Maker Pro
Maker Pro

Need help interpreting Mosfet turn-on delay / rise time

royalmp2001

May 20, 2014
11
Joined
May 20, 2014
Messages
11
Hi everyone,
Searching through datasheets to find the fastest mosfet for my application.
What is better? Shorter rise time or turn-on delay... or both.?
I am finding if one is short the other is longer...how best do I compare different combinations for fastest speed.
Same goes for turn off delay and fall time.
Thanks
 

Harald Kapp

Moderator
Moderator
Nov 17, 2011
13,799
Joined
Nov 17, 2011
Messages
13,799
What is better? Shorter rise time or turn-on delay... or both.?
"Better" in which sense? It depends on the application.

For example in a switched mode power supply turn-on and turn-off delay limit the max. frequency at which the power suply can be operated. As higher frequency means smaller transformer, this has a direct effect on cost of the power supply. When switching a pwm waveform for controling a motor at a few kilohertz, turn on and turn off delay may be insignificant.

Another example is rise time. Short rise time minimizes the time the transistor is in transient operation between fully on and fully off. This in turn minimizes dynamic power dissipation during this time. On th eother hand short rise time generates gigh frequency transients (high frequency voltage and current components) which need more filtering to keep emc below the allowed levels.

fastest speed
is not always "better".
 

Fish4Fun

So long, and Thanks for all the Fish!
Aug 27, 2013
481
Joined
Aug 27, 2013
Messages
481
Hey Royalmp2001!

Everything Harlod said is absolutely correct.....but may not help you make your decision....

If you truly want some help selecting a mosfet, I would urge you to post details about your circuit and anticipated use, market, or production quantity.

There are thousands of Mosfets to choose from, and yet only a subset of these are "Popular Choices" among hobbyist. In almost every case mosfet selection for any given circuit is a process of elimination. In most production designs (as Harold states above) the mitigating factor is total system cost ...... BUT, how "Total System Cost" is defined is intimately linked to production volume.....

If the long-term goal is a mass-market consumer product with anticipated annual sales of 10k-10M units per year, then devoting 100's or even 1000's of man-hours to engineering the lowest possible cost BOM has much more weight than it would if the anticipated annual sales were only 1k to 10k units per year. If you calculate engineering time @ $100/hour, and you allocate 1000 engineering hours to the project, your engineering cost is obviously $100k. If your anticipated sales are 1M units per year then the first-year engineering cost per unit is only $0.10/unit. In this scenario, if an additional 1000 engineering hours allowed a cost savings of $1.00/production unit, then you would net an extra $900k in the first year. If the same project only had an anticipated annual sales of 10k units per year, the engineering cost would be $10/unit for the first year, and doubling the engineering time to 2000 hours in an effort to save the same $1/production unit would cost you an extra $90k!

For a DIY project or a short-run production project the "Total System Cost" is frequently dominated by engineering costs, shipping costs and "set-up" costs. For instance if you only anticipate selling 100 units per year, every engineering hour adds $1/unit to the cost of the product. PCBs might be $1.00 each in lots of 1000 but require a "set up" charge of $500 and a shipping charge of $100, so your cost per PCB would be $1.60, but your up-front cost for the PCBs would be $1600....to recoup that in the first year you would need to add $16/unit just for the bare PCB! Another supplier might specialize in "short runs" and have a set-up fee of only $100, but quote the PCBs @ $6.00 each in quantities of 100 with a shipping cost of $25....Here your total cost per PCB would be $7.25ea, but your up-front cost would only be $725. Likewise with the various components, even if buying the parts 1000 at a time from various suppliers reduced the cost of each individual component by 60%, to minimize the up-front cost you might order all of the requisite components from a single source in the exact quantities required simply because the total cost combined with shipping costs saved you several thousand dollars up-front.

....and on it goes. As the number of units increase, the "fixed costs" become a smaller and smaller cost per unit and the importance of individual component costs become more and more dominant.

In the specific case of selecting the "Best" mosfet for your application, the "rise/fall time" and "turn on/off delay time" are important only in that they are chosen such that the mosfet will function nominally in your circuit, and that is simple and well-documented math....likely with hundreds of potential solutions. If this is a one-off DIY project then I would suggest you "work backwards" starting with what's cheapest on ebay in the voltage rating you require and then review the data sheet to verify that the current, Pd, Tr, Tf, Tdr, Tdf, di/dt, Rdson, Cgs etc all meet or exceed design requirements with a prejudice toward through-hole components for easier DIY prototyping/assembly. If you anticipate building 1k or more units per year, I would likely use a source like Digikey to find the lowest cost solution for several permutations of component selection and juxtapose the component costs with the variations in PCB area with a strong prejudice toward SMD components with minimal footprint area // easy Pick-And-Place assembly.

Good Luck!

Fish
 
Top