Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Using a battery to drive a mosfet gate.

Status
Not open for further replies.

rnorman3

New Member
Hello all, I'm trying to control an N-Channel mosfet in a 'high-side' arrangement using an isolated power source such as a 9V battery. So my main supply is about 80V DC, then that passes though 2 mosfets which I want to independently control, and perhaps later an H-bridge. I read about the many options for driving the high-side mosfet using charge pumps or driver chips but also a battery solution which I like but can't find any examples of that. Does anyone have an example ?
Thanks!
 
Using a battery is the most impractical "solution" around. However, as you asked, there are some circuits shown here: **broken link removed** that describe some methods for turning on the MOSFET. The voltage marked VGH is where you'd connect your 9V battery positive and the battery negative connected to your 80V supply. They are the first suitable drivers that came up in a search - you can likely find better ones if you look harder.
 
I don't see why a battery can't be used like this:

switch.png

R1 is to turn off the mosfet. Since you probably will not be switching quickly and often, I would use a value that did not drain the battery too much. The larger the resistor, the slower the switch off will be. I suggest something between 1K and 10K. R2 is a gate resistor. Its purpose is to inhibit gate ringing. In this case, you can probably eliminate it. If you include it, use a low value, such as 10 to 100 Ω.

You may have a problem with switch bounce, which means each time you turn on or off the mosfet, it will actually be transitioned several times. That could cause unnecessary heating. I would try it without debouncing first. Then add a suitable RC debouncer, if needed.

As a safety note, remember the negative terminal of the battery will be at approximately 80V relative to ground when the mosfet is on.

John
 
Thanks John, this is exactly what I was thinking of, it seems very simple and inexpensive.

doughy83 suggests this is a most impractical solution, but isn't it practical if you don't want to buy a bunch of components or fancy $5 IC's that will do the same thing? Or is there an even cheaper/easier solution than using a battery?

Thanks!
 
Both designs still have a mechanical switch, which is the most unreliable component.

Your original question mentioned an isolated supply, such as a battery. You can get isolation with a transformer, thus you will find transformer-based gate drives. They are not simpler.

The usual way with N-channel mosfets is to connect them between the load and ground (i.e., low side). Then you won't need an isolated supply. Something as simple as a voltage divider with a pair of resistors from your 80V source might work. More often you will see something a little more elegant with a zener to give you the 10V or so that you need for gate.

What are you making? Is this to drive a DC motor or other inductive load? Will it be mostly on or mostly off? If it is mostly on, for example, you could use the switch to turn off the mosfet, rather than turn it on. That could reduce your battery drain.

John
 
I'm taking the DC output from a welder which is about 80V, and attempting to put it though an H-bridge of power mosfets to create square wave AC at frequencies from 100-400Hz and with PWM. The bridge is controlled by a PIC. The welder output is constant current so the 80v will vary widely down to perhaps 20v, currents as high as 140A. To control the high side mosfets as simple and cheap as possible is my first problem... An isolated battery in a circuit like John suggests above seems to be very simple design. I will not be using a mechanical switch, it would be a TTL mosfet. The battery solution also provides quick switching for a good square wave, some other solutions introduce delays. Is the battery the best way? I'm not concerned with the battery going dead etc...I can easily charge it up.

RN
 
I would not use a 9V battery in that design. In your first post, I thought you were just switching the mosfet on and off occasionally, maybe to reverse a motor. You didn't say that, it was just my assumption in answer to whether a battery can be used. I don't believe a small, 9V battery will last very long in the application you now describe, since charging the gate of a large mosfet(s) at 400 Hz and, say, 50% duty cycle will take quite a bit of current.

I recommend a dedicated gate driver that will drive both high and low-side mosfets. Several are probably still available. They provide dead time, so both high and low mosfets are not on at the same time. The LT1158 and IR2110 are examples (https://pdf1.alldatasheet.com/datasheet-pdf/view/70323/LINER/LT1158.html ; https://pdf1.alldatasheet.com/datasheet-pdf/view/82793/IRF/IR2110.html ). Many of those older, complete gate drives are becoming obsolete, as the same functions can be done with a microcontroller. Individual gate drives that can be controlled from an MCU are not that expensive and are readily available.

John
 
Thanks again John. If I go the battery route i'd probably use a 11.1V Li-poly battery from an RC vehicle, i think it was 2200mAh or more, which should last as long as I need it for. But it might be a temporary test only with the battery until I can properly design it with gate drivers like you mention.

The LT1158 is only available at about $8 a piece and does only about 50V on the high side, and $5 a piece for the 2110 which does up to 500V...this would work for me but i'd rather an even cheaper solution. Do you have examples of the individual gate drives? I'd setup the MCU to control dead time etc... I found the Si9910 (https://www.electro-tech-online.com/custompdfs/2013/04/70009.pdf) driver. This driver it seems for high-side application still needs to be supplied with a floating power supply (bootstrap or charge pump) from what i'm reading in the datasheet. Is it correct to say that whatever method i use to control the high-side mosfets will require an isolated supply, and that supply will be either a charge pump, bootstrap or a battery? (in which case I need to start reading about charge pumps and boot straps I guess...)

Thanks!
 
As for gate drives, go to whatever distributor you use and do a parametric search. International Rectifier (IR) used to offer all sorts of high-gate drives.

If you will give the part number for the mosfets you plan to use, number used (if parallel), and duty cycle, I will give you a first order estimate of how much current capacity you will need at 400 Hz per hour. That is, the current capacity needed to charge the gates repetitively at 400 Hz for 1 hr.

John
 
Here's a simple high-side driver using a bootstrap capacitor rather than a battery. The simulated turn on characteristics are a 300ns delay and 320ns rise time (to 200V - will be proportionally less for 80V). Turn off prop delay of 50ns and fall time of around 22ns.
 

Attachments

  • high side driver scr pnp.png
    high side driver scr pnp.png
    5.3 KB · Views: 1,172
  • high side driver scr pnp.asc
    2.4 KB · Views: 177
Hi John, that's very kind of you. I'm thinking to use mosfet AOT410L (https://www.electro-tech-online.com/custompdfs/2013/04/AOT410L.pdf) because of it's 100V-Vds, price, and 150A continuous current capacity. The total gate charge is 129nC max which is what I assume you might use to calculate the current needed? I'm not an engineer so it's a little beyond my scope! I also like dougy83's solution below, but it seems a little more complex, it will be something to consider as the project progresses. 1 hour would be much longer than I would need, perhaps more like 1/2 hour of use total, but i guess the relationship is linear.

Thanks!
 
John, hang on a sec, looking over the datasheet this AOT410L will not work...after looking at the SOA I could only get about 5A at 80Vds. The devil is in the details... Do you know of a cheap mosfet that would handle 80Vds with 150A continous or am I dreaming? Might I have to double up?

Thanks!
 
The total gate charge is 129nC max which is what I assume you might use to calculate the current needed?
Coulombs (C) are in the units of Amp.seconds, so if you multiply 129nA.s by 400 Hz (1/s) you get 52uA. If the 9V battery has a capacity of 400mA.Hr, then it will last for around 11 months, if it were just charging and discharging the Gate capacitance. There will be some other losses in the driver which will reduce this, probably by a fair amount; though it will definitely be in excess of 1hr by a few orders of magnitude.

I also like dougy83's solution below, but it seems a little more complex, it will be something to consider as the project progresses.
It can be simplified, though at the moment it has pretty good performance and will take up much less space than your 9V battery and will be cheaper.

John, hang on a sec, looking over the datasheet this AOT410L will not work...after looking at the SOA I could only get about 5A at 80Vds. The devil is in the details... Do you know of a cheap mosfet that would handle 80Vds with 150A continous or am I dreaming? Might I have to double up?
Figure 9 is for a MOSFET that is operating in its linear region. From what I understand, you want to drive your MOSFET such that it is either fully on or fully off, right? In this case you should be looking at the details in figures 1 & 2. You are probably dreaming if you want to get 150A through it continuously, as you'd have to dissipate ~170W of power. This would need a large heatsink and fan. If you double up, each MOSFET dissipates a quarter of the original power, so you could get away with half the original heatsink.

You should realise, if you're switching at 400Hz, then the MOSFETs aren't running continuously.
 
Thanks again John and dougy83, yes it will be fully on and off. I'll take in all this info and spend some time doing mockups in my basement for a while with mosfets and batteries, let you guys get back to work lol...it'll probably take a few weeks. I'll update the thread with my results in case anyones interested!
Thanks!
 
Guys, quick question, is there an IC that can be used to control an H-Bridge (not the H-Bridge it'self) which also includes a frequency control and dead time all in one package?

Thanks!
 
Intermediate Design

OK guys, no problem with the last question, I'm going to use a PIC10F222 to do the entire job. In the mean time here's my schematic so far, first time using spice so bear with me. I know, it has 4 batteries right! (one for the pic not seen) For me my main goal is to keep this thing simple and cheap. + if I isolate everything as much as possible, making it modular you could say, it'll be easy to upgrade/replace parts.

If you have any advice, to make it more efficient, better opto-coupler, mosfet, circuit configuration improvement (but not more complex), give up, etc... please let me know. In the mean time i'm trying to get LTSpice to interface with some board design software like DIPTRACE or ExpressPCB or something...learning curve.

Thanks!
 

Attachments

  • Draft6.asc
    6.7 KB · Views: 186
  • Schematic.png
    Schematic.png
    52.3 KB · Views: 434
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top