The present invention relates to optical laser transmission systems in general and in particular to a sub-carrier multiplexed signals being launched into a laser. More particularly, the present invention relates to a solution for avoiding clipping of optical communication signals in an optical laser transmission system.
MSOs (multi-service operators) provide several services to end users through a fiber optic network, with the final connection to the user through a coaxial connection. The services provided by the MSO typically include broadcast analog video and narrow cast digital services (also referred to as sub-carrier signals), such as data, VoIP, subscription, pay per view and video on demand (VOD) services. The services are generally allocated a portion of an optical channel, which typically has approximately 1 GHz bandwidth available. While the bandwidth of a channel is generally constrained by the optical network (the optical network transmitters, optical fiber, channel filter bandwidths and coaxial connection), the number of users connected to the optical network continues to increase, which requires increased demand for bandwidth for the desired services.
In recent years wavelength division multiplexed (WDM) optical transmission systems have been increasingly deployed in optical networks to meet the increased demand for bandwidth by providing more than one optical channel over the same optical fiber. The WDM techniques include coarse wavelength division multiplexed (CWDM) and dense wavelength division multiplexed (DWDM) systems. Whether a system is considered to be CWDM or DWDM simply depends upon the optical frequency spacing of the channels utilized in the system.
Narrowcast signals may be RF frequency multiplexed into broadcast channels. The narrowcast signals are typically digital signals and are normally much lower in amplitude than broadcast video signals. The arrangement of sending the same broadcast signal and different narrowcast signals over multiple wavelengths (WDM) is a means of providing more segmentation in an optical network. Typically, the lasers 2121, 2122, 2123, and 2124 each receive a different narrowcast signal. The wavelengths carrying the combined broadcast and individual narrow cast signals, λ1, λ2, λ3 and λ4, respectively, are optically multiplexed onto optical fiber 240.
When the modulators 2101, 2102, 2103, or 2104 RF carriers over drive the laser, (e.g. drive the laser to produce laser pulses having an amplitude greater than the network or laser tolerances), clipping events will happen. The clipping events significantly impact a quadrature amplitude modulation (QAM) signal. Usually clipping will increase the bit error rate (BER) of the QAM signal. Although most QAM receivers use forward error correction (FEC), if the BER is too high, even FEC cannot recover the signal. When that happens the information may be lost.
One approach to avoid clipping is to use a delay line circuit and high speed peak detector to reduce the clipping events. The sub-carrier multiplexed signal passes through an RF delay line before being applied to the laser. Before the sub-carrier multiplexed signal goes to the delay line, a high speed peak detector is used to detect the clipping peak. The laser current is increased relatively slowly to prevent the bias current from producing signals at frequencies within the transmission band. Because the delay time is relatively long, the physical size of the coaxial delay line and the RF loss of the delay line make it difficult to be implemented.
Another approach is disclosed in U.S. Pat. No. 6,549,316 to Henry A. Blauvelt, which describes an anti clipping circuit which does not require significant delay of the main RF signal. The circuit design was based on statistical measurements of measured clipping events. The design includes a diode peak detection circuit that preferably generates a laser bias control signal that is proportional to the frequency and intensity of the clipping events, which is AC coupled to the lasers DC bias circuit. The approach also assumes that the clipping events happen randomly.
Accordingly, an improved method and apparatus for reducing clipping distortion is needed for a laser transmitter for reducing the BER of QAM signal transmitting.
The invention is directed toward a circuit design based on actual clipping measurements showing the specific time and duration of the clipping events are random. Once a clipping event occurs, the individual clipping pulse timing is fixed, such as by every 4 microseconds. When a clipping event is detected, the laser bias current/voltage gradually increased over about a couple of microseconds, so the pulse will not produce spurious signals within the CATV frequency band. Once it is detected that the clipping events are over, the laser bias current is gradually reduced to its original value. Thus the laser bias current is adjusted automatically according to the specific time of the clipping events and the duration of the clipping period of the events.
The increase of lasers bias current is DC coupled into the main laser DC bias circuit. One advantage of doing so is that when there are no clipping events, the PHEMT peak detected circuit noise can be isolated from the laser DC bias power supply by a reverse biased diode. Another advantage of DC coupling into the main laser DC bias is that when the clipping events are over, there is essentially no recovery time for the laser bias so that it will not cause additional clipping events during the laser current recovery time as may occur with AC coupling.
In accordance with an apparatus of the present invention, a laser transmitter may comprise: a modulator configured to receive a communication signal and provide a modulated signal; a laser configured to provide an optical communication signal based on the modulated signal; and an anti clipping unit configured to reduce clipping of the optical communication signal by controlling a parameter associated with the laser, wherein the anti clipping circuit is configured to periodically remove a clipping event after a random clipping event is detected. In the laser transmitter, the parameter associated with the laser includes a bias current or voltage of the laser. The anti clipping unit may include a pseudomorphic HEMT transistor. The anti clipping unit may include a missing pulse detector associated with a timer. The anti clipping unit may include transistors which convert a timer control signal associated with the missing pulse detector to increase a current to the laser. The anti clipping unit may be isolated from the laser during periods of no clipping.
In accordance with an apparatus of the invention, an anti clipping unit may comprise: a PHEMT peak detector; a missing pulse detector; and an isolation diode which isolates the anti clipping unit from a laser bias source, wherein the anti clipping circuit is configured to periodically remove a clipping event after a random clipping event is detected by controlling a parameter associated with the laser. The parameter associated with the laser may include a bias current or voltage of the laser. The PHEMT unit may include a pseudomorphic HEMT transistor. The missing pulse detector may include a timer. The anti clipping unit may further include transistors which convert a timer control signal associated with the missing pulse detector to increase a current to the laser. The anti clipping unit may be isolated from the laser during periods of no clipping.
In accordance with a method of the present invention for reducing clipping events in an optical communication system using an anti clipping unit comprising the steps of: detecting a random clipping event; removing clipping events after the random clipping event by controlling a parameter associated with a laser in the optical communications system; and isolating the anti clipping unit from the laser when clipping events are no longer occurring. The clipping events may include at least one pulse in a 4 microsecond interval. The parameter associated with the laser may include a bias current or voltage of the laser. The step of removing clipping events may include triggering a missing pulse timer for an appropriate duration. The step of removing clipping events may include increasing a bias current or voltage of the laser. The bias current or voltage of the laser may be increased gradually so that the increased pulse does not produce spurious signals within a predetermined frequency band.
The anti clipping unit of the invention may be used with a laser transmitter used to transmit sub-carrier multiplexed signals for CATV applications. When the anti clipping circuit is used, previously uncorrectable signals can be recovered so that an error free digital signal transmission can be realized. The laser bias current is adjusted automatically at the specific time of the clipping event and for the duration of all the clipping pulses in the event. The controlled increase and decrease of the lasers bias current is DC coupled back into the normal DC bias path via a diode. This gives the anti clipping circuit noise isolation when there are no clipping events and avoids laser clipping during the laser current recovery time as when using an AC coupling scheme. There is essentially no laser bias recovery time for this design and it can be used in any wavelength laser transmitters.
The anti clipping unit of the invention may be used with a laser transmitter used to transmit sub-carrier multiplexed signals for CATV applications. When the anti clipping circuit is used, previously uncorrectable signals can be recovered so that an error free digital signal transmission can be realized. The circuit design based on actual clipping measurements showing the specific time and duration of the clipping events are random. Once a clipping event occurs, the individual clipping pulse timing is fixed, such as by every 4 microseconds. A pseudomorphic HEMT transistor (PHEMT) peak detection circuit is used in conjunction with a missing pulse detector/timer to catch all the clipping events. Once it detects a clipping event, the missing pulse detector is triggered which will increase the laser bias current/voltage. The laser bias is gradually increased over about a couple of microseconds, so the pulse will not produce spurious signals within the CATV frequency band. The missing pulse detector will be triggered while the clipping events exist so the laser bias current stays higher to eliminate subsequent clipping events.
Once the pulse detector determines that the clipping events are over, the missing pulse detector timer is de-triggered. The laser bias current is gradually reduced to its original value. Thus the laser bias current is adjusted automatically according to the specific time of the clipping events and the duration of the clipping period of the events. The increase of lasers bias current is DC coupled into the main laser DC bias circuit. One advantage of doing so is that when there are no clipping events, the PHEMT peak detected circuit noise can be isolated from the laser DC bias power supply by a reverse biased diode. Another advantage of DC coupling into the main laser DC bias is that when the clipping events are over, there is essentially no recovery time for the laser bias so that it will not cause additional clipping events during the laser current recovery time as may occur with AC coupling.
The circuit design is based on actual clipping measurements showing the specific time and duration of the clipping events is random. A PHEMT peak detection circuit is used in conjunction with a missing pulse detector/timer to catch all the clipping events. Thus the laser bias current is adjusted automatically at the specific time of the clipping event and for the duration of all the clipping pulses in the event. The controlled increase and decrease of the lasers bias current is DC coupled back into the normal DC bias path via a diode. This gives the anti clipping circuit noise isolation when there are no clipping events and avoids laser clipping during the laser current recovery time as when using an AC coupling scheme. There is essentially no laser bias recovery time for this design and it can be used in any wavelength laser transmitters.
While anti clipping unit 220 is illustrated as being separated from modulator 210, the arrangement illustrated is only for discussion purposes. Those of skill in the art will appreciate that the anti clipping unit 220 may be integrally formed with modulator 210. Also, the anti clipping unit may be used with any laser transmission system, including single laser transmission systems, and is not confined to WDM laser transmission systems. Further, the anti clipping unit may be used to monitor and correct clipping events in both the broadcast and sub-carrier signals, anti clipping unit may be positioned to monitor and correct clipping events only in the sub-carrier signals or only in the broadcast signals.
Our experiment test results showed that when a laser is modulated by multi-channel signals the clipping events will happen at a random time. Once a clipping event does happen, every 4 micro sec the second clipping event will follow. How many clipping events (e.g. 4 micro sec pulses) will follow the first clipping event will be also random. Our circuit design is based on actual clipping measurements which show the specific time and duration of the clipping events are random. Once the clipping events happen, the individual clipping pulses timing is fixed, for example at every 4 μs.
A PHEMT peak detection circuit is used in conjunction with a missing pulse detector/timer to catch all the clipping events. Once it detects a clipping event, the missing pulse detector is triggered which will increase the laser bias current/voltage. The laser bias is gradually increased over about a couple of microseconds, so the pulse will not produce spurious signals within CATV frequency band. In a preferred implementation, the missing pulse detector will be always triggered while the clipping events exist so the laser bias current stays higher to eliminate subsequent clipping events. Once the pulse detector determines that the clipping events are over, the missing pulse detector timer is de-triggered. The laser bias current is gradually reduced to its original value. Thus the laser bias current is adjusted automatically according to the specific time of the clipping events and the duration of the clipping period of the events. The increase of lasers bias current is DC coupled into the main laser DC bias circuit. One advantage of doing so is that when there are no clipping events, the PHEMT peak detected circuit noise can be isolated from the laser DC bias power supply by a reverse biased diode. Another advantage of DC coupling into the main laser DC bias is that when the clipping events are over, there is essentially no recovery time for the laser bias so that it will not cause additional clipping events during the laser current recovery time which may occur with AC coupling.
An implementation of the invention uses the nonlinear effect of low noise high speed Enhancement Mode Pseudomorphic HEMT (PHEMT) transistor to detect the clipping events. A missing pulse detector of a timer may be used to automatically adjust for the random number of 4 micro second pulses, such as the missing pulse detector of a NE 555 timer. A transistor arrangement may be used to convert the timer control signal to increase the laser current so that best anti clipping results can be obtained. The DC coupling circuit may be used for the laser current.
When there are not any clipping events, the anti clipping circuit is disconnected to the laser bias circuit because they are isolated by (e.g. reverse biased) a diode. So the anti clipping circuit noise will not enter into the laser transmitter. Once the clipping events are coming, these two circuits are connected by the forward bias diodes. Once the clipping event is over, the laser current is set back to its original value. There is essentially no laser bias recovery time for this design and it can be used in 1310 nm and 1550 nm laser transmitters.
The upper trace in
In
A trigger point is denoted in
The PHEMT is biased so that non clipping RF levels do not bring Q1 into conduction, e.g. there is no current flow in the drain of transistor Q. Only when the RF input voltage is positive enough to drive the laser, 8, into clipping will the PHEMT start to conduct. The current flow in the resistor R6 will cause the voltage at N3 to go negative. The clipping negative pulse on N3 is passed to the emitter follower Q2, to N4. This is what we see in the negative pulses of
The adjustable resistor, R17, is used to adjust the triggering level for the 555 timer circuit shown in
When the negative pulse on N4 is passed to the Q3 base, the negative pulse will cause Q3 to conduct, thus discharging C19. The time constant choice for the C19/R18 is larger than 4 μs. As long as the 4 μs negative pulses exist, the voltage on the capacitor C19 will never get to the predetermined de-trigger point. The missing pulse timer will remain triggered until all the 4 μs clipping pulses are gone.
Once the 4 μs negative pulses exist, the missing pulse timer is triggered. The timer 555 pin 3 (N5) will output a high voltage (about 5 volts). It will cause the transistor Q4 to cutoff. When the transistor Q4 has been cut off, the negative 5 volt source will provide the additional laser current through resistance R25/26 and diode D3, (DC coupled), to the laser, 8. By controlling the value of the resistor, R25, the additional current to the laser can be controlled. Resistance R25 and capacitor C27 form a low pass filter with a time constant of about 1 μs. It makes the anticlipping current pulse increase and decrease gradually over a couple of microseconds.
When the missing pulse timer finds the input clipping event is over, the missing pulse timer will be de-triggered. In this case, the timer pin 3, (N5), will be at a low voltage (0 volts). The transistor Q4 will be conducting. The voltage on the N6 will be higher than the lasers normal forward bias voltage (about −2 volts), derived from the laser power supply, 10. In this case, the voltage on the N6 is higher than the voltage drop on the N7. Thus, diode D3 is reversed biased, so any noise created by the anti clipping circuit will be isolated from the laser. This also prevents any laser clipping during the laser current recovery time as when using an AC coupling scheme.
The anti clipping unit of the invention may be used with a laser transmitter used to transmit sub-carrier multiplexed signals for CATV applications. When the anti clipping circuit is used, previously uncorrectable signals can be recovered so that an error free digital signal transmission can be realized. The circuit design is based on actual clipping measurements showing the specific time and duration of the clipping events is random. A PHEMT peak detection circuit is used in conjunction with a missing pulse detector/timer to catch all the clipping events. Thus the laser bias current is adjusted automatically at the specific time of the clipping event and for the duration of all the clipping pulses in the event. The controlled increase and decrease of the lasers bias current is DC coupled back into the normal DC bias path via a diode. This gives the anti clipping circuit noise isolation when there are no clipping events and avoids laser clipping during the laser current recovery time as when using an AC coupling scheme. There is essentially no laser bias recovery time for this design and it can be used with any wavelength laser transmitters.
Number | Name | Date | Kind |
---|---|---|---|
6549316 | Blauvelt | Apr 2003 | B2 |
20020063930 | Blauvelt | May 2002 | A1 |
20040056720 | Jansen et al. | Mar 2004 | A1 |
Number | Date | Country | |
---|---|---|---|
20080292315 A1 | Nov 2008 | US |