Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

FET Driver Design

Status
Not open for further replies.

Kerrowman

Member level 4
Joined
Oct 12, 2021
Messages
75
Helped
0
Reputation
0
Reaction score
2
Trophy points
8
Location
West Penwith
Activity points
578
Hi there,

I have been building a generator that uses back EMF pulses to charge and repair lead-acid batteries and it works fine at the moment. However, as part of a series of upgrades I wanted to get the IRF840 FET to shut off as fast as possible to produce sharper voltage transients and for that I believe I need to use a gate driver.

My query then is what components and circuit design to use with this FET? I attach the present circuit layout.

Thanks

Jules
 

Attachments

  • Generator Circuit.jpg
    Generator Circuit.jpg
    152.8 KB · Views: 150

Hi,

There are dedicated MOSFET driver ICs.
Just do a search at your electronics distributor, the internet, here in this forum, at manufacturers.

Manufacturers provide datasheets as well as application notes and design notes that show schematics, what to care for and how to calculate things.

This is the way I´d go, because the informations are complete, detailed and very reliable.
*****

you say "as fast as possible". This is not a value one can calculate with.
I strongly recommend to decied values with units.
Best if you have two of them:
* "nice to have"
* and "at least"


Klaus
 

Thanks for info. I recognise that 'fast as possible' is vague but I don't have a value as yet. Probably in the 0.1-1 us range.

I will explore the various searches.
 

Hi,

That´s a good value to start.

The application notes will tell you how to calculate the necessary gate current.
With this information you can use MOSFET driver selection guides to find suitable devices.

Klaus

Btw: you are short circuiting the motor/generator coils. I don´t expect any usueful output power.
 

If you want highest dV/dt, dI/dt you should consider eGaN
devices.

These have gates you can drive with a stiff 5V driver - for
small ones, just parallel an octal bus buffer and you can
see ~2A peak drive from HCS or ACS series logic. For the
home experimenter I'd recommend GaN Systems or other
leaded-package types.

But I'd bet that busting sulfate off battery plates could be
done as well with microsecond risetimes, as nanoseconds.
Are you convinced (by data) that you're really being limited
by risetime, rather than pulse height or pulse energy (which
might argue to change the inductor, rather than the switch)?
 

If you want highest dV/dt, dI/dt you should consider eGaN
devices.

These have gates you can drive with a stiff 5V driver - for
small ones, just parallel an octal bus buffer and you can
see ~2A peak drive from HCS or ACS series logic. For the
home experimenter I'd recommend GaN Systems or other
leaded-package types.

But I'd bet that busting sulfate off battery plates could be
done as well with microsecond risetimes, as nanoseconds.
Are you convinced (by data) that you're really being limited
by risetime, rather than pulse height or pulse energy (which
might argue to change the inductor, rather than the switch)?
You raise some interesting points. Some background might help here.

I built this device about 3 years ago as a hobby/interest activity having retired from a profession in radiation physics. So electronics is not my specialty and I have had to learn a lot of new things and skills. Now that I’ve reached the stage where I can design and build a PCB, I decided to upgrade the device, also because I have become interested in the potential ‘energy harvesting’ properties of transient spikes.

I’m sure you’re right that heavy sulphate ions won’t respond well to ns duration pulses but my focus is now on creating high dV/dt spikes at the FET Drain when the magnetic field in the coils collapses rapidly. The faster the FET switches off the greater the rate of flux change in the coils and the stronger the CEMF pulses.

I understand that this can be achieved with a driver, hence my original query. I don’t know anything about eGan devices but am happy to find out about them. I also don’t know what HCS or ACS logic is and my primary trigger pulse comes from the Hall sensor so presumably CMOS or TTL which would work with a driver such as the TC1413NEPA? As I understand it the FET switch off can be accelerated with a few extra components.

Currently my IRF840 is getting quite hot which seems to impair its function after a while so if nothing else I need to fit a larger heat sink or change to a beefier device. Since I’m ready to build a new pcb then it’s a good time to consider such things (if you look at the board I originally built I’m sure you would agree :))

The attached pics show the setup and a scope trace (V=53.6V not kV as probe had been set up for another purpose). I have also modified the rotor magnets which has resulted in a 50% increase in pulse frequency, which will be useful for my purposes, but as the inductors are fixed changes will have to be in other areas and I agree that pulse energy, a function of both frequency and pulse height, is the focus.
 

Attachments

  • Rotor Generator 2.jpg
    Rotor Generator 2.jpg
    1.4 MB · Views: 94
  • 07D30A85-796C-4C5E-B92E-CF07AB543D61.jpeg
    07D30A85-796C-4C5E-B92E-CF07AB543D61.jpeg
    2.1 MB · Views: 121

Ok, so after some research, I have come up with an upgraded design for a replacement trigger circuit using a GAN FET and associated driver. The present circuit and the proposed one are shown in the attached images.

I have several queries about using the Driver and am not fully confident on whether I have all the pins I need connected, with the FET serving as a low-sided switch?

Firstly, I'm not sure if I need pin 16 connected in some way. Also, I'm querying if I can use the same Hall sensor setup, with the BC549C, as a signal to feed the STDRIVEG600 chip?

The use of the Ferrit bead and the DC-Link snubber (whatever that is) are both recommended in the spec notes for the GAN FET.

Thanks
 

Attachments

  • Present Trigger Circuit.jpeg
    Present Trigger Circuit.jpeg
    119.9 KB · Views: 138
  • Proposed Trigger Circuit.jpeg
    Proposed Trigger Circuit.jpeg
    241.6 KB · Views: 130

Hi,

you urgently need to re read the datasheet.

Some issues
* SGND needs to be connected
* PVCC needs to be connected, including capacitor
* LOFF needs to be connected.

The datasheet shows you how and what to care for.

Keep in mind: PCB layout is very important for high speed switching high currents.
Strictly keep on datasheet and example circuit / development board.

****
Generally I´d say this is more speed than you need. The IC is new and is not good available. Please check.
It is made for half bridge, while you only use low side. Not wrong, but you pay for something you don´t use.

****
It´s better to use a "symbol" that shows it´s function, than a package with pin numbers in a schematic.
It´s non statndard, causes confusion .. and misses informations.

Klaus
 

Hi Klaus,

SGND is connected at pin 7 from the Hall sensor ground. I’m not sure what to connect to the other two.

You are right that the GAN FET and driver can handle much more speed than I need which is only 200-300 Hz. I was only considering it due to the recommendation of a forum member and the FET’s beefy nature but, given that my coil (load) current is less than 3-4A, then I can probably do well with a different FET with a lower on-resistance. The IRF840 is 0.8 ohms I think so I need lower to reduce the heat generated in it. Also some suitable driver to shut off the FET quickly, which is the basis of my query. Can you suggest anything for these?

What is the symbol for a driver chip?

Thank you
 

Hi
SGND is connected at pin 7 from the Hall sensor ground.
SGND is the voltage reference for the input signals. It´s also the reference for VCC.
So in order to supply the logic section of the chip you need to connect VCC and SGND properly to a power source.

PGND is the voltage reference for the low side driver supply as well as the low side driver output.

Usually both PGND and SGND are connected via a low impedance path. There may be a current sensing resistor...But they must not float with respect to each other .. to keep the IC specification PGND to SGND must not exceed +/-21V.

Keep in mind:
1) Any voltage measurement needs two points: The signal you want to measure and the signal you refer to.
And so do all analog as well as digital inputs.

2) all currents flow in a loop. There is no "one way" current. Especially four your application it is urgent to know these loops of high dI/dt current flow. It needs to be short and wide traces and the loop should enclose a minimal area to be low impedance.

Thus: Follow the datasheet recommendations and keep as close as possible on demonstration board PCB layout.

***
I didn´t know the driver before. So I´ve read the datasheet. Every designer needs to do this.

* PVCC is named "Low-side driver supply ". It needs to be connected to a power supply that is meant to drive the gate. It must not exceed 21V.

* LOFF is named "Low-side driver sink output". It is used to switch OFF the Low side FET. This is urgent for you, because this is your main task. And for sure if you want to switch OFF something you need to swith it ON first, thus you need LON. See the example schematics and the description of LON and LOFF.

You are right that the GAN FET and driver can handle much more speed than I need which is only 200-300 Hz.
No. Don´t mix your PWM frequency with the high frequency component in the fast edge. A high rise rate in the edge (thats what you want) (mathematically: dV/dt) will have frequency components in the 100MHz range. PWM frequency has nothing to do with this.

I can probably do well with a different FET
I thought so. Thus I wrote my recommendation in post#2. This is the way i also had to go if I were in your place.

What is the symbol for a driver chip?
Usually a triangle.
The symbol in the datasheet is very detailled and even shows the internal FETs.
Imortant for a schematic: It should be as good as possible self explanatory .. no need for extra drawings and datasheets. Pin number s 1 , 2, 3 .. don´t tell anything about the function.
An alternative is to use tha signals names.

Just a simple example: You know the symbol of a BJT.
* The symbol tells you which pin is emitter, base and collector. It also tells whether it´s an NPN or a PNP.
* a box with three pins B, E, C .. at least tells which pin has which function.
* a box with 1, 2, 3 ... tells nothing.

Klaus
 

Thank you for the details. Electronics is not my primary skill. I will study what you have said and revise my proposed circuit and post it in the next day or so. Thanks
 

Hi Klaus,

I’ve taken your advice and reverted to a more traditional form of driver with a FET that has a lower On resistance to reduce heating.

I attach the revised circuit being triggered by a Hall sensor. I hope I have the resistance and cap values about right but please advise of any mods needed.

The second diagram is triggering the circuit with a 555 timer instead of a Hall sensor and rotor and I’m assuming I can just switch the feed without any additional modifications?
 

Attachments

  • Revised FET&Driver (using Hall sensor).jpeg
    Revised FET&Driver (using Hall sensor).jpeg
    190.4 KB · Views: 132
  • Revised FET&Driver (using 555 Timer).jpeg
    Revised FET&Driver (using 555 Timer).jpeg
    125.9 KB · Views: 132

"Normally" you want the diode you show connected to MOSFET Source and
Drain to be a fast recovery type. The 1N5408 is standard power diode and if
you look at its datasheet has no specs on reverse recovery. You want a fast
recovery diode.







Regards, Dana.
 
Last edited:

Thanks, I will look into that.
--- Updated ---

Here is the schematic based on my revised trigger/drive circuit shown above using readily available components. I have replaced the diode across the Drain/Source with a high speed one. Bear in mind that the purpose of this device is to access and tap off the CEMF spikes from the Drain so I hope it won't dampen them.
 

Attachments

  • Trigger & Drive Schematic.jpeg
    Trigger & Drive Schematic.jpeg
    372.9 KB · Views: 125

Also keep in mind not all Caps, for same capacitiance, yield equal
ESR performance, so choose wisely.

1634559434848.png


Polymer caps for bulk applications excellent. Thats the OS-CON type shown above.


Regards, Dana.
 

Also keep in mind not all Caps, for same capacitiance, yield equal
ESR performance, so choose wisely.

View attachment 172438

Polymer caps for bulk applications excellent. Thats the OS-CON type shown above.


Regards, Dana.
Thanks.

The various caps I have used in my first prototype have worked ok so no apparent problems. The major change will be the FET driver to switch it off more quickly and cleanly and the pulse rate from changes to the rotor (The spikes in the trace are 53.6V not kV).
 

Attachments

  • 35622B8F-98A1-45A8-B86A-F35867D72569.jpeg
    35622B8F-98A1-45A8-B86A-F35867D72569.jpeg
    3.3 MB · Views: 101
  • 3D06B0F4-2E47-4184-B71F-E3C9020F87D3.jpeg
    3D06B0F4-2E47-4184-B71F-E3C9020F87D3.jpeg
    2.1 MB · Views: 88

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top