User avatar
Hove
Posts: 1205
Joined: Sun Oct 21, 2012 6:55 pm
Location: Cotswolds, UK
Contact: Website

MOSFET switching recommendation, please. (solved)

Sun Oct 05, 2014 5:52 pm

As a follow up to this conversation http://www.raspberrypi.org/forums/viewt ... 8&p=622179, I need to find a mosfet capable of switching 5V (source to drain) with 3.3V (gate to drain) PWM signal - 300Hz carrier with microsecond granularity - although I suspect the pulses will only need to be in 0.1ms units. It'll be switching via a 25ohm resistor (to heat it up), so the current will be pulses of 0.2A (so max power is 1W which is about what most SMD resistors need to start overheating!)

To avoid a time consuming stroll through RS/Farnell vast range of mosfets, I'm asking for suggestions from the experienced electronic engineers out there. I think the only other requirement is that it's through the hole, not SMD.

Thanks in advance.
Last edited by Hove on Mon Oct 06, 2014 5:19 pm, edited 1 time in total.
www.pistuffing.co.uk - Raspberry Pi and other stuffing!

Tarcas
Posts: 740
Joined: Thu Jan 09, 2014 5:38 am
Location: USA

Re: MOSFET switching recommendation, please.

Mon Oct 06, 2014 12:50 am

I'm no electronics engineer, but is there any reason something like this wouldn't work?
http://www.ebay.com/itm/1pcs-RD16HHF1-S ... 3ce6069381

Ebay has a TON of cheap electronics which you can have shipped from China for pennies or a few dollars. With a well defined search, you should be able to track down the perfect part.

...and by the way, your MOSFET doesn't have to switch on 3.3v. If it needs more, you can always use the GPIO to switch a transistor to switch the MOSFET in something resembling a Darlington pair.

Obviously I haven't tried everything, or even any significant percentage, but what I've gotten and used so far (resistors, transistors, LEDs, various sensors, wires, breadboards, etc) have all worked great. And what I buy is usually among the cheapest per unit for a small number of units. (My definition of small depends on what I'm getting and my intended uses.)

User avatar
Hove
Posts: 1205
Joined: Sun Oct 21, 2012 6:55 pm
Location: Cotswolds, UK
Contact: Website

Re: MOSFET switching recommendation, please.

Mon Oct 06, 2014 5:20 am

That one's a bit bog for my needs - as I'm switching, I don't need a high power mosfet, but thanks for the pointer!
www.pistuffing.co.uk - Raspberry Pi and other stuffing!

User avatar
Hove
Posts: 1205
Joined: Sun Oct 21, 2012 6:55 pm
Location: Cotswolds, UK
Contact: Website

Re: SOLVED: MOSFET switching recommendation, please.

Mon Oct 06, 2014 5:22 pm

www.pistuffing.co.uk - Raspberry Pi and other stuffing!

grahamed
Posts: 277
Joined: Mon Jan 30, 2012 7:01 pm

Re: SOLVED: MOSFET switching recommendation, please.

Fri Oct 10, 2014 4:01 pm

Hi

I am thinking 2N7000 is maybe not good enough. Details aside RDS(ON) is measured in Ohms rather than milli-Ohms which is never a good sign.
and ....
N FET 2N7000 on-region characteristics.jpg
N FET 2N7000 on-region characteristics.jpg (17.01 KiB) Viewed 2779 times
As to RD16HHF1- my theory is that if the graphs on the datasheet don't go down to 3V then you shouldn't either.
N FET RD16HHF1 on-region characteristics.jpg
N FET RD16HHF1 on-region characteristics.jpg (14.95 KiB) Viewed 2779 times
IRL7833 is more capable with lower RDS and reduced VGS requirement, but bigger and more expensive.
N FET IRL7833 on-region characteristics.jpg
N FET IRL7833 on-region characteristics.jpg (18.65 KiB) Viewed 2779 times
Are you able to measure VDS when the device is being pulsed?

Of course this is only theory and if theory disagrees with practice then theory is wrong - so let us know how you get on.

User avatar
Hove
Posts: 1205
Joined: Sun Oct 21, 2012 6:55 pm
Location: Cotswolds, UK
Contact: Website

Re: SOLVED: MOSFET switching recommendation, please.

Fri Oct 10, 2014 4:18 pm

The BS170 seems to be working OK, though I will measure Vds with my 'scope to see how much of the 5V it's taking. I'm expecting only 0.1v based upon the load being 50 ohms.

If Rds turns out to be greater than the specified 1.2Ohms, then I'll keep looking, though currently my biggest problem is heat being conducted away via the chip pins and PCB.

Have a look here if you're interested: http://blog.pistuffing.co.uk/?p=4021
www.pistuffing.co.uk - Raspberry Pi and other stuffing!

User avatar
Hove
Posts: 1205
Joined: Sun Oct 21, 2012 6:55 pm
Location: Cotswolds, UK
Contact: Website

Re: SOLVED: MOSFET switching recommendation, please.

Sat Oct 11, 2014 8:37 am

grahamed wrote: Are you able to measure VDS when the device is being pulsed?
Oh dear, Vds is 4.5v when Vgs is pulsing at 3.3v on by BS176

That might explain why the 50ohm resistor between source and 5v is not working very well as a heating element :(

Trouble is the devices with the lower Rds and Vgs are the bigger ones that I have space for while prototyping but not on the final product.

Time to find a better mosfet :(
www.pistuffing.co.uk - Raspberry Pi and other stuffing!

User avatar
Burngate
Posts: 5887
Joined: Thu Sep 29, 2011 4:34 pm
Location: Berkshire UK Tralfamadore
Contact: Website

Re: SOLVED: MOSFET switching recommendation, please.

Sat Oct 11, 2014 9:51 am

Maybe not a useful suggestion, but ...
... have you considered ditching the resistor, and using the fet itself as the heating element?

Fets have a useful feature in that as the temperature goes up the current goes down.
That means that, unlike bipolar transistors, they don't suffer from thermal runaway (or atleast not in the same way)

User avatar
Hove
Posts: 1205
Joined: Sun Oct 21, 2012 6:55 pm
Location: Cotswolds, UK
Contact: Website

Re: SOLVED: MOSFET switching recommendation, please.

Sat Oct 11, 2014 3:07 pm

That could be a useful approach in other (analogue feedback) systems, but I need to be able to set the desired temperature, have a sensor read the temperature at the core of the chip, and use the sensor feedback to modify the width of the PWM signals to heat or cool the chip accordingly.

That then will allow the other sensors in the same chip to operate in a fixed temperature environment - they vary over temperature by 2% - I've got it down to 0.1% though interpolation but I think the only way to get another decimal point is to control the sensors' environment.
www.pistuffing.co.uk - Raspberry Pi and other stuffing!

grahamed
Posts: 277
Joined: Mon Jan 30, 2012 7:01 pm

Re: SOLVED: MOSFET switching recommendation, please.

Sat Oct 11, 2014 3:23 pm

Hi

Hove

BS170 eh.....
N FET BS170 on-region characteristics.jpg
N FET BS170 on-region characteristics.jpg (32.08 KiB) Viewed 2690 times
Graph shows VGS from 4V and up.

Extrapolation suggests FET will act as an approx. 100mA constant current source for VGS=3V3

Burngate

Good point. I have often used FET/BJT with no drain/collector resistor mainly to make a current sink but always within a feedback loop (usually a source/emitter resistor + OPA). But just a FET? Operation highly dependent on characteristics. I think maybe a 50R would work just same in this case.

grahamed
Posts: 277
Joined: Mon Jan 30, 2012 7:01 pm

Re: SOLVED: MOSFET switching recommendation, please.

Sat Oct 11, 2014 3:36 pm

Hi Hove

Having just read yours of 15:07 I don't understand why you want to use PWM? For a stable temperature I'd consider a BJT (plus collector resistor) in a feedback loop with a thermistor, or a diode, to sense chip temperature.

As an aside I have commented many times on this forum (as has Burngate) that people frequently over-estimate hFE and hence under-estimate Ib, I am thinking that equally many people under-estimate VGS requirement - even "logic level" FETs struggle at VGs=3V3. You are not alone.

Good luck.

User avatar
Hove
Posts: 1205
Joined: Sun Oct 21, 2012 6:55 pm
Location: Cotswolds, UK
Contact: Website

Re: SOLVED: MOSFET switching recommendation, please.

Sat Oct 11, 2014 4:13 pm

BJT / thermistor would work fine BUT...

The point of what I'm doing is to maintain the temperature right at core of my accelerometer / gyro. The sensor outputs vary across temperature by +/- 2%. Calibration at two temperatures and interpolating takes this down to 0.1% - but 0.1% of gravity is 1 cm/s/s acceleration which rapidly accumulates to vertical or horizontal drift. That limits my quadcopter autonomous flight times to 10s.

There is a temperature sensor right to the heart of the IMU chip too. Hence I get i2c feedback temperature exactly where I want it from. Then I need to adjust the warmth according to the sensor readings. I chose PWM from its use to control LED brightness - but in this case the 3.3v PWM is switching the fec @ 5V via a 50 ohm resistor. The resistor has enough heat capacitance to smooth out the pulses obviously.

OTT, yes, but when looking for 0.01% accuracy, my experience on this project has proved OTT is necessary again and again!
www.pistuffing.co.uk - Raspberry Pi and other stuffing!

grahamed
Posts: 277
Joined: Mon Jan 30, 2012 7:01 pm

Re: SOLVED: MOSFET switching recommendation, please.

Sat Oct 11, 2014 4:24 pm

Hi

OK I did not understand that the temperature sensor is I2C......

User avatar
Hove
Posts: 1205
Joined: Sun Oct 21, 2012 6:55 pm
Location: Cotswolds, UK
Contact: Website

Re: SOLVED: MOSFET switching recommendation, please.

Sat Oct 11, 2014 4:31 pm

grahamed wrote:OK I did not understand that the temperature sensor is I2C......
I'm only using that as it's there, and right at the centre of the chip whose temperature I'm trying to stablize.

The other factor is the chip is only 4mm x 4mm x 0.9mm so not much space - just about enough for an SMD resistor heating element! :)
www.pistuffing.co.uk - Raspberry Pi and other stuffing!

Return to “Automation, sensing and robotics”