FET gate driver resistor value?

[madlabs]

Hi All,

I've been working on my buck converter, and have a mock-up functioning. Before I solder up a proto version, I need a little help with figuring out the correct value series resistor for the gate drive. And I'm simply DYING to solder it up and see what I can do with it.

I will be using a IRFP250, which has a Rdson of .085 Ohms. I am supplying the LM5104 with 12V for VDD. The input voltage to the buck will vary from 16 to 14VDC, depending on the current MPP of my solar panels. I am PWM'ing the LM5104 at 23kHz.

Right now, I'm using a 2.2K resistor driving the gate of a RFP10N12 FET, simply because it will plug into a breadbaord, unlike the IRFP250.

I'm tickled that it is working at all on a breadboard.

Thanks!

Jonathan

[Flux]

As far as I can remember the LM5104 is in fact a true mosfet driver. A typical gate resistor would be 22 ohms and there is probably little point in going lower than that unless you can look at dv/dt. I certainly wouldn't go over 47 ohms. The ideal world may use a diode to have different values for turn on and turn off but I tend to use 22 ohms and forget it.

Flux

[madlabs]

Thanks flux, I'll use some 27 Ohms that I have laying around.

However, I just ran into another problem. I put a heat sink on the two fets I am using, and they got much hotter than without and smoked a LM5104. Is increased gate capacitance the issue here?

[ghurd]

Why an IRFP250?  To me, 0.085 ohms seems pretty high for an MPPT circuit.

The IR data sheet shows Rg=25 ohms (note #2).  Always a good hint.

G-

[Flux]

Can't really help much without knowing far more of what you are doing.

I have a feeling that LM5104 is a high and low side driver, are you using a low side mosfet or just a schottky diode?. If you are using high and low side with high gate resistors you may have shoot through with both fets on at once.

It certainly will be nothing to do with heatsinks unless they are connected directly to the tab and causing extra stray inductance somewhere. If this is a breadboard layout with large inductive loops then you will have lots of fry ups.

I also assume that you have an inductor, if you are trying pwm without an inductor like the series charge controllers then it's not going to work.

Flux

[madlabs]

I'm using an IRFP250 because I have some. This is a testbed for me to learn with, not a final product or anything. I looked in the IRFP250 data sheet, and I don't see Rg or note #2. Nor do I know what that means. I'm just learning about such things, and this is my first attempt.

Thanks for the help! I need it!

Jonathan

[madlabs]

Flux,

I am using a fet on the low side. I do have an inductor. I am using the basic schematic included in the data sheet.

The heatsinkks were directly conneceted to the tab. I take it I should have used a mica insulator?

Thanks for all the help folks!

[Flux]

The IRFP250 is a fairly high voltage device and will have a high Rds on value the low voltage ones are much lower.

Rg would be gate resistor. It depends where you get the data sheet, various manufactures sheets differ somewhat. It is not common for manufactures to specify things like gate resistors as they are circuit dependent. They sometimes give values in typical application circuits.

For a low voltage converter you would be better off with lower voltage mosfets with far lower Rds on, but you can use these for testing although the lower device may be little better than a schottky diode as far as volt drop goes if you drive it hard.

Flux

[Flux]

The high side heatsink is no problem. The low side one is driven up and down at your switching frequency and will have lots of capacitance to things it shouldn't. I think you would be wise to insulate it from the mosfet and ground it to keep it from interfering with neighbouring parts of the circuit. Normally with low Rds on fets and decent driving you will need little heatsinking unless you are looking for large currents.

Flux

[madlabs]

Flux,

First off, many thanks for the help! I appreciate you taking time out of your life to help some random noob!

I will order some more appropriate FETs. Any reccomendations? I will be working with input voltages from 12 or 24V panels. I know that I am better off making multiple lower power devices rather than trying to make one larger one. So lets say max amps of 30 -50?

I had the RFP250's laying around from a HV flyback driver project, for which they worked very well. I'm trying to get a handle on what I really need before I go ordering stuff shotgun style, I'm too poor for that.

Thanks again everyone, I'm having great fun and learning a lot, even if I am allowing the magic smoke to escape. And I burned the crap out of my thumb. Still, it's cheaper than a night out!

[ghurd]

I used the International Rectifier 170kb pdf here,

(note #2 is on the bottom of page #996)

http://www.datasheetcatalog.com/datasheets_pdf/I/R/F/P/IRFP250.shtml

Smoke is fun, if it is not overly expensive.

I like common IRFZ44 or IRFZ42.  Cost $2, or less.  Rdson=0.022~0.012 ohms.

I think you will learn more, the circuit will work better, and you will have more fun if the results are not skewed by using parts too far from the "final product's" parts.

I will always defer to Mr Flux on anything MPPT related. And blades, magnets, turns, etc.

G-

[madlabs]

Ghurd,

Thanks for the link and advice. The IRFZ parts look perfect. I have been using MUR120 diodes, any better reccomends while I am getting a parts order together?

And yes, I realize it would be better to be using the right parts. But I needed to play... And all I smoked so far is one sample chip, so I can live with that. I just wish I knew for sure why it smoked.

BTW, any better choices for the fet driver rather than the LM5104? I'd rather a DIP package if possible.

[Flux]

Sadly the world has moved into the high tech era and little future development will be in real components that can be handled. The days of one off production is over, the average person doesn't do this any more and manufacturers are concentrating on the millions plus market.

I though you were rather adventurous to be playing with that chip. There may be equivalent DIP about but most of the early stuff was lower voltage. I know of several up to 60v but I didn't find much higher.

I have reached the age where I can't see the surface mount things let alone deal with them and prototyping one off things these days is a nightmare. I have stuck to standard high side drivers rather than a dedicated chip but for synchronous buck you are struggling without the clever logic and matched components of a dedicated chip.

These days the trend is to polyphase buck converters with interleaved operation and this eases the ripple problem immensely but it really does need these hi tech chips to do it in a reasonably practical way. These things are developed for the computer industry and are essentially low voltage.

Flux

[madlabs]

Flux,

How true! The move towards SMT packages is distressing for the DIY guy. Playing with the LM5104 has been a pain in the a$$. I don't really need the high voltage rating of the chip, so if you could reccomend some lower voltage chips in DIP, I'd play with those. I know nothing about driver chips, and the LM5104 was used simply because I saw a schematic using it and decided to use it as a starting point.

Do you have a schematic for a standard high side driver, including component values? I'm not stuck on using the LM5104. While I would like to go sychronous buck for the obvious reasons, I'd go high side only and take the efficiency loss in order to have all DIP components.

I think I'm at a dead end with this chip until I get some better FETs anyway. It keeps getting hot when I try to pull more than a couple amps from the circuit, and I don't understand why.

Thanks again for all the help!

[TomW]

Flux;

I feel your pain. I had similar "withdrawals" when I was relatively fresh out of learning electronics on tubes and then they switched to solid state almost overnite at least on consumer gear. Surrendering my he man 500 watt soldering gun for a wimpy 40 or 50 watt soldering pen just seemed so wrong. Don't see much wire wrapping these days, either.

Now I don't do much circuit building because I haven't kept up with the advances under the hood of most devices now. The stuff is microscopic, too!

I do enjoy reading the posts on it, tho.

Tom