Hello There
I try to use the spread spectrum function of the i.MX7 to modify the frequency of the VIDEO_PLL.
First of all:
Q1: Are the formulas noted in the i.MX7 Reference-Manual correct?
My measurements on the output-clock have shown, that changes on the parameters (stop, step, B) behave more like the formulas noted in I.mx6 app. manual:
1) Modulation frequency = Fref * STEP/2*STOP
2) Max frequency change = STOP/DENOM * Fref
Q2: Which formulas should I take for my calculations i.MX7 or i.MX6?
Further, one can see that values with the same results for equation 1) and 2) behave different when measured:
Let’s say one want to spread the clock 1% of its original value.
The clock is set to 650 MHz -> fRange = 6.5 MHz and fmod = 500kHz.
So, when for example the parameters were set up like:
Setup | DEN | STEP | STOP | fmod | fRange |
A | 12.000E+3 | 136 | 3250 | 502.2E+3 | 6.500E+6 |
B | 15.000E+3 | 170 | 4063 | 502.1E+3 | 6.500E+6 |
One can see that the center frequency of the measured clock-output is not the same as defined without frequency spreading.
And even worse the center frequency of parameter setup A differs from setup B.
Q3: Are there any constrains on how the parameters can be set?
Q4: Why does the center frequency of the modulated output-clock changes?
Best regards Marc
A1. The formulas, as given in the i.MX7 Reference Manual document, are incorrect, there is a typo.
A2. The formulas should be read as given in the i.MX6 Reference Manual document, i.e. Frequency change range = 24MHz*stop/CCM_ANALOG_PLL_SYS_DENOM[B] and Frequency change step = 24MHz*step/CCM_ANALOG_PLL_SYS_DENOM[B].
A3. No, there are no any constrains of setting up the spread spectrum parameters except of generally reasonable ones, e.g. the frequency change range should not be larger than the PLL output frequency, the frequency change step should not be larger than the frequency change range etc.
A4. The initial PLL output frequency does not become a center frequency when enabling the spread spectrum modulation, it becomes the upper frequency limit, the lower frequency limit is initial PLL output frequency minus frequency change range.
Have a great day,
Artur
-------------------------------------------------------------------------------
Note:
- If this post answers your question, please click the "Mark Correct" button. Thank you!
- We are following threads for 7 weeks after the last post, later replies are ignored
Please open a new thread and refer to the closed one, if you have a related question at a later point in time.
-------------------------------------------------------------------------------
Hello Artur
Thank you for your answers.
A1, A2:
Are the formulas you've provided not the same as the formulas in the i.mx7 Ref. Manual? (there are no brackets)
A4:
My measurements have shown that the center frequency changes with changes in the range, and also my measurements do not show a spectrum which is downspreaded. I know how all of this behaves with the i.MX6 and there I would be agree with you, but not with the i.MX7. Do you have any further explanations for this strange behavior?
Q1:
Can you explain me @ which point the VIDEO_PLL gets the frequency modulation? Is it really the frequency from 650-1300 MHz or is the clock spreaded on an earlier stage or frequency?
I think there is no documentation for this.
Greets Marc
Let me get more details about i.MX7 PLL spread spectrum operation from the apps team.
Regarding the spread spectrum signal itself, it is generated within the PLL, so, the PLL output becomes already spread spectrum.