300v DC-DC Converter for CDI

[cybercheater]

hi..im planning to build a 12 -> 300v dc-dc converter for cdi..i found this circuit, but i dont know this circuit work or not

the transformer is 4:50 turn ratio..base on my calculation 12v x 12.5 = 150v
anyone here can explain how this circuit works?

 

[shahbaz.ele]

if transformer is generating 150V. there may be a doubler circuit at output.
Try to search doubler circuit. if anyone knows plz reply.

 

[rajudp]

for a flyback transformer (SMPS ) you have to calculate the dutycyle also

 

[cybercheater]

if transformer is generating 150V. there may be a doubler circuit at output.
Try to search doubler circuit. if anyone knows plz reply.

i think this circuit not have voltage doubler

for a flyback transformer (SMPS ) you have to calculate the dutycyle also

how to calculate voltage output if i know duty cycle?

lets say this circuit work at 50khz, primary turn @ 10uH, secondry turn @ 125uH

1/50khz = 20uS

Duty Cycle 50%
Ton = 10uS
Toff = 10uS
12v / 10uH x 10uS = 12amp
125uH x (12amp / 10uS) = 150v

duty cycle = 65%
Ton = 13uS
Toff = 7uS
12v / 10uH x 13uS = 15.6amp
125uH x (12amp / 7uS) = 214v

its correct?
sorry for my english

 

[FvM]

I don't know what you want to calculate, because the duty cycle will be varied by the controller IC. The problem is however, that the schematic won't work as shown, because it's missing the current sense, that is required by a UC3845 current mode controller. You may want to have also a loop compensation, consult the datasheet for details.

 

[cybercheater]

i know duty cycle will be varied by the controller IC..
but what a number of duty cycle will be generate by controller if output is 300v?


this circuit take from SportDevices. CDI Programmable Digital Ignition.

The SMPS UC4845 oscillator has two inputs:

* Voltage feedback is used to reach the desired voltage, when voltage in capacitor divided in the potentiometer reachs the reference voltage, the oscillator stops.
* Current sense input is normally used to control the maximum current in the mosfet. It depends on transformer inductance and oscillator frequency, but not on load (with this topology).
Here, this input is used to control the temporary switching-off the power supply while the PIC activates the SCR to discharge capacitor. When line is 0 volt, the oscillator is ON, when line is higher than 1 volt, oscillator stops. This is done for two reasons:

 

[FvM]

To operate the flyback converter in continuous mode at 12 V input and 300 V output voltage, a duty cycle of 66% would be needed. But UC3845 has a hardware limitation of 50%, so the converter will operate in discontinuous mode above about 150V output voltage. You'll get this by simply balancing the transformer flux and considering the 1:12.5 windings ratio.

If the converter will operate without damaging the switch transistor or the transformer depends e.g. on it's winding resistance and current capabality. But I strongly suggest to add a current sense resistor for reliable operation.

 

[cybercheater]

what do you mean about balancing the transformer flux?
it is possible to operate in discontinuous mode if i used EFD20 core with air gap?
what a value of current sense resistor i should use?


EFD20

 

[FvM]

what do you mean about balancing the transformer flux?

It means that the pulse areas V*t during on-state and flyback-state must be equal. If you put in the voltages and winding ratio, you get the time ratio respectively duty cycle in continuous mode.

As said, you can't achieve a voltage ratio of 1:25 in continuous mode with a duty cycle limited to 50% and the given windings ratio. In my opinion, continuous mode would be preferable to reduce the peak current. Considering the permanent variation of output voltage in operation, one should better use an UC3843 to allow a higher duty cycle (but only with current sense), instead of increasing the windings ratio

If you accept discontinuous mode, the transformer inductance has to be well considered. If it's too high, the converter won't achieve the intended power throughput. If it's low, you get high peak currents.

it is possible to operate in discontinuous mode if i used EFD20 core with air gap? what a value of current sense resistor i should use?

The circuit dimensioning has to start with a power calculation. You have to specify maximum ignition frequency and minimum input voltage for rated output. Then you set average and peak primary current, transformer inductance and sense resistor.

 

[cybercheater]

its possible if i use uc3842? this ic also can run higher duty cycle..but defferent in UVLO..do you know what a function of UVLO?

 

[FvM]

Yes, the UC38 types with high undervoltage lockout threshold (16 V) can't run at 12 V. By the way,, the UC3843 UVLO threshold also won't comply with the usual specifications for cold start, that requires the ignition systems to stay operational at low battery voltage.

 

[grittinjames]

Hi all,
In flyback converter if you not connect the te load voltage will keep increase in every pulse depends on your output cap.
No matter what is the turns ratio, duty cycle etc.
since primary inductor charges at ON time & release the energy to secondery at off time just like a charged inductor.

When a charged inductor is open. the voltage depends on its load!!

I dont know what is CDI, but still the scr is connected between o/p and to ground?

why second diode is required?

 

[FvM]

I dont know what is CDI

Capacitor discharge ignition - Wikipedia, the free encyclopedia

In flyback converter if you not connect the te load voltage will keep increase in every pulse depends on your output cap.
No matter what is the turns ratio, duty cycle etc.

It's not completely clear to me, which point of the discussion you're referring to. It's a trivial fact, that the output voltage can reach any value (after sufficient time). The problem with CDI is, that the nominal output voltage has to be achieved in time (before the next ignition pulse) without causing unacceptable inductor peak currents. The circuit dimensioning should consider this requirements.

why second diode is required?

Possibly to limit an overshoot of the discharge oscillation, that may retrigger the thyristor.

 

[RCinFLA]

It would seem you need to gate the boost switcher.

When the SCR fires it will remain in conduction until the current drops to near zero. With the flyback boost continuing to operate the SCR may never release. The object is to flush the high voltage cap charge through the ignition coil. That hv cap part of the circuit. discharging through ignition coil primary, will drop to zero current. You should also know what the ignition coil characteriztics are. 300v seems low to me for CDI. The HV cap must be a high quality mylar part with low Rs.

The SCR will also drive the flyback switcher's feedback control crazy when it short circuits the output. You need to have a primary circuit that limits the maximum duty cycle of the MOSFET driver or it may saturate the core and damage the MOSFET.

 

[FvM]

It would seem you need to gate the boost switcher.

It's actually done in the described system by the control input. Nevertheless, the thyristor will be most likely switched off during the oscillation of the LC circuit formed by capacitor and ignition coil. And the converter is only supplying current pulses.

300v seems low to me for CDI.

It's on the lower end of usual voltage range.

You need to have a primary circuit that limits the maximum duty cycle of the MOSFET driver or it may saturate the core and damage the MOSFET.

Yes. As pointed out before, operating the UC38xx controller in it's intended current mode would completely solve this issue.

 

[cybercheater]

if 300v cap is lower, what is the cap value should i use?
this converter need to fully charge the cap in 1ms..

 

[FvM]

1 µF/300 V means 45 mJ energy which is basically O.K. Regarding high current capability and low ESR, I would rather go for a polypropylene capacitor (FKP or MKP), as commonly used in power electronics.

 

[cybercheater]

may i know what a defferent between polypropylene and others type film cap? can i use polyester film?
the converter need to produce 0.3amp output to charge 1uf cap in 1ms right?
if freq about 50khz and turn ratio is 1:12, how much inductance is need?

 

[FvM]

The suggested Mylar trademark refers to polyester film capacitors. You should consult the datasheets of devices of interest, but I fear, that the dV/dt rate usually provided by polyester capacitors is insuffcient for a CDI application. Polypropylene capacitors are better in this regard by at least an order of magnitude.

0.3 A would calculate for constant current operation. Considering the input power supply, one would rather aim at constant power. It's not clear to me, which motor would achieve an ignition period of 1 ms, by the way.

The inductance choice - depends on. If you change to continuous operation mode, utilizing current sense and above 50% duty cycle, the value isn't critical, you may want to increase it to reduce the input peak current. If using the original circuit, it must be selected very carefully to avoid damage of the components. I won't suggest a number before verifying the operation in a simulation.

 

[cybercheater]

just see at sportdevice :wink:

This DC-DC converter charges a 1uF capacitor at 250 volt in aprox. 1 msec. This theorically lets the CDI spark near 1000 times per second, 60,000 rpm in a single cylinder engine (1 spark per rev), 30,000 rpm rpm in a 4 cylinder engine (1 spark each 2 revs per cylinder).

maybe i need about 4ms if the motor run @ 12000rpm..
total time is about 5ms for 12000rpm in 1 turn..
charging the cap in 4ms and 1ms for firing (discharge)..

the output current need to charge cap in 4ms is 0.075amp

Charging a Capacitor


how to calculate average current if duty cycle is ~67% or etc?

whether the calculation of the 50% duty cycle must be divide by 4?