GenBoard/PowerSwitcher/FET (2005-06-24 12:52:29)

This page describes different varieties of FETs (e.g. logic-level, DPAK) used by the VEMS group. It is primarily intended for use by developers.


Vanilla FET -- TO220

Used as cheap/easy to find switching drivers for injectors. Low RDSon requires the use of a FET driver:

non-isolated (forget them)

Part Number Description Datasheet
IRFZ34N Non-isolated (undesirable).
IRFZ44N "
IRF3205 "
BUP.... "

Logic-Level FETs -- TO220

Part Number Description Datasheet
FQPF20N06L 15A, 9.5nC, Rds=0.055 logiclevel. Available in the Web Shop. The recommended insulated logic-level FET for the injector drivers and WBO2 heater driver. http://www.fulcrum.ru/Read/CDROMs/FAIRCHILD/DISK3/PDFs/MOSFETs/TO-220F/N-Channel/FQPF20N06L.pdf
IRLI 530 N (Lomex appr. 1Euro, 50pcs, 2 weeks) logiclevel 30V/116A (130W)
IRL2203N 30V/116A (130W)
IRL2505 50V/10A/logic level FET/TO220
IRLZ24N "
IRLZ34N "
IRLZ44N "
RFP25N05L "
RFP50N05L "
MTP50N06VL "
IRLIZ24N TO220 Fullpack (Isolated) vclamp 55v, Rds ON 105mOhm (4.5v Vgs),Rds ON 60mOhm (10v Vgs) Ids 9.9A @ 100C, Total Gate charge= 15nC. Rth Junction to case= 4.8C/W, Rth Junction to ambient 65C/W. Power dissipation 26W.
IRLIZ34N TO220 Fullpack (Isolated) vclamp 55v, Rds ON 60mOhm (4.5v Vgs),Rds ON 35mOhm (10v Vgs) Ids 14A @ 100C, Total Gate charge= 25nC. Rth Junction to case= 4.1C/W, Rth Junction to ambient 65C/W. Power dissipation 31W.
IRLIZ44N TO220 Fullpack (Isolated) vclamp 55v, Rds ON 35mOhm (4.5v Vgs),Rds ON 22mOhm (10v Vgs) Ids 20A @ 100C, Total Gate charge= 48nC. Rth Junction to case= 3.3C/W, Rth Junction to ambient 65C/W. Power dissipation 38W.
IRLI2505N TO220 Fullpack (Isolated) vclamp 55v, Rds ON 13mOhm (4.5v Vgs),Rds ON 8mOhm (10v Vgs) Ids 41A @ 100C, Total Gate charge= 130nC. Rth Junction to case= 2.4C/W, Rth Junction to ambient 65C/W. Power dissipation 63W. http://ec.irf.com/v6/en/US/adirect/ir?cmd=catProductDetailFrame&productID=IRLI2505

More Information:

Note: most of the above are available as SMDs (e.g. TO-252 aka DPAK).


SMD FETs

Part Number Description Datasheet
RFD16N05LSM 16A, 50V, 0.047 Ohm, Logic Level, N-Channel Power MOSFETs DPAK http://www.fairchildsemi.com/pf/RF/RFD16N05LSM.html
HUF76121D3S Commercial 30 V 20 A 23 mOhm 13 nC 4.5v compatible DPAK
HUF76419D3S Commercial 60 V 20 A 37 mOhm 23 nC 4.5v compatible DPAK
HUF76423D3S Commercial 60 V 20 A 32 mOhm 28 nC 4.5v compatible DPAK
HUF76429D3S Commercial 60 V 20 A 23 mOhm 38 nC 4.5v compatible DPAK
HUFA76413D3S AutoCommercial 60 V 20 A 49 mOhm 17 nC 4.5v compatible DPAK
HUFA76419D3S AutoCommercial 60 V 20 A 37 mOhm 23 nC 4.5v compatible DPAK
HUFA76423D3S AutoCommercial 60 V 20 A 32 mOhm 28 nC 4.5v compatible DPAK
HUFA76429D3S AutoCommercial 60 V 20 A 23 mOhm 38 nC 4.5v compatible DPAK
ISL9N327AD3ST Commercial 30 V 20 A 27 mOhm 17nC 4.5v compatible DPAK
IRLR024N DPAK 65mOhm ...

Most frequently used (stocked in high quantity) products marked bold. IRLR024N is the mass FET, and HUFA76423D3S where very high current is needed, eg. switching supply for CdiSupply and IonSense. All DPAK of course


Smart FETs

These FETs have added protection like:

Part Number Description Datasheet
IPS021 150mOhm, Ishutdown 5A, Tshutdown 175C, Vclamp 50v, 3.3nC Total gate charge. http://www.irf.com/product-info/datasheets/data/ips021.pdf
IPS031 60mOhm, Ishutdown 12A, Tshutdown 175C, Vclamp 50v, 11nC Total gate charge. My preferred injection and misc driver (Jörgen). http://www.irf.com/product-info/datasheets/data/ips031.pdf
IR6216   http://www.irf.com/product-info/datasheets/data/ir6216.pdf

Multiple FETs In One Package

Part Number Description Datasheet
TPIC6A259 8 * 350 mA (used on v3.x)
MC33291 8 * 500 mA (uses SPI, more complex than 259)
ULN.... unipolar stepper drivers
SN754410 (L293) 4* 1A push-pull like several other bipolar stepper drivers (used on v3.x)

FET drivers:

The FET driver is simple inverting device. If input is above threshold (appr. 2V), the output will be close to 0V, below that threshold output will be close to supply (+12V). The input is 0 or 5V with the AVR driving it. Anything between means PWMing is active.

For lowZ injectors, which switch upto about 16000 times a second if PWM-ing is configured, fast switching is needed to keep dissipation (heat) down.

The FETdrivers do exactly this:

For highZ injectors, that only switch max 100 times a second FETdriver is not needed (but it works with installed FETdrivers).

See GenBoard/VerThree/LowZInjectors


Some scientific discussion on IRC (2004-03-04) about Smart FET vs. Generic FET

issue: lean-condition that can be created at serious errors with smart-switches (even though unlikely).

<Jorgen> The choice between a vclamped FET with isolated case and a fully protected FET without isolated case isn't easy for me...
<rich1> What is the worst case scenario for each?
<Jorgen> Both are vclamped and can live without any other flyback. (not sure if the isolated part is built for it but it has the clamp). The isolated part will fail to a short circuit if there is an overcurrent / overtemperature condition. The Smart FET will just shut down if either happen.
<rich1> A controlled shut down sounds like a nicer option to me...
<Jorgen> The Smart fet can cause the injector duty to decrease if driven outside the limits. The other will flood the engine when it fails, or if it's a hard shortcircuit there will be damage to the board and cabling.
<rich1> What is the downside of the Smart FET then? Higher cost?
<Jorgen> By choosing the IPS31 smart fet the limits will be very far away from the operation range.
<Jorgen> Very little difference.
<rich1> Ok, I see now. So ideal is a smart fet with tighter limits
<Jorgen> The current needed to charge the smart FET is also much smaller then what's needed for the other FET.
<Jorgen> No, tight limits can cause problems for us.
<Jorgen> Imagine that you run your setup with a few large injectors and you cause your injector FET to hit shutdown limit after 70% of the commanded duty. The injector shut's early and you loose the last 30% of fuel.
<Jorgen> The next injector opening the same thing will happen.
<rich1> Better a loss of power than frying them
<rx7turbo> Would that be a big concern if the limit was far away from normal range though?
<rich1> As long as there is some feedback of a problem, that should be bearable
<Jorgen> No, if the limit is far away from the normal range it will not be a problem
<Jorgen> You won't see loss of power, you will see a hole in your pistons...
<rich1> So if the choice is burned motor vs burnt fet, i know what i'd choose...
<rich1> Are there any fets with an overload signal as well?
<Jorgen> We would need to add a diagnosis circuit on the board or switch to a high side switch with a diagnosis signal.
<Jorgen> The Isolated switch with a fuse sounds good to me.

I suggest low-voltage FlyBack for injector drivers, using S2X (see below) SMB diodes and 0 Ohm 'resistors', or throughole diodes mounted SMD style with a bit art.


See also: