Transformer voltage resiliency

[joeyjoejoe]

I am powering my device off a 24VAC transformer. This is a very common voltage in north America for HVAC equipment.



This is my latest design. 24VAC is 34Vp, so for nominal operation it works well. However, transformers are quite poor at being over-voltage under no load. And the transformers used are outside of my control.

Do you think the protection is adequate?

The chosen SMPS - https://www.ti.com/lit/ds/symlink/lmr54410.pdf?ts=1731249147710&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FLMR54410%252Fpart-details%252FLMR54410FDBVR. I guess nominal up to 36V, but can go up to 45V. I like it and it was chosen for form factor and availability. I could choose another if absolutely needed, but it's hard to not then go into higher-power SMPS which is not needed, or lower switching frequencies which require larger inductors.

In short, I would prefer to put additional protection in front, if needed, rather then a different switcher. However, if cost/complexity is too large, I am open to a more robust switcher.

My 1 additional question - since PTC is still relatively slow to trip, would the fuse+zener be better in front of the capacitors? Or would inrush current still cause a trip or bring it too close to a trip condition, and it's best left where it is.

[Kleinstein]

Especially with small transformers it is normal to have a no load voltage that is quite a bit higher than the nominal voltage.
I would consider having the PTC fuse before the recitifer. Here the current is a bit higher, which can help a little with faster tripping. In addition the resistance before the filter capacitor helps a little to reduce the peak current / improve the power factor, even if not by much. Also the recitfiers are possible points of failure and may fail short. 

For 24 V nominal I would not use schottky diodes anymore as one would need quite some voltage rating (2 x peak voltage).

It looks reasonable. One could try out the combination of PTC and TVS in how high the voltge peaks for a reasonable test case (e.g. 40 or 48 V transformer).

[David Hess]

Doorbell transformers have high leakage inductance, so voltage regulation is not very good, and sudden changes in current will cause changes in input voltage.  The later can be a real problem with regulators because if the load is removed, the input voltage will spike high causing damage unless there is enough input capacitance to absorb it.

So your input zener or a power shunt regulator is needed no matter what else is used for protection.  I would use either a higher current TVS, or combine a TVS/Zener with a bipolar transistor to make a power shunt regulator as shown below.

[joeyjoejoe]

The higher current TVS interests me, as it would just be picking a bigger similar part - and the PTC should then trip.

The chosen one is currently a pulse of 60A/400W. What sort of rating am I trying to achieve? Or style of part should I be browsing? I assume that TVS's would typically only have these pulse ratings since they are designed for transients. Wouldn't I need more of a sustained rating?

[Siwastaja]

Sad to say it but that switcher is just out of question.

Consider line high tolerance combined with high no-load voltage of usual small transformers and you are already beyond the 36V input rating. And it's a terrible idea to shunt using TVS before a switcher; that works only for short peaks, you need ample margin so it never normally conducts, otherwise you have power dissipation nightmare and need to use large expensive TVS. This would conduct twice every cycle! TVS curve is soft and unit-to-unit variation large so if you want to protect the switcher even from the abs. max rating of 45V, it would then already conduct with normal line voltage peaks.

And remember, absolute maximum rating is nonfunctional rating.

But worry not, switchers with higher voltage rating are easily available. Just keep looking. Don't remember the exact part number but last time with similar requirements I picked up a 100V rated switcher, 63V rated electrolytics and a TVS which starts conducting somewhere around 70-80V.

[David Hess]

The higher current TVS interests me, as it would just be picking a bigger similar part - and the PTC should then trip.

The chosen one is currently a pulse of 60A/400W. What sort of rating am I trying to achieve? Or style of part should I be browsing? I assume that TVS's would typically only have these pulse ratings since they are designed for transients. Wouldn't I need more of a sustained rating?

What I am suggesting is that you cannot have a good idea of the transformer size and characteristics, so using a larger TVS than necessary is good insurance.

The SMDJ series has about twice the power capability of the SMDA series, albeit at twice the low cost.  These parts have continuous power ratings of 3.3 and 6.5 watts in the datasheets.

[joeyjoejoe]

The higher current TVS interests me, as it would just be picking a bigger similar part - and the PTC should then trip.

The chosen one is currently a pulse of 60A/400W. What sort of rating am I trying to achieve? Or style of part should I be browsing? I assume that TVS's would typically only have these pulse ratings since they are designed for transients. Wouldn't I need more of a sustained rating?

What I am suggesting is that you cannot have a good idea of the transformer size and characteristics, so using a larger TVS than necessary is good insurance.

The SMDJ series has about twice the power capability of the SMDA series, albeit at twice the low cost.  These parts have continuous power ratings of 3.3 and 6.5 watts in the datasheets.

Aha, thank you! The official Littelfuse datasheet had the information. The *ahem* overseas variant did not

I've changed it to a SMDJ33A.

[joeyjoejoe]

Ok. You have convinced me to switch to an LM2591HV  Also available in a 5V version, and it is popular enough to be copied so that brings the cost to the same range as the LMR54410. It will require a bigger coil since it's only 400kHz, but I have the space.

Sad to say it but that switcher is just out of question.

Consider line high tolerance combined with high no-load voltage of usual small transformers and you are already beyond the 36V input rating. And it's a terrible idea to shunt using TVS before a switcher; that works only for short peaks, you need ample margin so it never normally conducts, otherwise you have power dissipation nightmare and need to use large expensive TVS. This would conduct twice every cycle! TVS curve is soft and unit-to-unit variation large so if you want to protect the switcher even from the abs. max rating of 45V, it would then already conduct with normal line voltage peaks.

And remember, absolute maximum rating is nonfunctional rating.

But worry not, switchers with higher voltage rating are easily available. Just keep looking. Don't remember the exact part number but last time with similar requirements I picked up a 100V rated switcher, 63V rated electrolytics and a TVS which starts conducting somewhere around 70-80V.