Driving MOSFETs
- Thread starter Haku
- Start date
Thank you all for your lengthy & detailed replies to this subject, I am slowly learning more about dealing with these devices in relation to high currents, I've previously used individual MOSFETs to control things like fan speeds & a string of 300 5050 sized LEDs and had no problems as they run straight from the Picaxe's output (with a resistor) and heat hasn't been a problem due to the relatively low current through them.
I knew this would be a whole new ball game and I always enjoy a challenge I know I can achieve even though I don't know the necessary steps so have to learn them as I go.
Today, with a Picaxe set to pulse user-triggered signals of 20ms to the NPN+PNP transistor setup in an earlier post, along with a freshly setup array of 10 MOSFETs in parallel, capacitor charged to 20v, I achieved this:
The welds are strong enough for me to easily pick up the screwdriver by the welded on tab
But they aren't strong enough for the final purpose of making a battery pack, so along with the 2nd capacitor in parallel, a proper MOSFET driver which I'll be ordering tonight, thicker copper cables, and perhaps higher ampage MOSFETs at some point, I'm sure I'll be able to get the welds I want.
One very important thing I've been doing throughout is wearing eye protection, especially important when this happens:
Omlette & eggs
edit: we're gonna need a better bank of MOSFETs... I upgraded the cables to some decent 6mm diameter copper ones and in so doing, 2 of the 10 MOSFETs blew into pieces with just a 16v 33ms pulse, and one or more of the rest are dead. Testing is over for today. Rats.
I knew this would be a whole new ball game and I always enjoy a challenge I know I can achieve even though I don't know the necessary steps so have to learn them as I go.
Today, with a Picaxe set to pulse user-triggered signals of 20ms to the NPN+PNP transistor setup in an earlier post, along with a freshly setup array of 10 MOSFETs in parallel, capacitor charged to 20v, I achieved this:
The welds are strong enough for me to easily pick up the screwdriver by the welded on tab
But they aren't strong enough for the final purpose of making a battery pack, so along with the 2nd capacitor in parallel, a proper MOSFET driver which I'll be ordering tonight, thicker copper cables, and perhaps higher ampage MOSFETs at some point, I'm sure I'll be able to get the welds I want.
One very important thing I've been doing throughout is wearing eye protection, especially important when this happens:
Omlette & eggs
edit: we're gonna need a better bank of MOSFETs... I upgraded the cables to some decent 6mm diameter copper ones and in so doing, 2 of the 10 MOSFETs blew into pieces with just a 16v 33ms pulse, and one or more of the rest are dead. Testing is over for today. Rats.
Last edited:
I now have a TC4452 MOSFET driver (pdf datasheet) and 10x IRFP2907's (pdf datasheet), not a cheap combination but I believe will do the job as several people use the IRFP2907 for their capacitive discharge spot welders.
One thing that puzzles me is that people are using the IRFP2907's in their CD spot welders to control the output of the positive connection of the large capacitor, but I thought the IRFP2907 is an n-channel MOSFET which is suited to GND switching, what's up with that? is it to do with how basically the MOSFETs short circuit the capacitor? with the by-product of this shorting being the heating up of the part with the most resistance (the bit you're welding).
Take a look at the bottom of the pdf schematic at the bottom of this page: http://ultrakeet.com.au/index.php?id=article&name=cdWelder_p2
One thing that puzzles me is that people are using the IRFP2907's in their CD spot welders to control the output of the positive connection of the large capacitor, but I thought the IRFP2907 is an n-channel MOSFET which is suited to GND switching, what's up with that? is it to do with how basically the MOSFETs short circuit the capacitor? with the by-product of this shorting being the heating up of the part with the most resistance (the bit you're welding).
Take a look at the bottom of the pdf schematic at the bottom of this page: http://ultrakeet.com.au/index.php?id=article&name=cdWelder_p2
Last edited:
An N-Channel Mosfet can be used for high side switching as long as the gate to source trigger voltage is within
specification. This is where a high side driver comes in.
In the schematic, Q7 conducts through R20 putting a voltage > Vcap at the gate of Q6 turning it on. High side
driving is very common in many applications, we just don't see it much in hobby circuits due to complexity of
the driving scheme. Note that there is a separate 18v supply for Vdrive to turn on the Mosfet operating at
15v. This circuit would operate more efficiently if Vdrive was a bit higher, but it may be designed to keep
dv/dt ( rate of change) down a bit.
specification. This is where a high side driver comes in.
In the schematic, Q7 conducts through R20 putting a voltage > Vcap at the gate of Q6 turning it on. High side
driving is very common in many applications, we just don't see it much in hobby circuits due to complexity of
the driving scheme. Note that there is a separate 18v supply for Vdrive to turn on the Mosfet operating at
15v. This circuit would operate more efficiently if Vdrive was a bit higher, but it may be designed to keep
dv/dt ( rate of change) down a bit.
Last edited:
Thanks for the explanation, I just did a test and it appears using the IRFP2907 as a high side switcher needs 4v higher than the load it's switching, looks like I'll be needing a 20v PSU as one of the two power capacitors I have is only rated at 16v max, and I'll copy the charge/discharge capability to set the capacitor charge voltage.
You can do that obviously, but there are several other methods.
For an 'economy' motor bridge design I did recently I used one of the IR family drivers plus Ns all round.
A very common circuit design.
This uses a charge-pump and bootstrap method to derive the extra high-side gate voltage.
The advantage is that the high-side gate supply 'floats' so that the charge-pump generation is always X volts higher.
This allows you to drive a device (with caution wrt design, layout and handling) of hundreds of volts.
Compared to your 'average' driver it requires 2 or 3 extra passives (20p), no big deal really.
Sadly, it requires some extra calcs. and head-scratching for the designer - if you want something good, that is.
There is one fundamental limitation of the basic design, but I'll allow the student to ponder on that...
It's not difficult to get around, including a interrupt driven charge-pump or even the wonderful 555.
One of the main reasons for using this method is cost. Industrial designers are obliged in many case to get the design costs right down. There are other methods for doing this (including high-side P) but they tend to be a few pennies more expensive.
The very same driver can , by the way, drive P-chans with a little bit of twiddling.
Much of the decision is down to budget and knowledge and application and many hobby/amateur designers simply copy other people's designs so a 'universal' circuit is born .
For an 'economy' motor bridge design I did recently I used one of the IR family drivers plus Ns all round.
A very common circuit design.
This uses a charge-pump and bootstrap method to derive the extra high-side gate voltage.
The advantage is that the high-side gate supply 'floats' so that the charge-pump generation is always X volts higher.
This allows you to drive a device (with caution wrt design, layout and handling) of hundreds of volts.
Compared to your 'average' driver it requires 2 or 3 extra passives (20p), no big deal really.
Sadly, it requires some extra calcs. and head-scratching for the designer - if you want something good, that is.
There is one fundamental limitation of the basic design, but I'll allow the student to ponder on that...
It's not difficult to get around, including a interrupt driven charge-pump or even the wonderful 555.
One of the main reasons for using this method is cost. Industrial designers are obliged in many case to get the design costs right down. There are other methods for doing this (including high-side P) but they tend to be a few pennies more expensive.
The very same driver can , by the way, drive P-chans with a little bit of twiddling.
Much of the decision is down to budget and knowledge and application and many hobby/amateur designers simply copy other people's designs so a 'universal' circuit is born
.