Go Down

Topic: High Powered Motor Design (Read 329 times) previous topic - next topic

laveur

Hey everyone,

I have this following design for controlling a 12V 4A motor (well really a fan but same thing):



The code is fairly simple, it starts off with the fan shut off. You press the button it comes on low, press again medium, press again its high. After that it shuts off again.

It started off working fine but lately when you turn it on it kills the high side driver of the IR2184 Driver IC.

I have a feeling it has to do with high in rush current. But I am not 100% sure it is.

Any thoughts on whats going on please let me know.

jackrae

Have you confirmed the possibility that Q2 isn't turning ON at the same time as Q1
Your code would be helpful in diagnosis

MorganS

That's a pretty fancy driver. I don't expect that it would ever turn on Q1 and Q2 at the same time.

However driving a fan in one direction only Q2 isn't doing anything. Except it's required by the bootstrap power supply for the VB pin. This is one of those devices that can't operate at 100% duty cycle as it needs PWM operation to pump up the capacitor on VB. The datasheet doesn't list a maximum or minimum duty cycle but I don't see how it generates that voltage inside the chip without oscillating.

For just driving a fan on and off and PWM control, there's much better chips that have everything integrated. Something like a BTS716G will drive up to 6 amps per channel and it detects short-circuits and other faults too. It's only available as a SMD part but it's big enough to solder by hand and you can buy one of my breakouts from OSH Park. (I get nothing from OSH Park if you use my design. You only pay for the PCB.)

There's no other components required: no diodes or capacitors, no pullups or anything. Plug it in backwards and it will just sit there waiting for you to correct your mistake, although it cannot prevent the reverse-battery flowing through to the load, so your fan may run backwards or explode if you try the experiment.
"The problem is in the code you didn't post."

Wawa

Not sure why you use a push/pull driver for your swamp-cooler fan.
That circuit could brake (slow down) the fan during the "off" parts of the PWM pulse.
Maybe ok for a motor, but not for a fan.

A single logic level mosfet should be able to PWM that 4Amp load low-side.
First diagram here.
Note that the IRF3205 is NOT logic level.
Leo..

laveur

First thanks for all the suggestions.

I would like to address some comments that Wawa made. Yes a Logic Level n-channel mosfet would work but it would not work well enough. 5V's on the Gate would indeed let me control the fan BUT I would convert a ton of power into heat. The Rgs when the gate voltage is 5v is higher than when its supplied with a 10+v signal on the gate. This directly equates to higher power dissipation and thus more heat. One of the requirements for this circuit was to be able to run in a place where the ambient temperature is near 100*f (37*c) during the day. If the ambient temperature is that high it leaves very little wiggle room for thermal dissipation even with a heatsink.

Secondly I'm gonna go ahead and answer my own question as to what happened.

TL;DR Version my bootstrap capacitors where not sized right.

Longer Version:

I found this article in my quest to sort this out: http://tahmidmc.blogspot.com/2013/01/using-high-low-side-driver-ir2110-with.html

Although its not the same gate driver its in the same IC family as the one I am using and there a pretty good pretty good facsimile of what I had going on. Reading over the notes it gave a pretty straight forward answer to another question I had. How do you size your bootstrap capacitors. After conferring with a video from GreatScott! on YouTube I noticed some similarities to the aforementioned blog post. At this point I made two changes to my boot strap circuits. I added a 47uF Electrolytic cap and a 470nF ceramic one in parallel. Ordered some new boards because bread boarding this would only cause bigger issues. Got them gave it a test run and it worked... sort of. The low and medium speeds worked fine but the high speed instantly killed the board. Same as before but the low speed was killing it. I had extra boards so I swapped the 470nF ones for 100nF ones and everything work.

Now for the why this change was needed. Something I had done a while back after the original board design was working was I changed the duty cycle of the PWM that controls the fan speed. Bootstrap capacitors are inherently tied to the duty cycle and frequency of the PWM signal. When I changed these it also changed the requirements for the bootstrap components. The 100nF allows it to provide the quick initial burst of power to quickly start the fan, while the larger one provides a larger reservoir to keep it running. The lower one also re-charges quicker when the PWM signal goes low then high again. Hopefully someone else will find this useful.

-Laveur

Wawa

Yes a Logic Level n-channel mosfet would work but it would not work well enough. 5V's on the Gate would indeed let me control the fan BUT I would convert a ton of power into heat. The Rgs when the gate voltage is 5v is higher than when its supplied with a 10+v signal on the gate.
This doen't make sense.
I think you mean Rds(on), and that can be quite low with a modern logic level mosfet.
A 4Amp fan (don't know stall current), PWM-ed with ~500Hz is not extreme.
I would try the simple way before venturing into gate drivers (with fan brake).
Leo..

MarkT

#6
Jul 28, 2018, 03:17 pm Last Edit: Jul 28, 2018, 03:22 pm by MarkT
The capacitors on the mosfet drivers are the wrong way round.  The bootstrap cap must be much
smaller than the decoupling cap or you'll crowbar the supply on every cycle.

33nF is probably fine for the bootstrap, 1 to 10uF for the decoupling cap, both must be ceramic.

If the MOSFET has a total gate charge of N nC, then a bootstrap cap of N nF will be sufficient
(it will only droop 1V).  33nF will handle most MOSFETs fine.  Its good practice to add about
10 ohms in series with the bootstrap diodes to reduce noise injection on the 12V supply rail.

The diodes across the output  connectors do nothing, the bridge provides full free-wheel diode compliment
as you use MOSFETs with body diodes.

If you have 12V MOSFET supply then you need to condition it before presenting it to the
MOSFET drivers - the main supply rail will be very noisy and MOSFET drivers need a pretty clean
12V.  (Instant catastrophic failure can occur due to spikes on the 12V, note, I've been there)

You need some heavy duty decoupling on the supply to the MOSFETs, lots of electrolytic and some
ceramic is good.  You have none.
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

Go Up