Hello, I have a circuit that is turning 24VAC from a transformer into 5VDC (and then 3.3VDC)


The 24VAC goes through a bridge rectifier and has a couple smoothing capacitors to create approximately 30 volts DC (with ripple).

When I plug in the 24VAC transformer, it produces a voltage spike up to about 65 volts, then stabilizes (this is completely normal for a transformer to do this):

My problem is the LM2596 can only handle an absolute maximum of 40 volt DC.

I can't find an appropriate TVS diode, because the TVS diode would need a "breakdown (min)" value of like 35 volts, and "max clamping voltage" of 39 volts. This is too tight of a window for TVS diodes.

So I am thinking I need to use a Zener diode here.

The problem is I am not sure how to appropriately select a zener wattage, and a series resistor, to combat these spikes. Also, if I put a series resistor right before the Zener, isn't this going to cause problems for the voltage regulation down the line?

I should also mention that my current draw requirements is a maximum of 1 amp on that 5 volt rail.

So it looks like the 65 voltage spike is less than 1ms in duration, and then its roughly 50 volts for roughly 3ms. Is there a good way to calculate a zener diode and resistor for this spike in this scenario?

Thanks and any help or advice here is greatly appreciated!

 

Try replacing C1 C2 with a 100uF - 150uF.
What are you using to turn the transformer on?

 

Try replacing C1 C2 with a 100uF - 150uF.
What are you using to turn the transformer on?

This particular case was just a relay/contactor to activate the transformer. There are other transformers that will just plug into the wall as well (like those rainbird 24VAC transformers from home depot for sprinkler systems).

 

You could also place an RC snubber across the relay contacts, if it is switching the primary.

 

If you cant supress the spike with increased input capacitance,
add a bit of series inductance on the input, with capacitor on either side.

 

hi,
Try a 100n ceramic cap across each of the input electrolytic caps.
E

 

If you cant supress the spike with increased input capacitance,
add a bit of series inductance on the input, with capacitor on either side.

A transformer with a two-section bobbin does that for you, with its leakage inductance.
Also add some film capacitors. instead of the electrolytics as they have less series inductance.
With 10V peak-to-peak ripple, perhaps a bit more smoothing capacitance.

 

Will a larger capacitor here really prevent the voltage spikes? Or just make it less likely to occur?

If you cant supress the spike with increased input capacitance,
add a bit of series inductance on the input, with capacitor on either side.

Do you know how I might calculate what values to use for the inductor and capacitors?

 

hi mah,
Have you tried adding a100nF cap across the two input electrolytic caps.??? Post #16
E

 

Will a larger capacitor here really prevent the voltage spikes?

It should reduce them to a non-damaging level.

Do you know how I might calculate what values to use for the inductor and capacitors?

Unless you can determine the energy in the spike, there's no way accurately determine the value of the capacitor.
It's select and test.
And an inductor is likely to be large and expensive for this purpose.

 

How to calculate ?
yes, lots of equations, but in reality, we just see whats around.

Looking at the plot,
there is quiet a big voltage dip on the input of the regulator cycle to cycle,
this as @MaxHeadRoom states is indicative that the input capacitance is to small, which it does look like, try the values he says , or bigger,
as the size goes up, the start up current increases, so if excessive, might blow input fuse if you have one,
also the bigger capacitors will absorb more of the energy of that input spike.


as @ericgibbs says, the bigger input capacitors have a high ESR. This limits how fast they can charge / discharge.
the application of ceramic capacitors across the big ones, allows the ceramics to respond first to the spike.

Inductors,
I'd suggest that you look at a BLM,
or as a start , the same as you have for the DCDC switcher.

If yo look here at the eval board for that switcher you will get an idea of the different capacitors needed

https://www.ti.com/tool/LM2596S-ADJEVM



My overall thought is that the input glitch is due to the source ,
and as @MaxHeadRoom says in #4, a snubber on the relay might be of more effect.

One other thought,

Is your scope on the same supply as the AS source your using ?
could it be that your seeing the spike on the mains either radiated or conducted through the earth wire.

 

How to calculate ?
yes, lots of equations, but in reality, we just see whats around.

Looking at the plot,
there is quiet a big voltage dip on the input of the regulator cycle to cycle,
this as @MaxHeadRoom states is indicative that the input capacitance is to small, which it does look like, try the values he says , or bigger,
as the size goes up, the start up current increases, so if excessive, might blow input fuse if you have one,
also the bigger capacitors will absorb more of the energy of that input spike.


as @ericgibbs says, the bigger input capacitors have a high ESR. This limits how fast they can charge / discharge.
the application of ceramic capacitors across the big ones, allows the ceramics to respond first to the spike.

Inductors,
I'd suggest that you look at a BLM,
or as a start , the same as you have for the DCDC switcher.

If yo look here at the eval board for that switcher you will get an idea of the different capacitors needed

https://www.ti.com/tool/LM2596S-ADJEVM



My overall thought is that the input glitch is due to the source ,
and as @MaxHeadRoom says in #4, a snubber on the relay might be of more effect.

One other thought,

Is your scope on the same supply as the AS source your using ?
could it be that your seeing the spike on the mains either radiated or conducted through the earth wire.

How to calculate ?
yes, lots of equations, but in reality, we just see whats around.

Looking at the plot,
there is quiet a big voltage dip on the input of the regulator cycle to cycle,
this as @MaxHeadRoom states is indicative that the input capacitance is to small, which it does look like, try the values he says , or bigger,
as the size goes up, the start up current increases, so if excessive, might blow input fuse if you have one,
also the bigger capacitors will absorb more of the energy of that input spike.


as @ericgibbs says, the bigger input capacitors have a high ESR. This limits how fast they can charge / discharge.
the application of ceramic capacitors across the big ones, allows the ceramics to respond first to the spike.

Inductors,
I'd suggest that you look at a BLM,
or as a start , the same as you have for the DCDC switcher.

If yo look here at the eval board for that switcher you will get an idea of the different capacitors needed

https://www.ti.com/tool/LM2596S-ADJEVM



My overall thought is that the input glitch is due to the source ,
and as @MaxHeadRoom says in #4, a snubber on the relay might be of more effect.

One other thought,

Is your scope on the same supply as the AS source your using ?
could it be that your seeing the spike on the mains either radiated or conducted through the earth wire.

Thanks for the response! First off, what is a "BLM"? I tried google search and got nothing.

I can't solve the problem by putting something on the relay contacts. This will be powered by those wall-wart style plug in 24VAC transformers as well. The solution needs to be contained on the circuit board.

I have another part of my circuit that uses a "pi filter", consisting of a 33uH high current inductor, with a 10uF ceramic capacitor on both sides.

 

Would the pi filter work here since I can't put anything on the relay contacts? And what is a BLM?

 

Cin 470uF 50V table 3 38uH?
https://www.ti.com/lit/ds/symlink/lm2596.pdf

 

I could barely see the faint schematic so I added some contrast. The original was saved as a fuzzy JPG file type instead of as a clear PNG.

 

Cin 470uF 50V table 3 38uH?
https://www.ti.com/lit/ds/symlink/lm2596.pdf

Table 3 is output capacitor, not input capacitor. We are talking about voltage spikes on the input not output.

 

ok I will be more specific
The others are correct the slope is too steep the time constant too small the inrush too high and the voltage too high using wrong inductor.

Look at 11.2 Layout Example, it was mentioned before and others also said try larger capacitors. There is connectivity at 150kHz
for a given load there are specific values, and you stated you did not want a spike but you are not following basic datasheet guidelines.

The question and scope shot shows that the inductor should be much smaller like 47uH and table 3 shows that.
The datasheet example works your does'nt. If you work from known to unknown you can try something different but you will see what happens when you deviate and the math example in the datasheet tells you why.