Ferrite core transformer design step by step

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
:crazy: Chocoholic I must say you scared away my new friend Tony,

.....then you have to answer those :h_ache: thats the price you will pay for scaring away my friend :D.


I guess your friend was just a little bit shocked about changing direction every two posts and might come back after seeing that you remain stable in his playground.

In fact I did not promise my support, because I knew that I do not have the to time to act like a responsible personal teacher. Why should I make false promises to you?

90% duty cycle:
I understand your concern regarding resolution. In fact resolution can be a limitation in performance, but unfortunately power physics are pretty ignorant about what we think would be easy to implement by an uC.

In general it is a pretty advanced task to generate the PWM for the switches of a SMPS directly by the uC. You need to know every detail about the power stage, about your uC and about the relevant part of control theory. A few wrong pulses and your power stage will burn (sometimes already a single wrong pulse is enough).
I would clearly recommend to start with simple standard circuits and related control chips.

Have a look to this page. It offers information and simulation hints for multiple different dc/dc converters.
dc to dc converter design / switched mode power supply design

No matter if we are talking about the power stage or the control, I have to restrict my support to the basics.
 
OK... Chocoholic talked be back down off the ledge. Flybacks I understand well.

I concur with Chocoholic on the use of a standard controller for your first try at this. I can get you through the power stage but you would be on your own for the firmware because I have restrictions on what algorithms I could share.

Lets start with the basics.There are really 4 knobs to turn simultaneously on a flyback transformer. They are Bmax, DeltaI, Duty Cycle and the minimum voltage on the bulk cap. The turns ratio is a ouput of above... not an input. I will explain that a little bit below.

For Bmax you can assume 3700 gauss as an operating point. This is the flux density at the lowest input voltage and peak power. This will typically be the point you place the hardware OCP. If you look up most materials you will find a saturation flux density of about 4000 gauss at 100C. 100C operation is the lowest core loss point fore most materials because that is the way they are intended to run. You will probably hit 100-105 on the wire over the gap inside the transformer where the fringing flux lines cut through the wire. I use 3700 gauss because you will still need some margin due to lot-lot differences in the material.

DeltaI, is the current that you put through the primary winding during the MOSFET on time. Idc=Popk/(Vinmin*DutyCycle). For DCM operation DeltaI is 2*Idc. For CCM Operation DeltaI is a percentage of Idc (this is arbitrary but typically between 25-50%). CCM operation is the easiest on the transformer since Ipk is lower.

Duty Cycle should be around 50% and not more than 75%. Remember that for the Flyback portion of the cycle the on-time for the secondary side rectifier is 1-DutyCycle so when you skew the duty cycle too far to one side the peak current on the rectifier can become very large. This is why I try and make the ducty cycle no more than 65%. 62.5% is usually my target.

Finally we have Vbulkmin. For an AC rectified line the haversine (lowest point on the rectified bulk cap voltage) can be very low. For example for 85W and 90V operation you would drop to 61V at the bottom of the bulk cap voltage. This voltage in combination with teh duty cycle sets you turns ratio.

So, as you can see, we have 4 variables at the top but given the limits as I have explained above the only real variable is the minimum bulk cap voltage Vinmin.

Once you define Vinmin, Popk (Peak output power) and the output voltage we can start core selection.

Tony
 
Hello Tony, Thanks for coming back here.
So We start with learning example?

Mode = Discontinuous
Power = 85 Watts
Vmin = 90Vdc
Vmax = 100Vdc
Dutycycle = 60%

So first calculating DeltaI
Idc=Popk/(Vinmin*DutyCycle)
for DCM DeltaI = 2*Idc
85/(90vdc * 0.6) = 1.57
so DeltaI for DCM will be 2 * 1.43 = 3.14 A

Correct ?

Now to Vbulkmin. Is the lover voltage (61v) calculated from ripple voltage?

eg needed ripple voltage is 29V so 90-29 = 61V.please correct me .

how was Cap value calculated ?

you can also proceed with selection.

Thanks Tony :)
 
I think I made mistakes In taking above example kindly ignore it .

below is the corrected one. Tony please correct me if I am wrong.

Mode = Discontinuous
Power = 85 Watts
Vmin = 200Vdc
Vmax = 220Vdc
Vout = 90Vdc
Dutycycle = 60%

So first calculating DeltaI
Idc=Popk/(Vinmin*DutyCycle)
for DCM DeltaI = 2*Idc
85/(200vdc * 0.6) = 0.70
so DeltaI for DCM will be 2 * 0.7 = 1.41 A

if this time example is correct please tell me how we calculate the cap for Vbulkmin. I mean using the values in equation.
 
Last edited:
OK... 90Vdc isn't that bad. Though you will have to tap the output winding to get the TL431 regulation correct (max TL431 regulation is 32V).

So the bulk capacitance you use is a function of holdup time. I usually use 4ms but you may want 10 so that you can survive a single half cycle dropout. Server us as much as 20ms.

Cbulkmin=(2*Po/efficiency(85%)*Tholdup/((230*sqrt(2)-2*ForwardVfBridgeRectifier)^2-Vbulkmin(voltage you expect to run down too - usually your brownout point)

All calculations after this will use Vbulkmin and Pi=Po/0.85 expected efficiency.

Choose Cbulkmin>>than the answer above. I would huess somewhere around 100uF for 230V operation only.

Idc=Pin/(Vbulkmin*DutyCycle)

DeltaI=2*Idc for DC operation. (cor CCM operation DeltaI = Idc*25% as an example where 25% is that amound of ripple current you want to appear on the output).

Ipk=DeltaI; (For CCM Ipk=Idc+DeltaI/2)

Ton=DutyCycle/Fsw where Fsw=the switching frequency

Lp=Vbulkmin*Ton/DeltaI (Notice I didn't say Ipk. This is because for CCM operation you can significantly increase your inductance and lower Ipk)

Notice up to this point we don't care about the Ae of the core. As I said before you can run any power on any core if you can get enough wire on the core.

We will pick up core selection tomorrow. However... I will give you a hint, you will need an Ae of around 98 for 85W. we'll go over the reasons why in the next few days. See if you can find local material in that range. Not much smaller but bigger is OK.

Tony
 
So the bulk capacitance you use is a function of holdup time. I usually use 4ms but you may want 10 so that you can survive a single half cycle dropout. Server us as much as 20ms.

Cbulkmin=(2*Po/efficiency(85%)*Tholdup/((230*sqrt(2)-2*ForwardVfBridgeRectifier)^2-Vbulkmin(voltage you expect to run down too - usually your brownout point)
Tony

Is this bulk cap related to mains line frequency? cause 10ms is halfcycle for 50hz and 20ms is full cycle for 50Hz.

All calculations after this will use Vbulkmin and Pi=Po/0.85 expected efficiency.

Choose Cbulkmin>>than the answer above. I would huess somewhere around 100uF for 230V operation only.

Idc=Pin/(Vbulkmin*DutyCycle)

Does Pi mean to be Pin ?

Lp=Vbulkmin*Ton/DeltaI (Notice I didn't say Ipk. This is because for CCM operation you can significantly increase your inductance and lower Ipk)

Lp is primary inductance in henry ? or something else.

See if you can find local material in that range. Not much smaller but bigger is OK.

Yes, I am talking to a local dealer who can provide me TS of the core he sells.

but for the time lets use datasheet of some well known core mfg. for our reference. this will help me to understand the reading of core and material datasheet and its practical implementation.
 
...please do not kill yourself with high voltages.
I am not sure about Tony's view, but I would enjoy this learning project much more at voltages below 60V.

Hey Chocoholic,

For me any voltage is low, unless I knows the designing hazards for that specific voltage. Can you please tell me about it. so that I can be cautious in selecting voltage?

:radar: I am only in a scan mode now trying to learn this. And very much greatfull thankful to you and sply Tony.
 
Is this bulk cap related to mains line frequency? cause 10ms is halfcycle for 50hz and 20ms is full cycle for 50Hz.



Does Pi mean to be Pin ?



Lp is primary inductance in henry ? or something else.



Yes, I am talking to a local dealer who can provide me TS of the core he sells.

but for the time lets use datasheet of some well known core mfg. for our reference. this will help me to understand the reading of core and material datasheet and its practical implementation.

You are correct on the Mains Frequency. 50/60Hz if you are connected to a system with a batter then use 4ms otherwise use 10-20ms depending on what you can afford or how importatant it is to have uninterrupted behavior.

Pi=Pin sorry for the bad typing.

Lp is in henry.

A well known and fairly easy to get transformer is the RM10. Lets use Mag Inc or Feroxcube.

Tony
 
Hey Chocoholic,
For me any voltage is low, unless I knows the designing hazards for that specific voltage.

This was the answer I was afraid of.
The designing hazard is simple: :RIP:

Never got zapped when changing a light bulb, or when fingering bread out of your toaster, or by a defekt power outlet?
If no, you are really a virgin. And I can tell you: DON'T even think about trying! At 230V/50Hz the chance of a lethal accident is already pretty high!
It depends a little bit on the country and safety standard which levels are considered already as dangerous voltages.
Above 60V DC you can be sure to be in danger.
Up to 60V DC is still a typical DIY range, where I have the perception that lethal accidents happen not so often.
Since more and more experience is gained by accidents, the voltages that are considered uncritical are going down over the years.
As far as I know 24V DC are considered uncritical in almost every country and safety standard.

Here one general info that states above 50V to be critical.
High voltage - Wikipedia, the free encyclopedia

Finally it is up to every DIYer to take care on his own.
The five safety rules / Industrial safety - Economy-point.org
 
..... so that I can be cautious in selecting voltage?

Are you free in selecting the voltage?
To me it appeared that you and Tony were discussing offline applications for 120V/AC and 230V/AC as well.

I do not want to stop you from doing this, you just should be aware that you are entering the big boys playground. Tony is a pro, but he cannot protect you over a cyber remote chanel.
 
Ha Ha , Yes chocoholic, I am well baked of 230vac. its lot of times I had experienced it.
and now finally I got ELCB installed at my home.

I had designed some dc to ac converter for home use rating was 500W 230 Vac 50 Hz. using CRGO. the magnetics was done by local transformer winder. In this project i had used SG3524 With mosfet at the output and the stepup transformer in pushpull config.
this served me for 5 years and then the Leadacid battery gave up.

I very much understand your concern about safety. And thank you for that. I agree that although I am baked of high voltages many times, Playing with such high voltage is not a part of domain. I am into embedded domain. and into design of uc based systems. so very much into hardware and its firmware.

While you asked me to use low voltage, I thought you are talking of some insulation breakdown of the winding. and some special care that need to be taken.SO i said any voltage is low for me unless I understand the designing hazard.

:)

I am very much eager and waiting for Tony's next words. Sure this thread would be great learning for the many other diyaudio members as well.
 
I've done a few HV step-up switchers but it was long ago and I only remember a few issues. One is that core losses are published at 50% duty cycle. If you use a topology that runs at a high duty cycle, core losses may be far higher than anticipated. I like flybacks, but the requirements on the output diode can be extreme. Fortunately the selection is far better there than it used to be. Most of the design info you'll find is for low voltage supplies and works poorly for HV step-ups. McLyman is probably the best book I own, but it does have a couple formula errors and reveals little about optimizing HV step-ups. Expect to learn new things about alternate winding designs for transformers to control capacitance and stray inductance. Expect to make smoke a few times. Expect lower efficiencies than LV switchers and be sure you know exactly where that extra power gets dissipated! IMO, if you can account for every mW from input to output, and understand why the dissipation occurs, success is close by.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.