Pulling frequency or spread spectrum of Colpitts osc

Find one that works and keep buying them. If volume is high, have an understanding with the maker to not change the fab. Smearing out a spectrum is hardly a precise thing anyhow.

As far as his circuit goes, we don't know what it is.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

That was just a tangent discussion about generating true randomness. All you need is a triangle wave to modulate the frequency. Best-case, all you might need is a uP pin and one resistor.

Or maybe just the uP pin.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

It may be hard to find a diode with high enough capacity to work in his circuit. Shielding and proper oscillator adjustment may be all that is needed.

Klaus seems to like Colpitts, Clapp and Pierce. How many other types are there?

Diodes are not cheap? Do not believe a word of that bunk. Many power diodes will act a a varicap, and even in the FM band.

It occurs to me it may be a simple layout problem. Perhaps Klaus is using two-sided FR1 or FR2 with not much room for a good ground plane. There might be room for improvement.

It's not the power, but the voltage. High stand-off voltages are achieved with a PIN structure, and that gives you the varactor behaviour.

te:

ering a

jittered

a p

guts

an

could use these for the function.

o receiver, and as I wrote I do not have that luxury ;-)

That would be ideal

The Colpitts BJT base is biased with 6k and 2k from a 5V supply (AFAIC). So to move the midpoint to have a different trigger point, I need to inject a current into the midpoint

It seems to me, this will mostly just offset the bias point and not alter t he frequency much. I did a simulation of it, where I injected a 100Hz sine of 10V into the bias point with a 1nF cap. That did move the frequency some what, but did not seem to be consistent (the 10V 100Hz was from the ac ripp le on the DC link capacitor, since that is a place I can get a rampling vol tage without the aid of the microcontroller. Must try the micro instead, si nce the DC link voltage ripple is load dependant

A lot of weak words, I have not had much time to look into it, have a manag er that wants me to work on other stuff :-)

Cheers

Klaus

I like the Colpitts due to a single inductor circuit, whereas the Hartley needs two inductors

This link has some details of the workings and formulas:

Cheers

Klaus

We have 4 sided board. This is not a layout problem, but we have a parasitic capacitance over an barrier with a cable connected. THe oscillator is no symmetric, so it forces a current into this parasitic capacitor which spreads into the cable

I could filter it out, but that is expensive. I could make the oscillator symmetric around GND, but that is not easy for this design

Cheers

Klaus

The key distinctive of the Hartley is that the inductors are coupled, so that you get better impedance transformation from the emitter to the base. A single-layer tapped coil will give coupling coefficients of about 0.4 to 0.7, depending on the geometry. With uncoupled inductors, it becomes just a minor Colpitts variant.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

Funny how different people select different criteria to classify oscillators. To me, the distinction between Colpitts and Hartley is that the former has two capacitive reactances and one inductive one as principal components of its tank, while the latter has L's where the former had C's and vice- versa. In my mind, the fact that the Hartley inductors might be coupled is just a minor variant.

Those criteria are vague, anyway. A capacitor in series with the Colpitts inductor and it turns into a Clapp. Choose the node between the two capacitors of a Colpitts for GND and suddenly it gets called a Pierce. What if I do both? There so many ways to make oscillators and only a few names to choose from.

Jeroen Belleman

This may be the kind of off-label usage where simulation differs from real-life. In the 1960s folk made one-transistor bugs where the ca.

100MHz oscillator transistor base was jiggled up and down by the microphone signal - even microvolts audio made kHz of deviation.

piglet

Coupled inductors give you some autotransformer action, which makes the impedance-transforming behaviour of the tank circuit work better, especially at lower loaded Q. A Colpitts has only the tank Q to enforce the impedance transformation, and so falls apart at Q values where a Hartley is still working fine.

There's nothing analogous for capacitors, of course, so without mutual inductance there's little difference between tapping down the capacitance or the inductance.

I've never been a big fan of minute classifications of oscillator types either, except for Hartley vs. Colpitts. The practical distinctions have a lot more to do with things like tank impedance (especially with crystals) and the effect of bias resistors on Q--in a Colpitts the base bias string shunts the whole tank, whereas in some others it just shunts part of it. Even there, if you use n*V_BE biasing (C-B feedback at DC, emitter at DC ground), you can jack the impedance of the bias string way up compared with the usual CE bias network.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

A Colpitts runs fine at Q=1, as demonstrated by the vigorous parasitic oscillations commonly found.

I doubt a Hartley with a coupling coefficient less than 1 would oscillate as strongly.

Version 4 SHEET 1 880 708 WIRE -96 -112 -160 -112 WIRE -32 -112 -96 -112 WIRE -32 -96 -32 -112 WIRE -160 -80 -160 -112 WIRE -576 -32 -608 -32 WIRE -496 -32 -576 -32 WIRE -336 -32 -416 -32 WIRE -240 -32 -336 -32 WIRE -224 -32 -240 -32 WIRE -336 -16 -336 -32 WIRE -608 0 -608 -32 WIRE -32 0 -32 -16 WIRE -336 64 -336 48 WIRE -160 64 -160 16 WIRE -160 64 -336 64 WIRE -128 64 -160 64 WIRE -112 64 -128 64 WIRE -336 80 -336 64 WIRE -160 80 -160 64 WIRE -608 96 -608 80 WIRE -336 160 -336 144 WIRE -160 176 -160 160 FLAG -32 0 0 FLAG -336 160 0 FLAG -128 64 Vout FLAG -240 -32 Q1B FLAG -608 96 0 FLAG -160 176 0 FLAG -96 -112 VCC FLAG -576 -32 VBB SYMBOL voltage -32 -112 R0 WINDOW 123 0 0 Left 2 WINDOW 0 33 33 Left 2 WINDOW 3 33 80 Left 2 SYMATTR InstName V2 SYMATTR Value 5V SYMATTR SpiceLine Rser=0 SYMBOL npn -224 -80 R0 SYMATTR InstName Q1 SYMATTR Value 2N3904 SYMBOL cap -352 -16 R0 SYMATTR InstName C1 SYMATTR Value 1.2732nF SYMBOL cap -352 80 R0 SYMATTR InstName C2 SYMATTR Value 1.2732nF SYMBOL ind -400 -16 M270 WINDOW 0 32 56 VTop 2 WINDOW 3 5 56 VBottom 2 WINDOW 39 -16 53 VBottom 2 SYMATTR InstName L1 SYMATTR Value 1.5915uh SYMATTR SpiceLine Rser=50 SYMBOL voltage -608 -16 R0 WINDOW 123 0 0 Left 2 WINDOW 0 33 33 Left 2 WINDOW 3 33 80 Left 2 SYMATTR InstName V1 SYMATTR Value 2.5V SYMATTR SpiceLine Rser=0 SYMBOL res -176 64 R0 SYMATTR InstName R1 SYMATTR Value 560 TEXT -408 -240 Left 2 !.tran 0 20u 0 50p TEXT -408 -264 Left 2 ;'Colpitts Q=1

The Colpitts runs fine at Q=0.25

Change Rser in my previous post to 200 ohms and R1 to 50 ohms.

XL = 2 * pi * 1.59e-6 * 5e6 = 49.9513231921 ohms

Q = XL / R = 50/200 = 0.25

For sufficiently loose definitions of 'fine'.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

Troll.

You used "fine" first: "where a Hartley is still working fine." So you are criticizing yourself.

There is nothing wrong with the operation of the Colpitts. It starts at very low amplitude and increases exponentially until it reaches the maximum amplitude. This is determined by the current through the emitter resistor.

The frequency is determined by the LC product, with an offset due to the low Q of the tank. See Chapter 9. Tuned Circuits, Equation 3, Page 449, in the Radiotron Designers Handbook, 1954 edition, at

The Colpitts is operating normally in every sense of the word.

However, you are completely missing the significance of these demos. They show the Colpitts is perfectly capable of oscillating with extremely low Q tanks, such as in parasitic oscillations. This is why you should provide for base suppression resistors or beads for any circuit that may be capable of parasitic oscillations. See the Readme.txt file for 02.ASC Colpitts Q=1 in Oscillator.zip at