how hot can a toroid inductor go safely?

[eccerr0r]

Well, since I now have a grid tie inverter, I have to follow its input requirements. So, I designed/built a small switch mode supply to feed it.

After several iterations of pitfalls with a very uncooperative solar panel, I got the supply to behave - except the toroid inductor gets to 165°F (74°C) ... though granted I'd rather it get this hot than the 75N06 which only gets to 108°F (42°C) despite on average passing 5-6 amps when the panel is in full sunlight.

Not quite sure if it's due to core saturation, eddy losses, or just I*R losses yet but it is hot...but how hot is too hot?

(the smps works! Efficiency is nothing to write home about, probably because of that inductor, albeit if I don't use the panel at all it's 0% efficient...)

[diif]

Re: how hot can a toroid inductor go safely?

Surely the question should be "How does my design look ?", not how how is too hot.

[eccerr0r]

Re: how hot can a toroid inductor go safely?

Well, for one thing I used parts that I had on hand, so I don't have many options on redesigning so unless it's a gross swap, it is what it is. Just wonder how long this thing will last being that hot before it loses its characteristics and fries the device, or could I just run it as it is and hopefully I'll have another solution down the road.

I think this coil came out of a dead off-line PSU probably some filter inductor. I really should characterize the core to know exactly what I have, but well, it is what it is...

[sam_sam_sam]

Re: how hot can a toroid inductor go safely?

I would be concerned about if is getting so hot that the enamel paint degrades and you end up with a shorted coil but I could not tell you how hot it has to be for this to happen

[eccerr0r]

Re: how hot can a toroid inductor go safely?

I had a toroid filter core go short in a Antec PSU after it overheated so definitely a concern. I don't have a fan for the toroid (it's the energy storage element in a boost supply). Hoping to run fanless so I could pot it, it may need to be left outdoors someday...

Running through back of envelope math the component selection doesn't seem way too far off actually (inductor around 270µH), the troubles I was having initially was due to bad assumptions (555's reset LOW) and not taking brownouts into consideration. So right now I think it may be core loss or IR loss, don't think it should be saturating at 40KHz.

hmm....

[stj]

Re: how hot can a toroid inductor go safely?

I would be concerned about if is getting so hot that the enamel paint degrades and you end up with a shorted coil but I could not tell you how hot it has to be for this to happen

this is the concern.

[eccerr0r]

Re: how hot can a toroid inductor go safely?

Oh man... I totally screwed up on my oscillator frequency, was running around 10kHz (spied on gate drive with a scope) when I was expecting 40KHz. It was probably core saturating. Let's see if speeding it up a bit helps though now the mosfet probably will take some more heat...

[eccerr0r]

Re: how hot can a toroid inductor go safely?

Nope, still getting hot, so I*R loss is still highly suspect. The windings aren't that thin, probably at least 18 AWG if not 16 AWG. I've never seen a 14AWG+ inductors yet except probably welders and induction heaters...

For those of you who have measured temperatures on inductors under load... what are typical?
The varnish on the windings probably could take some heat but yes they will eventually melt, sublimate, or even burn... plus not sure of what the binding material in the core can deal with.

[momaka]

Re: how hot can a toroid inductor go safely?

For those of you who have measured temperatures on inductors under load... what are typical?

I've measured a few output toroids in ATX PSUs, but not under full load and with a type -K thermocouple... so not the most precise measurements. I was getting about mid-50's C @ medium load and and mid-40's C with a light load. Higher loads or highly unbalanced 5V-12V loads (crossloads) resulted in the output toroid running the hottest, reaching almost 60C in some instances.

Nope, still getting hot, so I*R loss is still highly suspect. The windings aren't that thin, probably at least 18 AWG if not 16 AWG. I've never seen a 14AWG+ inductors yet except probably welders and induction heaters...

I've seen them in some higher power PSUs (over 500W.)
Most manufacturers prefer to put two or more parallel runs of thinner wire than a single run of thicker wire - less losses due to "skin" effect at higher frequencies.

The varnish on the windings probably could take some heat but yes they will eventually melt, sublimate, or even burn... plus not sure of what the binding material in the core can deal with.

By the time the varnish starts to melt, the core will be long overheated and damaged.

IIRC, 80-90C is about where most cores will start to get damaged from heat. In contrast, the varnish / enamel on "magnet" wire is usually specified for at least 120C. That's why most "class II" line frequency transformers are specced for up to 120C - the steel laminate plates in the core can take more (being steel and all), but the wire enamel and varnish will start to soften and could allow a short-circuit to develop somewhere.

I think this coil came out of a dead off-line PSU probably some filter inductor. I really should characterize the core to know exactly what I have, but well, it is what it is...

Well, if you haven't done that part...

This is probably the most important thing to do here. First you have to identify if the core can even handle any kind of high switching frequency or not. And second, you need to size the core appropriately. Even if the core is right for the switch frequency, you can still saturate the core if you have too many turns or too much current through fewer turns. So you really need to hunt a datasheet for the core material.

The few common ones used in ATX PSUs are:

Micrometals -26 type (yellow core with one side painted white): good for up to 50 KHz and has decently high initial permeability.

Micrometals -52 type (light green core with one side painted blue): same permeability specs as -26 type material, but with lower core losses and thus could be used at 100+ KHz switching frequency.

Micrometals -45 type (completely black core): good for ~100 KHz, but has even higher permeability compared to -26 and -52 type cores... actually, the highest for iron powder cores.

Nowadays, I see many PSU manufacturers use completely blue cores, but I'm not sure what those are. I've seen them used frequently on some P4 / Athlon 64 -era boards, along with the all-black -45 type cores, so I imagine similar characteristics overall.

Surely the question should be "How does my design look ?", not how how is too hot.

Was thinking the same thing.

Plus, it would be interesting just to see the circuit design itself. Maybe someone here could even suggest improvements without having to change the circuit too much or too many parts.

[eccerr0r]

Re: how hot can a toroid inductor go safely?

hmm...another thought, well, this thing was intended for around 100W and all that power needs to be stored in it. This is starting to push it for some atx psu parts.

Now not sure how I can characterize cores without a high power frequency generator. I'm not even sure I can trust the color of the core, though it is yellow and estimated cross section (since there's a winding on it, have to guess) is around 0.7cm² or so, ID around 1.5-2cm (again winding is already on it). Really don't want to unwind it, because I'd eventually have to rewind it...

[eccerr0r]

Re: how hot can a toroid inductor go safely?

hmm looked at it again, ID is more like 1cm. Perhaps not quite big enough for the job?

suspecting that there should be a rule of thumb somewhere on size of transformer and wattage it can pass, but I thought this should be about right, though maybe a little small. I also have a huge toroid core that has like 3-4cm² cross section and 4cm ID, but that one is in use for another design... and that one does not get hot with even 200W passing through...

---

oh boy I totally forgot the capacitor to filter the diode output! Ack! I was wondering why it was completely bonkering out AM reception, worse than I thought it should be. Added a cap there and now I don't have a few meters worth of cabling radiating switching pulses!

[sam_sam_sam]

Re: how hot can a toroid inductor go safely?

I following this post to see what you come up with

[eccerr0r]

Re: how hot can a toroid inductor go safely?

Oh fun! I found another toroid that was definitely used as a filter coil for a multi-output flyback PSU. This toroid has at least another mm² of cross section and its ID is a bit larger. It unfortunately does not have as many windings on it, and it has multiple windings. I suppose I'll ignore the 26-gauge windings and use the thicker 18-ish AWG windings and see if this behaves any different...

This one since it has fewer windings clearly is yellow with one white stripe on one side, implying a micrometals -26 core...

[momaka]

Re: how hot can a toroid inductor go safely?

This one since it has fewer windings clearly is yellow with one white stripe on one side, implying a micrometals -26 core...

Oh I suppose I'll ignore the 26-gauge windings and use the thicker 18-ish AWG windings and see if this behaves any different...

You can also connect the 18 AWG and 26 AWG windings in series if they are wound in the same direction to get a higher inductance... though I think the 26 AWG windings probably won't be good for too much current.

This toroid has at least another mm² of cross section and its ID is a bit larger.

If it has about 27 mm (1") for the outter diameter, 14-15 mm for the inner diameter, and 11-12 mm thick, it's probably a T106-26 core - a very commonly used core for older cheapo group-regulated PSUs and even non-cheapo older PSUs rated at 250W or less.

Real limit for these cores is around 300 Watts at the lower output voltages that ATX PSUs usually output, or probably could do 400-450W with much higher output voltages (and lower associated output currents.)

Oh fun! I found another toroid that was definitely used as a filter coil for a multi-output flyback PSU.

If the PSU it came out of had an output toroid, then it didn't use flyback (discontinuous) topology. With flyback/discontinuous topology, the "transformer" primary side serves as the output toroid. I put transformer in quotes, because for flyback, the transformer is gapped and actually functions more like a coupled inductor. In contrast, the transformer of a continuous topology is not gapped and does not store energy. A pulse on the input / primary side directly produces a pulse on the output / secondary. With flyback, a pulse is produced on the output after the pulse on the primary is ended and the magnetic field in the inductor collapses.

So in fewer words...
Output toroid = continuous topology only
No output toroid = discontinuous topology (what's commonly referred to as flyback)

[eccerr0r]

Re: how hot can a toroid inductor go safely?

As a filter coil (why not use linear cores as filter inductor cores?) does it really handle that much power? I suppose that's the concern and perhaps I need an even bigger core. 100W is the target since that's one solar panel, and the thin 26-gauge windings are useless, they can't handle the multiple amps needed to deal with 100W at 18-36V... though even if they are wound in the wrong direction, just reverse the connections? A little loss here and there is okay and no worries about cross coupling or insulation limits for now.

[eccerr0r]

Re: how hot can a toroid inductor go safely?

Now what if I connected a second toroid in series with the existing toroid, what will happen?

[momaka]

Re: how hot can a toroid inductor go safely?

LOL, this is like watching an episode of Garage54, except instead of messing about with old Lada's, we're observing experiments with some mystic home-brew PSU circuit.

Now what if I connected a second toroid in series with the existing toroid, what will happen?

Inductance of the two inductors added up??
As for how the circuit will react... well, that I can't tell you. We have no details or schematic of what your circuit is exactly, so my guess is probably as good as that of a squirrel.

As a filter coil (why not use linear cores as filter inductor cores?) does it really handle that much power?

All of the output current does pass through it, so yes.

If you look at any regular half-bridge or forward topology ATX SMPS, you have square wave pulses produced on the primary side (by the MOSFETs / BJTs) and reduced amplitude square wave puses at the output side of the transformer corresponding to the turns ratio. These then go through the rectifying diodes and into the output toroid, where they get "averaged" into rectangular waves (with a DC offset) and further smoothed into DC by the output caps.

As for "linear cores as filter inductor cores", not sure what you mean by that. If you're implying of using a transformer's winding as an inductor, that won't work (too well.)

[eccerr0r]

Re: how hot can a toroid inductor go safely?

Well the problem is that filter cores indeed have a LOT of DC component/offset compared to one used for boost/buck which can go MUCH further and could go discontinuous or get close to saturation. Therefore there isn't as much of -something- that the core does not need to handle.

I see a lot of PSU filter inductors wound on toroid cores but why not use I-cores... and why not E-I cores (shape, not the laminated sheets used for 60Hz)?

BTW the curiosity of two series inductors, if one inductor is saturating, adding another identical one series, would they still both saturate or can they take more now since energy is now spread among two cores.