Signal generator instability cause?

Hi I have built this signal generator and I have noticed that is quite stable on the KHz region but as it passes to the MHz frequency starts to get more and more drift.
I was wondering what could be the source of this drift which appears only in the MHz region?

I have used molded chokes with ferrite cores, which could possible be the case, but is that really only that, or is it the design of the circuit that causes the drift itself?

You are talking about frequency drift? Can you quantify the observed drift?

It is not stable for communications usage. For example it cannot be used as a local oscillator, to test a mixer, as the signals at the output of this mixer would drift away too fast.
I would say be ear, a tone at the output of this mixer drifts from 1KHz down to zero in a few tens of seconds, maybe less.
So it cannot be used in such purposes, nor for testing a narrow IF filter response for example.
As an AM source (6-10KHz) it is probably ok, as the oscillator can be retuned less often.

It is possible that these ferrite core chokes I have used as coils really play a role on this, but why this does not happen in the KHz frequencies as well?
I want to try it with material 7 Amidon toroids to see the difference, but before doing so, I wonder if this circuit topology has the tendency to drift in frequency so much.
I have seen LC unlocked oscillators (half a MHz though) that are really stable, that is why I wonder this. Maybe the L/C ratio plays a role on this?

I am sorry I cannot provide any numbers but I do the tests by listening the signal on a nearby dds tuned RX.

Hi,

I asume it does.

But with 1MHz a drift of 0.1% is 1kHz drift.
With 1kHz and a drift of 0.1% is just 1Hz of drift.

Klaus

Thanks, that makes sense.

Do you think better inductors quality like these iron powder cores (T50-2, T50-6, T50-7) will cure the problem or do you think that it is the circuit topology that causes such drift? The L is driven by square waves in that circuit.

Hi,

what causes the drift?

* a drifting power supply? --> then stabiize the power supply.
* temperature? --> you may find out which device is sensitive with an coolant spray or a soldering iron (no need to touch the device, just put the soldering iron tip close to the devices)
* what else?

Klaus

I have used a regulated PSU, so I do not think it is the PSU that causes drifting.
Temperature variations could be the main source and these ferrites have awful characteristics related to temperature. That is why I considered these at first.
Usually the source of drift is temperature dependent in PSU-stabilizer driven circuit. I will do the test with the iron tip. near the components, that is a good trick.

Anything else that comes in mind let me know

The loaded Q of the tank circuit is what stabilizes the frequency "drift", or random FM noise. Obviously, as you enter the MHz region, the inductors have to change in style from those that worked well in the KHz region. Maybe something like that.

Note also that the tuning control is the only component of a capacitive nature that determines the frequency. As L gets smaller, the effect of C gets greater. I would doubt the tuning capacitor is temperature stabilized (image is too small to read it's value). The classic fix is to add a capacitor in parallel with the tuning control (~47pF) and to select a type with an opposing temperature coefficient.

Brian.

The tuning capacitor is 100pF. This is too small for LF to HF use but he uses this and he prefers switching more coils, than having a bigger capacitor and less coils. I am not sure why he does that, It may have to do with the constant output level of the generator, or for better fine tuning.

I changed the molded chokes to amidon material 7 cores (the most temco stable) and indeed the frequency stability was very good to 20MHz ( I haven't tried more than that).
I also wound an air coil and again the frequency stability was much better. However I forced to be 2 meters away from the variable capacitor and the coil, any movement detuned it, like a theremin.
So yes it seems the molded chokes are awful for the purpose...

How can I eliminate detuning by proximity, mayby by enclosing the whole thing into a metal enclosure, will that be enough?

I have also noticed that even with the better coils, there was much hum in higher frequencies above 10MHz, when listenning on AM. On SSB this hum was heard worse, like modulation. Why is that hapening?
Maybe the PSU I use pass noise into the circuit?, or maybe the LC is picking up a signal and modulates the oscillator? How to cure that?

Just place a 100nF capacitor in parallel with each rectifier diode on your power supply, and the hum it will disappear.

https://qph.fs.quoracdn.net/main-qim...5f69709eaba2e8

you need VERY stable capacitors to make an osc with low drift, e.g. potted polypropylene, with higher rated volts, else temp will give you drift ....

It is possible the rectifiers are producing low level hum (assuming you are using a linear supply!). It is an effect caused by the recovery time of the diodes as the polarity across them reverses that produces short, sharp spikes at twice mains frequency. The reservoir capacitors are good for holding the overall charge but less effective at stopping very short pulses. The extra capacitors 'soften' the spikes at source.

Shielding in a metal enclosure is the only way to stop the proximity effect but if you need absolute stability, use a metal and thermally insulated box and add temperature regulation to it to keep it's internal temperature constant. It isn't difficult to do but you do generally have to leave it turned on for 10 minutes or so to allow the temperature to settle before use.

Brian.

Hm... I suspected the lab PSU for the hum, it gave me hair pulling hum problems at the past on oscillators especially LC on shortwave.
Well, I will try it with a battery as a power source, if this eliminates the problem then it is definitely a PSU problem and I will place these capacitors you and FwM suggested there.
Did I mention, the original PSU published with the circuit DID have these capacitors in place.

Thanks for the info I will come back with results.




Ok, I replaced the PSU with a battery. The hum is much less now, so definitely this was a PSU problem to a big extent. However hum is still present in a good amount making it unusable when switching the nearby RX to SSB, as the tone is far from clean. Switching in AM mode the hum can be still heard, although it is much less as said.
Here is a closer look at the circuit, as well as how I built it. The PDF with the whole article is too large to post here.

I have replaced the air wound coil with a amidon material 7 toroid, to see if the air coil was picking any RF from the lab. It did not make a difference.

I would greatly appreciate if you could suggest me things to try to cure the problem.

I found what the problem was with the "hum". The PSU was already a source as I told you, but the hum was completely eliminated when I disconnected the oscilloscope probe from the circuit. It also made the circuit muck less sensitive to hand effects, now I have to put my hand very close to the circuit so as to alter the frequency.
I do not know why?
Maybe the probe cable was picking some LF signals and modulated the generator, or maybe there is something wrong with the scope/probe? I remember in the past I had the same problems with a crystal oscillator which when operated with high HF crystals was producing the same hum.
I am happy I found the solution, but I am not sure what is the case with the scope, is it malfunction or is this normal?

Hum when connected to other equipment has been discussed elsewhere on this Forum. Basically, instead of being completely isolated, you are tethering it to a potential (the scope ground) which itself may not be noise free. Capacitive coupling from all parts of the circuit to things around it then introduce unwanted signals, predominantly hum from house wiring and appliances.

The other things that worries me now I can see the markings on the schematic is the use of a 12V secondary on the mains transformer which loses ~1.4V in the bridge rectifier then goes to a 12V regulator that needs about 15V minimum at it's input. The peak voltage of 12V RMS is about 17V so you only have a safety margin on voltage peaks of 17 - 15 - 1.4 = 0.6V, in other words it relies heavily on low ripple across the reservoir capacitor.

Good to see you covered the LED, that could also be a source of hum.

Brian.

Ok, I will use a higher voltage transformer (maybe 15v or 24v) in my final setup then. Or probably I will add a "thrifty voltage regulator" https://www.eeweb.com/extreme-circui...tage-regulator better or maybe a capacitance multiplier regulator like this http://www.tubecad.com/2016/06/26/Ca...HV%20Zener.png which effectively multiplies the capacitor like having a bigger one.

Any ideas how to isolate the circuit from the measuring equipment?

Increasing the transformer voltage or using a low drop-out voltage regulator will give better results than both those circuits.

To minimize the effect of the scope, use two probes. Do not connect the ground clips on either one but connect the tips across the signal you want to measure. Use the scope's add and invert function to make the probes work as a differential input. If you have x10 probes, even better but make sure they are set the same.

Brian.

That's a great hint I have to try!
I do not have any other source of information about this, what each function (add/invert) will do with this setup?

It is quite logical, if you show a signal on trace A and another signal on trace B then select ADD mode, it shows the waveform A+B.
If you do the same with one channel inverted (lets say channel B), you get A+(-B) which is A-B.
Think about what you measure normally with a ground and tip probe, its the voltage difference between them. If instead you use two (ideally x10) probes and only connect the tips, it will also show the voltage difference between them but now your scope is isolated through the probe divider networks.

There is a good explanation here: https://www.youtube.com/watch?v=BHY4o7Iknes

It uses Tek scopes but yours will I'm sure have the same functions.

Brian.