Wireampient means Wireambient right ?
Does that mean Temp of wire at ambientTemp?
and WireTemp is ? operating temp of wire ( operating temp of transformer)
What is the thumbrule for calculating current/cmsq. ( chocoholic had said in previous thread that this goes by thumbrule.) or a equation exists?
so that we can calculate number of strands of 24AWG required for Ipk of 8.3A.
Is margin tape the one which is in between two winding?
How is minimum core size that would be required, calculated?
The Area of the wire is Awb=PI()(r)^2
The Lp resistance is RcuPri=(Rho.e/Awb)*MLT*Np (told you we would eventually use it).
We will approximate the RMS current (there are precise calculations to do that and I will let you find them) by using the PawbPri=(Po/(n(90Vac*sqrt(2))))^2*(RcuPri) where n is the efficiency. If this number is acceptable (about 115mW for the primary winding).
Is 115mw copper losses at primary?
So for secondary losses, Would it be
RcuSec=(Rho.e/Awb)*MLT*Ns
PawbSec=(Po/(n(Vout)))^2*(RcuSec)
Total cu losses = Priloss + SecLoss. right ?
in previous thread you said that the good design is the one where copper losses and core losses are equal.
How do we calculate core losses?
How is minimum core size that would be required, calculated?
The Area of the wire is Awb=PI()(r)^2
The Lp resistance is RcuPri=(Rho.e/Awb)*MLT*Np (told you we would eventually use it).
We will approximate the RMS current (there are precise calculations to do that and I will let you find them) by using the PawbPri=(Po/(n(90Vac*sqrt(2))))^2*(RcuPri) where n is the efficiency. If this number is acceptable (about 115mW for the primary winding).
Is 115mw copper losses at primary?
So for secondary losses, Would it be
RcuSec=(Rho.e/Awb)*MLT*Ns
PawbSec=(Po/(n(Vout)))^2*(RcuSec)
Total cu losses = Priloss + SecLoss. right ?
in previous thread you said that the good design is the one where copper losses and core losses are equal.
How do we calculate core losses?
Last edited:
Wire temp would be 100C and Ambite would be 25C.
The rule of thumb is 500A/cm^2 and 300 is considered conservative. This is covered fairly well in McLymans book. However I do not agree with it. The reason is that it does not distinguish between short wires and long wires. If I remember correctly it uses the RMS current for the density calculation.
I am not denegrating wither Mclyman or Chocoholic. I am just saying that I consider wire temperature more than I consider current density.
Tony
If I remember correctly McLyman uses an AreaProduct calculation to estimate what core will be required. It utilized the Winding Area (bobbin width x Bobbin window Height) and the Ae of the core. There is indeed a correlation that you can walk through if you want. But again it ignores some key parameters. As I said at the beginning, I can run any power level through any core if I can get enough turns on it.
So... What is correct? I can give you some general ranges (these are for Flybacks). For 45-65W is usually an Ae of around 98mm^2, for 25-35W an Ae of around 54mm^2, 10-25W would be a core around 25-35mm^2 respectively. For 90W with a PFC front end then a Ae of 98 is OK. For 130W you will need something around 110mm^2, and for above 180W you are on your own because I would have shifted to a different topology.
There is also one other way... that is to use the black body approach and assume that your core has to dissipate 10% of the processed power. Look at the black body model and see if your estimated Delta T rise is going to be a reasonable number. We can cover that if you want.
And finally... If I ever need a gap larger than 1mm on a flyback or 2mm on the PFC Inductor then I up the Ae of or the frequency.
If you notice there is a trend here... For every rule there are several exceptions and fudge factors, all of which you can trade off to get the desired results. The question is what are your knobs and how hard do you turn them to achieve what you want.
You are correct on the Cu losses. You will need to also calculate the Gap loss. Unfortunately for an accurate estimation of the gap loss there is a
Tony
So... What is correct? I can give you some general ranges (these are for Flybacks). For 45-65W is usually an Ae of around 98mm^2, for 25-35W an Ae of around 54mm^2, 10-25W would be a core around 25-35mm^2 respectively. For 90W with a PFC front end then a Ae of 98 is OK. For 130W you will need something around 110mm^2, and for above 180W you are on your own because I would have shifted to a different topology.
There is also one other way... that is to use the black body approach and assume that your core has to dissipate 10% of the processed power. Look at the black body model and see if your estimated Delta T rise is going to be a reasonable number. We can cover that if you want.
And finally... If I ever need a gap larger than 1mm on a flyback or 2mm on the PFC Inductor then I up the Ae of or the frequency.
If you notice there is a trend here... For every rule there are several exceptions and fudge factors, all of which you can trade off to get the desired results. The question is what are your knobs and how hard do you turn them to achieve what you want.
You are correct on the Cu losses. You will need to also calculate the Gap loss. Unfortunately for an accurate estimation of the gap loss there is a
Tony
That was someone else that said the Cu should = the core losses.
This is again another rule of thumb (Again McLyman goes on at some length about this). For me, I worry a lot about efficiency, EMI, and cost so I tend to use more Ae and less Cu as that leads to a lower leakage design (higher efficiency and lower EMI) and a cheaper transformer.
Tony
This is again another rule of thumb (Again McLyman goes on at some length about this). For me, I worry a lot about efficiency, EMI, and cost so I tend to use more Ae and less Cu as that leads to a lower leakage design (higher efficiency and lower EMI) and a cheaper transformer.
Tony
OK... Now lets cover the section that is really only a SWAG (scientific wild a** guess), gap losses.
Gap losses are basically the imperfect storage/release of the energy in teh gap of the transformer. The more gap you have the higher the loss. There is really no way to accurately estimate the gap loss other than by a finite element analysis with some really expensive software.
McLyman covers gap losses fairly well and we will use his equations for the estimate. I will not use the Core Loss data from the data sheets because they are for forwards and not Flybacks.
Ki=0.0388
Pg=Ki*WindowWidth*(Bac^2)*lg*Fsw=12.3W
DeltaT is 194C using the plack body model - Way too high.
If we use 35Khz as the minimum then we need 86 turns and 1.6mH for the primary inductance. This is certainly too many turns.
If we use 50KHz then we end up with 1.1mH and 60 turns. That sound about the right number of turns. lg=0.452. Bac=0.37, but, your fringing flux went up (gap loss) so Pg=15W. Still too high.
I am still not comfortable making any judgments about the suitability of the core based on Gap loss calculations alone
If someone has a better estimate on how to calculate the gap losses I would appreciate it.
Tony
Gap losses are basically the imperfect storage/release of the energy in teh gap of the transformer. The more gap you have the higher the loss. There is really no way to accurately estimate the gap loss other than by a finite element analysis with some really expensive software.
McLyman covers gap losses fairly well and we will use his equations for the estimate. I will not use the Core Loss data from the data sheets because they are for forwards and not Flybacks.
Ki=0.0388
Pg=Ki*WindowWidth*(Bac^2)*lg*Fsw=12.3W
DeltaT is 194C using the plack body model - Way too high.
If we use 35Khz as the minimum then we need 86 turns and 1.6mH for the primary inductance. This is certainly too many turns.
If we use 50KHz then we end up with 1.1mH and 60 turns. That sound about the right number of turns. lg=0.452. Bac=0.37, but, your fringing flux went up (gap loss) so Pg=15W. Still too high.
I am still not comfortable making any judgments about the suitability of the core based on Gap loss calculations alone
If someone has a better estimate on how to calculate the gap losses I would appreciate it.
Tony
I am still unhappy about the above equations for DCM mode. For CCM operation Bac would yield about 3W for the gap loss which is about what I expect.
Anyway to continue on...
Pxfmr=Pcore+Pgap+Pwirelossprimary+Pwireloss secondary
I usually don't use the core loss figures and teh gap loss figures are an extimate but even though there are some inaccuracies we can take some general assumptions on the overall power loss and estimate if the core size is correct.
Assumption: transfomer losses are 1/2 the total losses in the system or 7W of the 15 available.
DeltaT rise over ambient is calculated as follows:
DeltaT=(Pxfmrloss/AT)^8.33 where AT is the surface are of the dissipation surface.
At 7W that gives us a DeltaT of about 119C
A core (EE) that has a higher surface are would actually bring the overall temperature down some. Since we are trying to run the core at 100C then this is probably still an OK number.
Tony
Anyway to continue on...
Pxfmr=Pcore+Pgap+Pwirelossprimary+Pwireloss secondary
I usually don't use the core loss figures and teh gap loss figures are an extimate but even though there are some inaccuracies we can take some general assumptions on the overall power loss and estimate if the core size is correct.
Assumption: transfomer losses are 1/2 the total losses in the system or 7W of the 15 available.
DeltaT rise over ambient is calculated as follows:
DeltaT=(Pxfmrloss/AT)^8.33 where AT is the surface are of the dissipation surface.
At 7W that gives us a DeltaT of about 119C
A core (EE) that has a higher surface are would actually bring the overall temperature down some. Since we are trying to run the core at 100C then this is probably still an OK number.
Tony
OK... I am going to make some final comments about the gap losses. These are thing to keep in mind when you are picking the CCM/DCM/QR topology. Bac is the AC component of Bmax. A CCM converter running at 65KHz and 25% DeltaI would have only about 3W of gap losses as apposed to the 12W from the QR converter. Thoase are calculated values and not measured. I can tell you from personal experiance That I have converter several QR designs from other engineers into CCM designs just because of the one issue.
The gap loss are really a funny thing and most people completely ignore them. I have seen several exampled of people using the iintegrated magnetics and a phase shifted full bridge that forget to add teh fringing flux into the design equations and then can't understand why they were only able to reach 250mT before the core started to saturate. The reality is that they were running 1.4x as high because of the gap on the core. The same goes for a resoanant inductor on a LLC. You have to ccount for the gap effect in your design equations.
This pretty well concludes what I had to say about a flyback. Anyone else that wants to chime in please do.
Thanks for listening.
Tony
The gap loss are really a funny thing and most people completely ignore them. I have seen several exampled of people using the iintegrated magnetics and a phase shifted full bridge that forget to add teh fringing flux into the design equations and then can't understand why they were only able to reach 250mT before the core started to saturate. The reality is that they were running 1.4x as high because of the gap on the core. The same goes for a resoanant inductor on a LLC. You have to ccount for the gap effect in your design equations.
This pretty well concludes what I had to say about a flyback. Anyone else that wants to chime in please do.
Thanks for listening.
Tony
Very Important Input Thanks a lot.
I am having a walk through the previous posts. Re-calculating whole so to make sure that your student has been able to grasp it well .
Deeply thank you for the priceless experience.
If you are using margin tape then you do not need triple insulated wire. If you do not want to use margin tape then you have to use triple insulated wire. This applies to both cores equally. You either have to have mrgin tape or you have to have triple insulated wire.
Tony
From the Book, McLyman. Chapter 13. Design Example, Buck-Boost Isolated Converter Discontinuous Current, Step No. 25 Calculate the primary, the new micro Ohm/cm.
The equation is New micro Ohm/cm = (micro Ohm/cm) / (no. of strands at primary)
The value of micro Ohm/cm shown is 1345. I am not getting the reference for this value. also it was not calculated in any of the previous steps.
The equation is New micro Ohm/cm = (micro Ohm/cm) / (no. of strands at primary)
The value of micro Ohm/cm shown is 1345. I am not getting the reference for this value. also it was not calculated in any of the previous steps.
Hello Tonny,
As I said I am going through book examples and trying to understand . also I am reading your lesson. I have some doubts on older posts. but I would ask you when I reach that stage.eg I have some question about transformer construction. But i would ask when I reach there.
Well, I have doubt about calculating no. of strands required.
I think, we did not cover this calculations. (pl. correct me. cause I am loaded in understanding the working mechanism of transformer. what is the importance of each parameter.)
Is it true that if I keep the frequency same. voltage same, but change current to say 10 times then the copper wire should be thicker? And so possibly more no of strands. right ?
There is only one real reason to use multiple strands... Efficiency.
When you have a waveform that has very high harmonic content (like a DCM flyback) is is common to use multiple strands of wire so that the higher frequency harmonics can easily flow in the wire and are not crowded out at the surface. If you think of a triangle wave you will see strong odd harmonic content in the EMI spectrum. These higher frequencies need to flow somewhere just like the fundamental frequency. So if you wanted to make sure the 3rd harmonic has an easy path to flow through then you would use multiple thinner wires so that you fully utilize the cross sectional area of the wire. For TM mode PFC often use Litz wire to improve their efficiency by dropping the Rac at the higher switchign frequencies.
When specifying a tranfer to be built outside, you would usually specify the test frequency and votlage (100KHz and 1V as an example) and you would ask for the test data for the Rac so you could insude that the transformer was built correctly.
Was this what you were asking for or do you need additional details?
Tony
When you have a waveform that has very high harmonic content (like a DCM flyback) is is common to use multiple strands of wire so that the higher frequency harmonics can easily flow in the wire and are not crowded out at the surface. If you think of a triangle wave you will see strong odd harmonic content in the EMI spectrum. These higher frequencies need to flow somewhere just like the fundamental frequency. So if you wanted to make sure the 3rd harmonic has an easy path to flow through then you would use multiple thinner wires so that you fully utilize the cross sectional area of the wire. For TM mode PFC often use Litz wire to improve their efficiency by dropping the Rac at the higher switchign frequencies.
When specifying a tranfer to be built outside, you would usually specify the test frequency and votlage (100KHz and 1V as an example) and you would ask for the test data for the Rac so you could insude that the transformer was built correctly.
Was this what you were asking for or do you need additional details?
Tony
Tonny, you have added some important things to my knowledge.
But there was something else I was trying to ask. Well let me try again.
I am reading McLyman. To under it better I am using same example as in book.( Chapter 13. Design Example, Buck-Boost Isolated Converter Discontinuous Current)
I have created an microsoft excel sheet . So that I can play with variables. and see effects on each parameter. while doing this I keep frequency unchanged (100khz) and total wattage equal to 18.5W or less unchanged.
now when I changed the current requirement at secondary. I expected only thickness of wire(or number of strands required) should change and not turns. But it was not so.
Now correct me If wrong:
1) when I change total wattage of the transformer , the thickness of wire used for primary winding ( or the number of strands) but it does not happen so. I have verified the equations set in excel sheet.
2) If I change VA ratings for one of the two secondary, the other secondary too will change its winding parameters. which actually should not be.
The above things leaves me doubt on equations given in the book.
hope I could explain this time.......
But there was something else I was trying to ask. Well let me try again.
I am reading McLyman. To under it better I am using same example as in book.( Chapter 13. Design Example, Buck-Boost Isolated Converter Discontinuous Current)
I have created an microsoft excel sheet . So that I can play with variables. and see effects on each parameter. while doing this I keep frequency unchanged (100khz) and total wattage equal to 18.5W or less unchanged.
now when I changed the current requirement at secondary. I expected only thickness of wire(or number of strands required) should change and not turns. But it was not so.
Now correct me If wrong:
1) when I change total wattage of the transformer , the thickness of wire used for primary winding ( or the number of strands) but it does not happen so. I have verified the equations set in excel sheet.
2) If I change VA ratings for one of the two secondary, the other secondary too will change its winding parameters. which actually should not be.
The above things leaves me doubt on equations given in the book.
hope I could explain this time.......
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.