Not taken
You are talking about coupled inductor design (aka. flyback 'transformer', energy storage device).
I have given a formula for transformer as used in pushpull, halfbridge or full bridge topologies, which should be a good starting point for someone who's just asked how to design one on an internet forum.
Yes to the coupled inductor. They are wonderful things when done right. There is a thread I responded to a few days ago using Dixons I believe it is located here. http://www.diyaudio.com/forums/powe...ake-coupled-inductor-using-toroidal-core.html
Tony
You will use two synced PWM output 180 out of phase (hope i used the right digital terminology there). They must be locked together. Yes that is indeed 45% on and 55% off. You can do a phase shifted full bridge but that is a much more advanced control that I think you want to go for.
Whose simulation package are you using? PSIPCE (My favorite sim package is solder!)?
Tony
I have done using LTspice but that was without microcontroller. now for microcontroller I am using Protues.
ok lets hold this for a moment. lets not consider this whole microcontroller based design. I will put a simple schematic based on sg3525. this will have a pot to vary its duty cycle. give me an hour.
Here is a diagram.
so to simplify lets just follow this diagram. ( even keeping bridge conf aside).
please guide me how do I go with this.
Vin = 10V.
Vout = 100V ( peak)
VA = 100VA.
as you said before considering losses to be 20% lets say VA = 120
so current is 12Amp.
Sir, guide me for next.
so to simplify lets just follow this diagram. ( even keeping bridge conf aside).
please guide me how do I go with this.
Vin = 10V.
Vout = 100V ( peak)
VA = 100VA.
as you said before considering losses to be 20% lets say VA = 120
so current is 12Amp.
Sir, guide me for next.
Attachments
Last edited:
OK.. so what you have drawn won't work for a full bridge. You have more of a forward configuration (though that would need a demagnetization winding).
Take a look around and look for a full bridge schematic and then at a forward. We'll then discuss what might be the better approach and the pro's and cons of each.
Tony
Take a look around and look for a full bridge schematic and then at a forward. We'll then discuss what might be the better approach and the pro's and cons of each.
Tony
Yes Tony I agree, What I have drawn wont work in full bridge. yes its more of forward configuration. for fullbridge the reverse supply works to give demagnetization. right ?
also tell me is the 55% OFF in dutycycle used to self demagnetization ? did I understand it right ?
I will put other design of HF inverter. which is available on net. the only difference is I am achieving the same using micro-controller. The reference file
You can say I will need transformer for the design.
...going for the simplest thing and immediately 2kW as a start up project?
Note: The paper does not show where from they get the 170V DC.
Your questions sound to me like you would intend to run the full bridge from 12V and step up by the transformer behind the full bridge?
Theoretically possible, but not advisable. At this point of the circuit the transformer does not just have to handle HF but also the full amount of the intended LF sine wave. This heavy LF content is brought there on purpose by the PWM modulator.
Unfortunately it would need a transformer that can handle a huge time voltage product ==> high number of turns and/or huge core. Like a 50Hz transformer or 60Hz transformer for 2kVA.
Better do a two stage approach.
Stage 1: DC/DC from 12V to 170V
Stage 2: Sine wave inverter similar as described in the paper
@Toni:
... just had to point out that this is going to be an extensive project.
Sure you already asked for holidays in order to support this thread.
Note: The paper does not show where from they get the 170V DC.
Your questions sound to me like you would intend to run the full bridge from 12V and step up by the transformer behind the full bridge?
Theoretically possible, but not advisable. At this point of the circuit the transformer does not just have to handle HF but also the full amount of the intended LF sine wave. This heavy LF content is brought there on purpose by the PWM modulator.
Unfortunately it would need a transformer that can handle a huge time voltage product ==> high number of turns and/or huge core. Like a 50Hz transformer or 60Hz transformer for 2kVA.
Better do a two stage approach.
Stage 1: DC/DC from 12V to 170V
Stage 2: Sine wave inverter similar as described in the paper
@Toni:
... just had to point out that this is going to be an extensive project.
Sure you already asked for holidays in order to support this thread.
Hey Tony you don't need to say that, you were helping me wholeheartedly. Although Inverter transformer is not a part of your experience.It dosent matter to me.
I am here to understand the basic design of transformer My knowledge is total zero with this.the reason why I as sticking to flyback transformer. so that I get something to start with.
and flyback transformer is used in most of our powersupplies.I wish to understand the working of the transformer and the consideration while design with some practical design.
So can you please help me with a basic flyback transformer design?
we would go with basic discontinious type of flyback transformer.Is it ok? can you help? I need you .
I am here to understand the basic design of transformer My knowledge is total zero with this.the reason why I as sticking to flyback transformer. so that I get something to start with.
and flyback transformer is used in most of our powersupplies.I wish to understand the working of the transformer and the consideration while design with some practical design.
So can you please help me with a basic flyback transformer design?
we would go with basic discontinious type of flyback transformer.Is it ok? can you help? I need you .
Please go through this example design
I am not able to understand the 4th step here and last step ie finding the thickness of wire that would serve our current (ampears) need.
I have some doubts regarding topology.
1) can the duty cycle of flyback transformer operating in discontinuous mode not have duty cycle more than 50%
2) how is discontinuous different from continuous mode in terms of utilization, and application area, advantages and disadvantages.
I am not able to understand the 4th step here and last step ie finding the thickness of wire that would serve our current (ampears) need.
I have some doubts regarding topology.
1) can the duty cycle of flyback transformer operating in discontinuous mode not have duty cycle more than 50%
2) how is discontinuous different from continuous mode in terms of utilization, and application area, advantages and disadvantages.
Last edited:
Hi Acousticat!
I hope that I did not kick down this thread by pointing out that the intended project will be extensive.
It is absolutely great that you are looking for technical discussion.
Also very wise that you are not insisting on extreme power.
I would love to see thread going on.
If you follow my proposal of a two stage approach, then the first stage with the DC/DC converter might still fit to Tony's experience.
Hopefully he is coming back.
The second stage has tons of commonalities with class D amps, consequently support at DIYaudio is likely to come.
Your idea of going for a flyback is possible.
There are many others as well, which fit also.
I.e. Push pull converter like often used in car Hifi SMPS
or a single ended forward converter.
From terms of simplicity a flyback (check out for dedicated controllers) or a push pull controlled by the SG3525 should be both fine.
The push pull is likely to have slightly less losses in your application, I guess.
I hope that I did not kick down this thread by pointing out that the intended project will be extensive.
It is absolutely great that you are looking for technical discussion.
Also very wise that you are not insisting on extreme power.
I would love to see thread going on.
If you follow my proposal of a two stage approach, then the first stage with the DC/DC converter might still fit to Tony's experience.
Hopefully he is coming back.
The second stage has tons of commonalities with class D amps, consequently support at DIYaudio is likely to come.
Your idea of going for a flyback is possible.
There are many others as well, which fit also.
I.e. Push pull converter like often used in car Hifi SMPS
or a single ended forward converter.
From terms of simplicity a flyback (check out for dedicated controllers) or a push pull controlled by the SG3525 should be both fine.
The push pull is likely to have slightly less losses in your application, I guess.
Hi Chocoholic,
Its over long discussion but still no real start. I agree I ts my mistake that I did not present things properly. And so I am seeing this moment.
As said I am forgetting my inverter thing. I just want to start understanding the design of a flyback transformer. so for study purpose I found an example sheet on net. which I have posted in thread no.32
While studying the example I found that core selection was not explained well for understanding. so I have put it as a query here. Also I have some other doubts which I couldn't find on net readings. You too design transformer and has knowledge. please It would be nice if you can reply the thread no 32.
Its over long discussion but still no real start. I agree I ts my mistake that I did not present things properly. And so I am seeing this moment.
As said I am forgetting my inverter thing. I just want to start understanding the design of a flyback transformer. so for study purpose I found an example sheet on net. which I have posted in thread no.32
While studying the example I found that core selection was not explained well for understanding. so I have put it as a query here. Also I have some other doubts which I couldn't find on net readings. You too design transformer and has knowledge. please It would be nice if you can reply the thread no 32.
1) Basically, you can operate a flyback in dcm with a duty cycle above 50%.
It is just a question of the relations input voltage / output voltage and primary turns / secondary turns.
But why do you want to do this?
2) In continuous mode you never fully demagnetize the core, while in dcm you always reach zero (neglecting the remanence of the core material) magnetization in every cycle.
From perspective of control theory the continuous mode has one additional hidden integrator in the transfer function, requiring a more clever control loop - especially in designs that have to handle both...
From power transfer perspective, the continuous mode has the advantage that the ratio of required rms currents in the windings vs transfered power is more fortunate than in discontinous mode, but especially at higher output voltages this advantage can be eaten up by the poor properties of the diode on the secondary, because in continuous mode this diode is being switched hard....
I think for starting you should chose dcm, because it is easier to debug and more forgiving regarding short comings in control.
Theoretically you could even go for a design in critical conduction mode, means operating just at the max duty cycle which will not go into continuous mode....and you can even extend this into a lossless snubbered version, I think some years ago I posted a thread about this particular version... somewhere here in the forum.
Tony? Don't let me alone!
@ Tyco paper,
page 1 & 2 , they have just tried to explain discontinues mode and continuous mode in very short.
On page 3 they have provided with equations for the calculation for flyback transformers.It also gives Peak current and power output
calculations specific to discontinues mode.
on page 4 they have provided example design of discontinues mode. In this the 4th step is select core. its as given below.
4) Select core
In this example we will use a current density of about 300 c.m/A. Since lpp =
2.22A, a total c.m will be 300*2.22 = 666 c.m. From wire chart, 22 AWG has
diameter of .028 inch. We chose Magnetics, Inc material type "P" and from their
catalog selected BM = 500 Gauss. This will give us about 100mW/cm3.
Therefore,
AcAe = (6.33*4) ‘(.00018Hy)*2.22*(.028)1*(10"8) I (500) = 1.59 cm^4.
From the catalog PQ43230 (PQ3230) size has AcAe = 1.60 cm^4.
I coudnot understand how did they arrive to 300 c.m/A? which wire chart they refered to get 22AWG?
BM = 500 gauss.?
Regarding current density they just choose a rule of thumb.
From experience they assume that for the intended application this might work.
Please note usually current density is defined as current per cross section.
Tyco is using an inverse definition. Cross section per current, chosen unit: circular mils per Amp (never saw this in magnetics before, appears to me more common in PCB design).
Here a wire chart:
American wire gauge - Wikipedia, the free encyclopedia
500 Gauss is derived from the material properties of the core.
When running a medium sized core with micrometals material P at 100kHz with a peak flux density of 500 Gauss, then the material will show an acceptable temperature rise.
==> loss charts of the manufacturer or temp rise charts of the manufacturer.
In general application notes for magnetics often use very different units.
SI units or the older non SI units.
When using non SI units, then your formulas will be complicated by adjustment factors and you have to take care that you enter values in exactly the unit that is given for the use in this formula.
If you use SI units, you just need to know some basics about SI units.
Finally this is a matter of taste... and choice might also depend on the data sheets provided from the manufacturer....
Furtheron application notes for magnetics often use many rules of thumb, because always starting from the maxwell equations and discovering all implications from scratch, would let go any company bankrupt before the first transformer would be designed.
It depends on your application and experience how deep you have to dig into physics or apply a rule of thumb.
It is no wonder that you do not fully understand this topic at first glance.
One or the other thing in the tyco paper might give you additional headache,
because besides putting rules of thumbs without explanation (common in the world of magnetics), they additionally mess around with pictures and units (have a look to their duty cycle unit handling ....).
From experience they assume that for the intended application this might work.
Please note usually current density is defined as current per cross section.
Tyco is using an inverse definition. Cross section per current, chosen unit: circular mils per Amp (never saw this in magnetics before, appears to me more common in PCB design).
Here a wire chart:
American wire gauge - Wikipedia, the free encyclopedia
500 Gauss is derived from the material properties of the core.
When running a medium sized core with micrometals material P at 100kHz with a peak flux density of 500 Gauss, then the material will show an acceptable temperature rise.
==> loss charts of the manufacturer or temp rise charts of the manufacturer.
In general application notes for magnetics often use very different units.
SI units or the older non SI units.
When using non SI units, then your formulas will be complicated by adjustment factors and you have to take care that you enter values in exactly the unit that is given for the use in this formula.
If you use SI units, you just need to know some basics about SI units.
Finally this is a matter of taste... and choice might also depend on the data sheets provided from the manufacturer....
Furtheron application notes for magnetics often use many rules of thumb, because always starting from the maxwell equations and discovering all implications from scratch, would let go any company bankrupt before the first transformer would be designed.
It depends on your application and experience how deep you have to dig into physics or apply a rule of thumb.
It is no wonder that you do not fully understand this topic at first glance.
One or the other thing in the tyco paper might give you additional headache,
because besides putting rules of thumbs without explanation (common in the world of magnetics), they additionally mess around with pictures and units (have a look to their duty cycle unit handling ....).
Hi Chocolic,
Yes as you said I too found there is more magic than straignt calculations. Most things are based on thumb rules
and they dont give what the thumb rules are used. so I always felt that some secrete is always maintained. No examples provided
were complete with the required refrence. But when trying to understand where to start learning,I found this is no straight maths for me.
The reason why I was fumbling with what do I start with. In some of his old thread Tony had said the person don't know what question to ask, even that was true in my case too. to be true to you I am still not strongly clear about the utilization and differences between topology. but hope I will be able to be apart of this magic world.
Yes as you said I too found there is more magic than straignt calculations. Most things are based on thumb rules
and they dont give what the thumb rules are used. so I always felt that some secrete is always maintained. No examples provided
were complete with the required refrence. But when trying to understand where to start learning,I found this is no straight maths for me.
The reason why I was fumbling with what do I start with. In some of his old thread Tony had said the person don't know what question to ask, even that was true in my case too. to be true to you I am still not strongly clear about the utilization and differences between topology. but hope I will be able to be apart of this magic world.
Chocoholic I must say you scared away my new friend Tony,Well I have a big long homework. It will take a while for me to understand. Let me do that first. I would be asking some more stupid questions here but then you have to answer those
thats the price you will pay for scaring away my friend .but yes... jokes apart, I really need to do lot of homework.
Tell me which topology can be used to 90% dutycycle?
Well, I use microcontroller to get my pwm. If i use 8bit PWM then. I get 0 - 255 steps for 0 to 100% dutycycle. for getting better resolution it would be good if i can work till 90% or 99 % dutycycle.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.