[TriEmbed] Power FETs Vol 25, Issue 27

[Shane Trent]

Glen,

Here is a write up where I misused the IRF2804 by driving the gate with an
ATTINY85 and was surprised when it got hot! The things you can find hidden
in those curves in a datasheet.

http://fettricks.blogspot.com/2013/08/why-is-my-fet-hot-circuit-detective-work.html

Shane

On Thu, Jun 18, 2015 at 2:55 PM, Shane Trent <shanedtrent at gmail.com> wrote:

> Glen,
>
> The FET driver on my PCB can actually be used for logic levels down to 2v.
> The 5V limit refers to the load voltage being switched. I have used it with
> both a 5V Arudino and the Raspbery Pi's 3.3v outputs. If you replace the
> IRF2804 with a logic FET you can use a load voltage as low as 2.75v.
>
> I do not have much experience with logic level FETs. At low power levels
> you may be better off with a 2N3904 NPN BJT since they start turning on at
> 0.65V or so and are rated for 200mA.
>
> Finding a good 3.3v part is a challenge because most "Logic-Level" FETs
> are based on 5V logic. The datasheets specify a gate threshold voltage but
> that is just when the chip starts to conduct a small current. You really
> have to study the current vs voltages family of curves for various Vgs
> voltages. Basically you are going to have to put in some work to find a FET
> that can carry several amps and give you a low Rds at 3.3V. If you are
> going to be doing heavy lifting, more than 5A or so, it might be worth
> adding a gate driver.
>
> Shane
>
>
>
>
>
> On Thu, Jun 18, 2015 at 9:38 AM, Glen Smith <mrglenasmith at gmail.com>
> wrote:
>
>> Shane,
>>
>> I can't speak for anyone else, but I don't think it was YOU that was
>> stumbling. I know I was, but I will re-read it this evening when I can
>> devote my full attention to it.
>>
>> One question: it sounds like your little breakout board can be used with
>> its current FET (IRF2084) for a 5v signal and up to 30vdc load, but you
>> would use a different FET for a 3.3v signal. Is there a specific FET you
>> would recommend that has the right pinout for an R-Pi 3.3v application?
>>
>> Thanks
>>
>> On Thu, Jun 18, 2015, 8:55 AM Shane Trent via TriEmbed <
>> triembed at triembed.org> wrote:
>>
>>> Brian,
>>>
>>> Yes, the MIC5014 is under the FET. I fold over the FET to make it more
>>> compact and expose the TO-2220 tab for convenient mounting to a heat-sink
>>> if needed. Generally you can get 1W - 1.5W of dissipation from a TO-220
>>> package without a heat-sink. Keeping the I^2 * Rds power dissipation low is
>>> one of the reasons I like to use over-sized MOSFET's in my projects.
>>>
>>> There are two primary differences in using a FET driver vs using an IO
>>> pin with a logic-level FET. First, this FET driver will fully enhance a
>>> even a non-logic FET as long as the chip is powered by at least 5V (my
>>> application uses for 12V, note you can still power the FET driver from 30V
>>> and drive the input with a 2V logic signal) so I can use a wider range of
>>> FETs. Looking at Figure 1 on the FQP30N06L datasheet you can see that this
>>> FET is not fully enhanced at 3.3V which is not surprising since the Rds for
>>> the FET is quoted at Vgs = 10V (not an problem until you reach higher
>>> currents)
>>>
>>> Second, the FET driver will drive the gate of the FET faster that you
>>> can typically pull it up and down using a logic output. My application
>>> involved applying variable power to a solenoid via pulse-width modulation
>>> (PWM) so I wanted to make sure I could quickly switch the FET. This
>>> differences become more important at higher currents because high current
>>> FETs have a larger input capacitance and are harder to drive quickly. For
>>> example, the IRF2084 that I used has an on resistance of 2 milliohms vs the
>>> FQP30N06L with 35 milliohms (at 10V gate drive!). But this comes with a
>>> larger gate capacitance 6450 pF vs the FQP30N06L with a typical value of
>>> 800 pF. So even if the Pi could drive the IRF2804 to the higher gate
>>> voltage needed, the circuit would be slower since the gate capacitance is 8
>>> times higher.
>>>
>>> The PCB offers another advantage that may not be clear at first glance.
>>> You can mount the PCB at the load (or even on the load in my application)
>>> and switch it by supplying two wires (logic ground and your control
>>> signal). This means that the high current switching and the PWM of the gate
>>> drive are localized at the load and not traveling though long cables. This
>>> improves the performance of the switching by not having the gate drive
>>> loaded by long cables and not having power losses and reduced voltage to
>>> the load due to voltage drop in the power cables to and and from the FET if
>>> not mounted at the load.
>>>
>>> Surely something I said was not clear. Please feel free to ask for
>>> clarification if I stumbled through something.
>>>
>>> Shane
>>>
>>> On Thu, Jun 18, 2015 at 7:45 AM, Grawburg via TriEmbed <
>>> triembed at triembed.org> wrote:
>>>
>>>> Shane,
>>>>
>>>> So the FET driver is under the FET itself?
>>>> If the FQP30N06L I am using switches with the 3.3VDC from the Pi, what
>>>> advantage does your combo offer (disregarding the other components)?
>>>>
>>>>
>>>> Brian
>>>>
>>>>
>>>> _______________________________________________
>>>> Triangle, NC Embedded Computing mailing list
>>>> TriEmbed at triembed.org
>>>> http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
>>>> TriEmbed web site: http://TriEmbed.org
>>>>
>>>>
>>>
>>>
>>> --
>>> A blog about some of my projects.  http://fettricks.blogspot.com/
>>>  _______________________________________________
>>> Triangle, NC Embedded Computing mailing list
>>> TriEmbed at triembed.org
>>> http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
>>> TriEmbed web site: http://TriEmbed.org
>>>
>>
>
>
> --
> A blog about some of my projects.  http://fettricks.blogspot.com/
>



-- 
A blog about some of my projects.  http://fettricks.blogspot.com/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.triembed.org/pipermail/triembed_triembed.org/attachments/20150618/5e8ea05a/attachment.htm>