Yes, you are correct I checked it on google. But then in the data sheet where do i find Bmax?
Is it Flux density (min) = Bs = Bmax?
Hoping what I guessed is correct.
I would select HP-300 material.
Bs = 420mT = 4200 gauss.
This material has Powerloss of 300mw/cc.
As hinted by Tony we would need Ae of 98 mmsq.
We can use one of
EE 42x21x9.
EC 35
or RM10 (as directly suggested by Tony.)
Pl correct me if going wrong somewhere.
I would select HP-300 material.
Bs = 420mT = 4200 gauss.
This material has Powerloss of 300mw/cc.
As hinted by Tony we would need Ae of 98 mmsq.
We can use one of
EE 42x21x9.
EC 35
or RM10 (as directly suggested by Tony.)
Pl correct me if going wrong somewhere.
As a brief example of a 10-15W 1.5kV supply I did-
3019 pot core using 3C81 material
.0067" gap (x2 because I used a simple spacer)
9 turn primary #26AWG
240 turn secondary #38AWG triple build ins.
SG3524 control chip
The all-in-one chips for LV switchers don't lend themselves to HV. I prefer the old fashioned controllers like the 3524 because you can worry about each function of the supply separately and set the duty cycle to specific limits. This used a MOSFET to switch, and controlling the loop area of the circuit is critical. I'll try to find the page of McLyman that wasn't quite right.
CH
3019 pot core using 3C81 material
.0067" gap (x2 because I used a simple spacer)
9 turn primary #26AWG
240 turn secondary #38AWG triple build ins.
SG3524 control chip
The all-in-one chips for LV switchers don't lend themselves to HV. I prefer the old fashioned controllers like the 3524 because you can worry about each function of the supply separately and set the duty cycle to specific limits. This used a MOSFET to switch, and controlling the loop area of the circuit is critical. I'll try to find the page of McLyman that wasn't quite right.
CH
The McLyman transformer and inductor book is up to the fourth edition, so maybe the formulas got changed. In my third edition, page 13-31, the boost formula at the top of the page didn't work for me. I think it should be:
Ipk=(2Po)/(nVin(1-Dw)) where Dw=tw/T
then
Ipk=SQRT((2loT(Vo+Vd-Vin))/nL)
LIpk^2=(2IoT(Vo+Vd-Vin))/n
Hopefully those make sense without using special characters, based on what you see in the book. I emailed McLyman, but don't remember the outcome- if somebody has the fourth edition I'd like to know if it was changed. BTW, there are usually several ways to approach these designs, so no one formula should be considered the last word on the matter. Also, the transformer I described above is for a flyback circuit, as the turns get out of hand quickly with other designs- wire too fine or huge window required.
CH
Ipk=(2Po)/(nVin(1-Dw)) where Dw=tw/T
then
Ipk=SQRT((2loT(Vo+Vd-Vin))/nL)
LIpk^2=(2IoT(Vo+Vd-Vin))/n
Hopefully those make sense without using special characters, based on what you see in the book. I emailed McLyman, but don't remember the outcome- if somebody has the fourth edition I'd like to know if it was changed. BTW, there are usually several ways to approach these designs, so no one formula should be considered the last word on the matter. Also, the transformer I described above is for a flyback circuit, as the turns get out of hand quickly with other designs- wire too fine or huge window required.
CH
I have not seen al reactions, but I use this.
PowerEsim - Free SMPS Switching Power Supply / Transformer Design Software
you can download there the software who is a internet one.
PowerEsim - Free SMPS Switching Power Supply / Transformer Design Software
you can download there the software who is a internet one.
Sorry for the late replies everyone. I am finishing an ASIC design and the nights are running together...
OK... The HP-300 Material is Fine. To make things simple... we are going to design this for universal imput (90Vac-264Vac) in case someone else wants to do this design. We will use 120uF for the hold up capacitance. I believe you said you wanted 90Vdc and we will use 1A for the output current and a 60% duty cycle. Operating frequency is 100KHz (usually you use 65KHz so that the 2nd harmonic is below the 150KHz start of the conducted band for EN60950).
Core Parameters for the RM10 that you will use Ac (Ae in most books but McLyman uses Ac)-98mm^2, MPL, MLT, permeability (2200)...
HP-300 material is fine. I found it here for those that are interested.
http://www.hinoday.com/downloads/ferrite_core_brouchure.pdf
Most vendors have this or an equivalent. The saturation flux density is 420mT (4200 guass) at 100C which is close to the probable operating point. We will use 3700 gauss for Bmax since you will need to be able to handle lot-lot material variances. It is important to note the material loss curve is a minimum at 90C so that is a good point to actually target the operating point of the transformer. It also provides 15C headroom for the 105C safety limit (hottest point inside the transformer). In case you ever need to calculate the MLT by hand, you take the bobbin, select the midpoint up the sidewall and then calculate the length of 1 turn around the bobbin. This calculation will get you close... it's not exactly the same as what the manufacturers use but it is a good approximation. Magnetic Path Length is usually listed. One of McLymans other books has the tables for almost all teh different core types.
Tony
OK... The HP-300 Material is Fine. To make things simple... we are going to design this for universal imput (90Vac-264Vac) in case someone else wants to do this design. We will use 120uF for the hold up capacitance. I believe you said you wanted 90Vdc and we will use 1A for the output current and a 60% duty cycle. Operating frequency is 100KHz (usually you use 65KHz so that the 2nd harmonic is below the 150KHz start of the conducted band for EN60950).
Core Parameters for the RM10 that you will use Ac (Ae in most books but McLyman uses Ac)-98mm^2, MPL, MLT, permeability (2200)...
HP-300 material is fine. I found it here for those that are interested.
http://www.hinoday.com/downloads/ferrite_core_brouchure.pdf
Most vendors have this or an equivalent. The saturation flux density is 420mT (4200 guass) at 100C which is close to the probable operating point. We will use 3700 gauss for Bmax since you will need to be able to handle lot-lot material variances. It is important to note the material loss curve is a minimum at 90C so that is a good point to actually target the operating point of the transformer. It also provides 15C headroom for the 105C safety limit (hottest point inside the transformer). In case you ever need to calculate the MLT by hand, you take the bobbin, select the midpoint up the sidewall and then calculate the length of 1 turn around the bobbin. This calculation will get you close... it's not exactly the same as what the manufacturers use but it is a good approximation. Magnetic Path Length is usually listed. One of McLymans other books has the tables for almost all teh different core types.
Tony
...Ok, so you know that getting zapped it is pretty inconvinient.
ELCB : Yup, I am using the current sensing version of that for my hobby work bench at home.
Regarding isolation: There are multiple possible solutions.
You can be sure that Tony will point to this and together you will find a solution that fits to isolation materials and/or wires, which you can get locally.
I am not sure if India has a specific national safety standard for audio equipment or if you adopt IEC or UL.
I guess Tony will propose to follow an UL safety standard, which should be fine. As long as you are not going to sell your design, any reasonable safety standard should do the job, because they all have the intension to protect human beings against electric shock.
Ae=98, MPL=44, MLT=53, u=2200
I screwed up on the Minimum Vac. If we use 90Vdc on the output then I will need a turns ration less than 1 (we will talk about this later). This would mean a 1kV Mosfet so we will use Vac Min as 190Vac. That gives us a Min Bulk voltage of about 184V.
From previous post that will give us a Idc of about 0.965A so DeltaI=200%*Idc and Ipk=Idc+DeltaI/2 so Ipk=DeltaI=1.93A (triangle shaped DCM).
Ton=Dutycycle/Fsw so that is 6us.
Lp=Vbulkmin*Ton/DeltaI=571uH (derivation is V=L*dI/Dt)
Now the fun begins...
Tony
I screwed up on the Minimum Vac. If we use 90Vdc on the output then I will need a turns ration less than 1 (we will talk about this later). This would mean a 1kV Mosfet so we will use Vac Min as 190Vac. That gives us a Min Bulk voltage of about 184V.
From previous post that will give us a Idc of about 0.965A so DeltaI=200%*Idc and Ipk=Idc+DeltaI/2 so Ipk=DeltaI=1.93A (triangle shaped DCM).
Ton=Dutycycle/Fsw so that is 6us.
Lp=Vbulkmin*Ton/DeltaI=571uH (derivation is V=L*dI/Dt)
Now the fun begins...
Tony
I am only going to propose a triple insulated wire version. Furikawa TEX-E style (Rubidue also makes a version as well as some Asian manufactureres) wire. This and a standard bobin will pretty well guarentee compliance. Also, if you have a vendor do it for you they will already have the safety certifications and insulation systems in place. What I will propose will be guaranteed safe... As Chocoholic says... you just can't sell it. You will also need some enamel magnet wire (2UEw or equivalent).
I do not want to talk about margin tapes, teflon tubing for crossing the boundary, insulating tapes etc... We can talk about this at a later date if you are interested.
Tony
Lets next look at the energy handling capacity of the design...
Energy =Lp*Ipk^2/2 is a good approximation
Power=Energy*Fsw and that is 106W. for a 90W (pk) converter that would give me enough headroom. If I ran the numbers and I was less than 15% margin I would probably change the design.
Tony
Energy =Lp*Ipk^2/2 is a good approximation
Power=Energy*Fsw and that is 106W. for a 90W (pk) converter that would give me enough headroom. If I ran the numbers and I was less than 15% margin I would probably change the design.
Tony
Np=Ipk*Lp/(Bmax*Ae)=30.4 turns. Usually I would round up to the next turn. In this case I do not like 31 turns. We will change Bmax downward a bit to get 32 turns. 0.35182T (3518.2 guass) gives you 32 turns and a bit more Saturation Flux Density Headroom.
Now for the gap calculation... Fringing Flux is a function of the gap and is used in the gapped inductance calculation so this will be an iterative process of increasing teh gap until the gapped inductance meets our target inductance.
Start with an assumed lg (gap distance) and iterate it up or down until Lgapped=Lg above)
FFlux=1+(lg/swrt(Ae))*ln(2*WindowWidth/lg) where lg=gap distance
Lgapped=(0.4*PI()*Np^2*FFlux*Ae)/(lg+MPL/u)
You should end up with a gap of about 0.224mm
Next we will cover skin effect and wire selection...
Probably not until tomorrow.
Tony
Now for the gap calculation... Fringing Flux is a function of the gap and is used in the gapped inductance calculation so this will be an iterative process of increasing teh gap until the gapped inductance meets our target inductance.
Start with an assumed lg (gap distance) and iterate it up or down until Lgapped=Lg above)
FFlux=1+(lg/swrt(Ae))*ln(2*WindowWidth/lg) where lg=gap distance
Lgapped=(0.4*PI()*Np^2*FFlux*Ae)/(lg+MPL/u)
You should end up with a gap of about 0.224mm
Next we will cover skin effect and wire selection...
Probably not until tomorrow.
Tony
Chocoholic reminded me of one issue and that is that you may not have access to tripple insulated wire (rubedue or furicawa). Let me know if this is true and we will change the core type to an EE/EER type structure so you can uses standard wire and tape with Teflon sleeving.
Above all else, I want to make sure that everyone stays safe.
Tony
Above all else, I want to make sure that everyone stays safe.
Tony
Thanks CH,
I will just try to find it. I too got Third edition (Revised and expanded).
I am could not find the formula. I will PM you for further talk on this. Just to avoid any kind of copyright issue.(which we sometimes unintentionally break).
Thanks Tony,
Yes getting triple insulated wire is a problem at my place. Or if it would be available it would be available in bulk. for pro mfg.
Tony I would request you to keep the component selection simple, and the one that would be easily available in local shop. Something that would be in reach of DIYer.
I am enjoying the learning. Thanks. Everymorning I wake up. I run to my laptop to read here
(even before I brush my teeth ).You are right Tony, It would be nice If you would brief us on transformer construction, but later. For me all things at a time would be a overdose.
How did you find values of MPL & MTL using data provided by Hindoy?
was MTL = " In case you ever need to calculate the MLT by hand, you take the bobbin, select the midpoint up the sidewall and then calculate the length of 1 turn around the bobbin." ?
Sure you can take any values for example, rather I would you request you to select example values cause you know better what values would ease learning.
"This would mean a 1kV Mosfet so we will use Vac Min as 190Vac" is it regarding the back emf that mosfet needs to sustain ?
Can I request you to explain, how to ensure that the transformer size and the winding area is enough for our requirement?
Last edited:
Ofcourse your Work always goes on priority. If you would be unavailable for some time, just put oneliner here of your unavailiblity. we would wait for your return.
I haven't understood the above point. Ok, seeing the datasheet I understood at 90 degree Celsius, core losses is the lowest. but I didn't understand the thing about operating point & 105 degree Celsius where we got it from?
1) How do we calculate Ae ( I mean equation to calculate what Ae we require for the design. ) ? so we select core size based on this?
2) lg is gap distance, is it distance between two core ? eg. gap at the point where two EE core join together. Is 0.224mm = lg ? or Lgapped?
3) Whats Lgapped?
4) what length of core is windowWidth?
Sorry Tony, I Know I have asked lot of questions today. but pl don't get irritated.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.