What do these parts do?

[alank2]

In this circuit (LPRO-101 sine wave to square wave), I have two questions:

What does the 43.2 ohm resistor at the top do?

What does the 33.2 ohm resistor at the output do?

Thanks!

[Simon]

no much, the one one the supply would create a filter with the by pass capacitor and do very little. The output resistor is probably to help protect it. Omitting both probably won't be a problem.

[TimFox]

The 33.2 ohm resistor is probably there to make a "series termination" when driving a 50-ohm coax cable to the load (of indeterminate impedance).  Any reflection from the load will be (partially) absorbed in the 33.2 ohm resistor plus (ill-defined) output impedance of the CMOS device.  This reduces the ringing in the cable.

[Gyro]

A typical example of someone taking the output of a simulator and not correcting for real world values, nothing in the circuit is THAT critical.

The input resistor (49.9R) can be 47 or 51. The 43.2R filter resistor on the 5V supply can be 47, 51 or whatever (probably anything up to 100R would be ok). The 33.2R series terminator on the output can be 33R.

Edit: Actually, as the circuit doesn't specify what 'cmos logic' family the output series terminator is a bit academic. Ideally it should be 50R minus the output resistance of the logic device. 33R is a reasonable ballpark value. Also, as the circuit is to convert sine to square wave, the inverter really ought to have a schmitt trigger input.

[dom0]

Also, as the circuit is to convert sine to square wave, the inverter really ought to have a schmitt trigger input.

Absolutely not. This circuit uses the gate in it's semi-linear region around Vdd/2 where transconductance is very large (for a MOSFET circuit).

[alank2]

The input to the above circuit is coming right off a LPRO-101 connector plugged into the same pcb. There is no RF cable.  Given that, do I still need the 49.9 (or 51) to ground?  I ask because that reduces the amplitude about half as expected, but with the IC's I tried, it drives them better when it is stronger.  I had a few IC's to try, perhaps I tried ones that were not ideal : SN74AC14N worked when I stopped the 51 ohm to ground termination so the amplitude was large enough (3.5V).  I tried this IC along with 74HC02N, SN74HCT125N, SN74AC04N with the termination and it didn't seem to be enough amplitude (1.75V), maybe I was doing something wrong.

The output will go to a couple of devices, one is the CLKIN for an AVR, does it need the 33 ohm output resistor?

Absolutely not. This circuit uses the gate in it's semi-linear region around Vdd/2 where transconductance is very large (for a MOSFET circuit).

Why do you feel this way?  What will the Schmitt hurt?

[dom0]

Basic Schmitt triggers like a 74HC14 will degrade phase noise. This is why you will only find circuits like the one shown above operating in semi-linear mode in the app notes of crystal osc. manufacturers.


The parameters of the different CMOS logic families differ quite a bit. There are also app notes about linear usage of logic gates by the manufacturers where such things are detailed. 74HC, however, will be totally fine for 10 MHz.

[Dave]

no much, the one one the supply would create a filter with the by pass capacitor and do very little.

I think not.

This circuit uses the gate in it's semi-linear region around Vdd/2 where transconductance is very large (for a MOSFET circuit).

If the oscillator input is left floating, the biasing resistors would hold the logic gate right where the transistors conduct the most, possibly shorting out the supply and destroying the inverter. A resistor in series with the power supply would limit the current and prevent that.