configuration shown in the schematic I have since inserted a 270 ohm resistor in series with the LED. Everything goes down the drain even when I use a bleeder to properly develop the signal in technicaly correct terms across the LED. It's such a robust compressor that I would rather keep changing the opto (every few years that is)than to change how I drive it.

Analag
 
analg, I am sure that the way you are driving it is important to the proper functioning of your compressor. It makes the circuit work more like a peak detector which I think is advantageous.

There are ways to get similar performance but they are more complex.
 
You are precisely right, I spiced it extensively, and then tested it thoroughly and I have been critically listening to it. It squeezes transparently I must say. It drives the ADC to my DAW and the waveform is controlled, no overs and no sound like it's there compressing the hell out of anything. I have a stereo unit strapped across my mix bus, with a stepped up circuit design, with the same attenuator amp but a different tube, one capable of less current than a 12AU7A and it still works, very well.

Analag
 

Latest posts

Back
Top