There was some concern that the protection could impact signal integrity at high speeds. So I did some tests with a spectrum analyzer with tracking generator.
- small scrap of protoboard
- copper foil on the back as ground plane
- side mount SMA connectors
- No diodes or other components yet
- just straight through connection between the SMAs
- compared to a proper SMA through
Results:
- not too shabby up to 1 GHz
- but already a wobble of about 1 dB down just because of the non-optimal transmission line over about 3 cm
I used this as reference for all further measurements, so this result is set as 0 dB there.
Because of a hunch I tested just adding the diode to ground first, no protection fets yet.
As I already guessed this makes a shitty rf diode. Already 3 dB down at 556 MHz. Note that I had to rescale the amplitude to make it fit the screen.
These are my first go-to diodes if I want to replace BAT54 without the downsides of Schottky. RF performance much better.
But I don't know if the higher forward Voltage of silicon diodes isn't a problem when protecting against negative voltages. The internal protection diodes of the level shifters could begin to conduct at about the same voltage level.
TODO: test negative voltage performance with BAV99 and 74LVC1G45.
Diode, 2 depletion FETs and 10R, like in the original schematic. Without the TL431 as it won't impact rf performance (too low peak voltage to activate it).
- at DC you see the 36 ohms
- with higher freq the attenuation slowly drops, capacitive coupling starts to become a thing
- slight wobbles with 1.37 dB max
- same order of magnitude as just the degradadtion due to the test board
I would consider this a success, the effects are minimal and about the same as due to wiring or signal routing.
The DUT is shown with the protection circuit installed. I reworked it during the progress of the test.