I'm building a TDR pulse generator using a single GaN FET (TP65H300G4LSGB) in high-side configuration, driven by NCP51820. The drain is at 40 V, and the source goes through a 50 Ω resistor to GND. Output is taken from the FET source. My Final target is upto 200v.

Pulse specs: ~40 V amplitude, 135 ns width. Test cable: 200 m general wire (not coax), open at the far end. I expected full reflection (Γ = +1), but I only get ~3–5 V reflected.

Scope probing is at the FET source with a 10× probe. The incident pulse is clean (Square), but the reflection is very weak.

Need help with:

1)Why is reflection so weak?
2)Better probing point or output method?
3)Tips for improving reflection visibility?

I think something is wrong while genrating pulse using mosfet that why me getting less Reflection

i have attaced Circuit just check it.

and Thanks

 

hi At,
Welcome to AAC.
Do you know the characteristic impedance of the wire pair?

Why have you chosen 50R?

E
For unshielded twisted pair (UTP) cables, the characteristic impedance is often specified as 100 ohms. However, other common values can be found in the 100-150 ohm range.

 

Thanks for the welcome!


Yes, I measured the wire’s parameters using an LCR meter. For a 200 m general-purpose unshielded wire, I measured:

• Capacitance: ~53.6 nF
• Inductance: ~45 µH

Using Z0=Sq. Root of (L/C), I calculated the characteristic impedance to be approximately 29 Ω. So I initially used a 50 Ω resistor as an approximation for source matching.

Later, I also tested with a 30 Ω resistor, but the reflected pulse actually decreased in amplitude, which confused me — I expected the reflection to increase as matching improved.

 

What is your pulse width? And your wire AWG or copper diameter?

Why would you expect your reflection to be 100% on a lossy transmission line?

 

First: " Test cable: 200 m general wire (not coax)," does not tell us much about the cable as far as losses. And of course, by the time the pulse reaches the other end, the rise time is a bit slower. Next, how did you measure the inductance?? I have rolls of alarm-sensor wire that would be very lossey at audio frequencies. So they could work for a short speaker connection, but quite adequate for a 24 volt switch sensing application. If you can connect your scope to the far and of that 200M cable you will see a slower rise time, and fall time as well.