TIDA-070002: TIDA-070002

Rajan,

Daniel and Javier are out this week taking needed vacation :-) .  One of them will respond to this query early next week.

Regards,

Wade

Hi Daniel,

I did wonder when reconciling the rating of the inductor current with the specified load current of the DC-DC.

It's insightful that you were not able to source an off-the-shelf coupled inductor for a 50W DC-DC - I too have had a similar issue!

How did you calculate the total amount of input and output capacitance that would be required? The former comprises low-ESR ceramic to reduce input ripple voltage and larger bulk capacitance for minimise deviations caused by load transients and to provide enough charge for the inductor and FET during the ON time.

Output capacitance obviously keeps the load voltage flat, assists stability and also has to comprise some low-ESR/ESL capacitance to keep the output ripple low.

Kind Regards,

Rajan.

Hey Rajan,

There are 2 equations in the design guide that show how to get output capacitance (section 2.4.5 page 7 bottom).

What I try to do is pick the capacitance value based on transient requirements and then figure out the frequency response accordingly.

The equations are by no means exact however since voltage ripple has more components than just capacitance (notably ESR and ESL).

Input capacitors can be quite difficult to pick as they are actually based on what the input filter looks like.

I don't believe I did anything fancy for the TIDesign, I mostly just added what I believed was "enough" especially since the input capacitors had much higher ESR since they were electrolytic. You can also use the same equation 22 (section 2.4.5 page 7 bottom) for the output capacitors for the input. Simply use it for input voltage ripple instead.

Something you have to make sure with input capacitors is that they are stable for a given input filter. Below are equations that pertain to the issue. Damping factor (d) = 1 is perfect while > 1 is over damped. Generally I find 1 to be unobtainable without drastically increasing the ESR. I have used ~0.6 in the past and been fine.

Note that the capacitance/ESR that the equation uses is at the resonant frequency of the filter.

Thanks,

Daniel

TI's SLTA055 is a useful reference ...

Once the total capacitance has been calculated and realised, one must also check the rating of each capacitor, e.g. rms ripple current, voltage etc ....

Hi Daniel,

How did you reconcile eqns. 22 and 25 which produced two very different capacitances?

Kind Regards,

Rajan

Hey Rajan,

One of the equations is based on a voltage ripple requirement and the other is based on a transient requirement.

The equations are to show the minimum capacitance needed to meet each requirement.

Thanks,

Daniel

Hi Daniel,

For the reference design, must the duty cycle be 50% to ensure all the energy stored when the FET is ON equals all the energy delivered when the FET is OFF, i.e. continuous mode.

Can other duty cycles be used and would these saturate the coupled inductor?

Kind Regards,

Rajan.

Hey Rajan,

It is not necessary that the duty cycle be 50%.

The core saturation you are worried about is related to balancing Volt-Seconds.

The average voltage across the inductor must be 0 to prevent saturation, for a flyback this means the following in equation:

Vin*D = (Vout +Vdiode)*(1-D)*Nps where Nps = Npri/Nsec

Notice the duty cycle can be controlled in order to balance the equation. This duty cycle happens to be the steady state duty cycle that the converter will operate at.

The outcome of this balancing can be seen as the current through the coupled inductor on the primary side always starts a the same value in steady state. For CCM, this value will simply not be 0.

Thanks,

Daniel

Hello Daniel,

The use of TI's SM74611 diode would have allowed you to achieve higher efficiency because of its lower forward voltage drop. Would you agree?

Neither the SM74611 nor the commercial Vishay part that you used are space-grade parts by the way!

Kind Regards,

Rajan.

Hey Rajan,

You are correct on the efficiency.

If you notice the only parts that are space grade are TI parts.

The TIDesign was meant to show off our space grade parts and the cost and export control of doing something fully space grade showed to not be worth the effort.

Thanks,

Daniel

I am considering using the sm74611 to build a flyback ...

any thoughts ...

Hey Rajan,

Do you not have a radiation requirement?

I don't see anything wrong with it if you don't have a radiation requirement, however I can only comment on the space grade or enhanced product grade power devices.

If you have questions specifically on the sm74611 I suggest making a new thread so the correct expert can answer any questions you have.

Thanks,

Daniel

I note your comment about radiation hardness, however, I do not think any of TI's space-grade diodes offer such a low forward drop ....

Hi Daniel,

Why did the TIDA-070002 reference design use T2 to isolate the feedback control loop when there is an auxiliary winding on T1?

Kind Regards,

Rajan.

Hey Rajan,

I actually did something what you are talking about for internal testing to see how it would go.

We found that without something to blank the parasitic spike of the Aux winding, the regulation was often too poor to work.

Regulation using the Aux winding usually has a very high dependency on output current as well.

Simply put, the way I isolated the feedback was more accurate, that isn't to say there aren't improvements I would make looking back on the design.

One of the changes I would make is noted in the design for the Vcc of the UC1901-SP. The other one would be having the output of the feedback isolation transformer go through the diode and onto a capacitor in series with the feedback resistors. This is something that I have seen that I would like to test out and see if it is a better method of sending over the isolation signal.

Thanks,

Daniel