# Series vs Parallel termination load for transmission lines

I have a source with 5V 10Mhz Squarewave output, then I have an OR logic gate then 20m coaxial cable and then I have another OR logic gate. I guess I will need a matched load (now it is 100Ohm (random resistance)) for this system. Should I add the termination load in series or using voltage divider?

I must note - I do not need any numbers, I am just curious about pros and cons in the two following systems.

Here is the updated picture. Coaxial cable: 20m, 50Ohm Voltage Source: 5V, 10Mhz, Square Wave

• For what it's worth, I don't think the transmission line simulation in the Falstad circuit simulator is really accurate to any great extent. The falstad circuit-sim is great for playing with ideas for circuits, but when you get to edge-cases, or complicated interdependencies, it makes a lot of simplifications of how the circuits actually behave, which makes it a poor assessor of real-life behaviour. – Connor Wolf May 22 '12 at 0:33

I would series terminate the line assuming the receiving end is a single receiver, if its hitting multiple devices you basically have to end terminate.

Your source resistor should be output_impedance + R = Transmission_line_impedance. You may need to fiddle with this as your going to have to compensate for impedance changes caused by the connectors, etc. Side note, are you sure your logic gate can drive this load? Most can't, you may need a line driver on the sending side.

As for your pictorial end termination, its not correct. That resistor should be as close to the receiver as possible, you should also consider Thevenin termination as it uses much less power than the parallel termination.

In general (this can vary a bit by application):

Source Termination: Lowest power, slowest rise/fall times, single destination only.

Parallel Termination: Highest power, only 1 resistor

Thevenin Termination: Middle power, 2 resistors

Parallel AC Termination: Middle-low power, fastest rise/fall times, 1 resistor, 1 capacitor

For most signals what i would say is this:

If its point to point and rise/fall times are not required to be especially fast (most point to point applications): source terminate

If its point to multiple point: Thevenin terminate AFTER the last device.

If rise/fall times are absolutely critical (this is fairly rarely needed): Parallel AC terminate.

• Thank you. It seems I needed to know is a bit of terminology. – Arturs Vancans May 22 '12 at 8:49
• I used source and series interchangeably since there is no series termination possible at the receiving side, sorry if that is a bit confusing, many may (correctly) call it series termination. – Mark May 22 '12 at 20:54

As far as I understand you have a low-impedance source driving a cable with a high-impedance reciever at the far end, correct?

AFAIK the propper way to get the same waveform at the receiving as you are sending is to match the impedance at both sides, so put a series resistor at the sending side, and a parallel (load) resistor at the receiving side, both equal to the characteristic impedance of your coax cable. This will of course attenuate your signal, but (ideally) it won't affect the shape.

If the form (shape) of the wave is not the issue you might get by with only the series resistor at the sending side.

• If shapes are not an issue, he doesn't need resistors at any side. – Telaclavo May 22 '12 at 2:02
• Series resistor does not give a bad shape, it is by far my favorite form of termination but has situations where it cannot be implemented. – Kortuk May 22 '12 at 4:14
• @Kortuk: will a series resistor at the receiving side do anything when the receiver is high-impedance?? – Wouter van Ooijen May 22 '12 at 6:27
• @WoutervanOoijen, no, when I talk about series termination it is installed at the source side. – Kortuk May 22 '12 at 6:27
• I've never seen anything with series and parallel resistor, what would be the goal of this? – Mark May 22 '12 at 20:57