Since I don't have experience using isolated DC/DC,

It would be great if anyone could give me some advice.

(sorry to ask many questions at the same time)

#1 Question,

I'm trying to calculate Transform ratio by using the following formula

Ns/Np=(Viso+Vd)/(Vdd1min*D*2)

But I don't know what duty cycle should be.

it says  Duty should be 0.3, but I dont where this number come from.

#2 Question,

In order to select Trans inductance, Cout and L1 inductor,

Are there any calculating formulas?

The datasheet recommend value like 400uH for Trans, 47uF for Cout and 100uH/400uH for L1.

Should I use these figures?

Since the output current for my model could be less than 10mA so I'd like to use the smaller capacitor and inductor.

#3 Question,

The datasheet says that it's okay to use 5V to 3.3V for Ns/Np=2.

But I don't think Ns/Np is calculated in 1.2 for 5V to 3.3 V.

Could you make sure if it's correct?

=================================================================

[From page 15]

For a similar 3.3 V input to 3.3 V output, isolated single power

supply and with V DD1 (MIN)  = 3.0 V, the turns ratio is also N S /N P  =

1. 2. Therefore, the same transformer turns ratio N S /N P  = 2 can be

used for the three single power applications (5 V to 5 V, 5 V to

1. 3.3 V, and 3.3 V to 3.3 V).

=================================================================

#4 Question,

It looks like the most low power isolated DC/DC incorporate Fryback converter.

But ADUM3070 is push-pull converter(using two FET inside the IC).

I think Flyback can reduce external components like Inductor on the second side.

So I just wonder what is the merit of push-pull converter for the low power isolated DC/DC.

Best regards,

Tarzan

• Tarzan: My response to your questions in order below:

1. The max duty cycle is 0.4 for the ADuM3070 but in the Transformer Design section of the datasheet it is advised to use 0.3 typical for most applications to allow for variations in the design.  The exception is the doubling circuit, where 0.35 is advised to use so that too large a turns ratio Ns/Np does not cause a large voltage stress.

2. Transformer inductance is advised in the datasheet to be 60 to 100uH min per winding.  There are 2 primary windings,and if one winding is 60uH, the total primary inductance would be 4 times this or 240uH. The transformers in the datasheet meet this requirement.

3. For 5V in to 3.3V out:

Ns/Np = (Viso + Vd)/ (VDD1min*D*2) = (3.3 + 0.5V)/(4.5V*0.3*2) = 1.4

Since Ns/Np = 2 is > 1.4, this transformer can be used for 5V to 3.3V
Note, even though 5V in to 3.3V out is a step down, due to losses in the diode and the typical minimum duty cycle of 0.3, a turns ratio of at least 1.4 is needed.

4. The merit of using the push-pull topology in the ADuM3070 is to increase the efficiency over using the flyback topology.

Regards, Brian

Speaking of #2,

I understand what you mean, for ADUM3070 transformer inductance should be 60 to 100uH.

But I just wondered how the switching frequency affects the selection of transformer.

Because this device has flexible Fsw setting (200kHz-1MHz).

And for the selection of the LC components on the secondary side,

Don't you have any calculating formula?

I don't know why the datasheet says Cout=47uF,and L1=47uH.

I'd like to use smaller capacitor and inductor to adjust Fsw since the load current will be less than10mA in our system.

Do you have an application note about ADUM3070, which explains more about how to select external components at the different Fsw?

Best regards,

Tarzan

,

Tarzan

• Hi Brian,

Do you have any updates on my questions?

but it would be very helpful.

Best regards,

Tarzan

• Tarzan: A response from the designer:

It is a little bit complex topic.  Small output current doesn’t meant to choose small inductance.  Small inductance tends to make converter work in DCM mode at low load current; this will make the transient response bad; if they don’t care load transient response they can try some inductances with small values. Attached is one document regarding magnetizing inductance and filter inductance for your information; both of them will affect the work mode (CCM or DCM), that is the transient response.  The switching frequency affects the work mode, also while choosing transformer we need pay attention to its saturation characteristic.

Typically the filter inductance decides the work mode and the bypass capacitance decides the ripple (V= I*DT/C).  Also the L and C relates to the loop compensation design, if they change the values, they can change the small signal response  of the converter; but this is tricky, they can change the L and C, however try to maintain the stability of converter.

Regards, Brian