Question regarding FET with low Drain to Source capacitance - FET_termination.jpg (0/1)

A 2N7000/7002 has about 2 ohms on resistance at +5 on the gate, and 11 pF typical drain capacitance. They cost a few cents each.

John

As an alternative, you could try using three diodes in series fed by a Rf choke and a 100 ohm resistor., that should let any legal video signal through, then when you apply 5 volts that should make for a pretty low value shunt to ground.

What about one of these bus switch ICs which contain 4 MOSFETs. Other manufacturers make similar things too. FST3125M

Chris

What about the logic level FDV301N (6pF)

or perhaps just a plain BC847 instead of the FET (add base resistor)

Regards

Klaus

I can't picture it in my mind. Can you sketch a schematic? I am very interested.

Thanks for the info

Caesar

FET off and the output capacitance value for the suggested Fets [Coss] will be at a maximum as there is no drain to source voltage. Best so far is the

So John, you're saying that if I can bais the FET' Drain to Source with the 5V while it is OFF, this will help to lower the Drain to Source Capacitance.

Caesar

Yes. (but ........ :) As arrangement seemed so intriguingly "simple", thought I'd run the spice and oh-so casually figure out a working arrangement. A whole 3 hours! later, figured the FET idea is a time waster. Now been driven home to me the significance of the Ciss,Coss,Crss value interactions under this linear-switching condition. Horrible little things are constantly on the move, resulting in significant DC shifts, signal loss variations and harmonic distortion. Tease apart one complicated interaction and another strolls in to fill it's place. Tried various bias arrangements but at 20MHz, the FET harmonic distortion (5% at best), the loss of output voltage (-10% at best) and variable DC offset shifting with signal level suggested I was on a loser.

Eventually just fitted a normal NPN small signal transistor like Kraus K suggested (I used a BC547 and 1k base R)). Magic ... . Distortion mostly 3rd at the 1% level. No mysterious DC shifts and NO signal loss. (confirmed on a spectrum analyser and tracking gen).

I enjoyed that, Ta!. john

Klaus suggested BC847. Is this the same as BC547?. Also, is there a negative bais voltage on the emitter of NPN or is the emitter connected to ground? So basically the circuit is

COAX 75ohm Line video source------------------------------+---->video receiver | | | R=75 | | C TTL-------R=1k---B BC547 E | | -V=????

By the way, I notice that when I connected a -5v to the source of the MOSFET, I did see a DC variation. I had to connect the source of the FET to GND in order to stop the variation. It is interest that in your analysis that you commented that the MOSFET arrangement will have "...significant DC shifts". Maybe the DC shifts are still in the MOSFET arrangement with the source connected to GND, but I don't see the shift under the GND connection condition.

Caesar

BC847 is just the SMD version of the leaded BC547 (only major difference is the thermal resistance of the junction-case)

Regards

Klaus

(As Klaus' reply)

The DC shifts were occuring due to the FET generating even harmonic distortions due to assymetrical shifts in it's operating point at the rate of the incoming frequency. This distortion in turn was created via the leakage paths through the FET internals. Which in turn also depended on series impedance in the gate driver circuit (drop the gate impedance and then lose out on drain signal level). But, it gets messier, as the FET starts to 'float' on it's self created DC levels which in turn alters the operating point again. This now shifts the internal cap values and a new cycle of leakage-distortion-operating-level change ensues. Static level shift also happened every time the video inputs level changed. DC shift may be only a 100mV or so but this is noticable on a video screen before the monitor AGC has time to kick in At 20MHz a -lot- of precious signal current is running through the FET parasitics to ground. Not a good situation. john

Have you got room for a SIL reed relay?

eg, The Meder Electronic Micro SIL relay has a footprint of about 15.2x3.9mm, about 2pF, coil to contact, and takes less than 20mA from a 5v supply. Switching speed is about 0.5mS. Their SIL05-xxx is slightly larger, but takes only 10mA at 5V.

-3dB down from what? Are you making an amplitude or impedance measurement?

That's not right- the ZVN4106F should be something like 20pF.

The ZVN4106F is one of the lowest available.

The problem is most likely that the video is causing some amount of conduction through the FET body diode on the negative excursions.

That's not right, unless you bias the FET to Vds = 25V. Like all MOSFETs, the capacitance increases severely below about 10V. The datasheet curves show Coss climbing rapidly, nearing 100pF at 0V.

It's in the same territory as the 2n7000 and 2n7002.

LRCR can do much better with MOSFET switches contained in cmos ICs, or with some bipolar photoMOS optocoupler relay switches, such as Aromat's AQY221, with 1pF of capacitance at 0 volts. LRCR can use a T configuration for optimal off attenuation.

Tony,

Thanks for your response. Using a reed relay is not a bad idea. I have not consider this as an option. I will have to see if I have room on the PCB I am working on.

Thanks Caesar

The blasted thing...