Hartley oscillator problems

[mpegjohn]

Hi,
I am building a hartley oscillator that will tune beween 15.24MHz and 31.44MHz.

I have built the circuit based on a QST article, image attached.

However, I am using a padder and trimmer setup, as I need this to track in a superhet.

The Inductance is 277nH Padder is 4.03nF the trimmer is 70pF.
The tuning cap goes from 16pF to 356pF.

The circuit does oscillate, but as the tuning cap is reduced to almost its minimum value, the waveform on the anode changes to a very strange shape, and is about 240KHz, and very large.
While this is happening on the anode, on the grid, and cathode I het a nice 31MHz
Increasing the capacity brings back the much smaller correct sine wave at about 28MHz.

Any idea what is going on here?

John

[Nuvistor]

Is the paddler really 4.03nF? It is a bit late in the evening for my maths but that value seems high for a padder at those frequencies.
The posted circuit is a bit blurred though.
Frank

[TrevorG3VLF]

The padder should be in series with the coil?

4nF could be running above its resonant frequency, try it without the padder to start with.
The padder is normally connected to earth but here with a Hartley you need a path for the cathode current.

[mpegjohn]

Hi,
I managed to get the anode waveform looking good, after thinking about it. I just reduced the feedback, and it seemed to make things settle down.

However I can't get the frequency span I need. I can get upto 31MHz fine now, but it stops below 16MHz.

Any ideas?

John

[ms660]

Ironically more feedback...maybe? It can get critical sometimes.

Lawrence.

[Ed_Dinning]

Hi John, one of the usual dilemmas with oscillators with wide frequency range, the level of feedback to give a good wave shape. I have see a pot used ganged with the variable C to achieve this.
You will probably find that stray capacitance adds to the feedback above a critical frequency helping it to oscillate.

Reduction of coil Q may help, or winding it with resistance wire will also work; you will then need to increase the feedback. The downside of this may be more drift.

Ed

[kalee20]

Have you been able to scope the anode and the grid at the same time? I can't believe that the thing can oscillate cleanly at 31MHz and at 240kHz simultaneously. Ill bet that adding the scope to the grid damps it a bit, taming the oscillation so it does what it should; removing the scope (or adding it to the anode) provokes a change to a squegging mode. But unless you can look at both points, you'll never know.

[Tyso_Bl]

For such a wide frequency range might I suggest that the Vackar configuration would be more appropriate.

[Radio Wrangler]

Squegging?

Bias-component-tuned oscillation?

Also I see no padder, just lots of capacitors across the coil.

APC I thought meant "Aperiodic Choke"
N950 sounds like a temperature coefficient, and you're too early in the design process to go trying to compensate temperature drift. You have to find out how much drift you get first, before you can know how much overall temperature coefficient you need to handle it.

Note that the circuit has no cathode resistor (and decoupler). It relies on self-rectification at the grid to make negative grid bias as a kind of feedback to regulate the operating point.

Oscillators are susceptible to giving LF oscillations or bursts of RF in an on/off mode as they look like negative resistors to the anode supply, and you can get a relaxation oscillator mode running in addition to the RF oscillation mode. I once wrote a lengthy piece on how oscillators work and how to 'read' them. It formed the first several pages of the ARRL handbook oscillators chapter (all editions since 1995) and .pdf files of it seem available on the web.

David GM4ZNX

[Radio Wrangler]

It looks like I overlapped with Tyso.

There's no reason why a Hartley can't do the job. In fact they are quite advantageous for wide range oscillators because you can make the variable C constitute a larege fraction of the total C seen by the inductor more easily. The Vackar starts winning for narrower tuning ranges, and where you need toisolate the amplifier more from the tank in order to minimise tempco.

David

[mpegjohn]

Thank you all for your help, I looked at the grid voltage on a scope, and as the freq is decreased the amplitude also decreaes.
At the highter freq I am guessing that I was somehow overdriving the grid, and causing the anode waveform to look strange. Is that possible?

[Radio Wrangler]

The way oscillators work, the waveforms often look very distorted.

There's an awful lot written on their frequency control mechanisms, but damned-all on their amplitude control mechanism.

Running, the circuit will be driving the valve into cutoff for part of the cycle, so the anode current will be very distorted, and as this circuit has a resistive anode load, the anode voltage waveform will be the same.

The way to get a sinusoid signal from most oscillators is to have a cathode follower driven from the tank circuit, not from the anode.

David

[mpegjohn]

Dear all,
After after many nights scratching about for answers, I have finally hit on the culprit.
I managed to use TinyCad to draw my schematic, and I have attached it to the thread.
As you can see I actually switch ranges with the aid of 2 relays per range.

These relays are SIL 12V SPNO reed relay(not the SPDT shown in circuit).

What I found is that with the relay in the circuit the oscillator would happily go upto 32+ MHz but as already said, stopped below about 18MHz.

I found by taking the relays out and hard wiring the connections all was well 13MHz upto 30MHz approx.

The feedback relay has little effect, and can be in or out, as I guess it is the low Z part of the circuit.

The real problem is the HOT end relay.

Any ideas what the problem is here. I tried a 100nF cap across the coil, but no change.
Only thing I can think to try is a sub min relay, where the contacts arn't within the coil. What does anyone think?

Regards,
John