Several questions

Hello Nick,

You can use a toroid core if your topology is forward and if you do not have too many turns. You can not have air gap and for this reason the current you can pass trough your toroid transformer is very limited. For further information, see: http://info.ee.surrey.ac.uk/Workshop/advice/coils/index.html
I hope we will be able to provide more details for magamp design, for now see this link:
http://www.powerint.com/community/papers-circuit-ideas-puzzlers/circuit-ideas/ferrite-magamp-flyback-power-supply
I assume you refer to this article: http://www.powerint.com/community/papers-circuit-ideas-puzzlers/circuit-ideas/floating-constant-current-source-allows-ultra-
For LinkSwitch-LP you must have a minimum of 50V Drain Voltage in order to insure a minimum of 84uA for BP pin (internally extracted from Drain in OFF state). You can try to decrease this 50V limit using an external current source, but this is not a test we used for LinkSwitch-LP and I can not tell you if you can go under 50V. If you run this test, let me know your results.

Cheers,
PI_Crusher

hello Crusher

Thanks. I am working on low power designs. One is less than 1 watt and another is about 12 watt. which is why I am investigating toroids. I dont understand why you say only for forward topologies. I really want to use linkswitch LP in flybackmode

I have made calculations that can avoid saturation and can provide about 2000 uH primary inductance with about 50 turns and I thought of using foil for the secondary which may only be about 2 turns.

I was looking at primary current limits which brings my next question. if I do a quick linkswitch LP PI expert design for 85-220 VAC in and 5v at 100 ma out I get a primary current limit of 130 ma min and and 150 ma max and the primary is specified for 38 SWG wire. Yet my wire tables give me a 22 ma maximum current for 38 SWG wire. So how come it doesnt burn up.

Can you explain the relationship between current limit mentioned in the PI expert design documents and the actual current flowing through the inductors... I presume this is common to all PI devices.

Mag amps
========
yes I had seen this. But there is no winding info for the mag amp.

Ok. If I ever try this I will let you know.

Regards
Nick

I should point out that the 12 watt design is for the DPA switch and the 1 watt design is for Linkswitch LP. I want to use small toroids for both designs.

My interest in the magamps is because I want a very low component count approach to regulating extra ouputs and I have a small SOC-cpu in the design which can directly control the magamp reset current without any other components being used.

Hello Nick,

If you checked your flyback toroid transformer and you are sure you can avoid saturation you can try it. I would advise you to double check your assumptions, look here:
http://info.ee.surrey.ac.uk/Workshop/advice/coils/gap/index.html
I have no access to your design file, but 38 AWG for your 0.5W output power (5V and 100mA) is realistic. For 85Vac input you will get more than 110V on the input DC rail. With a current limit of 150mA you get 16.5W peek power. As you guessed already, 150mA is peek current, your RMS current is very low. For 60% efficiency at 0.5W output power you need 0.83W input power. This translates into RMS 7.5mA input current. You can consider this RMS input current quite close to RMS current trough your transformer, even if the peek current could go a lot higher. Your table gives you 22mA RMS for 38AWG and is actually what you need.

Cheers,
PI_Crusher

hello crusher this will probably be my last question on this subject.
It concerns the calculation of the magnetic field strength in amp/turns developed by my toroid transformer so that I can check for saturation.

As we have established above the linkswitch family switch off when current reaches a preset peak limit. In the case of the Link 526 this is about 130 ma. Now, should I be using peak current to calculate my mag field strength or the RMS current, or average current - both of which will be much lower.

Because of the large difference if I use the peakcurrent I will get a different design and enter saturation with a smaller number of turns.

The way i see it is that even if the core saturates or lets say permeability starts to decrease (ie not fully saturated and within reversible limits) the current limit is going to halt the process anyway. the only thing that can be happening is the di/dt will be faster and the duty cycle less. Also because of the very lower power of this design the system will probably be cycle skipping as well.

any thoughts about this ?

My second questions is about magnetic core reset. What mechanism is being used to reset the core in a linkswitch LP clamp-less design? Since current cannot circulate in the primary after the mosfet has opened. And does the core reset completely when in discontinuous mode, which I believe is the normal mode for all linkswitches.

regards

nick

Hello tangray,

You must use the maximum peek current in order to verify your transformer is not saturating. To make things even worst, you must allow for extra margin to that limit, let’s say 10% or even a more conservative 20%, to account for a whole multitude of parameters you can not control: core section, magnetic and thermal variations, etc. You can not count on the internal current limit or cycle skipping to compensate for saturation because all you will get is dramatic output power limitation.
A clamp-less design is not actually “clamp-less”. Even if you do not use special components to implement a clamp circuit, you have a real and invisible clamp circuit for free. This is because you have parasitic elements (most important capacitors and resistors) associated with each component: MOSFET, transformer, PCB. After the MOSFET switches from ON to OFF, the leakage inductance of your transformer is still connected to those parasitic elements acting like a real clamp circuit. This is the reason you will see sometimes a small capacitor placed from Drain to Source when the clamp-less design has dangerously high voltage spikes.

Cheers,
PI_Crusher

You can use toroidal cores for both, forward and flyback SMPS.
In forward, you should simlpy design your transformer with PI Xls
and take ferrite toroid with similar cross section area.
In flyback, you can design your transformer with PI Expert, and take insteed of calculated core, toroidal core with similar cross section area, but you can use only toroidal cores with air gap (Ferroxcube) or toroids with distributed air gaps
for example Arnold MPP or MSS cores.
In application up to 20 W it works good.
If you choose cores with the same Al, and the same cross section area, you will wind the same numbers of turns.
For meeting leakage inductance and safety standards, you can use TIW wires for all: primary, bias and secondary windings.
We use toroids in SMPS, when dimensions (height) of SMPS must be as low as possible.

hello Feryster,
Since you seem to have this experience already, maybe I should tell you what my objective is. I want to design an ultra small 220 VAC to 5 vdc standy supply providing less than 1 watt, but with the possibility to provide addition outputs.

I am wondering why you say I need to use a gapped toroid for a flyback design.
I am aware that this is common pratice in alot of cores and understand why a gap can be required. Is this because di/dt will be higher in a flyback design and cores may saturate sooner...

The design approach I was using was to take the calculated inductance from a PI expert/XLS design and recreate this based on a specified toroid. I have run into the problem that for smaller xsection area toroids (which I wanted to use) the amp/turns takes me into saturation with very small currents. But if I use a larger cross section I can just find a combination that will provide me with the correct inductance,and turns ratio and current limit of 130 ma. I am using a amps/turns value of 250 as the saturation limit based on high permeability ferites (>2000 ) without having to use a air gap or special type of ferrite. I need an OP power less than 1 watt.

Actualy I am trying to find some ferrites with a relatively small ID and therefore a greater xsectional area for their OD size - but havent found anything yet.

It seems to be that if I use the required xsection area at the low power I require then I can use a high permeabilty ferrite without the need for a gap.

I havent been able to verify this yet as I am still waiting for parts.

Finally I have thought about increasing the inductance from a normal PI/Expert design so that the additional reactance would reduce the peak current below the specifc chip current limit and therefore also allow me to use a smaller core.
This would mean running at full duty cycle and regulation only by cycle skipping.

Any ideas welcome.

Hello tangray, in every cases you design flyback transformers, you must use core with airgap or cores with distributed airgap, becauce gaps in cores are necessery for starage energy in first step of work of flyback, and deliver this energy to secondary side in second step.
In attached file you can find new design using your link swith LP and smallest core EF-12.6 with cross section area 0,13 square cm and AL=34.
Insteed of this core, you can try to use MSS core:
MSS-050075 (diameter 0,5") with cross section 0,11 sq. cm and Al=34
see on website:
http://www.feryster.pl/polski/rtmss.php?lang=en
and wind all windings using wires and numers of turns as attached design.

Thank you for your information.
I have checked your website. Your company is interesting.
Can you supply me an email address.
I have used your design file to ask for a quotation.
regards
Nick