Flyback with LAN transformers

Ok, most of us have used LAN tranformers for power transfers. However, their datasheets are skimpy at best and usually contain no information about the core. Of course it won't contain any regular air gap and you cannot easily install one later because it's all potted up.

So, did anyone use these for a flyback design yet? Low power, something well under 100mW.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg
Loading thread data ...

Joerg a écrit :

Which voltage from which voltage? Do you really need flyback? I once did a funny self-oscillating one for that kind of power, from 50V supply. Worked pretty well with about 60%, maybe 70, efficiency IIRC. I can dig up the drawing.

Anyway how about measuring your xformers? That's so easy...

--
Thanks,
Fred.
Reply to
Fred Bartoli

I've used oscillators as well but then there was no DC component and everything was just fine. It's regular voltage levels, 12V, 24V etc. Flyback is very easy from a topology point of view, and regulates very fast. If there is no other way I'll just do the oscillator thing again.

Sure. But you know how it goes after that. A couple of years later the manufacturer "improves" the design and ... POOF.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Then take enough safety margins. Given the power level it's probably not difficult. And then what can be improved? Ferrite won't change much and evolves at a very slow pace, and turns/geometry will stay the same or it'd be another part number.

--
Thanks,
Fred.
Reply to
Fred Bartoli

We use the Tamuras as forward-mode dc/dc converters, driven p-p ct by a pair of 2N7002's, in the 60 KHz ballpark. They are toroids, ungapped, so probably don't store a lot of energy. They *are* cute and cheap, and come with all sorts of interesting turns ratios.

John

Reply to
John Larkin

I'll check those out. Forward p-p would be a bit inconvenient here because there is no high side drive capability. What power do you push through? If it's more than a watt a flyback might be possible since even p-p is never exactly 50.000% duty cycle.

At Digikey the Tamura xfmrs seem kind of pricey.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Well, I've seen it a few times that a manufacturer switched to another source for the ferrite and problems arose. But you are right, if we had something like 90% margin we'd be ok.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Oops, it's an Alfamag/Talema ISDN part I was thinking of, SMJ-140B. It has 4 separate windings, 1:1:2:2, but you can get other ratios. We're paying $1.18 at 1K. It's a nice cased sugar-cube surface-mount thing.

I think we're running about a watt, but it's probably good for more. We're driving it with two open-drain 2N7002 fets into essentially a center-tapped winding with the ct at +5 or +12, depending.

I posted a pic a while back, of a VME board with 16 of these.

It would probably work flyback, but a watt might be tricky. Just guessing. It does look hunkier than the typical Ethernet xfmr.

John

Reply to
John Larkin

Purpose designed miniature flyback coupled inductors ('transformers') are readily available a equally low prices, so why bother? Paul Mathews

Reply to
Paul Mathews

Nice, but most of my apps can stomach >$1. I've used some Asian ISDN transformers for that (Murata?) but it's been a while and now it seems ISDN isn't exactly flourishing anymore. Just like those cute modem transformers that are probably going to almost vanish one day.

That's the thing, they do not say how much DC they can take. I need less than 100mW here but still, if it works on the bench that doesn't mean it'll still work in three years because there ain't nothing about that in datasheets and when I ask all I get is a shrug. Because it's the first time someone asked.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Where?

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Question: Can someone teach them how to design a web site? And how not to design one? It doesn't even let me get to the SMJ series and I don't know which distributor sells their stuff. Because they don't say.

I hope they didn't hire the web designer from National...

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

And nice low dcr's too.

I think you can buy less, but they're made in India, so leadtime is terrible. We have 1200 in stock if you need a few.

John

Reply to
John Larkin

Is 1K the MOQ? Can you buy just 100? It has a nice high primary inductance.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
Reply to
Spehro Pefhany

Some sources with very aggressive pricing on miniature flybacks include Xfmrs, Inc, Premier Magnetics, Vishay/Dale, GB Int'l. Quantity requirements for low prices are surprisingly low....you basicallly have to cross their threshold for having the parts batch produced in Asia. Prices in lower quantities are about 6x higher. We buy UR 1W flybacks for about $.50 ea in 1K lots. Paul Mathews

Reply to
Paul Mathews

$0.50/1k is expensive. Non-gapped is much less. But still witing for an answer from Xfmrs, let's see. Not holding my breath though ;-)

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Expect lower pricing for non UL-Recognized parts that don't have inspection and creepage/clearance margins, taping, and related. What's your target? PM

Reply to
Paul Mathews

Can't tell. But it doesn't have to be UL cert'd. However, the premium versus a non-gapped core cannot exceed the cost of the parts we'd save going from forward to flyback, maybe 10-15c.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.