blocking signal cause DC-offset in receiver and 1 minute or more will become normal

[helen.liu]

Hi friends,


I meet a problem as title, could you please help to give me some ideas? :-?:-?
This issue is found in GSM base station radio board, the receiver architecture is two stage LNAs->RF filter->RF mixer->IF saw-> one chip (which include IF amplifier, I/Q modulator, and ADC), the receiver band is 880MHz-915MHz, and the IF frequency is 71MHz. the gain before RF mixer is about 23dBm, and the mixer used MAX9985.


When we do blocking test, the blocking signal is 400kHz offset the receive signal, one case, receive signal is 8PSK, 897.6Mhz, strength is -84dBm, blocking signal is CW, 898MHz or 897.2MHz, strength is -19dBm, then observe the Constellation diagram(this is after baseband processing), the DC point of diagram will remove when first time add blocking or switch blocking from 898MHz to 897.2MHz, or from 897.2 to 898, and the BER is large. After 1 minute or more, the DC point is remove to center, and then test the BER, it shows good.

Actually, I have do some experiments, calculate the strength based on gain, add useful and blocking signal inject after mixer, then found the performance is good. But when I inject signals at mixer input, the situation is same as inject at receiver input port, so I think the problem maybe caused by mixer, but I don't know why and how to solve it. :|:|

Thanks ^^

 

[Baas Rietrot]

Floating DC point? Sounds like a large cap and resistor or something (RC constant).

 

[tony_lth]

Do you have DC bolck cap for each port of Mixer?
Do you add the two bolck signals to Mixer at the same time? If so, there should exist DC current in the mixer.

 

[helen.liu]

Floating DC point? Sounds like a large cap and resistor or something (RC constant).

to Baas Rietrot,

what't the meaning of "Floating DC point? Sounds like a large cap and resistor or something (RC constant). " could you please to give some clearly explaination? thanks very much, ^^.

- - - Updated - - -

to Tony_lth

It have DC block cap for each port of mixer, and only add one blocking signal, +400kHz or -400kHz offset the useful signal. so it very strange for DC point remove, and about 1 minut the DC point will remove to centre, thanks ^^

 

[helen.liu]

:smile::smile:
The root cause is found, the reason is because LO has a spur at 400kHz offset, and the blocking signal is so strong, then is mix with spur and produce a 71MHz CW, it is cause the DC-offset, and has this phenomena.