Intro

    The isee3 satellite has 2 transmitters

• 2217.5 Mhz  (transponder B)
• 2270.4Mhz (transponder A)

• measured the received tones
• The frequency resolution used was about 1.2 hz.
  
• doppler correct it using the velocity from the ephemeris
  
• compute  freqMeasured - (tx0Freq  Doppler corrected  using  ephemeris)

• Let FreqErr=Measured Frquency - doppler corrected freq (using ephemeris)
  
• top: transponder B 2217.5 freqErr
• 2nd: transponder A 2270.4  freqErr.
  
• 3rd: FreqerrA/FreqErrB
  
• Bottom: Shows the difference in the satellite transmit oscillator errors.
  

• 22may14: jump in 2217.5 tone after first file
• the pointing was updated after the first file. The signal went from 7db to 50 db above the noise floor. The1st freq is probably off.
  
• 28may14: there is a drift in the 2217.5 and the 2270.4 tone starting with the 3 file
• transponder A drifted by 60 hz in 1.4 hours
• Transponder B drifted by 6 hz
• 06jun14: no change in frequency when we commanded coherent mode.
  

What can the frequency measurements tell us:

Terminology:

• name	equation	Notes
  F0a	constant	2270.4e6 hz . the nominal frequency
  F0b	constant	2217.5e6 hz
  Da(t)	?	drift in oscillator A
  Db(t)	?	drift in oscillator B
  Ftxa	F0a + Da	transmitted freq A
  Ftxb	F0b + Db	transmitted freq B
  Veph	ephmfile	velocity from ephemeris file
  Verr	?	error in ephem vlel
  V	Veph + Verr	correct velocity. (neg is approaching)
  Frcva	Ftxa(1-V/C)	received frequency
  FrcvB	Ftxb(1-V/C)
  FErcva	FrcvA-F0A(1-Veph/C)	remove doppler corrected freq (using ephemeris) from received freq.This is  what was plotted  above.
  FErcvb	FrcvB-F0b(1-Veph/C)

  
  

• FErcvA = FrcvA- F0A(1-Veph/C)
• =(F0a + Da)(1-V/C) - F0a + F0a*Veph/C
• =F0a  - F0a(Veph + Verr)/C + Da - Da(Veph +Verr)/C  - F0a + F0a*Veph/C
• = -F0a*Verr/C  + Da (1 -(Veph+Verr)/C) 
• FErcvA - Da(1-(Veph+Verr)/C) = -F0a*Verr/C
• Using transponder B gives:
• FErcvB - Db(1-(Veph+Verr)/C) =  -F0b*Verr/C
• dividing the two equations gives FEratio
  
• K= F0A/F0B=1.02386=(FErcvA - Da(1-(Veph+Verr)/C))/
•                                   (FErcvB -Db(1-(Veph+Verr)/C))
• FErcvA is about -9800 Hz
• FErcvB is about  -7000 Hz
• The velocity (from the ephemeris) is  -2 to -4 km/sec
• so (Veph + Verr)/C ~ 1e-5  can be ignored relative to 1..
• K=1.02386= (FErcvA - Da)/(FErcvB - Db)
• Putting the oscillator errors on one side give:
• K*FErcvB  -FEa = (K*Db - Da) 
• plotting this shows the difference in the oscillator errors.
  
• fitting for the Da,Db oscillator errors
• 0. =(FErcvA(t) - Da(t))/(FErcvB(t) - Db(t)  - 1.02386
  
• if the oscillator changes slowly we might fit for  Da(t),Db(t)
  

 home_~phil