XTAL Oscillator

Groundloop · 2012.04.20, 16:48:06

Hi,

Can anybody please help me with this oscillator?

The oscillator runs but the output after the low pass filter is only 100mV.
The low pass filter has a starting cutoff point of approx. 81KHz.

I need at least 5 volt peak to peak at the output.

GL.

poynt99 · 2012.04.20, 17:10:13

You could try a switched capacitor filter instead of the multi-pole passive low-pass. You could also take the 100mV and run it up through an amplifier.

If what you want is a crystal controlled, or tightly controlled sinusoidal source, I would suggest a DDS chip.

Groundloop · 2012.04.20, 17:26:23

.99,

Thanks, I will take a look at the link.

GL.

poynt99 · 2012.04.20, 18:30:19

I tried the switched capacitor trick about 25 years ago. I think I still have the pcb. It may have been a MF10 chip though.

Anyway, it did a nice job of filtering out all but the fundamental of the square wave on its input. The output level didn't drop much if I recall correctly.

In addition to the LMF100 linked above, this LTC1063 also looks promising (5th order).

Groundloop · 2012.04.20, 19:23:50

.99,

Thanks, the LTC1063 looks good. But that will be on the next project. :-)

I got the oscillator up to speed now. I did forget to put a series resistor
inline with the xtal. Now I get approx. 5 volt ptp out of the low pass filter.
And it looks like a nice sinus.

Thanks,
GL.

Groundloop · 2012.04.26, 10:42:36

Hi,

I still have problems with my oscillator. The oscillator starts up but is running
on two different frequencies at the same time. One is the xtal frequency of 77.5 KHz.
The other is some low frequency (around 5 Hz). The output looks like a 5 Hz amplitude
modulated (with 77.5KHz) signal.

Anybody know how to get rid of the low frequency oscillation?

GL.

Peterae · 2012.04.26, 10:52:40

maybe a coupling capacitor will knock the low Hz out if you can tolerate loosing the dc coupling.

Groundloop · 2012.04.26, 18:16:17

Quote from: Peterae on 2012.04.26, 10:52:40
maybe a coupling capacitor will knock the low Hz out if you can tolerate loosing the dc coupling.

P,

Thanks. I will try that solution. But the real question is why this happens?

I have built many xtal oscillators over the years, but this is the first time
I have seen such a double frequency oscillations. I also noticed that
the oscillation was sinus but that is maybe because this inverter (4069)
is not an Schmitt Trigger? Any idea on what capacitor value that will pass
the 77.5 KHz but not the 5 Hz?

GL.

Peterae · 2012.04.26, 19:27:01

As long as your power rail is decoupled i would think it the internals to the chip, there's not much else really from what i see, you could try switching to a 74HC and see if that cures things

If this calculator is right then a 2.2uF would have an impedance of 0.93 Ohm at 77.5 kHz and 14.469 KOhm at 5Hz
Just google capacitive impedance calculator if you want to play, it can also depend what you are driving as well.

Groundloop · 2012.04.26, 19:58:47

Quote from: Peterae on 2012.04.26, 19:27:01
As long as your power rail is decoupled i would think it the internals to the chip, there's not much else really from what i see, you could try switching to a 74HC and see if that cures things

If this calculator is right then a 2.2uF would have an impedance of 0.93 Ohm at 77.5 kHz and 14.469 KOhm at 5Hz
Just google capacitive impedance calculator if you want to play, it can also depend what you are driving as well.

Peter,

Thanks. I will try the capacitor in my circuit. I can not use 74HC because my power rail is 12 Volt.
But the HEF40106BP is pin compatible with the CD4069 that I'm using now, so I can try that IC
also. The largest capacitor I can fit onto my PCB is a 100nF SMD size 0603 so I did the math
for that one.

C = 100nF
Impedance (Z):
0.02053612903225806 Ohm at 77.5 KHz
318309.99999999994 Ohm at 5 Hz.

So I get 318K Ohm at 77.5 KHz. That is not bad at all.

The output of the oscillator is buffered to one inverter and that inverter is feed to a low pass filter.
The low pass filter starts to cut at approximately 81 KHz.

Thanks for the help, Peter.

GL.

EMdevices · 2012.04.26, 20:00:20

Quote.... But the real question is why this happens?

Groundloop, are you sure that the low frequency is not 7.3 Hz? LOL

Very interesting.

EM

Peterae · 2012.04.26, 20:41:00

Np Groundloop, worth a try anyway. The value sounds good

Groundloop · 2012.04.26, 22:25:58

Quote from: EMdevices on 2012.04.26, 20:00:20
Groundloop, are you sure that the low frequency is not 7.3 Hz? LOL

Very interesting.

EM

EM,

:-) I just eyeballed it on the scope so it could be 7.3 Hz. :-)

GL.

Groundloop · 2012.04.27, 07:28:03

All,

This oscillator is really starting to piss me off.

There is no way I can get this oscillator to work as it should.

The low frequency sinus is going right through my low pass
filter. The low pass filter have two capacitor is series but
the low frequency sinus has the same amplitude as the
high frequency sinus. No damping at all!

I think that there must be some issue with the crystal I'm using.

GL.

Peterae · 2012.04.27, 09:49:24

Sounds very strange, what's the impedance of the item it feeding.

Groundloop · 2012.04.27, 09:59:12

Quote from: Peterae on 2012.04.27, 09:49:24
Sounds very strange, what's the impedance of the item it feeding.

Peter,

The oscillator is at inverter 1 of the 4069 IC. The output of the oscillator
goes to inverter 2 of the same IC. The output of the inverter 2 goes
via a capacitor to the low pass filter. So the first inverter output is feeding
a impedance in the very high Ohm range.

I have tested the circuit on a 12 volt battery also but the oscillation is
still approx. 5 Hz and also 77.5 KHz. The 77.5 KHz is amplitude modulated
onto the 5 Hz so the 5 Hz also goes straight through the capacitors and
low pass filter.

So it is impossible to filter out the 5 Hz with a capacitor.

GL.

Peterae · 2012.04.27, 11:39:59

Yes the load impedance is important, your capacitor is in effect acting as a 318K at 5Hz so you will only kill it if the load is smaller, so if your load is 318K also then you would expect the 5Hz component to be half the amplitude it was previously.

I think you are going to end up with an op amp oscillator circuit.
The only other way i can think of is to make a trap for the 5Hz, maybe a tuned circuit to ground that is set for 5Hz, i am not so good on theory of this horrible analogue stuff

, which seems you are in the same boat, we are both digital people

Groundloop · 2012.04.27, 12:50:11

Quote from: Peterae on 2012.04.27, 11:39:59
Yes the load impedance is important, your capacitor is in effect acting as a 318K at 5Hz so you will only kill it if the load is smaller, so if your load is 318K also then you would expect the 5Hz component to be half the amplitude it was previously.

I think you are going to end up with an op amp oscillator circuit.
The only other way i can think of is to make a trap for the 5Hz, maybe a tuned circuit to ground that is set for 5Hz, i am not so good on theory of this horrible analogue stuff , which seems you are in the same boat, we are both digital people

P,

I have already designed a new separate oscillator, but I wanted so much to know
what is the reason why this digital oscillator did what it did. It looks like the xtal
is oscillating on 5 Hz with a 77.5KHz on top of that. I have tried to draw it.

GL.

Peterae · 2012.04.27, 13:36:07

Wow man that 5Hz is way bigger in amplitude than your 77.5kHz i thought it was the other way around $:-\$
Are you sure it's 5Hz and 77.5kHz because if it was you would not be seeing both waveforms clearly, because there would be 15500 full wave cycles of 77.5kHz in one cycle of 5Hz

Groundloop · 2012.04.27, 14:31:11

Quote from: Peterae on 2012.04.27, 13:36:07
Wow man that 5Hz is way bigger in amplitude than your 77.5kHz i thought it was the other way around $:-\$
Are you sure it's 5Hz and 77.5kHz because if it was you would not be seeing both waveforms clearly, because there would be 15500 full wave cycles of 77.5kHz in one cycle of 5Hz

Peter,

It is very hard to explain. The two sinus frequencies are the same in amplitude on my o-scope.
I will be off my lab until middle of next week, but I will bring my camera to my lab and take some pictures
of the o-scope for you then.

Have a nice weekend.

GL.

EMdevices · 2012.04.28, 16:07:57

Than it sounds like multiple overtones present, which produce the illusion of a 5 hz frequency, so you need L C filtering to keep the correct frequency and discard the rest. Or most likely you have the wrong crystal and you are driving it too hard and it creates distortions.
EM

PS I was in the lab a month ago and the guys were trying to fix this very same problem on an old UHF radio. The procedure was quite involved, step by step adjustment of capacitors in different tank circuits. That's the problem with overtone circuits. Fundamental xttals are easier some what.

Groundloop · 2012.04.28, 17:55:50

Quote from: EMdevices on 2012.04.28, 16:07:57
Than it sounds like multiple overtones present, which produce the illusion of a 5 hz frequency, so you need L C filtering to keep the correct frequency and discard the rest. Or most likely you have the wrong crystal and you are driving it too hard and it creates distortions.
EM

PS I was in the lab a month ago and the guys were trying to fix this very same problem on an old UHF radio. The procedure was quite involved, step by step adjustment of capacitors in different tank circuits. That's the problem with overtone circuits. Fundamental xttals are easier some what.

EM,

I think you are right about the above. I will try out different values on the crystal series resistor.

GL.

Groundloop · 2012.04.30, 09:35:45

Peter,

Here is the scope shots I promised.

GL.

Thaelin · 2012.04.30, 10:11:55

You have to be looking at harmonic hell here. Have you looked this over in a spectrum analyzer?
Bet its rich..
thay

EMdevices · 2012.04.30, 14:36:55

GL,

I retract what I said before, now that I'm looking at your scope shots.

You have a beautiful and pure frequency! I don't see any harmonics! The last graph is so common to see with digital scopes when it's zoomed out, so there is no 5 hz modulating frequency. That patern you're seeing is the beat frequency between your digital display and your pure signal.

Good job, you had it working all along!

EM