Tube DIY Asylum Do It Yourself (DIY) paradise for tube and SET project builders.

---

[ Follow Ups ] [ Post Followup ] Thread: [ Display  All  Email ] [ Tube DIY Asylum ]

---

[ Alert Moderator ]

---

RE: c core output transformers vs E&I

Posted by mqracing (M) on July 28, 2008 at 16:24:03

In Reply to: RE: c core output transformers vs E&I posted by sser2 on July 27, 2008 at 13:50:05:

Hi Sser:

thought I might offer some comments...

::::1. Because coils are much longer, windings are much closer to the core. As a result, for the same number and arrangement of turns, stray inductance is much lower. A transformer thus can be made with less interleaving, translating into less parasitic capacitance between primary and secondary.::::

of course, this is not always the case--- it depends on the window's aspect ratio--- and the particular dimensions of the c-core employed.

Long windows (and hence long winding lengths) have advantages and disadvantages. All other things being equal the longer the winding length the less leakage L you will get--- but---- your C (capacitance) from layer to layer will be commensurately much larger--- so you leakage goes down and your C goes up as a first approximation.

To understand how the C goes up--- think of each winding layer as being a plate of a capacitor--- it you have two plates and a voltage difference btwn them--- you have a capacitor--- the larger the plates (if distance btwn the plates remains constant) the more C you will get.

::::2. Because average turn length is shorter, winding DCR is lower.::::

some truth to this--- but I question assigning it too much importance in a design--- given that on a well designed EI core we can (on output transformers) obtain insertion losses of one quarter of a DB or less.
This quarter of a db insertion loss is often less than the insertion losses I've seen on many commercial C-cored transformers.

::::3. Much more room for copper than EI core.:::::

again--- this is too much of a generalization--- c-cores come in a near infinite variety of dimensions--- so it would depend on the particular dimensions that were used on a specific design--- and bear in mind that there are some special EI shapes with commensurately huge windows--- look at the EI 537 for example.

Another interesting EI shape to look at is the EI-7 and EI-8---- not your
"normal scrapless" ratios of window height to window length--- and the overall area of the window is quite large compared to a scrapless EI shape.

::::4. Thinner laminations -less Eddy current losses.::::

the most common thickness for c-cores is 11 thousandths thick--- this thinner stock is often used because of the aspect ratio of the window and the longer sides make the bending of the core material over the mandrel much more difficult because of the often times shorter sides (compare the D and E dimensions)----

we had a custom c-core made for us--- with the standard 11 mil thick grain oriented silicon steel--- and given the length of the window versus the depth of the window--- (remember the first turn of coil stock forms the window)--- the c-core tended to bulge in the middle on the long axis----
the solution was to go with a thinner guage for this particular design--- not because of wanting to lower eddy losses--- but so that the design could be built well and be square and not have bulges in it.

You might also take into account that EI's can be ordered in several different lamination grades and thicknesses--- you can easily get 7 mil, 9 mil, and 11 mil in most of the common catalog EI shapes.

The question is--- the thinner the lamination or material used to make the c-core--- the poorer the stacking factor---- thin lams have poorer stacking factors almost universally. Stacking factor is a very important criterion design wise.

::::5. If made of grain-oriented steel, magnetic field lines are always parallel to grain, resulting in higher magnetic permeability.::::

True, true. But the question then becomes--- how does this pan out in the real world?

Much of the efficiency of the c-core is going to be lost because you have to cut it into two parts---- so you have a relatively large airgap--- and it is a contiguous airgap (not distributed)--- it is mechanically similar to "butt stacking" an EI core. this airgap kills permeability in a hurry and raises the exciting current of the core.

C-core manufacturer's try to combat this by offering (but it is an option--- and somewhat expensive) of polishing the two surfaces resulting from the core being cut into two---- but this leads to another problem of the polishing "shorting" the lams against each other--- so then you etch the individual plies to "insulate" them from each other----

c-cores are wound under tension---- sort of like winding a coil--- you have an arbor and the material is slit to the proper width---- it is then wound on the arbor to the right build thickness---- with the start and the finish being welded together--- later these most outside and inside turns will be stripped off---- after it is wound--- it is then impregnated--- then cut into two---- and since it was wound under tension you get some spring action from the tension being released---- and they will have some tendency to spring apart---- good c-core companies can combat many of these "problems"--- and you can get good products---- I'm just trying to give some sense of how c-cores are made.

C--cores also have problems with "tilting"--- where the cut and the further processing (annealing) makes the two halves of the core not seated together perfectly---- so that one side will lean.

Old time engineers who I've talked to will sometimes mention "c-core chatter"--- alluding to the tendency of the laminations to sing and dance )i.e., be noisy) where they have been cut--- since they are not supported like an EI would bee.

there is no magic to c-cores---- you can take any EI design and have a c-core made to replace it. You can use a single c-core or a double c-core.

::::Advantage of EI cores:

-More compact transformer. For the same size of transformer, two times larger core cross-section. So, if high output power at low frequencies is a goal, C-core transformer may be too big. :::::

this throw me off a bit--- when we size transformers we usually size by the core's cross-sectional area (it's net area not it's gross area)--- so the core with the larger cross sectional area is the core I would ordinarily call the larger core---- irrespective of it's window dimensions.

and if your after high output power at low frequencies then the transformer with the largest core cross-sectional area (disregarding number of turns for the moment) will have the greatest power handling capacity irrespective of whether it is an EI or a C-core.

One of the relative (again--- if we are speaking in generalities) is that an EI core is more efficient because it will have (in many or most cases) a shorter magnetic path length.

Over on the MQ forum Voltsec had just posted up a resource---- a really good paper on transformer design--- with a fairly good treatment of core design requirements and a look at several different core types and etc.
His post is still near the top--- I would recommend a read of it and the url he provides a hot link to.

Here is an interesting thought experiment--- take a really good PP output---- say one of the unit's from Peerless' top line series---- which, coincidentally, were made on EI's almost exclusively---- and substitute in a c-core (your choice of a single or double c-core)---- then bench it--- do some measurements of exciting current, of L, and any other factors you might want to---- and tell me if the c-core made any improvements at all or if the performance suffers. My educated hunch is that the EI will do quite well--- thank-you.

MSL

Edits: 07/28/08

This post is made possible by the generous support of people like you and our sponsors:

Kimber Kable

---

[ Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2025, Audio Asylum, all rights reserved ]

---

[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]

---