I need help with my first flyback converter

[bonedoc]

Hey guys, I am new here. Im used to writing assembly for pic microchips, but have more recently started to learn about other things....more specifically DC to DC conversion. I have a situation where I need to convert 6V DC into 300V DC. Originally, I was going to use the LT3750 capacitor charger, but when I ordered them, I realized they were microscopic. Not a chance I am soldering that.

SO....I figured I would learn how to do this without the LT3750 device. Also, I have a transformer winder, but I hate using it. So, I ordered a few coilcraft transformers, which were actually designed for that LT3750 cap charger. Just so you know, that cap charger allows for a 3v-24v DC input and does 300V out. Here is a link to the transformer I am using, the da2032:

http://www.coilcraft.com/pdf_viewer/...ore:da2032.pdf

The primary windings are soldered in parallel. The turn ratio is 1:10 and the inductance is .013 on the primary, and 1.6 on the secondary.

I attached a picture of the circuit I am trying to build. Instead of the switch, I am pulsing the transistor with a pic ccp module at 4000Hz. I verified that this is accurate with my scope.

Here are my problems, and excuse me if they are stupid.

-On hand, I have some 2sd965 NPN, and some IRF530 N channel power mosfets:

http://www.datasheetcatalog.org/data...JC00200BED.pdf

http://www.datasheetcatalog.org/data...ild/IRF530.pdf

These are both massively overheating. I dont understand why if I am simply pulsing 6V through them. I have a 1k resistor to the base / gate from the pic. I have 6V going to the primary winding an then to the NPNs I am testing. In the case of the 2sd965, the output of the transformer primary goes to the collector and the emitter goes to the ground. In the case of the IRF530, the output of the transformer goes to the drain, and the source goes to the ground.

Am I just trying the wrong transistors? If so, what is a good selection?

-My last question is this, regarding the secondary side. If I have everything running on the primary side, and I dont have a load on the secondary side, should I not be able to measure an AC voltage across the windings? When I test, do I need to measure from secondary end to secondary end, or from one secondary end to ground? Also, if I have no load but have a high speed diode on the secondary, should I not see a large DC voltage from the anode to ground?

Sorry for all the stupid questions...Im going crazy and new to this! I think I have read everything I can, but dont get something.

 

[davenn]

HI there

you circuit doesnt show how you have the transistor connected into circuit
show us the complete circuit on how you wired it up

Dave

 

[bonedoc]

Hello, thanks for the response. I have an IRF510 in place of the switch. The transformer connects to its drain, and the sink to ground. I dont have any load or cap on the secondary side, just a high speed diode. I am hesitant to place a cap because I didnt want to overload anything yet.

I want to use this circuit to charge a 300v and 120uF cap. Would a capacitor of this size also be suitable to reduce ripples?

Also, I am pulsing at 4k Hz. I just read that low frequencies can cause overheating of the transistor, and that I should go for 100k Hz.

Thanks for any advice.

 

[Resqueline]

The coilcraft link is not working. Actually none of the links are good.
Primary inductance of 0.013 what? Microhenries or millihenries?
You can pulse it with 1Hz if you like, but the important value for flybacks is the pulse width.
If you drive it with a 4kHz 50% square wave then that is a pulse width of 125 microseconds.
6V in 13 µH gives a slope of 0.46 A/µs. This results in a (theoretical) peak current of 57.7A at the end of the 125µs period.
This high current, together with a low gate drive voltage, plus a possibly slow turn-on/off (due to the 1k gate series resistor), can give rise to overheating.
Notice also that the energy stored in the inductor by each pulse has to go somewhere when the transistor turns off. It'll enter the transistor if not conducted elsewhere.

 

[bonedoc]

Thats strange, maybe the inks were disabled.

Sorry, I left out the details. I was trying a 4000Hz pulse with a 50% duty cycle. I will give it a go at 100kHz.

I made a big mistake. I meant the primary IMPEDANCE is .013 ohms and the secondary is 1.6 ohms.

I attached an image of its data below. Thanks for helping. Could you possibly show how you get your calculations? It seems as though many of the tutorials never really get to the point, and I am lost in all the calculus.

 

[Resqueline]

The links contains 3 dots where it should've had a lot more text. You have copied an abbreviated link text line instead of the link itself.
0.013 is not the primary impedance, it's the primary DC resistance (Ohms), that's a big difference.
Also, an inductor impedance reference (in Ohms) needs to be accompanied by a frequency if it's to have any value.
The datasheet says the primary inductance is 10µH, dropping to 9µH at 3A (which is the peak current level you should design for).
Max pulse width with a 6V supply is thus 3A * 10µH / 6V = 5µs. (With a 12V supply it'd need to be half of that, of course.)
This could seemingly be accomplished by a 50% 100kHz but it could run the flyback into continous-mode operation (and hence saturation & overcurrent).
A flyback is usually operated in discontinous mode - that is the secondary is given sufficient time to dump all the stored core energy before a new pulse follows.

 

[bonedoc]

Ok, that makes more sense. Thanks. Nobody has put it that way.

I will try this out later today. How does one determine the duty cycle to cause discontinuous use? Is there a general rule of thumb?

With a max pulse width of 5uS, that is 200kHz. Should I hit the frequency instead?

 

[jackorocko]

You say flyback converter, but isn't this what they also consider a dc - dc boost converter? I actually find these sorts of threads very entertaining, tons of information. Resquline is a god amongst mere mortals when it comes to power supply design.

 

[duke37]

Most power supplies are happy when feeding an open circuit and do not like a short circuit. The switch mode supply that you have is the opposite of this, it is happy running into a short circuit and does not like an open circuit. You should therefore always use a load to provide somewhere for the energy to go.

As Resqueline says, you can use a very short pulse with a low repetition rate to start with.
The LT3750 has a current measuring resistor to turn off the switch when the current has risen to the maximum permitted value. I believe that your computer drive does not have this so that excessive current can result. Also the LT3750 measures the output voltage and stops pulsing when the capacitor voltage reaches the desired value.

 

[(*steve*)]

You say flyback converter, but isn't this what they also consider a dc - dc boost converter? I actually find these sorts of threads very entertaining, tons of information. Resquline is a god amongst mere mortals when it comes to power supply design.

This is a particular type of switchmode DC-DC converter.

 

[bonedoc]

Thanks for the advice guys! I am getting pretty decent charge rates.

I tried 100kHz at 30% duty and it was not that bad. I then tried 40% and 45% duties..and even faster. I can charge a 120uF 330V cap in a few seconds. No sign of heat. You mentioned a 5uS frequency so I tried that at 40% duty, and it was slower. I also tried 200kHz. Maybe this core and transformer were designed at an optimal frequency of 100kHz....at least that is what the datasheet specs were tested at.

You mentioned that when charging stops, that the IRF730 would see a large voltage. How should this be handled? Should a I put a cap from the drain to the ground? If so, how do I choose a value? Is it pretty arbitrary?

What about snubbers?

You said that my 1k base resistor was too high. What should I use? Other circuits had 220v with a 1.5v input, so I thought 1k was fine for 6v in.

I dont need anything fancy, but I will most likely turn off the charging when the cap reaches a certain voltage. I will probably sense it with a voltage divider to the ADC.

Thanks for the advice!

 

[Resqueline]

A flyback delivers a specific energy packet [Joules] after the end of each conduction cycle, so more frequent conduction cycles means more power delivered [Watts].
It takes a certain number of "Joule packets" to fully charge the big capacitor. W = J / second, J = W * seconds, W = Joules per cycle * number of cycles / seconds.
The frequency in itself is not important other than as a measure of repetition rate. It's the time periods involved that matters.
For 6V 5us is the max conduction time for a 10uH 3A coil, for 12V it's 2.5uS.
5us on & 5us off = 100kHz & 50% duty. 2.5us on & 2.5us off = 200kHz & 50% duty.

As the capacitor charges the transistor also sees higher peak voltages. With no load there's almost no limit to the voltage produced.
But if you just stop the base/gate pulses when the cap is fully charged there's no ill effects.

Snubbers are used if the transformer leakage inductance is large enough to cause spikes exceeding the transistor voltage ratings.
Faster switching (risetime) exaggerates the problem. If the spikes exceeds transistor ratings it may fail after some time (a number of cycles).
A MOSFET has a certain gate capacitance (around 2nF) and this requires a certain current to be charged in a certain time. A 1k resistor severely limits this current.
1k * 2nF = 2us which is a considerable portion of the cycle time. A tenth of that would be more acceptable.
Then there's the gate voltage itself needed to turn on the transistor fully. 5V is not quite enough for standard MOSFETS. They need 10-15 or even 20V to meet spec's.
If you drive it directly from a 5V PIC you may need to use a logic-level MOSFET instead.
On the other hand, a Darlington transistor may be quite happy with a 1k base resistor driven from 5V.

But as long as the circuit does the job and stays cool you're probably ok, and it may be good enough.

 

[bonedoc]

Thanks! That really does help. I was reading the datasheet for the IRF730:

http://www.datasheetcatalog.org/datasheet/SGSThomsonMicroelectronics/mXtxuvq.pdf

It gives a gate charge value, but I do not see any farad values. However, it does give a "Turn On Time" of 11.5nS with 4700Ohm gate resistor. So, .0000000115 / .000005 = 2.3% of the cycle time. Is this calculation correct?

As fat as voltage goes, it has a Max Gate Threshold Voltage of 4V. Should my 6V with a 1k resistor therefore be fine?

My last question is this. How can a person tell how long it should take on the secondary side to charge a capacitor to a given charge?

 

[Resqueline]

No, it says it gives a Turn-on Time of 11.5 ns with a 4.7 Ohm gate resistor..
It seems to have an input (gate) capacitance of 700pF.
It is a high voltage device so it has a lower capacitance than the low voltage devices I was thinking about.

As you can see from the Output and Transfer Characteristics it will barely conduct 5A at 6V gate (and drain). It needs at least 9V gate to conduct fully.
And at 3V drain it can only conduct 4A no matter what gate voltage. To me it would seem to limit your transformer performance somewhat.

Google capacitor energy for formulas (like these for example). W [J] = 1/2 * C [F] * V squared [V], and P [W] = W [J] / t
A given capacitance charged to a given voltage contains (& requires) a given energy [J].
If you know the energy transferred by the transformer by each cycle you can calculate the number of cycles required and hence the time it takes.

 

[bonedoc]

Thanks. That really helps. I was looking a flyback by LT. It is the LT3750:

http://www.linear.com/product/LT3750

I notice that on the primary transformer input there are 3 caps. One 10uF and 2 56 uF caps. What are the functions of these?

I did some experimenting with my frequency generator for the heck of it. It turns out that I could charge extremely faster at about 50kHz. As I approach 100kHz, it becomes slower but remains cool. If I go lower, it becomes slower and extremely hot.

So, I set my pulse at 50kHz and then tried different duty cycles. It turns out that it also charges a lot faster at about a 30% duty cycle.

I think this shows how the equations and theories are working. However, LT says their unit charges 100uF to 300v in 1/3 sec. Off a 6V battery supply, I hit this in 9 seconds. When I use my signal generator at 10V, it is even quicker. Surely the charge time difference is from the input voltage differences? In other words, their unit allows up to 24v. I imagine they are getting this performance largely from a higher input voltage?

Thanks for all of your help. I think I am about done asking questions ;-)

I am just wanting to fine tune this a little more so that it is simple yet fast. I know you mentioned increasing the input voltage and base voltage. Should I select a different transistor that allows more current flow at lower voltages? I suppose I could use a 9-12v AA input instead of a 6V D battery pack, but I hate to do that. I also ordered the other transformers that have an Ipk of 5A and 10A. I dont know if this matters since my current flow is limited by my transistor.

Anyway, just wanting to find out the purpose of the caps on the input, and any other suggestions. Maybe I am just maxed out on charge speed because of my input voltage and I could charge faster with a higher input. Ian

 

[Resqueline]

The capacitor functions are explained in the datasheet. It's called power supply decoupling and is used to ensure a low power supply impedance.
A 1-10µF is supposed to go close to the chip to ensure a stable supply to it, and a bigger close to the transformer to ensure a sufficient current pulse capability.

The transformer has a 1:10 ratio, so if the cap is charged to 300V the transistor will see 30V (plus a short spike due to the leakage inductance).
Using a 400V MOSFET is overkill - and even detrimental to the performance. In the LT3750 design note they used a 100V 34A 0.019Ohm transistor.
Such a MOSFET will only have a 0.06V drop at 3A. A darlington transistor is an option but will cost you around 1.5V c-e drop.

I guess a slow switching speed together with a high Rds-on resistance slows down the current ramp-up in the transformer so that it benefits from a lower frequency.
A better transistor and a proper gate drive circuit would negate this effect, except for long pulse widths which will still saturate the transformer and cause current spikes.
Varying the gate voltage varies Rds-on which varies the d-s voltage drop, which varies the current ramp-up time.

Doubling the supply voltage will charge the transformer twice as fast, and so you can attain a higher switching frequency, resulting in a faster capacitor charge.
You'll need a better transistor (+ a better gate drive circuit) to benefit from a higher current transformer (and to get faster charge rates).

 

[bonedoc]

Thanks! you have been a huge help!

 

[GonzoEngineer]

The gate drive of your device, whether a MOSFET or an IGBT, is the most important thing you you need to check if the device is getting hot.

The most common mistake people make is not understanding that the gate drive determines the turn on time, because you are essentially driving an capacitor.

The slower the turn on time, the more heat you get......it's called "Switching loss"

"Google" TC 427......power the 427 with 15 volts, and have at least 10uF backing it up, and you should be fine!

The lower the gate resistor.....the faster you switch!

 

[bonedoc]

Hey guys, good advice. Everything is running pretty good, and have a couple new questions now that I am playing.

-As far as the PWM signal to the transistor, should there be an optocoupler to protect the microchip providing the PWM signal?

-Everything is charging good. When the cap is charged on the secondary side, and I discharge it, I notice something odd. If I discharge the cap to ground, nothing odd happens. But, I want to learn how to discharge high voltages with a 2N6509 SCR. This is where the problem is. I have the load on the secondary side going to the ANODE of the SCR. Ground is directly to the CATHODE of the SCR, and I apply 6V (no resistor) to the gate of the SCR to control it. When I apply 6V to the gate, the load flows through the SCR, but when I let the 6V signal off of the gate, I see a problem. My capacitors no longer charge and the flyback transistor on the primary side overheats. Everything "resets" if I turn everything off and on again. I am not sure what is happening. Does anyone have an idea? This ONLY happens when I discharge my load through the SCR, so it has to be related. It almost acts like the SCR is stuck on the on state, and the load/secondary side of the transformer in the DC to DC converter is overheating the transistor on the primary side. Any thoughts?

http://www.datasheetcatalog.org/datasheet2/5/0q9t16ckj945igaokck1l9pyxo7y.pdf

 

[bonedoc]

I just read this, and I think this is the problem, since I do have induction. Can anyone tell how I prevent it from being stuck in the on state?

------------------

"Commutating dv/dt and di/dt

The commutating dv/dt rating applies when a TRIAC has been conducting and attempts to turn off with a partially reactive load, such as an inductor. The current and voltage are out of phase, so when the current decreases below the holding value, the triac attempts to turn off, but because of the phase shift between current and voltage, a sudden voltage step takes place between the two main terminals, which turns the device on again.

In datasheets, this parameter is usually indicated as \left ( \frac{\operatorname{d}v}{\operatorname{d}t} \right ) _c and is generally in the order of up to some volts per microsecond.

The reason why commutating dv/dt is less than static dv/dt is that, shortly before the device tries to turn off, there is still some excess minority charge in its internal layers as a result of the previous conduction. When the TRIAC starts to turn off, these charges alter the internal potential of the region near the gate and A1/MT1, so it is easier for the capacitive current due to dv/dt to turn on the device again.

Another important factor during a commutation from on-state to off-state is the di/dt of the current from A1/MT1 to A2/MT2. This is similar to the recovery in standard diodes: the higher the di/dt, the greater the reverse current. Because in the TRIAC there are parasitic resistances, a high reverse current in the p-n junctions inside it can provoke a voltage drop between the gate region and the A1/MT1 region which may make the TRIAC stay turned on.

In a datasheet, the commutating di/dt is usually indicated as \left ( \frac{\operatorname{d}i}{\operatorname{d}t} \right ) _c and is generally in the order of some ampères per microsecond.

The commutating dv/dt is very important when the TRIAC is used to drive a load with a phase shift between current and voltage, such as an inductive load. Suppose one wants to turn the inductor off: when the current goes to zero, if the gate is not fed, the TRIAC attempts to turn off, but this causes a step in the voltage across it due to the afore-mentioned phase shift. If the commutating dv/dt rating is exceeded, the device will not turn off."