Hi,
ok the simplest way of course would be to plug one in and try it out, but our mosfet bridge design is still a proto and doesn't go beyound a few 100W yet, so...
Two questions to you class D output inductor design gurus here:
- do you know if the Amidon T200-2 or T200-6 non-gapped toroidal powdered iron core is useable in 1000Wrms (4ohm) ~400kHz class D applications, too, as opposed to their traditional use in multi-kW RF applications?
- how well are gapped toroidal cores really suited for >500Wrms output inductors? do they leak just too much or are they still useable?
- Jan
ok the simplest way of course would be to plug one in and try it out, but our mosfet bridge design is still a proto and doesn't go beyound a few 100W yet, so...
Two questions to you class D output inductor design gurus here:
- do you know if the Amidon T200-2 or T200-6 non-gapped toroidal powdered iron core is useable in 1000Wrms (4ohm) ~400kHz class D applications, too, as opposed to their traditional use in multi-kW RF applications?
- how well are gapped toroidal cores really suited for >500Wrms output inductors? do they leak just too much or are they still useable?
- Jan
No, is the simple answer.
However, since you have asked the question I felt the need to try and explain how to design filter inductors.... it's something I'll need to try sometime so this seems like a good time to start.
The beginning is at
http://www.genomerics.org/inductors/inductors.html
Don't get too excited.... I'll let you know when there is actually something worth reading. In the mean time I'm just playing.
Cheers
DNA
However, since you have asked the question I felt the need to try and explain how to design filter inductors.... it's something I'll need to try sometime so this seems like a good time to start.
The beginning is at
http://www.genomerics.org/inductors/inductors.html
Don't get too excited.... I'll let you know when there is actually something worth reading. In the mean time I'm just playing.
Cheers
DNA
I can only find permeability figures for those cores, being 12nH/turn^2 for T200-2 and 10nH/turn^2 for T200-6. That would give suitable inductance values around 30uH with 50 turns, but I can't find data on energy storage capabilities and core losses versus frequency and flux density.
As a reference, I have standard yellow-white iron-powder cores of similar dimensions whose permeability is around 200nH/turn and whose measured energy storage capability is 8mJ (.5*L*Isat^2). That energy storage figure seems quite insufficient for most applications, and these cores will become quite hot if even operated above 50Khz, so I'm sticking to gapped ferrite inductors. However, the Amidon ones feature much lower permeability so they should feature considerably higher energy storage figures and lower losses allowing for operation at higher frequencies.
By the way, could anybody recommend me affordable iron powder materials with high energy storage capabilities? High frequency operation is not a requirement for some of my projects.
As a reference, I have standard yellow-white iron-powder cores of similar dimensions whose permeability is around 200nH/turn and whose measured energy storage capability is 8mJ (.5*L*Isat^2). That energy storage figure seems quite insufficient for most applications, and these cores will become quite hot if even operated above 50Khz, so I'm sticking to gapped ferrite inductors. However, the Amidon ones feature much lower permeability so they should feature considerably higher energy storage figures and lower losses allowing for operation at higher frequencies.
By the way, could anybody recommend me affordable iron powder materials with high energy storage capabilities? High frequency operation is not a requirement for some of my projects.
That's the old...... 'Not enough useful Data' problem.
I'm not certain but I think you might find that Arnold are doing lots of business because they provide 'useful' data.
Arnold
You might find that companies that provide finished products use a lot of this stuff.
Interpreting it is another thing, it is in there somewhere though. Would be nice to have some idea of thermal resistances.
Cheers
DNA
I'm not certain but I think you might find that Arnold are doing lots of business because they provide 'useful' data.
Arnold
You might find that companies that provide finished products use a lot of this stuff.
Interpreting it is another thing, it is in there somewhere though. Would be nice to have some idea of thermal resistances.
Cheers
DNA
Next page is up. First part of the Magic Sum.
First page has some links on it for good data.
It's linked on the front page under Other Bits, Filter Inductor Design.
Genomerics
Cheers
DNA
First page has some links on it for good data.
It's linked on the front page under Other Bits, Filter Inductor Design.
Genomerics
Cheers
DNA
Thanks everyone for your input!
Micrometals.com indeed had some "specs", http://www.micrometals.com/appnotes/appnotedownloads/ipcs4rfp.pdf. Page 9 figure K says T200-2 (material 2) is supposed to have a "power rating" of 800W @ 1..2 MHz (for 25degC temp rise due to core losses). OTOH, figure D there claims ~3000gauss 500kHz => 0.5W/cm^3 core losses - 0.5 seems ok, AFAIK comparable to power ferrite...!
http://toroids.info/T200-2.asp : Material = Carbonyl E. More googling on that material yielded "high volume resistivity", "rf inductors", "resonant inductors above 50kHz".
And further, this looks actually quite good for T-200-2: http://www.41hz.com/downloads/TA2022.pdf : Page 22: "Tripath recommends that the customer use a toroidal inductor with a Carbonyl-E core for all applications of the TA2022."
In that TA2022 tech info paper, Tripath has used and recommends the significantly smaller core T68-2, for ~100W. So I'd guess the T-200-2 should be suitable, and capable to pack a lot more of them little watts
Anyways. The Arnolds pdf is very comprehensive, too, thanks for the link!
Your genomerics.org inductor design page is nice. plus has some attitude there... Hmm - somehow reminds of sci.electronics.design DNA / Genome. Btw there is also some sort of automatic calculation that selects suitable inductor cores, not sure but you might find it useful/interesting: http://schmidt-walter.fbe.fh-darmstadt.de/smps_e/l_smps_in_e.html
with infos http://schmidt-walter.fbe.fh-darmstadt.de/smps_e/etd_hilfe_e.html
Micrometals.com indeed had some "specs", http://www.micrometals.com/appnotes/appnotedownloads/ipcs4rfp.pdf. Page 9 figure K says T200-2 (material 2) is supposed to have a "power rating" of 800W @ 1..2 MHz (for 25degC temp rise due to core losses). OTOH, figure D there claims ~3000gauss 500kHz => 0.5W/cm^3 core losses - 0.5 seems ok, AFAIK comparable to power ferrite...!
http://toroids.info/T200-2.asp : Material = Carbonyl E. More googling on that material yielded "high volume resistivity", "rf inductors", "resonant inductors above 50kHz".
And further, this looks actually quite good for T-200-2: http://www.41hz.com/downloads/TA2022.pdf : Page 22: "Tripath recommends that the customer use a toroidal inductor with a Carbonyl-E core for all applications of the TA2022."
In that TA2022 tech info paper, Tripath has used and recommends the significantly smaller core T68-2, for ~100W. So I'd guess the T-200-2 should be suitable, and capable to pack a lot more of them little watts
Anyways. The Arnolds pdf is very comprehensive, too, thanks for the link!
Your genomerics.org inductor design page is nice. plus has some attitude there...
Hmm - somehow reminds of sci.electronics.design DNA / Genome. Btw there is also some sort of automatic calculation that selects suitable inductor cores, not sure but you might find it useful/interesting: http://schmidt-walter.fbe.fh-darmstadt.de/smps_e/l_smps_in_e.htmlwith infos http://schmidt-walter.fbe.fh-darmstadt.de/smps_e/etd_hilfe_e.html
grr, mixup with materials... All reads in the TA2022 pdf, didn't read it too closely the first time.
Amidon's carbonyl E is apparently material -06 and not -2, so the core might actually be T200-6 (or, hmm!? old -2 but new -06? go figure ). and not T200-2. And Micrometals carbonyl E material code is -2 (e.g. T94-2). Wouldn't be the least bit surprised if the color codes were different, too, at least for ferrite toroids the colors mean zip...
But odd that many x kW <3MHz RF baluns use the T200-2. So, hmm... dammit, this has me confused...
Wrt "T200-2 like toroids", good observation there Not sure, but from http://www.coldamp.com/opencms/opencms/coldamp/en/ it seems this is painted wine red. Amidon pages at http://www.amidoncorp.com/aai_ironpowdercores.htm say T200-2 color is red. /Might/ be a T200-2, then, provided the core is from Amidon. Argh. So if coldamp also maybe uses T200-2.. Oh well, I'll just have to buy one of those T200-2's or a smaller -2'one and give it a try!
Amidon's carbonyl E is apparently material -06 and not -2, so the core might actually be T200-6 (or, hmm!? old -2 but new -06? go figure
). and not T200-2. And Micrometals carbonyl E material code is -2 (e.g. T94-2). Wouldn't be the least bit surprised if the color codes were different, too, at least for ferrite toroids the colors mean zip... But odd that many x kW <3MHz RF baluns use the T200-2. So, hmm... dammit, this has me confused...
Wrt "T200-2 like toroids", good observation there
Not sure, but from http://www.coldamp.com/opencms/opencms/coldamp/en/ it seems this is painted wine red. Amidon pages at http://www.amidoncorp.com/aai_ironpowdercores.htm say T200-2 color is red. /Might/ be a T200-2, then, provided the core is from Amidon. Argh. So if coldamp also maybe uses T200-2.. Oh well, I'll just have to buy one of those T200-2's or a smaller -2'one and give it a try!Can you have too big a core?
Here's a question...what's more important, the material or the size of the core? Obviously, having too small a core can lead to saturation, or bad thermal stuff, but is it possible to have too large a core? e.g. if I have a plan that calls for a T106-2 core, is there any detriment (other than size, and possibly DCR, depending on the turns required) to using a larger core of the same material, e.g. T200-2 core?
--Greg
Here's a question...what's more important, the material or the size of the core? Obviously, having too small a core can lead to saturation, or bad thermal stuff, but is it possible to have too large a core? e.g. if I have a plan that calls for a T106-2 core, is there any detriment (other than size, and possibly DCR, depending on the turns required) to using a larger core of the same material, e.g. T200-2 core?
--Greg
Jan, have you tried Micrometals inductor design software?
http://www.micrometals.com/software_index.html
With Idc=20A, 30uH 400khz and 3A inductor ripple current T175-2 should be enough, calculate with your own values to see if T200 suits your needs.
Amidon specs are quite difficult to interpret in this kind of like use(mostly RF-stuff) , I would go for micrometals just to get more meaningful specsheets.
btw:
How about using gapped ferrite?
Or is MPP/Sendust/hi-flux too expensive ?
http://www.micrometals.com/software_index.html
With Idc=20A, 30uH 400khz and 3A inductor ripple current T175-2 should be enough, calculate with your own values to see if T200 suits your needs.
Amidon specs are quite difficult to interpret in this kind of like use(mostly RF-stuff) , I would go for micrometals just to get more meaningful specsheets.
btw:
How about using gapped ferrite?
Or is MPP/Sendust/hi-flux too expensive ?
The OP is asking about 400kHz Class D application. My application is also for output inductor of Class D amp. Not SMPS application.
Well, a class D amp is more or less a switching power supply with a variable reference... Anyway, you might want to look at type 8/90 (yellow/red) material as well. It not quite as high frequency capable as the #2 (red) material, but it is still reasonably low loss (if I remember correctly, it's rated for use out to 8 MHz or so), and the permeablity is higher. It is definitely much better than the type 26 (yellow/white)or 52 (blue/green) materials one would normally consider for SMPS applications. You'd be trading off some core loss for lower copper loss (fewer turns or a smaller mean length per turn). For the ultimate in low core loss material, you can try #0, which is just phenolic ( I'm joking here, but the stuff does exist). I've also used #18 (red/green) material in an inductor for a 400kHz DC-DC converter output indutor and it seemed to work OK, but it was a physically small inductor with high surface to volume ratio. A much larger inductor may be different. Type 18 is intermediate in permeability and loss between the SMPS materials and type 8/90 (also intermediate in cost).
BTW, the Amidon powdered iron stuff is mostly Micrometals cores renamed, so you can get the complete scoop on the materials from the Micrometals web site. Micrometals is also pretty reasonable about samples if you find something at their site that Amidon doesn't stock. Depending on the flux swing, I'd also consider using gapped ferrite at 400kHz. It all depends on how much inductance you're shooting for.
BTW, the Amidon powdered iron stuff is mostly Micrometals cores renamed, so you can get the complete scoop on the materials from the Micrometals web site. Micrometals is also pretty reasonable about samples if you find something at their site that Amidon doesn't stock. Depending on the flux swing, I'd also consider using gapped ferrite at 400kHz. It all depends on how much inductance you're shooting for.
I don't know how applicable this is to this application, but people who have used powdered iron cores in PFC applications ( continuous conduction) have been bedeviled with long term failure issues. The AC flux causes local hot spots in the core material at inhomogeneities, causing progessive breakdown due to deterioration of the core binding material. The whole choke eventually fries to a crisp as the binding material breaks down and allows the iron particles to clump togehter, increasing losses astronomically. The nasty thing is the deterioration is slow enough that the problem will not crop up in the course of a typical design evaluation.The choke would run dead cold in the lab, then eventually burn up in the field. The problem also happened in supposedly low loss materials. This didn't happen when the material was used in applications like an output choke with relatively small AC flux on top of a large DC bias.
Micrometals responded to the problem by developing a high temperature material. It might be a good idea to email Dale Nicol at Micrometals with a description of your application, including the proposed AC flux density, if known. He should be able to tell you if there is a potential problem in your application. Maybe it's not a problem here, but it's worth an email to find out. It may be less of a problem wih the lower perm materials that have a higher proportion of binder, and possibly a different type of powdedred iron as well ( carbonyl iron vs iron flake manufactured in a ball mill). The thing to remember, though, is that the mix may not be perfect, and an imperfection like a larger piece of iron can act as a seed point for this type of failure, yet not show up in a typical inductance or Q measurement. This particular issue slipped my mind when I first read this thread, but the combination of high frequency and possible high AC flux density rang an alarm bell a little later.
This failure mechanism doesn't happen with ferrites.
Micrometals responded to the problem by developing a high temperature material. It might be a good idea to email Dale Nicol at Micrometals with a description of your application, including the proposed AC flux density, if known. He should be able to tell you if there is a potential problem in your application. Maybe it's not a problem here, but it's worth an email to find out. It may be less of a problem wih the lower perm materials that have a higher proportion of binder, and possibly a different type of powdedred iron as well ( carbonyl iron vs iron flake manufactured in a ball mill). The thing to remember, though, is that the mix may not be perfect, and an imperfection like a larger piece of iron can act as a seed point for this type of failure, yet not show up in a typical inductance or Q measurement. This particular issue slipped my mind when I first read this thread, but the combination of high frequency and possible high AC flux density rang an alarm bell a little later.
This failure mechanism doesn't happen with ferrites.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.