Hello everyone, I’d like to ask your opinion on the design of an input stage.

I’m designing the input stage of a 48V three-phase inverter for PMSM motor control and I’d like your opinion on my reasoning. I’ve chosen this diode and this fuse.
The input and output parameters are shown in the photo (the ranges for Vi, Ii and the corresponding inrush current at the input), and two solutions are also shown:

• the first is ‘simpler’ as it only uses a diode and a fuse;
• the second (I haven’t worked it out in detail; I’ve just included it amongst the options) is more sophisticated.

What I’m asking you to do is to check the diode and fuse chosen and, if they are wrong, help me choose a TVS diode and fuse that will ensure the circuit operates correctly under normal conditions and shuts down properly in the event of either of the following two conditions:

• reversed input polarity (the diode must conduct in the forward direction and the fuse must blow)
• overvoltage exceeding the Vi,max of the DC-DC converter used (the fuse does not blow and the diode conducts in reverse, limiting Vi to its Vclamp)

It’s a delayed ‘slo-blow’ fuse with a high I2T; from my calculations, it doesn’t trip with the current inrush spike.

However, I’m afraid I may have made a mistake, as the Vclamp=93V of the TVS diode I’ve chosen is considerably higher than the Vin,max=66V of the DC-DC converter I’m using.
So can you confirm that I should look for a diode with Vrwm > 50V but Vclamp < 66V (assuming such a diode exists)?
.. although I believe that, unfortunately, almost no standard TVS meets this requirement


Does it make sense to connect a Zener diode with a breakdown voltage of 60V in parallel with the TVS?

I’ve also read online about more advanced crowbar circuits that include a MOSFET, but I want to understand all the pros/cons and limitations of this simple diode-and-fuse solution first, so please let’s not stray into circuits other than the simple one shown in the photo in this post. So, if you think it is possible to meet the two conditions highlighted earlier simply by choosing a diode and a fuse, then I would like to work with you to size them correctly

LET’S ASSUME we can’t reduce the inrush current… so please let’s focus on the situation described above; I want to learn how to select the correct fuse and diode in the event of an inrush current.

 

I've never liked the dead-short-blows-fuse approach to reverse polarity protection. The fuse can be in a holder that makes it easy to replace after an event, but swapping out a power diode in the field is not fun. Plus, you don't know what kind of damage the brief current surge is doing to whatever the power source is.

Better to turn off the source than to hope you don't kill it. My vote is a 20 A (minimum), 100 V power MOSFET, driven by one of Linear Technology's input control chips. They have a bunch. Granted, this is an expensive approach, but I put it in hundreds of battlefield systems.

ak

 

A "transient voltage spike" and an "input over-voltage application" are two quite different things.
There is a very different approach that has been used on at least one very special "radio."system
The alternative scheme uses a electromechanical RELAY to switch the power on, and with a series diode in the relay operate feed, current can only flow if the polarity is correct. The protection is very good .

 

Choosing the type of fuse and finding a TVS which will definitely survive long enough for the fuse to blow is going to be tricky? What happens if someone uses different length cables and reduces the fault current? The current is lower but the fuse takes longer to blow - does that make the diode’s survival more or less likely?
What happens if someone has already connected up back-to-front and blown the fuse? Did they replace it with the correct fuse, or whatever they had to hand?

I’d go with @MisterBill2 ‘s suggestion. Relay or DC contactor, but be careful in choosing the relay if there is plenty of smoothing capacitance! You might need a small Albright contactor.

 

I've never liked the dead-short-blows-fuse approach to reverse polarity protection. The fuse can be in a holder that makes it easy to replace after an event, but swapping out a power diode in the field is not fun. Plus, you don't know what kind of damage the brief current surge is doing to whatever the power source is.

Better to turn off the source than to hope you don't kill it. My vote is a 20 A (minimum), 100 V power MOSFET, driven by one of Linear Technology's input control chips. They have a bunch. Granted, this is an expensive approach, but I put it in hundreds of battlefield systems.

ak

Hi, do you mean something like the TPS2663?

 

A "transient voltage spike" and an "input over-voltage application" are two quite different things.
There is a very different approach that has been used on at least one very special "radio."system
The alternative scheme uses a electromechanical RELAY to switch the power on, and with a series diode in the relay operate feed, current can only flow if the polarity is correct. The protection is very good .

Is this what you mean? If so, which of the two?

 

If the part marked "30A" is a fuse, then you need only one of them. Add one input fuse to the upper schematic.

Also, neither circuit does anything about surge/transient protection.

ak

 

Put a bridge rectifier on the input and never worry about polarity again, nor having to replace fuses.

 

With a bridge rectifier you are assured of two series voltage drops for your load voltage. THAT might matter!