In Bowick's "RF Circuit Design" page 110 you can see a page from a Motorola 2N5179 datasheet with the following oscillator:-This is a tricky base grounded oscillator. Because oscillator start-up depends on the load impedance, therefore the author used a "Sliding Tuner" to provide this condition.
(Attachment Link)
Questions
a) Is the network essential to the oscillator? (looks like it)
b) If yes, what would an equivalent circuit be for simulation?
c) How is this type of oscillator called?
I did try to simulate that in LTSpice using a bunch of t-lines with open/short ends, but I don't really know what I'm doing and I don't get stable oscillations, and nowhere near 500 MHz.
(Attachment Link)
Bonus points: What's a good way of extracting parameters (hybrid or scattering) in LTSpice from a transistor that's already in a circuit?
-This is a tricky base grounded oscillator. Because oscillator start-up depends on the load impedance, therefore the author used a "Sliding Tuner" to provide this condition.Thanks, besides the miswired emitter bias, adding a parasitic capacitance to the chokes did it, it's now oscillating in LTSpice, although not very cleanly.
The feedback is supplied by CCE capacitance ( maybe interwinding capacitance of RFC ) . L1 and Sliding Tuner Impedance both contribute into Oscillation Frequency
-I hope you have already found the nonlinear model of 2N5179I used the model that was already in the "extended" LTSpice library that's floating around. The input impedance vs. frequency matches what's given in Bowick's book so I guess it's alright.
-In order to predict the oscillation condition, you have to simulate this circuit by opening Feedback Circuit ( this is not possible in this configuration) and you have to examine the Open Loop Nyquist Criteria. Open loop gain must have a circle and crossing point 1+j*0 point.Would a Middlebrook loop gain probe work in this situation?
-Only nonlinear models are used in oscillator analysis, small signal scattering or hybrid linear parameters are useless.But you talked about the Nyquist criteria which is defined for linear systems? You mean the small signal parameters are useless for determining starting or maintenance conditions?
-Transient simulation of oscillator circuits are tricky and sometimes misleads you. Start up conditions might be satisfied but oscillator may not seem oscillating.Interesting, I hadn't heard of harmonic balance. Is that somehow related, more general or more specific than Manely-Rowe relations (which apparently are useful for analyzing frequency multipliers)?
-HB analysis gives you more deeper insight about the oscillators.
-Would a Middlebrook loop gain probe work in this situation?
But you talked about the Nyquist criteria which is defined for linear systems? You mean the small signal parameters are useless for determining starting or maintenance conditions?
Interesting, I hadn't heard of harmonic balance. Is that somehow related, more general or more specific than Manely-Rowe relations (which apparently are useful for analyzing frequency multipliers)?
BigBoss: -Only nonlinear models are used in oscillator analysis, small signal scattering or hybrid linear parameters are useless.
-In order to predict the oscillation condition, you have to simulate this circuit by opening Feedback Circuit ( this is not possible in this configuration) and you have to examine the Open Loop Nyquist Criteria. Open loop gain must have a circle and crossing point 1+j*0 point.
a) Is the network essential to the oscillator? (looks like it)I think the main goal for this circuit is to allow a hand adjustable method to measure/prove the oscillator output power quoted (at the operating point) in the datasheet. I don't think this is meant to be a practical application circuit because it uses huge and old RF 'tee' connectors in that network. This network would be bad for microphony as well as being physically huge and wobbly. I think it is just a means to explore how much power can be extracted from the 500MHz oscillator by tweaking the line lengths and stubs etc. I don't think it's worth building it unless you want to take a nostalgic trip back about 50 years or so. I've seen this circuit used with other manufacturer's datasheets and I think the aim was the same. To be able to tweak the line lengths to achieve the oscillator output power at the operating point quoted in the datasheet for that particular transistor type. A different transistor type would require slightly different line lengths to get the power output. So the output network is big and clunky and adjustable.
-I don't know LTSpice what that probe is. I have no idea.It's not specific to LTSpice, it's a way of measuring loop gain without manually opening the loop, see here
Thanks for your replies!Quote-I don't know LTSpice what that probe is. I have no idea.It's not specific to LTSpice, it's a way of measuring loop gain without manually opening the loop, see here
https://www.edn.com/middlebrooks-and-rosenstarks-loop-gain-measurements (https://www.edn.com/middlebrooks-and-rosenstarks-loop-gain-measurements)
I don't have a five-digit software budget unfortunately so no Microwave Office for me...
I was actually looking for a simple, realistic circuit to check calculations against while reading the book, so I jumped onto that oscillator circuit that looked simple, but clearly it's not and it's not even well-defined (hidden choke parasitic capacitances are required.)
But the circuit simulates OK in LTSpice which tells me that the transistor model is good enough at least for learning purposes.
What I want to do next is to learn how to extract small signal parameters from a transistor that is in an existing circuit. I expect this to boil down to running AC analyses at the proper operating point with sources at the input and the output and computing V/I ratios.
The big advantage of using VNA derived small signal parameters is that it's possible to predict the negative resistance of the BJT much more accurately compared to using a non-linear library model of the BJT. That's why I always try and do it the old school way using a VNA or with the manufacturer's s-parameter data.I've just been trying to extract a negative resistance value in this CB configuration with LTSpice to no avail.
If you aren't familiar with negative resistance then it can be a bit of a mind bender to try and work out what it actually means. However, in reality it actually makes perfect sense once you realise how it is defined and calculated.Thank you very much for all the detailed write-up.
Have you ever used a CB radio and measured the antenna with a VSWR meter? Hopefully you have. The first thing you do is transmit and then calibrate the meter to full scale. Then you stop transmitting and change the meter to indicate VSWR. Then you transmit again and take the reading for the VSWR.
y_i=.164-5.61e-3*i
y_f=-.156+20.9e-3*i
y_r=-72.4e-6-624.9e-6*i
y_o=1.35e-3+993.9e-6*i
S11=0.718 < 1°
S12=0.001 < 84°
S21=0.270 < -7°
S22=0.997 < -0°
I think my problem was with identifying the inputs and outputs of the common-base configuration which is what I'm not too familiar with. Turns out you apply the input at the emitter and take the output at the collector.
COMMON BASE AMPLIFIER
Vsupply vcc 0 12
Vbias Vb 0 2.5
RC Vcc Vc 820
RE Ve 0 330
Q1 Vc Vb Ve 2N5179_Fairchild
Vsource Vin 0 DC 0 AC 1 PORTNUM 1 Z0 50
Voutput Vout 0 DC 0 AC 0 PORTNUM 2 Z0 50
Coutput Vc Vout 10n
Cinput Ve Vin 10n
.MODEL 2N5179_Fairchild NPN (Is=69.28E-18 Xti=3 Eg=1.11 Vaf=100 Bf=282.1 Ne=1.177 Ise=69.28E-18 Ikf=22.03m
+ Xtb=1.5 Br=1.176 Nc=2 Isc=0 Ikr=0 Rc=4 Cjc=1.042p Mjc=.2468 Vjc=.75 Fc=.5 Cje=1.52p Mje=.3223 Vje=.75
+ Tr=1.588n Tf=135.6p Itf=.27 Vtf=10 Xtf=30 Rb=10)