What is output of quartz crystal across MCU

Thread Starter

Vindhyachal Takniki

Joined Nov 3, 2014
586
1.What is output type of quartz crystal we connect across microcontroller? Is it sine, square, clipped sine?
2. I have 16Mhz crystal connected across(MCU) with two capacitors to GND from each leg. MY DSO is 60Mhz & probe is at 10x
3. I tried probing across one leg of probe at Crystal In & another at cyrstal output. It seems sine to me? Or is due to my DSO is less freq? I think DSO should be atleast 5 times for square & 10 times for sine wave?
4. Three probe I tried:
a) One probe at XTIN & another at XTout
b) One probe at XTIN & another at GND
c) One probe at XTOUT & another at GND
 

Ian0

Joined Aug 7, 2020
3,524
You'll generally find a sinewave (probably a bit distorted), larger amplitude (and more distorted) on XTOUT than on XTIN, centred on half the supply voltage.
 

MrChips

Joined Oct 2, 2009
24,457
The shape, amplitude and frequency will be affected by the probe attached to the legs of the crystal.

The shape is not important. If you need to measure the frequency you need to do so downstream, i.e. by measuring an output from the MCU.
 

Ian0

Joined Aug 7, 2020
3,524
The change in frequency caused by attaching a x10 probe is probably less than the accuracy to which a scope can measure it.
In other words, if you see a signal and the scope says it's 15.9something MHz, then you've probably got a functioning 16MHz crystal!
 

Papabravo

Joined Feb 24, 2006
17,029
1.What is output type of quartz crystal we connect across microcontroller? Is it sine, square, clipped sine?
2. I have 16Mhz crystal connected across(MCU) with two capacitors to GND from each leg. MY DSO is 60Mhz & probe is at 10x
3. I tried probing across one leg of probe at Crystal In & another at cyrstal output. It seems sine to me? Or is due to my DSO is less freq? I think DSO should be atleast 5 times for square & 10 times for sine wave?
4. Three probe I tried:
a) One probe at XTIN & another at XTout
b) One probe at XTIN & another at GND
c) One probe at XTOUT & another at GND
This is a picky point; but the crystal is a PASSIVE device. As such it doesn't have an input or an output. You can insert a crystal in the circuit either way. The two pins on the processor however consist of an input and an output for an inverting amplifier. Consult the datasheet for clues as to which is which. Hint: an external oscillator will connect to the input and the output will be left open.
 

Papabravo

Joined Feb 24, 2006
17,029
I think Vindhyachal Takniki might have been referring to the labelling of the pins on the microcontroller.
He said: What is output type of quartz crystal we connect across microcontroller?
On it's face this statement suggests that at least one crystal pin has an "output type".
Don't you think he can defend himself if that is the case? Why does he need you to do that for him?
 

MisterBill2

Joined Jan 23, 2018
9,529
He said: What is output type of quartz crystal we connect across microcontroller?
On it's face this statement suggests that at least one crystal pin has an "output type".
Don't you think he can defend himself if that is the case? Why does he need you to do that for him?
In most circuits where a crystal has direct connections like that, with two capacitors to common, one pin of the IC is driving the crystal and the other is the input from the crystal. Probing the drive pin will probably affect the frequency less, but certainly any change in the capacitance to either crystal lead will affect the frequency some. Usually that only natters where the crystal frequency is used for timing or frequency generation.
 

Papabravo

Joined Feb 24, 2006
17,029
In most circuits where a crystal has direct connections like that, with two capacitors to common, one pin of the IC is driving the crystal and the other is the input from the crystal. Probing the drive pin will probably affect the frequency less, but certainly any change in the capacitance to either crystal lead will affect the frequency some. Usually that only natters where the crystal frequency is used for timing or frequency generation.
Absolutely no argument at all on that point. It applies not only to crystals, but any passive component(s) connected to a pair of pins where one is a low impedance source and the other is high impedance. It might be instructive to look at a simulation for a crystal where they use "motional" parameters to model the crystal as a network of L's 'n C's and R's.
 

Kjeldgaard

Joined Apr 7, 2016
456
When measuring the input of the CPU oscillator, in some cases it may be necessary to switch to the AC coupling on the oscilloscope, so as not to pull the input towards ground with the 10 MΩ resistance of the probe.
 

MisterBill2

Joined Jan 23, 2018
9,529
The AC/DC coupling switch doesn‘t alter the scope input impedance. It‘s always 1MΩ.
The switch DOES change the DC resistance by inserting a capacitor. That is easy to verify, and in every instance I have switched to AC coupling the DC component is no longer passed through the probe. OF COURSE, I did not exceed the probe voltage rating.
 

Ian0

Joined Aug 7, 2020
3,524
Interesting. Is yours a digital or analogue scope?
This one is an ISO-Tech and Ac/Dc coupling is switched from a menu, and it measures 0.999MΩ on a Fluke meter regardless of whether It is set to AC or DC.
I’ll check my Tektronix (also digital) when I get home.
 

Kjeldgaard

Joined Apr 7, 2016
456
Then I learned something today about oscilloscopes.

An analog Kikusui COS6100M has the capacitor right after the input socket in AC mode, and a digital Tektronix TDS2014B always has 1 MΩ measured to ground on the input socket regardless of AC/DC mode.

But I still think a DC coupled probe can suffocate some low power oscillators.
 

Ian0

Joined Aug 7, 2020
3,524
But I still think a DC coupled probe can suffocate some low power oscillators.
Maybe. Depends whether it is the 1MΩ load that kills it, or the DC bias shift.
The 1MΩ load will still be there, but coupled through a very large capacitor which looks like a short circuit at 16MHz.
 

Ian0

Joined Aug 7, 2020
3,524
.An analog Kikusui COS6100M has the capacitor right after the input socket in AC mode, and a digital Tektronix TDS2014B always has 1 MΩ measured to ground on the input socket regardless of AC/DC mode.
Does that mean that there really is no such thing as AC coupling on a digital scope? Does the software just compute the average level and subtract it from the samples?
 

Thread Starter

Vindhyachal Takniki

Joined Nov 3, 2014
586
So if we dont connect any probe, since it effects the meausrement and distort the signal by loading.
What exactly is shape of signal at XTIN & XTOUT of MCU when a crystal is connected acrss it with each pin having capacitor to GND?

Is it sine, square, distorted sine?
 

MrChips

Joined Oct 2, 2009
24,457
So if we dont connect any probe, since it effects the meausrement and distort the signal by loading.
What exactly is shape of signal at XTIN & XTOUT of MCU when a crystal is connected acrss it with each pin having capacitor to GND?

Is it sine, square, distorted sine?
Why do you care?
 
Top