Hartley oscillator: amplitude not stable

Bordodynov

Joined May 20, 2015
3,177
I made a simple shunt model with parameters. Name is ShuntRef. Due to a simple model, placing such an element on the circuit does not greatly complicate the calculation and therefore practically does not increase the counting time. There are many such items on offer. Suitable is the cheapest. And yes, TLV431 will do.
Real: ZXRE125, LM4041CIM3-1.2, LM4051CIM3-1.2, LT1004...1.2
 

Thread Starter

patpin

Joined Sep 15, 2012
401
I made a simple shunt model with parameters. Name is ShuntRef. Due to a simple model, placing such an element on the circuit does not greatly complicate the calculation and therefore practically does not increase the counting time. There are many such items on offer. Suitable is the cheapest. And yes, TLV431 will do.
Real: ZXRE125, LM4041CIM3-1.2, LM4051CIM3-1.2, LT1004...1.2
Thanks, I guess this is it:
.subckt ShuntRef 1 2 Vref=1.22 rd=1 inom=1m imin=50u
d1 2 1 MyD
d2 1 2 dp
.model MyD D(Ron={rd} Roff={Vref/imin*1.05} Vfwd={Vref-rd*inom})
.model dp D(Ron=1 Roff=1G Vfwd=0.6)
.ends
Why do you call it irrealistic over here? (http://www.diyaudio.com/forums/solid-state/293245-ltspice-model-2v-zener.html)

And do you haven more info on the 3 NPN oscillator that you designed. How to calculate/parametrise the frequency e.g.
Is this your own design?
 
Last edited:

Bordodynov

Joined May 20, 2015
3,177
My reference source model reflects only the basic parameters. But it does not simulate generation at certain values of the capacitor, which are connected in parallel to the reference. Also does not simulate the complex character of dynamic resistance.
Yes this is my own design, which I designed for you. This generator operates on a series resonance. Its disadvantage is that it consumes a lot of current. For larger inductances, a smaller current can be obtained.
 

Thread Starter

patpin

Joined Sep 15, 2012
401
My reference source model reflects only the basic parameters. But it does not simulate generation at certain values of the capacitor, which are connected in parallel to the reference. Also does not simulate the complex character of dynamic resistance.
Yes this is my own design, which I designed for you. This generator operates on a series resonance. Its disadvantage is that it consumes a lot of current. For larger inductances, a smaller current can be obtained.
I have a lot to learn about LTspice etc...
You should publish yr Bordodynoscillator ! If you don't mind please explain how you calculate/design it. Don't hurry, I have time. Many thanks.
 

Bordodynov

Joined May 20, 2015
3,177
Neither of which is original. It is a non-inverting amplifier with low input and output impedance.
A series LC circuit is connected between the output and the input.
The circuit is a cascade with a common base and an output stage with a common collector on a composite transistor.
In order to not strongly depend on the currents of transistors, I applied a thermostable reference source of stable voltage.
I have a lot of experience in circuit design and therefore I do not bother with theory.
The theory is in my blood. I set such a current of transistors so that their dynamic resistance
(the reciprocal of the slope) was approximately 1 ohm, which is less than the series resistance of the inductance.
Then, trite using the selection of the resistor in the collector, I choose its value.
I do this using parametric analysis (Step analysis).
I developed many electronic circuits that are steeper than this one.
It's enough for me to realize what a good designer I am and for this I do not need publications. Especially for this, additional efforts are needed, and I am lazy.
 

Thread Starter

patpin

Joined Sep 15, 2012
401
Neither of which is original. It is a non-inverting amplifier with low input and output impedance.
A series LC circuit is connected between the output and the input.
The circuit is a cascade with a common base and an output stage with a common collector on a composite transistor.
In order to not strongly depend on the currents of transistors, I applied a thermostable reference source of stable voltage.
I have a lot of experience in circuit design and therefore I do not bother with theory.
The theory is in my blood. I set such a current of transistors so that their dynamic resistance
(the reciprocal of the slope) was approximately 1 ohm, which is less than the series resistance of the inductance.
Then, trite using the selection of the resistor in the collector, I choose its value.
I do this using parametric analysis (Step analysis).
I developed many electronic circuits that are steeper than this one.
It's enough for me to realize what a good designer I am and for this I do not need publications. Especially for this, additional efforts are needed, and I am lazy.
 

Thread Starter

patpin

Joined Sep 15, 2012
401
Thanks for yr explanation. Yr designing passion and knowledge is indeed the most important in life! Only yr last sentence... I have my doubts on...
 

Thread Starter

patpin

Joined Sep 15, 2012
401
Thanks for yr explanation. Yr designing passion and knowledge is indeed the most important in life! Only yr last sentence... I have my doubts on...
One more question... concerns making the L1, the pickup coil . I see the current I(L1) are +30mV to -30mV; as agreed upon, I plan to wind it on a ferrite rod. Giving the dimensional constraints, how would you realise it and with which rod diameter? and which wire gauge? and windings over each other? go for 60µH? I guess the distance between the rod with the windings and the passing small ferrite rod will have to be minimal. There are passing magnets at about 15mm (for the injection timing) and high tension sparks of up to 20KV (in distributor; gap <1mm) at about 30mm of the pickup coil.
 

Bordodynov

Joined May 20, 2015
3,177
I have a question for you. Why did you choose such an IGBT transistor? After all, it is already obsolete. Are you sure that this transistor will work the way you want? Do you have any information that it was used in ignition systems?
 

Thread Starter

patpin

Joined Sep 15, 2012
401
Previously B
I have a question for you. Why did you choose such an IGBT transistor? After all, it is already obsolete. Are you sure that this transistor will work the way you want? Do you have any information that it was used in ignition systems?
Previously BJT where used. Modern systems tend to use IGBT's, I read. I haven't personally used an IGBT. I know that the voltage really needs to go to 0 to cut off and that it is very sensitive to static electricity etc. Seems driving it with a totem pole circuit is advised. Tests with one (a cheaper one) are on my to do list but that is for later. If it doesn't function I'll have to use a power BJT.
 

Bordodynov

Joined May 20, 2015
3,177
I guess... you mean you put all windings one besides the other one in one layer? Was your design optimized for 60µH? Which AWG is thick enough?
At frequencies of ~ 500 kHz, it is better to use not a simple wire, but a litcandrite.
I meant not a single-layer coil and not a two-layer coil. As I thought there were difficulties translating. You can say roughly, I meant the chaotic winding. Not one layer, the turn to the coil closely, but scattered. I do not know what it's called in English. This is an alternative to the winding universal. In this way, the parasitic capacitance of the coil decreases, and the Q-factor also increases.
 

Thread Starter

patpin

Joined Sep 15, 2012
401
At frequencies of ~ 500 kHz, it is better to use not a simple wire, but a litcandrite.
I meant not a single-layer coil and not a two-layer coil. As I thought there were difficulties translating. You can say roughly, I meant the chaotic winding. Not one layer, the turn to the coil closely, but scattered. I do not know what it's called in English. This is an alternative to the winding universal. In this way, the parasitic capacitance of the coil decreases, and the Q-factor also increases.
OK Thanks. chaotic winding on the mentioned ferrite. litcandrite wire is unknown to me (and google). Did you optimised for 60µH?
 

Bordodynov

Joined May 20, 2015
3,177
Литцендрат (litzendraht in German) - this wire is twisted many, isolated wires.
This wire is used to reduce the skin effect.
Winding in bulk - not a regular winding (ie no layers of winding), reduces the capacity.
 

Thread Starter

patpin

Joined Sep 15, 2012
401
Литцендрат (litzendraht in German) - this wire is twisted many, isolated wires.
This wire is used to reduce the skin effect.
Winding in bulk - not a regular winding (ie no layers of winding), reduces the capacity.
If it is isolated and parallel... how does one connect it? Like this: https://de.wikipedia.org/wiki/Hochfrequenzlitze? Still need 60µH I guess? And is it necessary to trim the LC? Is there a specific formula to calculate the induction in function of litze windings, lenght, diameter, (gauge), mu ?
 
Last edited:

Bordodynov

Joined May 20, 2015
3,177
As for the litz-wire. At the end of this litz-wire it is necessary to cover every thin wire with solder, and then it is necessary to solder them all together. The same thing should be done with the second end of the wire.
Of course, the inductance depends on the diameter of the wire, but this dependence is weak. When the inductance of the core coil is calculated, the coil inductance of one coil can be calculated first. The conductor must have a section of the real coil (the real coil must be inscribed in this conductor). The real inductance is obtained by multiplying the inductance of this single-turn inductor by the square of the number of turns (W ^2).
 

Thread Starter

patpin

Joined Sep 15, 2012
401
As for the litz-wire. At the end of this litz-wire it is necessary to cover every thin wire with solder, and then it is necessary to solder them all together. The same thing should be done with the second end of the wire.
Of course, the inductance depends on the diameter of the wire, but this dependence is weak. When the inductance of the core coil is calculated, the coil inductance of one coil can be calculated first. The conductor must have a section of the real coil (the real coil must be inscribed in this conductor). The real inductance is obtained by multiplying the inductance of this single-turn inductor by the square of the number of turns (W ^2).
Thanks, You have an idea of the needed and commercially available AWG. And I have to go for 60µH? And trimming?
 
Last edited:

Bordodynov

Joined May 20, 2015
3,177
Your main goal is to get the right frequency.This you can do by picking up the right amount of capacitor.For example, placing several capacitors in parallel or using a tuning capacitor.Although I do not understand why you need this frequency.My circuit is not critical to the frequency.
 
Top