# Problem of two SAW fitler?

When I series two SAW filter, why the attenuation less than sum both of them in stop band region. Each of them (in stop band) have 60dB attenuation but when series them, the attenuation is equal to 85dB. What happening? Why didn't reach to 120dB?

Then, I put a 20dB attenuation between them ( SG --> SAW1 --> 20dB attenuation --> SAW2--> SA ), and the attenuation is equal to 90dB, this meaning only 5dB attenuate with 20dB attenuation. What happening again?

However, the insertion loss is true when series. Each of the have 8dB insertion loss and when both of them is series the insertion loss equal to 16dB.

• Hi! Welcome here. I honestly think the answer to your question is "measurement errors; can't tell you how to do it right without standing right next to you in your lab". But: I wonder what even benefits from more than 60 dB of suppression; the system that can actually make use of that much isolation would probably also be a system you want to use to measure your isolation. So: what are you doing this for? What benefits from 120 dB of isolation? is something we'd need to know for more useful answers :) But it very well be that you should be asking the underlying question in a new post: May 6 '20 at 10:18
• I want to do {whatever it is that you want to do}, and an isolation of 60 dB as supplied by a single SAW filter {complete type} doesn't suffice due to {why 60 dB isn't enough}. An approach I'm investigating is concatenating filters, but it seems hard to measure the resulting isolation. How can I solve the original problem? instead of why am I not measuring 2×60dB when concatenating filters?, which omits the context and purpose :) May 6 '20 at 10:20

So, probably a combination of different things:

• You physically can't isolate things arbitrarily much. My experience with well-designed RF PCBs is that they have -60 dB to -49 dB crosstalk – simply through the board, shields, grounds, power supplies
• You're not telling us how you're measuring isolation. Measuring more than 90 dB in dynamic range is tricky, and requires you think a lot about signal leakage.

I'd mostly chalk this up to measurement error, because if an attenuator doesn't attenuate according to your measurements, then it's probably your measurements.

First, I measure it by RG316 cable.The test structure following blow:

SG --> RG316 --> SAW1 --> Thru male SMA --> SAW2 --> RG316 --> SA

Signal Generator output power is 0dBm and when applied to SAWs, only 85dB~90dB attenuated. So, change RG316 cable to N Type cable and measure again and achieve to 120 dB attenuation.

It was cable leakage and measurement error.

• This is not a new answer in itself, but only confirms the other answer you've already gotten. May 9 '20 at 14:54