GFSK passband modulation in simulink

I'm trying to implement (using Simulink) a communication model based on a GFSK modulation with a source bit rate @2.4kbps.

To generate the GFSK I used a CPM block with gaussian filter (the spectrum is as I was expecting).

Now I need to bring the signal from baseband to passband @169Mhz as central frequency.

I tried to multiply the modulated signal by a sine wave with carrier frequency equal to 169Mhz, but analyzing the output with the Spectrum Analyzer the signal does not go further than 900Khz. I mean, if I generate a sine wave with 0<f<900KHz the signal is correctly shifted, but if I increase the carrier frequency than the output signal fluctuates in the range 0-900Khz.

Here are some parameters of the blocks:
Bernoully Binary Generator
-sample time:1/2400
-sample per frame: 1.

CPM:
-modulation index: 2
-sample per frame:100

SineWave:
-carrier frequency: 169e6
-sample time: (1/2400)/(100)
-sample per frame: 100

Is there any error in what I do?

Thanks,
Giuseppe

Dont implement passband simulation of such a high frequency.
I think it is a solver problem (menu Simulation - Configuration parameters). The internal default matlab sampling frequency.
I placed in an empty model sine wave generator and scope. Set your frequency and set the configuration parameter for simulation:
Solver - Solver Option - Type - "Fixed step". And the step value that corresponds to such a high frequency - 10^9.
The scope shows the sine wave. But in such simulation you will meet another problem - it is very-very slow and after some millisecond there will not be enough memory to store the logged data.

Mixing your signal with a carrier is just a shift, it influence nothing and if you succeed in baseband simulation, the same will work on passband.
You really dont need that passband.
But if you wish - make some experiments with the solver options.

Thanks for your reply, Mityan.
You are right when you say that is not useful working at passband, but I need to implement a system with more sources around that frequency, and just for coherence I would like to move there.

I tried your solution, but it didn't work. In order to reach 169MHz seems that a lower sample time for the sine wave is needed. The problem is that doing such operation, the bandwidth of the resulting signal (baseband signal multiplied by the sine) increases (from KHz to MHz).
I just want to shift the signal in frequency without increasing its bandwidth.
Here a snapshot of my current simulink model, maybe useful to understand where I'm wrong.



Thanks again,
Giuseppe

any suggestion?