Parallel/Series Transformers

[dknguyen]

If I neeed a 16:1 transformer with certain frequency characteristics and could only find 1:1 transformers with those frequency characteristcs, could I get 16 transformers and put all the primaries in parallel and all the secondaries in series to get the same effect (to pass and setup up a ultrasonic signal from 12Vpp to ~200Vpp). Or would doing this adversely affect the frequency responses. 16 transformers in series-parallel is still a heck of a lot of transformers though...with a lot of winding resistance and a huge inductance.

There is another transformer I coudl use but my frequency range of interest is right at the -3dB point while all the harmonics I'm don't want to pass are sitting across the optimal range of the transformer at the -0.5dB range.

THanks.

THe third alternative is to wind my own transformer. It's not very demanding but I need it in a strange frequency range and I cna't seem to find one. I'm still trying to find info and parts on winding your own transformers, but it sounds like a fairly hit or miss affair in that you can't really predict the bandwidth and attenuation in the bandwidth ahead of time. Any resources would be appreciated.

200Vp@25m-50A
-0.5dB to -1dB from 50kHz-100kHz
preferable to have -3dB up to 500kHz
16:1 ratio

 

[shaneshane1]

normally i would keep to myself, but

 

[dknguyen]

Wha????????????

 

[shaneshane1]

sorry, my bad, your avatar distracted me,

 

[stevez]

I obtained some design books from Amidon Associates about 10 years ago - very low cost. Jerry Sevick published a couple of books on transformers - mostly RF applications but the concepts apply in your case. Sevick's books were for radio amateurs. You might also find some help in amateur radio references.

 

[kchriste]

Have you tried any of the transformer manufacturers sites like:
**broken link removed**
??

 

[Hero999]

You probably already know this but parallelling the primaries will divide their inductance by 16 and putting the secondaries in series will multiply their inductance by 16. Also not that the power handling capabilities of the transformer is also multiplied by 16.

Connecting secondaries in series normally only becomes a problem when the insulation between the primary and secondary becomes an issue but 200V is still a low voltage so I can't see this being a problem.

Whay don't you pick a suitable core and wind your own transformer?

If the isolation transofmers you have can handle the power you could just rewind the secondary with 16 times as many turns wire with 1/16 of the cross-sectional area.

 

[dknguyen]

If the isolation transofmers you have can handle the power you could just rewind the secondary with 16 times as many turns wire with 1/16 of the cross-sectional area.

Ignoring the fact that the transformers are little toroids stuck inside a small box-like SMD package with an open base, I thought about that. But since the two transformers I am talking about are part of the same series, and their frequency characteristics degrade as the number of turns is increased, I think they are all wound on the same core anyways and that it's the extra winds that causes the degradation:
**broken link removed**

THe two transformers are:
TTWB1010 vs TTWB-16-A

So right now, I can either:
1. Deal with the -3dB loss of the 1:16 transformer
2. Use 16 1:1 transformer with the desired 0.5dB bandwidth in parallel-series
3. Use two 1:16 transformers in parallel to overcome the power limitation since it's not a problem at all for the driver (but all the frequencies I'm not interested in will still be passed twice as well as as my frequencies of interest...)
4. Do the learning and wind my own

 

[Styx]

yes you can parallel the primaries and series the secondaries

BUT breakdown of insulation is going to be your biggest problem!!!

 

[dknguyen]

yes you can parallel the primaries and series the secondaries

BUT breakdown of insulation is going to be your biggest problem!!!

THey are all rated for 400-500V coincidentally, and since my signal is 200Vpp at any one time, it should be a problem I don't think. I'm more worried about them not behaving the way I think they will behave due to some parasitic.

The 1:1 and 16:1 are practically the same, just the 16:1 has too much loss at the frequency I want. I'm mainly wondering if I parallel-series the 1:1 transformers if it will degrade the frequency response to be the same as the 16:1 (I'm fairly sure they use the same core and it's the higher winds ratio that degrades the frequency response of the one with the ratio I need).

 

[Hero999]

Don't forget that if you use more transformers then the losses will be higher.

 

[mneary]

Add a new primary with 1/8 as many turns, and put the two existing windings in series.