Planning to build this mosfet H-bridge motor control

[Triode]

I'm planning on trying out this design (By Eugene Blanchard)

**broken link removed**

I have most of the parts, but I'm going to have to order the transistors and the diodes. Its all specified except the mosfets. The first one I build will be wired up to control lego motors. The best site Ive found so far for their ratings is here. At 9 volts the maximum stalled current of the motor types I'm going to use will be 700 mA.

Since I'm not real experienced with this I want to check that I'm choosing good mosfets, I'm pretty clear on the rating, they say what current and voltage they can handle. But I've found with some the mosfets, particularly P type, the RDS (on) is fairly high for the low current ones. Now intuitively one assumes that you want the smallest that will work for energy efficiency, that would be the case with some relays, but if I get a to-220 p channel mosfet that handles 1 amp, the best I can find is about 2 ohms RDS, If I go up to 13 amps, I can find one thats just 0.009 ohms RDS, and it isnt even that expensive, 91 cents (heres a transistor of that description, datasheet). So from what I've learned so far, this would be a good choice, its go a wide factor of safety over the 700ma max im going to draw, and it should eat up a lot less power as heat because of its low RDS. While I probably misunderstand something here, I can learn from you guys telling me what I got wrong. I'll come back and put the datasheets for the specific P and N type mosfets I plan to use once I get them picked out, for now I want to make sure I've got the theory right.

Heres the ones I've picked out at the moment:
FQP8P10 100V P-Channel QFET (mouser page) (Data sheet)
FDP55N06 SINGLE N-CH 150V ULTRAFET TRENCH (mouser page) (Data sheet)

as far as I know they should work fine, and be fairly efficient, but even though the $8 it'll cost for the amount I'm going to order isnt much, it would get annoying if I have to pay that much and wait repeatedly to get it right, so I just want to make sure.

 

[ccurtis]

There is nothing to prevent the "not allowed" state from frying the FETs if the power supply doesn't fold first. Be careful about that.

 

[Boncuk]

I guess you're better off using an integrated H-bridge.

Four power mosfets with all circuitry around them require quite some space.

The L298 is a dual full H-bridge capable of 46V at 4A max. It doesn't care for the "not allowed" status, moreover it takes care of fast motor stop if inputs A and B are equal, no matter if they are high or low. An additional option is free running motor stop using the ENABLE input.

If you wire both halves of the chip you'll also have a complete backup driver circuit.

The L298 is manufactured in a Multiwatt15 and PowerSO20 package, both occupying 1/5 of the area of discretes.

Boncuk

 

[Triode]

This is why I'm getting so confused, I already bought an L298, and someone told me "Dont use that on those motors, it'll burn up, better build a mosfet bridge" But you know, since what I know of the ratings tells me that the L298 can handle this easily, and a few people here agree, I'm going to try it despite that guys advice despite the fact that he should know what hes talking about. I'm refering to my cousin who's a technician for testing missile guidance systems, just so no one here thinks I mean them and gets offended. If I can use the L298 it would make the board very small and cheap, and it would work nicely since I'm planning on controlling two motors. The L298 can do variable speed right? with a pwm input?

By the way thanks for the input, both of you.

 

[Triode]

Looking at the L298n again (datasheet) specifically the bidirectional motor control on page 6. I am not sure what rating of resistor to use for the sensing resistor, it says only that it depends on the load and that it cant be wire wound. Does anyone see where it says how to determine what resistance is needed, or can anyone explain it? And do i need a specialized type of resistor, or just one that isnt wire wound?

I see on another diagram it says, 0.5 ohm, I'm not sure if it would be different for the other schematic.

if found **broken link removed** that explains it some, i guess it is something specific that I need to order. I didnt have any resistors less that 10ohms anyway

 

[Boncuk]

Hi Triode,

according to the ST Micro data sheet (electrical characteristics) the voltage at RSA and RSB should be nominal at 1V, max 2V.

Use Ohm's law to determine the voltage drop across the current sensing resistor at rated motor full power.

E.g. motor full current = 2.5A and you want a maximum voltage drop of 1.0V the current sensing resistor should have a value of 0.4Ω rated 2.5W.

2.5W and higher power resistors are not available as carbon (the worst choice since resistance increases considerably with in increasing temperature) or metal film resistors. Using a 5W wire wound resistor will be OK if nothing else is available.

Applying PWM special care must be taken not to have both inputs at the same level simultaneously between pulses since this condition will result in fast motor stop between pulses with Ven high. This can be avoided taking Ven low during pulse pauses (free running motor stop).

Boncuk

 

[Triode]

So, since it says nominal, should I use the stall current? because that wouldn't be "nominal" it would be max. Ideally this would work with a range of motors as well. The largest one has a stalled current of 700mA at 9v and under a normal load draws 320 mA. The smallest drawing 360 mA stalled and 120 mA normally loaded also at 9v.

So, guessing that I should choose a resistor based on the requirements of the larger of the motors I will be using, and use the normal current (320 mA) I would need a 3.215 ohm 0.32 amp resistor, or as close as possible. But if I should use the stall current of the larger motor (700 mA) I need a 1.4 ohm 0.7 A resistor. I'm not really sure what I should base it on.

About wire wound resistors though, i think I was told that as they heat up their resistance changes. Cant remember if it goes up or down, but that's why I was told they shouldn't be used for sensing.

Thanks for the help!

 

[Boncuk]

About wire wound resistors though, i think I was told that as they heat up their resistance changes. Cant remember if it goes up or down, but that's why I was told they shouldn't be used for sensing.

This problem can be overcome using a resistor capable of double power.

 

[Triode]

Does that just mean a resistor with double the power capacity of what is called for, or is that a feature, I did google it before asking, and it didnt come up as meaning anything but double the capacity.

So since it is a sensisng resistor, I'm thinking that I want one in the median range of what the motor normally does. I mean it shouldent be made for the stall current if, as I understand it, its purpos is to limit the power if stall occurs. Would it be right to pick a resistor based on the ratings of the average of the normally loaded current of my motors?

Thanks again.

 

[Sceadwian]

You should value your resistor for the maximum stall current plus a little bit of headroom just to be safe. A single bad overheating event can permanently change the resistance, although I'm not sure about wirewound resistors. Be aware of the inductance of the resistor if you're PWM'ing your motor.

 

[Triode]

Thanks, which rating I should go with is what I was wondering about. So I should go with the stall current of the larger motor and then round up a bit? Will it be an issue at all that it is about 2x the stall current of the smallest motor i will use this circut with, and almost 4x the nominal current of the smaller motor? I know that with a resistor for most purposes exceeding what you need is a good idea, I just thought that since a sense resistor is used to measure current draw it might be more about exact ratings than the meets-and-exceeds ratings as you normally want.

 

[Sceadwian]

Nope I prefer a lot of overkill, even in sense resistors. The better the power handling the less heat will be an issue, the resistance of the sense resistor itself should be as low as possible. Using a very low sense resistor voltage and an opamp to step that up to a useable sense voltage would reduce heating worries as you could use incredible small resistances for the sense resistor which wouldn't be effected by current draw. But it also adds complexity to the circuit.

 

[Triode]

Oh, you mean overkill as in how much current it can handle, I understand that. But as for determining the resistance, I'm still unsure which motor rating I should base that on.

 

[Sceadwian]

It should be rated for the highest stall current you'll experience.

 

[Boncuk]

Does that just mean a resistor with double the power capacity of what is called for, or is that a feature, I did google it before asking, and it didnt come up as meaning anything but double the capacity.

So since it is a sensisng resistor, I'm thinking that I want one in the median range of what the motor normally does. I mean it shouldent be made for the stall current if, as I understand it, its purpos is to limit the power if stall occurs. Would it be right to pick a resistor based on the ratings of the average of the normally loaded current of my motors?

Thanks again.

The current sensing resistor is not meant for chip protection of the L298. It is meant to control the motor power using your control circuit. You might as well omit it and connect the emitters (inside the chip) directly to ground. If you approach the current limit of the L298 it is advisable to use a current sensing resistor to prevent overheating and possible destruction of the device.

For least losses I suggest to use a resistor causing ≥500mV voltage drop.

It won't prevent motor stall if you don't use a motor control circuit which uses the sensed voltage drop to cut back on power or interrupt power completely.

In my example the resistor should have a power rating of 2.5W. Using a 5W resistor with the same resistance means double the power rating.

Using the highest motor current you are planning for, 700mA current flow, the resistor should have a value of 0.714Ω (for 500mV drop). The resistor has to dissipate 357mW in that particular case. Using a 2W wire wound resistor (5.6 times the dissipated power) it won't get warm and change resistance. (A 1W carbon film resistor will get warm and change resistance)

Boncuk

 

[Triode]

I haven't been able to experiment with this chip much because I was worried that if I didnt use the right resistor I might damage it. So with the highest draw on the largest motor being 700mA and the capacity of the chip being 4A it sounds like it should be ok to hook it up without a sensing resistor, for now atleast. Which would be pretty nice because I have all the other components I need. But I'll put some resistors that will work in my next parts order, at the least they will allow me to test what effect they have.