# Hartley oscillator: amplitude not stable

#### patpin

Joined Sep 15, 2012
401
Another question, can I simulate the arrival and absence of the ferrite rod of 1mm thickness at 0.5mm distance of the E shaped ferrite ( 13mm height). Perhaps with the transfo coupling factor, and what K-value at 0,5mm and at 20mm distance?

Last edited:

#### Bordodynov

Joined May 20, 2015
2,906
As the ferrite rods approach, the inductance increases.
I made inductance, which depends on time.
In the circuit node named "L", the voltage V (L) is numerically equal to the inductance,
which depends on time. I added the "uramp" function so that the inductance could
only increase. I do not know the true law of the change in inductance.
I know that this function is intermittent. But for a demonstration, this is not important.
The value of the function is the voltage V (L) (see upper graph).
Since the current is more or less stable, the voltage on the circuit is proportional to the inductance.
Detector performs detection. You only need to add a separating capacity
(this compensates for the temperature) and send a signal to the Schmidt trigger
(with offset by the resistive divisor of the operating point in the middle of the range).
Next you will receive a digital pulse signal.
I do not know why you need it and you can have an analog signal.

#### Bordodynov

Joined May 20, 2015
2,906
Another question, can I simulate the arrival and absence of the ferrite rod of 1mm thickness at 0.5mm distance of the E shaped ferrite ( 13mm height). Perhaps with the transfo coupling factor, and what K-value at 0,5mm and at 20mm distance?
Why do you complicate everything. Why do you need a coupling factor ?! Measure two inductances - at the minimum and maximum inductance distance from the ferrite rods, and that will be enough. With simulation, I took the maximum change in inductance of 20% (0.2). As far as I understand, I took a large number. Having made the measurements you will be able to refine the coefficient.

#### patpin

Joined Sep 15, 2012
401
C7 - A real capacitor, for tuning to a frequency roughly equal to the working one.
Do you mean a trimming cap.?

#### Bordodynov

Joined May 20, 2015
2,906
It is not necessary to fine tune. Firstly, your frequency changes. Secondly, this resonant circuit has a low quality factor and, correspondingly, a broad bandwidth.

#### Bordodynov

Joined May 20, 2015
2,906
See

#### Attachments

• 5 KB Views: 2

#### patpin

Joined Sep 15, 2012
401
Why do you complicate everything. Why do you need a coupling factor ?! Measure two inductances - at the minimum and maximum inductance distance from the ferrite rods, and that will be enough. With simulation, I took the maximum change in inductance of 20% (0.2). As far as I understand, I took a large number. Having made the measurements you will be able to refine the coefficient.
Fantastic result. One last question. Can your output directly drive an IGBT (with Zener incorporated) or does one need eg. a totem pole driver.

#### Bordodynov

Joined May 20, 2015
2,906
With transistors IGBT there is a problem: they have a large threshold voltage of ~ 5V. I specifically limited the output voltage to not more than 5V (resistive divider at the output of 1kΩ and 3kOhm.) It is possible to increase the output voltage by throwing out a resistor of 1 kOhm and increasing the supply of the operational amplifier to the battery voltage, but the maximum voltage at the output of the operational amplifier is less than the power supply by 1.2 V. Operational amplifiers do not like the capacitive load.What are you going to control and whether the control phase is suitable? In general, you need to simulate.My model of the operational amplifier LM358 correctly reflects the behavior at the capacitor oh load (as opposed to the factory model). That's why I made this model.

#### patpin

Joined Sep 15, 2012
401
With transistors IGBT there is a problem: they have a large threshold voltage of ~ 5V. I specifically limited the output voltage to not more than 5V (resistive divider at the output of 1kΩ and 3kOhm.) It is possible to increase the output voltage by throwing out a resistor of 1 kOhm and increasing the supply of the operational amplifier to the battery voltage, but the maximum voltage at the output of the operational amplifier is less than the power supply by 1.2 V. Operational amplifiers do not like the capacitive load.What are you going to control and whether the control phase is suitable? In general, you need to simulate.My model of the operational amplifier LM358 correctly reflects the behavior at the capacitor oh load (as opposed to the factory model). That's why I made this model.
I want to fire the plug by means of an automotive coil with 2 windings: one is 1.5mH and the one where the plug is on is 15H. I bought a ixgr32n170h1 for it.

#### patpin

Joined Sep 15, 2012
401
Please allow me one more question concerning the coil L1 #46.
Which core would you prefer (maximum height 13mm and thickness 8mm). The reluctor wheel (with the 11mm heigh and 1mm thick ferrite rods) passes at 0.5 mm from the core.
And how to the wind the coil on it? 60µH I guess?
For the initial diagram I had used an E shape one with 3 windings on the legs but then it was a transformator, (pickup coil at the center E-leg and a 60µH winding on upper E-leg and a serial 10µH winding with opposite phase at lower E-leg; this gave me a good lowering of the amplitude at the pickup coil when the ferrite rod passes), but I guess with yr diagram on post 46, there are better and easier solutions.

#### Bordodynov

Joined May 20, 2015
2,906
I understand that you want, that when approaching ferrite
the amplitude decreased. But in my last scheme there is an increase
amplitude. Those. At the moment of passing the rods, a pulse is generated.
To control the transistor, you need a quick turn on and off. It can be done
logical elements.
Do you want to control the ignition with these pulses?
But the question arises of getting started (start)!

#### patpin

Joined Sep 15, 2012
401
I understand that you want, that when approaching ferrite
the amplitude decreased. But in my last scheme there is an increase
amplitude. Those. At the moment of passing the rods, a pulse is generated.
To control the transistor, you need a quick turn on and off. It can be done
logical elements.
Do you want to control the ignition with these pulses?
But the question arises of getting started (start)!
Indeed first start wont be easy but with a strobe I can adjust the timing. The voltage coming from the pickup coil is only for tachometer and firing of the plugs. On the modulated signal I filter 600Khz away and feed a comparator detecting the voltage going lower and interrupting a positive pulse to the gate of the IGBT (which was loading the coil). You think I better use a totem pole as described in http://microcontrollerslab.com/totem-pole-tutorial/ ?
And if you have a second... please look at #50. That seems to be your expertise domain!

#### Bordodynov

Joined May 20, 2015
2,906
Please allow me one more question concerning the coil L1 #46.
Which core would you prefer (maximum height 13mm and thickness 8mm). The reluctor wheel (with the 11mm heigh and 1mm thick ferrite rods) passes at 0.5 mm from the core.
And how to the wind the coil on it? 60µH I guess?
For the initial diagram I had used an E shape one with 3 windings on the legs but then it was a transformator, (pickup coil at the center E-leg and a 60µH winding on upper E-leg and a serial 10µH winding with opposite phase at lower E-leg; this gave me a good lowering of the amplitude at the pickup coil when the ferrite rod passes), but I guess with yr diagram on post 46, there are better and easier solutions.
I prefer the maximally open core, then the inductance will change as much as possible when the ferrite rods approach.But the size of the core of the coil should be close to the dimensions of the rods. I like it maximum height 13mm and thickness 8mm.
Tell me the order of work automotive coil with 2 windings: one is 1.5mH and the one where the plug is on is 15H.

#### patpin

Joined Sep 15, 2012
401
I prefer the maximally open core, then the inductance will change as much as possible when the ferrite rods approach.But the size of the core of the coil should be close to the dimensions of the rods. I like it maximum height 13mm and thickness 8mm.
Tell me the order of work automotive coil with 2 windings: one is 1.5mH and the one where the plug is on is 15H.
What do you mean by "maximally open core", a 13 mm ferrite rod?
I do not understand what you mean by "Tell me the order of work". The coil is a simple oil cooled automotive coil found in shops and to be used with a wire resistance at the +(the resistor is shunted when starting the car);

Last edited:

#### Bordodynov

Joined May 20, 2015
2,906
I do not have a car (I walk) and I did not do automobile electronics.I need to know the circuit of the coil and how this coil is loaded.Also I need to know how the processes are synchronized (the approach of the cores to the test coil and the inclusion of transistors).And yes, I meant that the winding of the test coil (in my opinion) is better on the rod.

#### patpin

Joined Sep 15, 2012
401
I do not have a car (I walk) and I did not do automobile electronics.I need to know the circuit of the coil and how this coil is loaded.Also I need to know how the processes are synchronized (the approach of the cores to the test coil and the inclusion of transistors).And yes, I meant that the winding of the test coil (in my opinion) is better on the rod.
I like to walk a lot too... Besides I work on an oldtimer.
But I like to replace the Tr by an IGBT.
I include a picture of the reluctor wheel (coupled mechanically with motor; this coupling is variable in order to adjust the advance;this is timing the spark relative to the cylinder piston positions) with the small ferrite rods (11mm height) and the E-ferrite with the 3 mentioned coils.

#### Attachments

• 112.4 KB Views: 5
Last edited:

#### Bordodynov

Joined May 20, 2015
2,906
See

#### Attachments

• 6.3 KB Views: 4

#### patpin

Joined Sep 15, 2012
401
Very nice result Bordodynov. Many Thanks!
Can this OUTD (or its inverted signal) directly drive the IGBT? The IBGT will be conducting most of the time, only to block when firing has to occur and then during about 1.2ms. (is about the average plug "spark duration" (This depend on the fuelmixture)). The rest of the interval between two firings, the coil has to load.
I guess it is safer to use an optocoupler too. Just in case the IGBT protecting Zener dies. Voltage goes to 400 V on primary windings of the car coil when IGBT stops conducting.

Last edited:

#### Bordodynov

Joined May 20, 2015
2,906
IGBT is controlled by positive pulses, i.e. it will be open for a small part of the period ~ 0.1ms. As soon as the control signal is quickly switched to zero (~50ns), the transistor will begin to close and a large voltage surge will appear on the collector of the transistor (this is the EMF of self-induction). The amount of ejection is determined by the closing speed of the transistor. The voltage can be lowered using snubber (R + C).
If the change in inductance is large, i.e.the signal at the output of the operational amplifier is large, then it is possible to throw out the separation capacitor at the input of the digital circuit (i.e., to use a jumper instead of the capacitor).Then, in the absence of rotation, the transistor will be reliably closed.I think you should start with this option (without a capacitor).

Last edited: