|
Homebrew Equipment A place to show, design and discuss the weird and wonderful electronic creations from the hands of individual members. |
|
Thread Tools |
29th Jan 2021, 9:53 pm | #1 |
Pentode
Join Date: Jan 2015
Location: Greenock, Inverclyde, UK.
Posts: 161
|
Help understanding oscillator circuit
On the “projects to build” page of vintage-radio.com there is a design for an RF signal generator. I’m trying to get my head around how the oscillator section works. I’ve copied the relevant section below. I believe this is a Colpitts circuit but using a pair of inductors in the tank circuit rather than a single inductor / two series capacitors.
My understanding is that positive feedback is taken from the junction of C2 and L6 and applied via C4 to the gate of TR1. TR1 then drives the C1/L5 combination at resonance. What I don’t get is how the signal from L5/C1 is coupled to L6/C2 given that the junction of C1/C2 is connected to ground. What am I missing? |
30th Jan 2021, 4:39 pm | #2 |
Pentode
Join Date: Jan 2015
Location: Greenock, Inverclyde, UK.
Posts: 161
|
Re: Help understanding oscillator circuit
To aid my understanding I have built the circuit on breadboard using 330 uH inductors and 1nF caps. It will only oscillate if the inductors are within an inch or so of each other. Moving them further apart or placing a screen between them stops it. So my version appears to need stray coupling to work. This seems to confirm what I found when building the signal generator, but it seems that others have got it working ok. Mine kind of works works but only at higher frequencies. I had been trying to improve it by separating the inductors around the rotary switch without much success, hence my need to further understand the circuit. Has anyone else built this?
Link to the project is here https://www.vintage-radio.com/projec...generator.html |
30th Jan 2021, 5:43 pm | #3 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,894
|
Re: Help understanding oscillator circuit
Don't know. It would depend on circuitry outside of the snippet posted. THe centre tap of the inductors and a 100pf capacitor go off-screen.
David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
30th Jan 2021, 5:46 pm | #4 |
Dekatron
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,087
|
Re: Help understanding oscillator circuit
Are the coils adjacent, so they are coupled?
Alternatively, it could be a TATG oscillator (or, being FET, TDTG). But as variable-frequency oscillators, these could be rather temperamental, as the slight tuning offset needed between the two tuned circuits could be difficult to maintain over a wide range. |
30th Jan 2021, 5:59 pm | #5 |
Pentode
Join Date: Jan 2015
Location: Greenock, Inverclyde, UK.
Posts: 161
|
Re: Help understanding oscillator circuit
David,
The complete circuit is shown below. I've looked at it so many times I assumed the snippet was more complete than it is! The centre tap goes to a positive supply (that can be modulated or not). The 100pF cap couples the oscillator to the next stage. Kalee20, the project shows a layout for the inductors that seems to keep them separated around a rotary switch. I don't think coupling is intended. |
30th Jan 2021, 6:06 pm | #6 |
Dekatron
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
|
Re: Help understanding oscillator circuit
I might be wrong but the schematic as drawn...I would have expected equal value inductors for the range selected.
Lawrence. |
30th Jan 2021, 6:11 pm | #7 |
Pentode
Join Date: Jan 2015
Location: Greenock, Inverclyde, UK.
Posts: 161
|
Re: Help understanding oscillator circuit
Lawrence, I think thats just a drawing error. The switch sketch in the project page shows the correct pairing.
|
30th Jan 2021, 6:43 pm | #8 |
Dekatron
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
|
Re: Help understanding oscillator circuit
|
30th Jan 2021, 6:47 pm | #9 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Help understanding oscillator circuit
It's drawn in a strange way which doesn't help but it looks to me to be a negative resistance oscillator with tuned circuits in the drain and gate. The feedback path is via the drain to gate capacitance.
The signal tap capacitor C5 looks suspiciously large in value at 100pF. I think the loading of this and the TR2 buffer stage could be enough to spoil the tracking of the two tuned circuits when the ganged 365pF caps are tuned towards the lower end of their range. This could starve the circuit of negative resistance and this would prevent oscillation. I can load the circuit into a linear simulator if that helps? I think it's best to keep the ganged capacitance within the range of 50-365pF for any initial analysis. Are you using 1nF caps here?
__________________
Regards, Jeremy G0HZU |
30th Jan 2021, 7:02 pm | #10 |
Pentode
Join Date: Jan 2015
Location: Greenock, Inverclyde, UK.
Posts: 161
|
Re: Help understanding oscillator circuit
Jeremy,
I have built the circuit as shown using a 270 pF variable cap, but also I have breadboarded just the oscillator circuit with fixed 1nF caps. The breadboard version does not have the 100pF coupling cap (C5). I would be very interested in seeing results from a simulation if its not too much trouble. Thanks, Ian |
30th Jan 2021, 7:30 pm | #11 |
Dekatron
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
|
Re: Help understanding oscillator circuit
Similar circuit is referred to as a Colpitts in the Leader LSG-16 signal generator (FET osc.) Having said that the split coil pairs in that one are on the same former:
https://nvhrbiblio.nl/schema/Leader_LSG16.pdf https://www.radiomuseum.org/r/leader_lsg_16_lsg16.html Lawrence. |
30th Jan 2021, 9:34 pm | #12 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Help understanding oscillator circuit
The Leader circuit looks similar but the coil tap point on the Leader circuit is via a 4k7 resistor meaning that this circuit looks a bit more conventional in the way the feedback is arranged.
The circuit in post #5 has a low impedance point here due to the shunt 2n2 cap C10 and the low impedance of the emitter of TR3. Is there a series resistor missing here in the circuit of post #5? As drawn I think the post #5 circuit will oscillate on parts of the ranges that I just looked at with a simulator. I do think there could be a circuit error somewhere?
__________________
Regards, Jeremy G0HZU |
30th Jan 2021, 9:46 pm | #13 |
Dekatron
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
|
Re: Help understanding oscillator circuit
In a Leader valve version (LSG-10) the supply resistor isn't fitted on the export version:
https://nvhrbiblio.nl/schema/Leader_LSG10.pdf Lawrence. |
30th Jan 2021, 9:58 pm | #14 |
Pentode
Join Date: Jan 2015
Location: Greenock, Inverclyde, UK.
Posts: 161
|
Re: Help understanding oscillator circuit
There is a definite similarity to the Leader circuit. If I add a 4k7 series resistor between TR3 and the inductors on my breadboard circuit the oscillation is much less affected by coil coupling, but the waveform becomes distorted dependant on supply voltage. I'll try the same mod on the complete circuit tomorrow.
|
30th Jan 2021, 10:35 pm | #15 | |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Help understanding oscillator circuit
Quote:
It looks like the only mechanisms for feedback are coupling between the coils and the anode to grid capacitance on the (Leader) valve circuit. On that circuit it looks like the coil tap point is heavily decoupled as there is more than just the 1nF cap there. Is the modulation fed here? I can't tell as I'm not good with valve circuits. I tried simulating the original circuit in post #5 and I do struggle to get it working across all the ranges. My simulation only allows feedback via the drain to gate capacitance so I don't model any coupling between the coils. It might need this coil coupling to work across all the ranges? I'm nervous to post up any simulation results or even any basic theory because the circuit may well rely on coupling between the coils and also there may be subtle component errors in the circuit. Connecting C5 (100pF) and the buffer circuit to the simulation really does cause problems across some ranges for example.
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 30th Jan 2021 at 10:40 pm. |
|
30th Jan 2021, 10:45 pm | #16 | |
Dekatron
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
|
Re: Help understanding oscillator circuit
Quote:
Lawrence. |
|
31st Jan 2021, 1:20 am | #17 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,894
|
Re: Help understanding oscillator circuit
It's not an overt Colpitts, because both the centre point of the inductors and the centre point of the tuning capacitors are at RF ground.
It needs either Cdg or stray field between inductors in order to operate. Given the variance in these strays and the variation in Cdg and Gm of FETs this isn't going to be a sure-fire circuit. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
31st Jan 2021, 1:46 am | #18 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Help understanding oscillator circuit
For a bit of fun I entered the whole circuit for the AM modulator and the oscillator and the output buffer into Simetrix and simulated it.
It did actually oscillate on all ranges although I only simulated top and bottom using 35pF and 365pF for the tuning capacitance. Range A 164kHz - 345kHz Range B 370kHz - 902kHz Range C 820kHz - 2.17MHz Range D 1.76MHz - 4.97MHz Range E 3.80MHz - 11.5MHz There is a slight gap between range A and B but this could be due to the fairly simple models I used. At the output buffer there was a very large variation in output voltage across the ranges and I did try and put a realistic Q in for the inductors to try and get realistic levels in the simulation. The JFET spice models in Simetrix are reasonable and often predict oscillator performance quite well. However, they can't be fully trusted on a fairly dodgy circuit like this. By contrast, the linear simulator showed that some parts of the tuning range were very marginal and it predicted that oscillation wouldn't happen in some places. However, the linear models for the JFET aren't optimised for the startup conditions for this circuit so this might explain why things didn't look so good for the simple linear simulation.
__________________
Regards, Jeremy G0HZU |
31st Jan 2021, 6:35 am | #19 | |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,894
|
Re: Help understanding oscillator circuit
Quote:
I wondered why anyone would do something so marginal, especially as circuits lighter on the cost of inductors are common? The application of AM to the oscillator itself is best avoided, you run into starting and stopping problems limiting the depth you can apply. An oscillator is an infinite loop with a circulating signal. It has noise addition, gain, non-linearity, and filtering within the loop. You can consider the circulating signal to have many components at any one instant, and each component to have a different history in terms of how many times it's been round the loop. Noise keeps entering more signal (uncorrelated between components) and the non-linearity compresses the total. So i thought, why not unroll the Swiss roll? Open it out into a linear structure of almost infinite length. Each stage is the oscillator circuit with the loop broken to make an input and an output, all dasy chained to hundreds of stages. Each oscillator circuit being modified to include a model of a realistic level of noise generator, and care taken to make the model realistic in terms of non-linearity. Sounds daft? sounds like an immense task? No, it works. People don't see how powerful modern computers are. Most processor horsepower is lost in amazingly inefficient operating systems and applications (I still found I sometimes had to wait for the word processor on the dual core PC at work to respond to typing!) But give them an arithmetic task and watch them go. A few hundred stages of unrolled oscillator simulated quickly in Spice. I could probe down the chain and see stage after stage with noise building up, then the noise taking on a filtered spectral shape and eventually the amplitude flattening off. With realistic component Q I got a useful indication of phase noise performance. Using Spice hierarchically let me edit circuit parameters in one place and hit 'run' It's most interesting with circuits that won't run, you can step along the stages and see the lack of increase in noise. You can inject a turn-on transient to the first stage and see if that kicks anything off. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
|
31st Jan 2021, 11:50 am | #20 |
Dekatron
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
|
Re: Help understanding oscillator circuit
Avo, Advance, Leader and Heathkit versions all call it a Colpitts or a modified Colpitts.
Lawrence. |