This circuit was designed some years ago for a specific need. My Micromega Stage 2 CD player had only a coax SPDIF output, and my Minidisc recorder only an optical input.
It may be of interest to others because the optical transmitter is a DIY affair and the method of driving the LED is unconventional.
A TOSLINK coupler was used to make the transmitter with a small LED carefully epoxied into one half. The LED chosen is a "perfect" fit. I experimented with a few LED types and found there were marked differencies in efficiencies. The one chosen works well even if the fibre optic lead is held within a couple of centimetres of the coupler... it's that good. Most common high brightness red LED's will work however.
The method of drive was developed after seeing how the non linear load/capacitance of an LED junction distorted the voltage drive waveform
across it. How much of an effect this has in practice with regard to jitter I can't say but it's satisfying to know the design drives the LED very cleanly indeed.
The DIY optical transmitter and drive circuit (R/C/Diode) proved so successful that I actually built another and incorporated it into the Micromega running it from the TTL drive that fed the original coax output feed in the player.
The waveforms show the drive (measured from ground to the LED anode) at a frequency of 2Mhz. The difference of conventional drive with just a resistor feeding the LED, and the modified drive can be clearly seen. The scope was set at 1 volt/div with the "zero" or ground centered. The modified version pulls the LED cleanly down below zero, the actual negative voltage limited by the 1N4148 clamp.
It may be of interest to others because the optical transmitter is a DIY affair and the method of driving the LED is unconventional.
A TOSLINK coupler was used to make the transmitter with a small LED carefully epoxied into one half. The LED chosen is a "perfect" fit. I experimented with a few LED types and found there were marked differencies in efficiencies. The one chosen works well even if the fibre optic lead is held within a couple of centimetres of the coupler... it's that good. Most common high brightness red LED's will work however.
The method of drive was developed after seeing how the non linear load/capacitance of an LED junction distorted the voltage drive waveform
across it. How much of an effect this has in practice with regard to jitter I can't say but it's satisfying to know the design drives the LED very cleanly indeed.
The DIY optical transmitter and drive circuit (R/C/Diode) proved so successful that I actually built another and incorporated it into the Micromega running it from the TTL drive that fed the original coax output feed in the player.
The waveforms show the drive (measured from ground to the LED anode) at a frequency of 2Mhz. The difference of conventional drive with just a resistor feeding the LED, and the modified drive can be clearly seen. The scope was set at 1 volt/div with the "zero" or ground centered. The modified version pulls the LED cleanly down below zero, the actual negative voltage limited by the 1N4148 clamp.
Attachments
Last edited:
Interesting stuff. Thanks for posting it.
I'm not a diode person, so my question is whether you were able to observe the diode output: specifically, how does the light output from the diode with the conventional drive compare to the conventional voltage?
Is it possible that although the conventional voltage drive is distorted the light output may still be "square"? (Or at least more "rectangular" if the duty cycle is distorted)?
And what are you using to drive the LED? I don't understand the code for those IC's.
I'm not a diode person, so my question is whether you were able to observe the diode output: specifically, how does the light output from the diode with the conventional drive compare to the conventional voltage?
Is it possible that although the conventional voltage drive is distorted the light output may still be "square"? (Or at least more "rectangular" if the duty cycle is distorted)?
And what are you using to drive the LED? I don't understand the code for those IC's.
Hi,
The drive to the LED is the SPDIF signal from a CD player which is the digital "code" carrying all the audio/clock/sync information. That is the signal that appears on the electrical SPDIF coax or SPDIF phono socket output that many players have.
The IC's on the board are hi speed CMOS invertors and just buffer the signal.
On a normal 'scope all you can see is a randomly changing signal (not a square wave which I used to design and test) but you can make out that the signal (on the LED) is cleaner in the same way that the test signal is.
Hard to say whether the light output is more square or not... I think it must be in the sense that the transitions are cleaner. If the LED voltage falls slowly the light output must also fall rather than be abruptly cut off and so I think it makes the receivers job a lot easier in that the signal is more definite.
The drive to the LED is the SPDIF signal from a CD player which is the digital "code" carrying all the audio/clock/sync information. That is the signal that appears on the electrical SPDIF coax or SPDIF phono socket output that many players have.
The IC's on the board are hi speed CMOS invertors and just buffer the signal.
On a normal 'scope all you can see is a randomly changing signal (not a square wave which I used to design and test) but you can make out that the signal (on the LED) is cleaner in the same way that the test signal is.
Hard to say whether the light output is more square or not... I think it must be in the sense that the transitions are cleaner. If the LED voltage falls slowly the light output must also fall rather than be abruptly cut off and so I think it makes the receivers job a lot easier in that the signal is more definite.
- Status
- Not open for further replies.