How do I read this transformer datasheet?

[ants]

I am looking to test a new transformer, I've attached the datasheet. I am looking for some kind of indication what the primary impedance would be. I will operate it at 100KHz.

Thanks,

Ant.

 

[keith1200rs]

Ant,

If you don't load it, it will be the impedance of 165uH at 100kHz = 104 ohms. However, I assume your reflected load impedance will change that.

Keith.

 

[ants]

Hi keith,

I see. Transformers are quite difficult for me to understand. At a guess some kind of LC circuit is set up between the piezo which is the load and the transformer primary. So I should try to get a transformer which makes a good tank circuit.

I have a transformer which works, it is 1:19, but I want to use a different piezo now which has a lower drive voltage so I want more current to compensate to keep the same power, hence trying a new transformer.

Thanks,

Ant.

 

[keith1200rs]

Ant,

If you put a piezo on the secondary of a transformer with a 1:19 ratio then the piezo capacitance will be reflected by the turns ratio squared. So, you will end up with a huge reflected capacitance on the primary and resonance of the primary inductance with that reflected capacitance could be an issue. The other way to look at it is that the secondary inductance is the primary inductance multiplied with the square of the turns ratio and that resonates with the piezo capacitance. Same result - one is transferring everything to the primary side, the other to the secondary side.

Keith.

 

[ants]

The piezo's I use have around 2nF in capacitance and I have just checked and the transformer which I usually use is 1.2mH. Exactly an LC tank at 100KHz, it is coincidence, now I know why I usually use that transformer.

I could try to unwind the secondary on my usual transformer, it is possible because I tried before, there look to be thousands of turns though, it was a disincentive to carry on.

---------- Post added at 13:09 ---------- Previous post was at 13:04 ----------

Ant,

If you put a piezo on the secondary of a transformer with a 1:19 ratio then the piezo capacitance will be reflected by the turns ratio squared. So, you will end up with a huge reflected capacitance on the primary and resonance of the primary inductance with that reflected capacitance could be an issue. The other way to look at it is that the secondary inductance is the primary inductance multiplied with the square of the turns ratio and that resonates with the piezo capacitance. Same result - one is transferring everything to the primary side, the other to the secondary side.

Keith.

I think LTspice will help there, it will be difficult to predict on paper calculations. I will try.

Thanks.

Edit. The transformer is 1.2mH primary and 600mH secondary. I was getting that mixed up.

 

[keith1200rs]

It is easier to wind a transformer than unwind one!

One thing you can do if you want resonance is to add capacitance to the primary or secondary. Usually the problem I have had is getting the primary inductance low enough for a given piezo and turns ratio. For example, if you need 10:1 to get the output voltage from a low voltage supply and the piezo is 7nF you only need 1.6uH to resonate at 150kHz. That can cause problems as you end up with just a couple of turns on the primary depending on the core.

Keith.

 

[ants]

I've been reisisting winding my own transformer for months I bought some coilcraft transformers which let you use lots of different combinations of series and parallel to get close to what you need but haven't got round to trying that yet. I find transformers more confusing than anything I've come across, I pretty much designed my circuit around the 1:19 transformer.

Thanks,

Ant.

 

[keith1200rs]

Time to get stuck in and wind some! I never seem to be able to buy anything even close to what I need.

Keith.

 

[ants]

I'm going to wind my own. I've spent more time avoiding it than it would take to actually get on with it!

I was under the impression coilcraft had kits that would let you wind transformers, but all I seem to be able to find now are deisgner kits that contain a range of values already made up.



Do I need the transponder coil kits?

Thanks,

Ant.

 

[keith1200rs]

I think you first need to find out how much power you need to put through the transformer and do some calculations. I have usually ended up using EFD or ETD cores - you can get them from RS or Farnell. Smaller transformers could use RM6 or RM10 type cores. I would have thought the transponder transformers would have been a bit feeble for your application.

Keith.

 

[ants]

Well the old piezo was around 5 watts, that was 60Vrms and 80mA, The new one is around 25Vrms and 200mA and about 5 Watts. I doubt I'd need over 10Watts but should leave some headroom.

The primary coil needs to be around 5Vrms and 1A and the secondary 25Vrms and 0.2A. So the turns ratio is 1:5. I will have a look on Farnell at some cores and wire.

Thanks,

Ant.

 

[ants]

I have put some RM6, RM10 and ETD cores into a Farnell shopping basket, I'm currently looking for the enamelled wire, I wil probably have found some by the time anyone has an opportunity to reply.

Is there anything else I'd need to wind my own, for example a helpful tool?

Thanks,

Ant.

 

[keith1200rs]

Make sure the cores come with bobbins, otherwise you will need to buy some separately. I think the RM series usually come as kits. Wire sizes down to 0.3 are useful. Anything bigger than 0.8mm gets tough to wind.

Keith.

 

[ants]

That was a good call, I didn't have any bobbins. I haven't come across any wire yet and the cores I've seen don't seem to come with any. I'm re-reading trying to understand the concept of reflection but it is difficult, I do understand inductance and impedance being the squre of the turns ratio.

I think I should try for resonance because I need as much energy transfer as possible, the piezo is dynamic/moving and so does operate at resonance from a mechanical point of view.
Some approximate calcs:

Piezo is 10nF so secondary inductance to create LC tank at 100KHz is 250uH

With a turns ratio of 1:5 --> 1:25 for squaring

So the primary inductance is 10uH. It is a bit low but a good starting point, I think.

 

[FvM]

I see the problem, that you apparently don't know the impedance of the piezo tranducer near it's operation frequency. You're assuming a capacitive impedance, but it's only valid for frequencies far below it's resonance frequency. The most interesting point is the real part of the transducer impedance, that has to be matched to the generator's optimal load impedance by the transformer. Another important parameter, that can be influenced by the transformer design, is it's leak inductance.

 

[ants]

Hi FVM,

That is a good point, as the piezo operates at resonance its impedance drops to the minimum, this is usually less than 100 ohms but i don't know whether it is 10 or 100 ohms.

The most interesting point is the real part of the transducer impedance, that has to be matched to the generator's optimal load impedance by the transformer.

I find this bit difficult. The generator is an opamp which should in theory have a low output impedance and supplies the transformer.

Thanks,

Ant.

---------- Post added at 23:00 ---------- Previous post was at 22:24 ----------

I have had a thought. If the transformer which works for me (the 1:19 ratio) is 1.2mH primary, I should build one which is 1.2mH primary and for 1:5 stepup a secondary of 30mH. I think this would work but I'd be greatful if someone could confirm. I can't say I understand why the 1.2mH primary works with my setup perhaps in building this transformer I will understand it better.

The only enamelled core I can find are these, the reels are a bit long though.

Browse for Products | Farnell United Kingdom

 

[keith1200rs]

The only enamelled core I can find are these, the reels are a bit long though.

Unfortunately 0.5kg seems to be a standard size so you get a lot, particularly of the smaller diameters. It will last you a lifetime - I still have reels which must be 35 years old! You can get smaller reels but not from RS/Farnell. It is expensive to buy small reels though, for the amount of copper you are getting:

enamelled copper wire - Google Product Search

Keith.

 

[ants]

Hi Keith, thanks for the link, I will buy some smaller reels. I hope I'm not still building this circuit in 35 years

 

[FvM]

I find this bit difficult. The generator is an opamp which should in theory have a low output impedance.

I sayed optimized load impedance rather than source output impedance. It's simply the impedance that allows maximum output power. The basic assumption is, that you are able to cancel the reactive load at the transformer primary side and set a transformer voltage ratio, that transfers maximum power to the real part of the transducer impedance.

I hope I'm not still building this circuit in 35 years

It depends on, if you are understanding the basic theory behind piezo transducer and related AC network bevaviour.

The only enamelled core I can find are these

Some catalog districutors and DIY shops should have smaller (e.g. 100g) coils. As a student, I got my magnet wire to build transformers from a motor repair workshop around the corner.

as the piezo operates at resonance its impedance drops to the minimum, this is usually less than 100 ohms

Transducers have a series resonance (minimum impedance) and a parallel resonance (maximum impedance). Usually the series resonance is utilzed for transmitters. In any case, it would be good to know the real impedance. Depending on the application, it may vary with the mechanical transducer load, e.g. for an ultrasonic cleaner. It can be even useful, to allow a generator frequency tuning for maximum transmitted power in these cases.

 

[ants]

It's simply the impedance that allows maximum output power. The basic assumption is, that you are able to cancel the reactive load at the transformer primary side and set a transformer voltage ratio, that transfers maximum power to the real part of the transducer impedance.

I see, this thread makes sense now, the capacitance on the load is reflected back with the square of the turns ratio. So essentially I add some capacitance to the load to set up a tank. I presume it doesnt matter if the added cap is in series or parallel and I should use a low ESR one.

I have ordered the parts and may return to this thread when I'm actually wiring it.

Thanks,

Ant.