Ferrite Inductor Tolerance
in [Design]
|
From: John Larkin on 13 Jun 2010 22:41 On Sun, 13 Jun 2010 17:51:38 -0700, Joerg <invalid(a)invalid.invalid> wrote: >Ian Bell wrote: >> On 13/06/10 21:26, Joerg wrote: >>> Ian Bell wrote: >>>> On 13/06/10 16:02, Joerg wrote: >>>>> Ian Bell wrote: >>>>>> On 12/06/10 23:00, Joerg wrote: >>>>>>> Ian Bell wrote: >>>>>>>> On 12/06/10 16:07, John Larkin wrote: >>>>>>>>> On Sat, 12 Jun 2010 10:22:59 +0100, Ian Bell<ruffrecords(a)yahoo.com> >>>>>>>>> wrote: >>>>>>>>> >>>>>>>>>> When winding modest inductors of a few hundred milliHenries on a >>>>>>>>>> ferrite >>>>>>>>>> core, given Al and a number of turns, what is the typical >>>>>>>>>> tolerance on >>>>>>>>>> the actual value of inductance when these are made in quantity? >>>>>>>>>> >>>>>>>>>> Cheers >>>>>>>>>> >>>>>>>>>> Ian >>>>>>>>> >>>>>>>>> An ungapped ferrite core could be all over the place. 25% wouldn't >>>>>>>>> surprise me. They will vary with temperature, too. You can buy >>>>>>>>> gapped >>>>>>>>> pot cores in tolerances around 2-5%, I think. Or use a pot core >>>>>>>>> with a >>>>>>>>> slug adjuster if you need 1% or better. See the datasheets. >>>>>>>>> >>>>>>>>> Powder-type cores can be bought with better tolerances. >>>>>>>>> >>>>>>>>> The people who wind inductors commercially get the exact number of >>>>>>>>> turns every time. >>>>>>>>> >>>>>>>>> John >>>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> It seems to me there are quite a few factors that could affect the >>>>>>>> actual inductance achieved and perhaps the least of them is the >>>>>>>> accuracy >>>>>>>> in counting the number of turns. I would expect there to be some >>>>>>>> tolerance in the Al value of the ferrite, that its exact dimensions >>>>>>>> would have an effect along with how neatly or otherwise the turns >>>>>>>> are >>>>>>>> wound. I have absolutely no idea if these are the major factors >>>>>>>> nor of >>>>>>>> the likely size of the actual major factors affecting the actual >>>>>>>> inductance. I am just trying to get a feel for the likely >>>>>>>> tolerance of >>>>>>>> ready made inductors. >>>>>>>> >>>>>>>> The reason I ask is am am designing some passive audio filters and I >>>>>>>> know exactly what tolerance of resistance and capacitance I can >>>>>>>> obtain >>>>>>>> but not a clue about inductance. It is no good me using 1% >>>>>>>> capacitors >>>>>>>> and resistors if inductors normally fail to achieve 5%. >>>>>>>> >>>>>>> >>>>>>> You can get 5% catalog inductors from Delevan, Miller, TDK and >>>>>>> several >>>>>>> others. If it needs to be more precise then you'd be off to boutique >>>>>>> lines, meaning $$$. >>>>>>> >>>>>>> Example: >>>>>>> >>>>>>> http://www.delevan.com/seriesPDFs/1782.pdf >>>>>>> >>>>>> >>>>>> Those devices seem to be in the sub milliHenry range, the one I >>>>>> need are >>>>>> in the humfreds of milliHenries. >>>>>> >>>>> >>>>> Then you'll likely have to live with 5%, for example: >>>>> >>>>> http://www.yuden.co.jp/us/product/pdf/elhl_e.pdf >>>>> >>>>> There are other ideas but you'd have to let us know about the nature of >>>>> the product. Things such as yearly qty, why it must be passive, how >>>>> many >>>>> inductors per unit, whether end-test trimming is ok, et cetera. >>>>> Otehrwise it'll all be speculation. >>>>> >>>> >>>> Thanks for the input but I think you have misunderstood me. I already >>>> have a source of suitable inductors. What I am interested in is the >>>> factors that govern the tolerance of a production inductor (as it comes >>>> off the line and before any selection process) and what the resultant >>>> overall tolerance is likely to be for inductors around 1H. I am not >>>> asking for help designing a product or in selecting parts. >>>> >>> >>> Having designed several custom inductors and transformers, the largest >>> tolerance contribution came from the core material. Not the dimensions, >>> those are very precise, but from the materials properties. >>> >> >> Thanks Joerg, that is exactly the sort of thing I was looking for. I >> have heard the Al can vary by 20% for some core types. has this been >> your experience too? >> > >Yes, 20% is quite common, often 30%. It depends on where you get the >core material from. It can be quite constant for hundreds of cores and >then all of a sudden there is a jump to another value. IOW, you cannot >rely on measurements and extrapolate. At least not for longer or larger >production runs. Ground-gap pot cores can be pretty good. Not cheap. John
From: John Larkin on 13 Jun 2010 22:39 On Mon, 14 Jun 2010 00:12:58 +0100, Ian Bell <ruffrecords(a)yahoo.com> wrote: >On 13/06/10 19:10, John Larkin wrote: >> On Sun, 13 Jun 2010 18:40:25 +0100, Ian Bell<ruffrecords(a)yahoo.com> >> wrote: >> >>> On 13/06/10 16:02, Joerg wrote: >>>> Ian Bell wrote: >>>>> On 12/06/10 23:00, Joerg wrote: >>>>>> Ian Bell wrote: >>>>>>> On 12/06/10 16:07, John Larkin wrote: >>>>>>>> On Sat, 12 Jun 2010 10:22:59 +0100, Ian Bell<ruffrecords(a)yahoo.com> >>>>>>>> wrote: >>>>>>>> >>>>>>>>> When winding modest inductors of a few hundred milliHenries on a >>>>>>>>> ferrite >>>>>>>>> core, given Al and a number of turns, what is the typical tolerance on >>>>>>>>> the actual value of inductance when these are made in quantity? >>>>>>>>> >>>>>>>>> Cheers >>>>>>>>> >>>>>>>>> Ian >>>>>>>> >>>>>>>> An ungapped ferrite core could be all over the place. 25% wouldn't >>>>>>>> surprise me. They will vary with temperature, too. You can buy gapped >>>>>>>> pot cores in tolerances around 2-5%, I think. Or use a pot core with a >>>>>>>> slug adjuster if you need 1% or better. See the datasheets. >>>>>>>> >>>>>>>> Powder-type cores can be bought with better tolerances. >>>>>>>> >>>>>>>> The people who wind inductors commercially get the exact number of >>>>>>>> turns every time. >>>>>>>> >>>>>>>> John >>>>>>>> >>>>>>> >>>>>>> >>>>>>> It seems to me there are quite a few factors that could affect the >>>>>>> actual inductance achieved and perhaps the least of them is the accuracy >>>>>>> in counting the number of turns. I would expect there to be some >>>>>>> tolerance in the Al value of the ferrite, that its exact dimensions >>>>>>> would have an effect along with how neatly or otherwise the turns are >>>>>>> wound. I have absolutely no idea if these are the major factors nor of >>>>>>> the likely size of the actual major factors affecting the actual >>>>>>> inductance. I am just trying to get a feel for the likely tolerance of >>>>>>> ready made inductors. >>>>>>> >>>>>>> The reason I ask is am am designing some passive audio filters and I >>>>>>> know exactly what tolerance of resistance and capacitance I can obtain >>>>>>> but not a clue about inductance. It is no good me using 1% capacitors >>>>>>> and resistors if inductors normally fail to achieve 5%. >>>>>>> >>>>>> >>>>>> You can get 5% catalog inductors from Delevan, Miller, TDK and several >>>>>> others. If it needs to be more precise then you'd be off to boutique >>>>>> lines, meaning $$$. >>>>>> >>>>>> Example: >>>>>> >>>>>> http://www.delevan.com/seriesPDFs/1782.pdf >>>>>> >>>>> >>>>> Those devices seem to be in the sub milliHenry range, the one I need are >>>>> in the humfreds of milliHenries. >>>>> >>>> >>>> Then you'll likely have to live with 5%, for example: >>>> >>>> http://www.yuden.co.jp/us/product/pdf/elhl_e.pdf >>>> >>>> There are other ideas but you'd have to let us know about the nature of >>>> the product. Things such as yearly qty, why it must be passive, how many >>>> inductors per unit, whether end-test trimming is ok, et cetera. >>>> Otehrwise it'll all be speculation. >>>> >>> >>> Thanks for the input but I think you have misunderstood me. I already >>> have a source of suitable inductors. >> >> Then why don't you ask them? >> > >I have asked them. What I am asking here is what is typical/normal and >what are the contributing factors. What do they suggest as normal tolerances? What can they do for more money? John
From: Archimedes' Lever on 13 Jun 2010 23:57 On Sun, 13 Jun 2010 19:41:03 -0700, John Larkin <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote: >Ground-gap pot cores can be pretty good. Not cheap. > >John Specified pre-gapped core orders ARE cheap. We used to get them all the time. We would also buy zero gap, and gap the set ourselves. Gap is important when conditioning oscillator circuits, particularly at zero crossing points of transistor/FET pairs.
From: Michael A. Terrell on 14 Jun 2010 03:08 Nico Coesel wrote: > > Archimedes' Lever <OneBigLever(a)InfiniteSeries.Org> wrote: > > >On Sun, 13 Jun 2010 13:23:29 -0700, Joerg <invalid(a)invalid.invalid> > >wrote: > > > >> > >>It seems you have never dealt with the financials in production? > >>Component culling is heavily frowned upon by CFOs and accountants, for > >>obvious reasons. > > > > > > It seems that you have never dealt with the electronics industry for > >the last 50 years. > > > > Matching and culling was REQUIRED in many instances due to so many > >variables that were around back then. You have no clue. You also have > >no clue as to how such needs and tasks (and designs) were optimized to > >minimize losses. > > One of my former teachers had an interesting statement where it came > to tolerances: > > "Electronics is the art of cancelling component variations." > > IOW: design smart and tolerances are not a problem at all. I suppose he designed radios in the '30s? -- Anyone wanting to run for any political office in the US should have to have a DD214, and a honorable discharge.
From: Michael A. Terrell on 14 Jun 2010 03:10
"krw(a)att.bizzzzzzzzzzzz" wrote: > > On Sun, 13 Jun 2010 14:50:22 -0700, Archimedes' Lever > <OneBigLever(a)InfiniteSeries.Org> wrote: > > >On Sun, 13 Jun 2010 14:43:31 -0700, Joerg <invalid(a)invalid.invalid> > >wrote: > > > >>krw(a)att.bizzzzzzzzzzzz wrote: > >>> On Sun, 13 Jun 2010 14:08:03 -0700, Joerg <invalid(a)invalid.invalid> wrote: > >>> > >>>> Archimedes' Lever wrote: > >>>>> On Sun, 13 Jun 2010 13:24:21 -0700, Joerg <invalid(a)invalid.invalid> > >>>>> wrote: > >>>>> > >>>>>> Oh, and how do you suppose you get an "original design" into production > >>>>>> without an ECO? > >>>>> > >>>>> You must also be an acronymical retard as well. > >>>>> > >>>>> There is a difference between a design release and a change order of an > >>>>> existing design. > >>>>> > >>>> So you guys release new designs without due ECO process? I sure hope you > >>>> don't design anything that can harm people. > >>> > >>> AlwaysWrong doesn't design anything, so no he doesn't release ECOs. Everyone > >>> else here who does, uses an ECO process, certainly. > >> > >> > >>Yup. > >> > >>Maybe they have a different release process for new stuff even though it > >>typically _changes_ a product from previous to next generation. Having > >>two different release processes doesn't strike me as particularly smart, > >>but who knows :-) > > > > ECO is NOT for a "release", and not all releases are next gen designs > >of previous work, idiot. > > AlwaysWrong is , *SURPRISE*, wrong again. It is an engineering change to the > database so an Engineering Change Order is necessary. It's really that > simple, AlwaysWrong. You should know simple, by mirror. > > > Two proofs that you are chasing your own tail. > > DimBulb just can't avoid the hind-end references. If his parents had read that last ECO, dimmie would have been born human. :( -- Anyone wanting to run for any political office in the US should have to have a DD214, and a honorable discharge. |