Bilal Amin wrote: > Hi Everyone, > > Some days back I emailed on the group regarding the jitter in the sampling clock of ADC. I have tried the phase modulation method for producing the jitter. > I used the HP 8648B Signal generator for this experiment. I modulated the RF wave (at 10 MHz) with noise using the modulating input of the Signal Generator. For noise I used a Noise gen. from General Radio Company. It has three options for frequency. 10K, 100K and 5 M Hz. > I observed the output using an oscilloscope. The output wave form was jittery. BUT when changed the noise from 5M to 100K there was an increase in jitter. Can any one help me out there, why this is happening? According to my understanding decreasing the bandwidth of noise results in decrease in jitter as well. > Also I want to calculate the jitter at the output. i.e. the amount of noise I am adding. I think I wont be able to do it over the oscilloscope. Any ideas ? > > Regards > > Bilal > _______________________________________________ > time-nuts mailing list > time-nuts@febo.com > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > > Bilal The 8648B phase modulation input has a bandwidth of 10Hz to about 10kHz. The resultant phase modulation dependency on the noise generator settings depends on the noise generator specifications. What was the noise generator model, or can you provide a circuit diagram so the effect of the noise generator settings can be deduced? Bruce

----- Original Message ----- From: "Dr Bruce Griffiths" <bruce.griffiths@xtra.co.nz> To: "Discussion of precise time and frequency measurement" <time-nuts@febo.com> Sent: Monday, March 19, 2007 9:25 PM Subject: Re: [time-nuts] Producing jitter with Phase Modulation > Bilal Amin wrote: >> Hi Everyone, >> >> Some days back I emailed on the group regarding the jitter in the >> sampling clock of ADC. I have tried the phase modulation method for >> producing the jitter. >> I used the HP 8648B Signal generator for this experiment. I modulated the >> RF wave (at 10 MHz) with noise using the modulating input of the Signal >> Generator. For noise I used a Noise gen. from General Radio Company. It >> has three options for frequency. 10K, 100K and 5 M Hz. >> I observed the output using an oscilloscope. The output wave form was >> jittery. BUT when changed the noise from 5M to 100K there was an increase >> in jitter. Can any one help me out there, why this is happening? >> According to my understanding decreasing the bandwidth of noise results >> in decrease in jitter as well. >> Also I want to calculate the jitter at the output. i.e. the amount of >> noise I am adding. I think I wont be able to do it over the oscilloscope. >> Any ideas ? >> >> Regards >> >> Bilal >> _______________________________________________ >> time-nuts mailing list >> time-nuts@febo.com >> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts >> >> > Bilal > > The 8648B phase modulation input has a bandwidth of 10Hz to about 10kHz. > > The resultant phase modulation dependency on the noise generator > settings depends on the noise generator specifications. > What was the noise generator model, or can you provide a circuit diagram > so the effect of the noise generator settings can be deduced? > > Bruce > > _______________________________________________ > time-nuts mailing list > time-nuts@febo.com > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > Hi Bruce, I am attaching the schematic with this email. It is actually Random Noise Generator by General Radio Company 1390 B. It is a very old model but I have a limited budget so I am trying to do the experiment using this. I need to produce 10 MHz jittery clock and the maximum frequency for this noise gen. is 5MHz as you can see from schematic. Regards Bilal Bilal

Bilal, The GR noise generators were specified for total noise voltage in a specified bandwidth. With the HP signal generator input limited to 10KHz, your observation of increasing noise as the GR source bandwidth is reduced is as expected. You are simply applying more of the GR's output to the HP signal generator by better matching the source bandwidth to the "load". An oscilloscope is a difficult tool for the measurement you propose. If that's what you must use, then get a storage 'scope or an equivalent digital 'scope which will allow you to integrate up the jittering waveform. This is essentially what folks do to create an "eye" diagram for estimating high speed data link jitter. The difficulty with this approach is that the probability observing maximum jitter is quite low & the 'scope waveform tends to have very low intensity for these "rare" events, so it's tough to know where the outer edges of the jitter are. The digital "scopes are superior for this application, but rather expensive. Regards, Pete Rawson

Pete, Thanks for your valuable information. You have explained exactly what is happening and I have got it finally. I have a digi. oscilloscope with a jitter measurement option, but it is only for square waveforms. Unfortunately the Sig. Gen. I am using has no option for square wave. I think I have to use a comparator to get the square wave form and then measure jitter with oscilloscope. Cheers Bilal ----- Original Message ----- From: "Pete" <peterawson@earthlink.net> To: "Discussion of precise time and frequency measurement" <time-nuts@febo.com> Sent: Tuesday, March 20, 2007 2:42 AM Subject: Re: [time-nuts] Producing jitter with Phase Modulation > Bilal, > > The GR noise generators were specified for total noise voltage in a > specified bandwidth. With the HP signal generator input limited to 10KHz, > your observation of increasing noise as the GR source bandwidth is reduced > is as expected. You are simply applying more of the GR's output to the HP > signal generator by better matching the source bandwidth to the "load". > > An oscilloscope is a difficult tool for the measurement you propose. If > that's what you must use, then get a storage 'scope or an equivalent > digital > 'scope which will allow you to integrate up the jittering waveform. This > is > essentially what folks do to create an "eye" diagram for estimating high > speed data link jitter. The difficulty with this approach is that the > probability observing maximum jitter is quite low & the 'scope waveform > tends to have very low intensity for these "rare" events, so it's tough to > know where the outer edges of the jitter are. The digital "scopes are > superior for this application, but rather expensive. > > Regards, > Pete Rawson > > > _______________________________________________ > time-nuts mailing list > time-nuts@febo.com > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts >