Signal Amplitude drop due to TVS diodes and Impedance matching

[Zainiii]

Hi everyone, I hope you are fine.
I have a question.
I have generated a 4Vpp signal with a 5MHz frequency using the following sequence: AD9851 (DDS) → AD8001 (Op Amp) → Ferrite bead (2KΩ) → Microfluidic sensor
This signal is connected to a microfluidic device whose impedance varies from 3KΩ to 10KΩ.
When I attach any through-hole TVS diode with a Reverse Stand-Off Voltage (V_R) rating of 6V to 30V across the signal, with one pin grounded and the other pin across the signal for circuit protection, it drops the signal to less than 1Vpp.
My questions are:
1.   Why does my signal amplitude drop significantly, even though the Reverse Stand-Off Voltage (V_R) is much higher than the signal amplitude?
2.   I want to match the impedance between my microfluidic device, which has an impedance ranging from 3KΩ to 10KΩ, and my signal, which currently has an impedance of 50Ω without the ferrite bead (2KΩ).
a)     How can I increase the impedance of the signal?
b)      When I add some series resistance across the signal, it greatly reduces the amplitude. Please suggest a solution to match the signal impedance with the microfluidic impedance without dropping the signal amplitude.
C)     does connecting a ferrite bead of 2KΩ in series to the signal add to the overall impedance of the signal?
The AD8001 has high input impedance and low output impedance, making it perfect for this application.
Your response would be highly appreciated.

[tom66]

TVS diodes are built on a process which maximises their ruggedness against brief electrical overstress events.  This process is not optimised for low leakage current because normally there is no need to do so; to build the device to be rugged in such a way requires the junction area to be enlarged compared to a typical device.  If it is the case that your circuit has 2kohm series impedance (do not confuse ferrite bead impedance with resistance though) then it is possible the TVS is distorting the signal excessively.

I would recommend a zener diode instead of a TVS diode.  You can find low leakage current zener diodes (they do not leak much current below Vz).  Be aware all diodes experience significant leakage current changes with temperature, so if your circuit needs to operate over a wider range of temperatures, you may need to consider the change in leakage current there.

At the frequencies you are talking about, 5MHz, unless you have very long cable lengths (>1% of wavelength, i.e. >60cm), matching the impedance should not be necessary.

[fzabkar]

I would recommend a zener diode instead of a TVS diode.

A TVS diode has an extremely low leakage current, typically 1uA at its rated standoff voltage at 85C and 0.2uA at 25C.

https://www.st.com/resource/en/datasheet/smcj12a.pdf

However, the capacitance of these diodes is typically 1nF - 2nF. That would correspond to 15 - 30 ohms at 5MHz.

TVS diodes that are incorporated into ESD suppressors for high-speed data links will have capacitances in the low pF range, typically 2pF. That would be 15Kohms at 5MHz.

[Smokey]

Ya.  I'm guessing the capacitance.  Try a tvs designed for USB. 
Also, being a zener, they start to turn on before the clamp voltage.  So make sure the operating voltage is below where the tvs starts to conduct.
See if there are spice models for the devices you are considering.  That helps.

[T3sl4co1l]

Also cross-posted at: https://electronics.stackexchange.com/questions/720720/signal-amplitude-drop-due-to-tvs-diodes

Tim