Best 100v TO220 package FET made? It's cheap too. :-)

[bowlofsalad]

Hey. Just saw this thread! I'd be up for some! Anyone that can resell them in matched sets?
Better yet, I'd like to see tests vs 4110's first, if anyone has done it!

Yeah, I'd really like to see some testing, a lot. By testing, I am mostly interested in the gains in efficiency by using these. I have trouble imagining what going from 3077 to CSD19506KCS, for example, would gain.

 

[liveforphysics]

Hey. Just saw this thread! I'd be up for some! Anyone that can resell them in matched sets?
Better yet, I'd like to see tests vs 4110's first, if anyone has done it!

Yeah, I'd really like to see some testing, a lot. By testing, I am mostly interested in the gains in efficiency by using these. I have trouble imagining what going from 3077 to CSD19506KCS, for example, would gain.

The smaller gate charge would get you the potential for more precise switching control. The lower RdsOn would give you ~20-30% reduced silicon conduction losses, and they have roughly matched thermal performance. This means it would enable some amount higher continuous torque from your controller if you were already so hot-rodded you were pushing the limits of what you had. If you currently have plenty of margin in your controller, then it would do essentially nothing for you but reduce controller heating to a minor extent.

 

[auraslip]

Can someone tell me how to calculate the differences in efficiency we would see from using these fets over the 4110s? Is it as simple as adding up to the RdsOn per phase and multiplying it by amperage to get the waste heat produced?

 

[Futterama]

auraslip, no, you would also have to calculate the switching losses, and I don't know how to do that, sorry.

 

[Lebowski]

and don't forget to add the losses due to the diode reverse recovery...

 

[auraslip]

I saw in another thread that the differences in losses between 4110 and 3077 is 25% which seems like a lot, but at 30a that comes out to like 1 watt. So IDK. Probably not worth switching out all you 4110s for this new FET from texas.

 

[Shenta]

This mosfet looks great, but from controller manufacture point of view, we also look at Pulse Drain Current and EAs... which IRFB4310 is slightly better.

Pulse Drain Current can save ur controller from noise caused shot-through, and EAs is the 2nd most important number for motor application...
No doubt this Mosfet has alot improvements, I will get some samples and give it a try too.

 

[John in CR]

Recommending a 4310 instead...WOW!

 

[zener]

What FET uses the Kelly KEB series?

 

[liveforphysics]

This mosfet looks great, but from controller manufacture point of view, we also look at Pulse Drain Current and EAs... which IRFB4310 is slightly better.

Pulse Drain Current can save ur controller from noise caused shot-through, and EAs is the 2nd most important number for motor application...
No doubt this Mosfet has alot improvements, I will get some samples and give it a try too.

If your design has noise induced shoot-through, why are you concerning yourself with something as comparatively not-important as MOSFET performance?

 

[zombiess]

This mosfet looks great, but from controller manufacture point of view, we also look at Pulse Drain Current and EAs... which IRFB4310 is slightly better.

Pulse Drain Current can save ur controller from noise caused shot-through, and EAs is the 2nd most important number for motor application...

If you are concerned about shoot through and are counting on the devices pulse current rating to survive this event, you have much larger design issues that need to be resolved. I don't know what power level you are designing with your controllers, but if they are meant to be above the typical junk toy level controllers used in ebikes it should follow proper design rules and have the ability to protect the MOSFETs from a catastrophic event.

I've personally tested a single IRFB4115 MOSFET to 96V at close to 300A and the gate driver shut everything down in < 5uS leaving the MOSFET undamaged during a shoot through event. I then did this dozens of times at this power level, no failure.

If noise is causing an induced Miller effect causing shoot through, the solution to fix it can be pretty simple depending on your desired power level.

 

[liveforphysics]

Shoot-through events from a controllers perspective are like crashing your car into a wall at speed from the cars perspective.

The solution to the problem is not found in stronger bumpers, because the bumper was not the problem. The solution is to stop smashing into walls at speed.

Preventing shoot-through should be a higher priority than any other aspect of controller design, because if you have it occurring, no other part of your controller matters. All the magic in a controller is in the gate drive being done right, then your car stops hitting walls randomly. Only after you have optimized gate control can you begin to realize the potential in the MOSFETs.

 

[fechter]

I think one of the more common fail modes in cheap controllers is over dissipation by the body diode. Better on resistance is good, but the forward drop of the body diode is going to be pretty close for any of the heftier FETs out there. A good controller will have synchronous rectification to greatly reduce body diode heating. I wish it was easy to implement.

 

[teslanv]

So has anyone built a Xie Change with these TI FETs? Reports/Reviews?

Thinking of doing an order of 18FET Controllers with these, if it's just a simple swap of FETs.

 

[johnrobholmes]

I've been trying to find the FET you posted luke, that was a fatty 3 phase all in one package with fantastic heat transfer and gate charactaristics. Still have that kicking around, or maybe there is another package now that is the jam? Just got into a position to design hardware with the software/ programming suite being handled by another. He wants to start with a 32bit IC so we can wifi program it easy. Sure man, if you want to write code for that I'll go along I said :lol: