Cores for 400Hz sine inverters.

[vimalkhanna]

Need to build an inverter 150watts pure sine wave .
However , core materials to be used are not clear .CRGO tape wound cores have poor retention and Bmax beyond 100Hz.
Ferrites dont function below 5KHz.as losses are heavy.
What should go for .?Suggestions invited
Advise whether to opt for transformerless designs?

 

[mtwieg]

I think ordinary laminated electrical steel should be suitable for 400Hz. Core losses can be higher than with 60Hz, so you'll have to increase the number of turns to balance things out. Ferrites don't have a lower cutoff frequency, but they're not really suitable for this simply due to their low Bsat and permeability.

 

[dick_freebird]

This is a standard shipboard power frequency and you ought
to be able to find surplus transformers, and follow the model
number and manufacturer down to core material details (if
one transformer isn't all you needed).

 

[vimalkhanna]

I agree with both .However , even a tapewound C Core falls insufficient and size grows .I tried getting an army /navy scrap transformer BUT the epoxy embodiment damages the airgap and bobbin to prevent reuse .
Best bet is going for transformerless design with Mosfet bridge modulator using a 350VDC supply .
Can honourable experts suggest a simple design for assembly?

 

[mtwieg]

I agree with both .However , even a tapewound C Core falls insufficient and size grows .I tried getting an army /navy scrap transformer BUT the epoxy embodiment damages the airgap and bobbin to prevent reuse .

I'm not sure what you're talking about here. Is it just a matter the transformers still being too big, too heavy, or too inefficient? What kind of requirements do you have?

Best bet is going for transformerless design with Mosfet bridge modulator using a 350VDC supply .
Can honourable experts suggest a simple design for assembly?

Are you sure you don't need the isolation? If not then you don't need the transformer on the inverter output, and that doesn't really change anything else. Besides having isolation, the transformer doesn't really change the design (unless you need to step up or step down the AC output).

 

[FvM]

Core losses can be higher than with 60Hz, so you'll have to increase the number of turns to balance things out.

I guess you mean an increased number of turns relative to frequency respectively reduced flux. Reducing the flux to 40% of a standard 50/60 Hz transformer should result in about similar core losses and a core size of a 60 VA 60 Hz transformer. Not really big.

 

[vimalkhanna]

When flux density is reduced to 30-35% the relative windings need to be increased by n2 times .The stepup needs to be done for generation of 230VAC /400Hz from a battery source of 24VDC.As such we require Pri:sec turns ratio of 1:11 as it is .Increase the same to square and you get prohibitively large no. of turns ...
When the transformer is replaced by a mosfet brigdge ,, the isolation is no doubt lost .
However , this can be done by having small ETD34 ferrite for the generation of the 350VDC from 24VDC battery..The same can be pushpull mode osc. using 3842 PWM ,,or better TL494 in which the oscillator drive can be PWM modulated with the pure sine to achieve a better duty cycle and efficiency....help me/guide if I am incorrect ....

 

[FvM]

You didn't tell anything about room or weight restrictions and if you are prepared for a state-of-the-art high frequency DC/DC "transformerless" design.

 

[vimalkhanna]

Thanks .However ,I had written about core losses and selection/rejection criteria as well as selecting a transformerless design in case no core fit into my frequency range.

 

[FvM]

400 Hz is a classical domain of "metal sheet stack" transformers, e.g. allover aviation electrics. In so far, they clearly fit the frequency range. If they also fit your application depends on the additional points that I asked for.

 

[mtwieg]

I guess you mean an increased number of turns relative to frequency respectively reduced flux. Reducing the flux to 40% of a standard 50/60 Hz transformer should result in about similar core losses and a core size of a 60 VA 60 Hz transformer. Not really big.

Could you clarify this? I'm only saying that in a high frequency core you want the flux swing to be lower than in a lower frequency core, thus requiring a lower number of turns as opposed to a transformer design which is saturation limited (most 50/60Hz transformers are saturation limited, not core loss limited).

When the transformer is replaced by a mosfet brigdge ,, the isolation is no doubt lost .
However , this can be done by having small ETD34 ferrite for the generation of the 350VDC from 24VDC battery..The same can be pushpull mode osc. using 3842 PWM ,,or better TL494 in which the oscillator drive can be PWM modulated with the pure sine to achieve a better duty cycle and efficiency....help me/guide if I am incorrect ....

Yes, this is certainly an option. I was under the impression that most inverters made nowadays used this design.

Thanks .However ,I had written about core losses and selection/rejection criteria

Uh, I'm not seeing any criteria anywhere.

 

[FvM]

Could you clarify this?

I think, it has been sufficiently clarified in the core size example: A reasonably designed 400 Hz transformer will have considerably reduced number of turns and core size compared to a 50/60 Hz transformer, even with reduced flux.

 

[mtwieg]

I think, it has been sufficiently clarified in the core size example: A reasonably designed 400 Hz transformer will have considerably reduced number of turns and core size compared to a 50/60 Hz transformer, even with reduced flux.

When I referred to adding more turns to a 400Hz transformer, I wasn't comparing to a 50/60Hz transformer of the same size, but rather to a 400Hz transformer with the same flux swing as a 50/60Hz (saturation limited) transformer. It's harder to compare to an actual 50/60Hz transformer of the same size (such a comparison is kind of moot since the goal is to shrink the transformer size). I imagine that a small optimized 400Hz transformer with reduced flux swing might have more primary turns than a larger 50/60Hz one. Hard to really say without some solid numbers to go on.

 

[vimalkhanna]

Dear FvM ,,
,Kindly clarify "400 Hz is a classical domain of "metal sheet stack" transformers, e.g. allover aviation electrics." and give me an idea where and what to buy ..I need to make a proto initially and then about 100+ qty after I finalize .
Can I get C tape wound core with lapped finish to minmize core losses in the airgap.

 

[FvM]

I just wanted to mention the fact, that 400 Hz equipment mostly uses regular transformers. Unfortunately I don't have design data for 400 Hz transformers.

 

[ALERTLINKS]

Use thin silicon metal sheet EI core which is readily available. You should know audio transformers deliver more than 10KHz using metal cores. As frequency is doubled, turns of winding are half. So turns for 400Hz will be 1/8 of 50 Hz winding. And the wattage will be 8 timers than that of the same size. This is in theory. Practically consider over stress of core and other increased losses so give it an edge.

 

[vimalkhanna]

Dear FvM & Alertlinks ,Thanks
I had thought of using CRGO /CRD /sheet silicon versions for my application but these are available with N67 core which gives 3600 as Bmax....For audio implementation this has lower relevance and more important is the response purity (distortionfree signals )
For high efficiency power conversions input power to output delivery ,,,curtailment of harmonics and flux density to 5500 has a large bearing .
The pure sinewave means a distortion level ranging to 2% max .meaning no clipping of the even -2nd/4th,6th /8th harmonics and no peaking of the odd 3rd /5th harmonics .
As such my search isstill open......

---------- Post added at 06:23 ---------- Previous post was at 06:22 ----------

Dear FvM & Alertlinks ,Thanks
I had thought of using CRGO /CRD /sheet silicon versions for my application but these are available with N67 core which gives 3600 as Bmax....For audio implementation this has lower relevance and more important is the response purity (distortionfree signals )
For high efficiency power conversions input power to output delivery ,,,curtailment of harmonics and flux density to 5500 has a large bearing .
The pure sinewave means a distortion level ranging to 2% max .meaning no clipping of the even -2nd/4th,6th /8th harmonics and no peaking of the odd 3rd /5th harmonics .
As such my search isstill open......

 

[FvM]

I seems like you have a lot of unsaid constraints and also very special ideas about the design implementation. The points addressed in your latest post go far beyond the scope of the original post. That's quite normal for a non-trivial design, but also in my view sets an end to the previous discussion.

In my opinion, the discussion has clarified that a regular transformer core is at least an option. My suggestion for the specific "400 Hz core" problem: Assemble your specifications, ask one or two manufacturers in your region what they can offer for it and decide about 400 Hz transformer or transformerless.

 

[mtwieg]

Dear FvM & Alertlinks ,Thanks
I had thought of using CRGO /CRD /sheet silicon versions for my application but these are available with N67 core which gives 3600 as Bmax.

3600 Gauss? Where are you seeing that number? Electrical steel should have a Bmax/Bsat near 1.7T, just like normal steel and iron.

 

[vimalkhanna]

This measurements are 1.2T to 1.4Tesla for 50Hz ...However , with sustantial losses at 400Hz ,,we need to get to 35% of this .which gives this figure.