electrical controls

[turbochris]

I'm rewiring the "entertainment" section of the little turbine bike. I want that thing done and out of my life before it maims me any more... but I must complete the vision.

Right now it's all computer case lights and relays and not very efficient. Im ripping it out and Richard said if i promised to hook all the junk up properly he would program it.

So I'm adding another arduino, completely separate from the arduino I use for the throttle for safety reasons.

I'll wire the hotstreak and AB pump buttons to the arduino and I'll have some safety interlock and a hotstreak timer. I have a strip of 144 addressable LED's for my lighting needs. I'll ad a few more buttons to the inputs to the arduino to trigger different programmed effffects.

When it's all done i'll give the whole thing to richard so he can sit in his apartment for a few days geeking out over it. I'll just bring him food and energy drinks and pretty soon I'll be starting it from my cell phone and blowing a little smoke n fire!

What would you guys use to trigger one of Ernies 10 amp N2O solenoids from an arduino? I'm guessing just a stout n channel mosfet and a resistor between it and the arduino? Do I need any back emf protecton? BTW Ernies niods take PWM nicely from what he tells me. What's a good tough comon mosfet w leads? I prolly have one, someone cleaned out a TV repair shop and gave me 10 pounds of components.

Today i'm working on power distribution. I have a hard pack 3s lipo pack for the entertainment w fused distribution and a master cutoff already, I need to add a 5v switching regulator, I have some samples from Texas Instruments some make like 25 amps!

The hot streak solenoid drawsa lot of current but the design is intrinsically safe, a timer would just keep the current draw time down.

The AB pump is from a GSXR, way way too big, it will be changed with a pump similar to the fuel pump for the turbine.

[cursorkeys]

Yeah, flyback diodes and snubbing are a must on FETs and BJTs. Most of the power FETs (especially those designed for chopper service) have beefy body diodes so snubbing is generally all that is needed for them. The datasheets are usually fairly clear on the FET body-diode capabilities, if nothing is mentioned then it's usually just the poor-performance intrinsic diode present and you need an external diode as well.

Personally I'd use a Darlington transistor as I've seen FETs fail in this kind of application unless the snubbing is very carefully designed (sometimes to the gate as well rather than just drain-source short). You can snub the heck of a FET though using an ultra-fast TVS diode across it.

Gate resistors on FETs are usually bad news unless you are deliberately trying to increase the time the FET stays in the linear region as the gate charges (Ciss) and thus limit di/dt for noise reasons or because the snubbing is inadequate. As noise shouldn't be an issue in this circumstance the less time you spend in the linear region means less heat. Above a few kilohertz you really start seeing the gate capacitance start to slow switching down unless you drive the gates hard.
If you just want to protect the driving circuit from a FET gate-source/gate-drain short circuit failure then there are loads of really nice FET driver ICs like the FAN3111 but pretty much every power semi manufacturer offers single-channel low-side drivers too.

For Darlington’s the BDV67B would be suitable.
If your Arduino was driving it with a 5V output you'd need a ((5-1.2)/(10/1000)) 380 ohm resistor, allow a bit for tolerance and that's 240 ohm.
If your Arduino was driving it with a 12V output you'd need a ((12-1.2)/(10/1000)) 1080 ohm resistor, allow a bit for tolerance and that's 820 ohm.

If you want to use a FET and your Arduino output is 5V then make sure you get a 'logic-level' FET like the IRL530 as most FET parameters like the channel resistance is specified at least 12V Vgs and the channel may not even ever fully open unless designed for low drive voltages.


Dunno if any of that if useful to you but I hope it helps

[turbochris]

Exactly what I was looking for, thanks buddy!
I'm sure there's some beefy darlingtons in this box I have.

[turbochris]

found these they were the closest I could reference.-

www.fairchildsemi.com/ds/1V/1V5KE16A.pdf

www.fairchildsemi.com/ds/2S/2SC5200.pdf[url/]

www.fairchildsemi.com/ds/FQ/FQP30N06L.pdf

my company gets free samples from Fairchild and TI

[cursorkeys]

Free samples are the best perk of a job involving electronics...my 12KW ion laser driver was full of top-of-the-line Fairchild sample FETs and diodes, very nice of them indeed

The 1V5KE16A is a bit close to the supply if you're on 12V. Something between the supply and the Vce/Vds breakdown voltage might be better like the 1V5KE33.

That FET is avalanche rated so it would be the best sort for the job if you use a FET.

The BJT is a monster but it isn't a darlington and has a really low Hfe (current gain) so you'd probably need a smaller BJT with a large Hfe as the first transistor to make a darlington pair (like the KSD2012 if you wanted to stay in a TO220 package). Or they do have a power darlington in one package the BDW93A as it's a darlington in one package the Hfe at 10A load isn't great but it's still 300 which means a sensible base drive current.


Edit: Something like this for the two circuits? The darlington here has a gain of roughly 4000 at 10A load. The cross-hatched bit on the FET circuit is completely optional, if you have huge ground bounce then as the gate has charge you can see voltages above below supply and that just clamps those transients before the internal ESD protection on your microcontroller conducts and you get nasty things happening.

[turbochris]

thanks my friend! You will save me some magic smoke/mixed up 1's n 0's for sure.
Thanks for explaining it too, I'm trying to wrap my head around how to spec this stuff out.

I made a breakout board for the arduino last night. I used old computer connectors and cables as i have a piles of them. The only thing left is to get w richard on the pin assignment but the power/ground is done.

[turbochris]

OK i'm starting to wrap my head around this linear/saturated response on transistors. I have a crapload of industrial IO modules w optoisolators. I think I'll use one of them to trigger the mosfet/darlington off the 12 v and isolate the output of the AVR at the same time. I saw how using a transisitor to trigger it helped but the isolator seems like a better idea to give some flexibility on driver selection. Since I've been reading about this i've been ripping apart anything that works a solenoid/motor off logic n learning by example. Everythng from a keruig to an allen bradley devicenet node that took a crap. I think I could use the whole output of one of these burnt PLC nodes n trigger it off ttl like a relay board. Not for the little bike but for the van.

[cursorkeys]

Yep, even FETs have a linear region and if you go far enough into the semiconductor physics you can see why the current gain changes at both high and low current extremes...that might for masochists only though

For switching duty as the frequency increases the dynamic losses (the energy you lose by passing though the linear region) become a bigger and bigger part of the energy loss than the static loss (for FETs that's Rds_on which is the intrinsic channel resistance, and Vce for BJTs) and thus heat in your switching device. That's why Rds_on is usually a red herring in FET datasheets, unless you're using it as a static or very low frequency switch the gate and leakage capacitance determine your effective device impedance. I've seen devices with Rds_on in the tens of milliohms but with enormous gate and miller capacitances...lies, damn lies and datasheets!

The circuits I posted are the most basic forms of a BJT and FET low-side switch. You can speed both of them up considerably and reduce losses.
The FET by using a low-side driver (discrete like a totem-pole design or an IC) so you can push and pull current _fast_ into that gate capacitance and thus minimise time in the linear region. This is a great doc on FET gate drive design: www.ti.com/lit/ml/slup169/slup169.pdf
The BJT darlington by adding a couple of carefully chosen resistors to help with something called drift charge and you can even add a negative voltage generator to drag the base closed quicker.

Also the TVS diode method I showed is the bullet-proof way of snubbing the switching device but has an EMI downside. You can use Zeners or MOVs across the switch instead or proper snubbing networks which usually require tuning with a scope after doing the math. If you just use a normal diode/schottky diode it has to be moved to across the solenoid rather than across the switch or it won't do anything.


Anyway, re-using commercial drive bits sounds like a solid idea

[turbochris]

heh I'll work my way up to this soon!-

www.fairchildsemi.com/ds/FT/FTCO3V455A1.pdf

[turbochris]

I ripped open 2 different 24 volt industrial I/O modules. They had these awesome optoisolators PC817. One is in a rail mount housing w lots of space inside. All the outputs were darlingtons w diodes hooked up in a manner you described.

I'm putting a header inside so I can plug n arduino in or one of the variants. I don't even know what I'll use if for but when I get an idea it's there.