Is it okay to put 6 IRF540N's (130W max power output each) to music synchronize some lights that have a max (completely on) load of 70W? My mindset is that it would be okay, because the Vgs threshold variances are between 2V and 4V - and at those Vgs voltage levels, the power load would be small enough such that 2 or even just 1 of the 6 FETs can output the power by itself. I have 3 x 1 x .5 cm heatsinks on each in a cold room.
If this works, would parallel BJTs still be a better idea?

 

Although I saw some people in parallel with mosfet pin to pin directly, but I would like to in series with two resistor for each mosfet, whatever the n or p type all connected to the S pin, the method was the same with E of bjt.

Another thing what you concerned was to in parallel with the Vgs, about this point you may in series a resistor 10~500Ω with g pin and then in parallel, the values are according to the input Vgs and frequency.

So here you have some different methods to use --
1. Connected the pins pin to pin, but I don't like this way.
2. Using one resistor -- in series a 10~500Ω resistor with g and in parallel with input Vgs.
3. Using two resistors -- the method 2 and in series a 0.2Ω/5W resistor with s pin and in parallel to Ground for n type and in parallel to +V for p type.

How to choosing the method, sometimes are depend on the current and voltages.

You can google "in parallel with mosfet" to get some ideas and to try them.

 

I see high powered voltage controlled switching devices like power Mosfets and IGBT's paralleled all the time( quite often in large numbers in switching type power supplies) so apparently there is a some 'magical mystical' way of making them work together with all of their drains and sources tied to each other.

 

I can't make your circuit work as shown, so I replaced the cap with a resistor and it kind of works.
You can replace the transistor with the IRF540, but you will need to increase the gain of the 386 (5k to 10k, 10k to 510 ohms) so it can reach the threshold voltage. A logic level FET would be better. That way it is the FET is just a switch so not to worry about how to parallel them. It will need a small heat sink to supply 10 amps.

 

Just a thought, if you want your leds to dim according to the music. Would that be more efficient to convert the (music) voltage to a PWM signal and switch the LEDs? I would be totally wrong tho.

 

Great idea! Then the brightness would change with the volume of the music instead of just blink with the beat.

 

Is it okay to put 6 IRF540N's (130W max power output each) to music synchronize some lights that have a max (completely on) load of 70W?

I think you've misunderstood the FET spec. The 130W rating refers to the maximum power the FET itself can dissipate as heat, and does not refer to the power dissipated in the load.
For example, let's assume the load is a 12V 60W lamp. That means the lamp draws 60/12 = 5Amps. Now assume the FET when switched fully on has a resistance Rds(on) of 0.077 Ohms (the max value as per the datasheet). The voltage Vds developed by 5A flowing through 0.077 Ohms will be 5 x 0.077 = 0.385V and the steady-state power dissipated in the FET will be 0.385V x 5A = 1.925W. So a single IRF540 (with a small heat sink) can easily handle your 70W load.

 

So I am still wondering if parallel FET is better (assuming the use of the best configuration) or parallel BJT is better (assuming the use of the best configuration)??

ScottWang: Okay thanks!!! Now I know my parallel FET options and will google which is the best for me.

Tcmtech: Yes because those transistors are used as switching devices which is much easier to parallel than when they're used as amplifying devices, because Vgs threshold variances (which is my problem) do not affect switches. Since I'm making an amplifier, that does not apply to me, sadly

Ronv: I don't have a cap in my circuit. Also, I cant change my configuration from an amplifier to a switch, because I want my light brightness to dim/bright in sync with my music, not just be on/off. Btw, I don't need a pwm to vary my led brightness - that happens already.

Bug13: sorry but no. My LED brightness already varies with the music - no need to implement a PWM.

Alec_t: So are you saying that one IRF540N can handle 70W because it can dissipate 128.075W and not 130W (at 5A)? If yes, then I think there are many much more significant factors that reduce the power capacity.
Or are you saying that, since the IRF540N can dissipate max 130W as heat and its Rds(on) is .077, then it can handle a max current of 130W/.077 = 1688.3A? If yes, I doubt you are correct.

 

About dissipation...

All of the specifications are limits in one application or another. If your mosfet was passing 41 amps with an Rds of .077 ohms, it would have 3.157 volts across it. That would cause 129.437 watts of heating in the mosfet. That heat is entirely separate from whatever else is using the 41 amps.

It is obvious that your mosfet can not pass 1688 amps because the wires would melt and the heat in the mosfet would be about 220,000 watts. The IRF540N is labeled for 33 amps if you can keep the case temperature down to 25 C. Not likely unless you include some kind of refrigeration circuit to keep it cool. That's why it is also specified for 23 amps if you can keep the case temperature down to 100 C.

So, if you told it to pass it's rated 23 amps through its rated .077 ohms, it would suffer 40.733 watts of heat. That is way below 130 watts, so we have to assume that the only way to get it to 130 watts is to pulse the current at high levels. This mosfet has a limit of 110 amps as a pulse. That would be 931.7 watts if it was DC, so you can see that some duty cycle less than 100% would stop at 130 watts. (A bit less than 13% "on" time if I did the math right.)

Still, the heat in the mosfet is entirely separate from the heat developed in the load. A 20 ohm load being slapped with 110 amps for 13% of the time would suffer 4089.8 watts. The point is that the load can have way more power applied to it than the mosfet that is in series with it.

 

Bug13: sorry but no. My LED brightness already varies with the music - no need to implement a PWM.

I understand your LED brightness already varies with music. The reason that I suggested using a PWM to drive your load is it has a lot better efficiency than your current design. (Less heat from the MOSFETs/transistor)

If you are driving the LEDs with PWM, you only need a single IRF540 with a small heat sink. (Ref to #27 by Alec_t and #29 by 12)

 

Your circuit may vary with the loudness of the music now, but it will only blink if you change to a FET. The current you can pass thru the transistor is a function of how much current you supply the base of the transistor. FETs are voltage controlled so as soon as the voltage goes a little above the threshold voltage they are all the way on - blink.

 

If you use the music to vary the pulse width in a PWM arrangement then the blinks will be too fast for the eye to notice but the perceived LED brighness will vary with the music and a single FET could cope with your 70W LED load.

 

#12: Oooooh I see. Wow, that was really helpful - thanks! Looks like I was wrong about a very fundamental concept.

bug13: Oh okay, sorry about that misunderstanding; I was thrown off by ronv's response. I still can't use a PWM though because then I would see the slow pulses when I just want the lights to be very dim. Otherwise, you would be right about it being more efficient.

ronv: Interesting.. you might be right, but I tried the FETs and it worked just fine - no blinking. I did not change my circuit either. Still using the 386 with the same gain.. I'm about to increase the gain though because of the high Vgs - I can tell you how it goes if you'd like.

Alec_t: True but then my minimum brightness level will have to increase by too much in order to maintain blinks too fast for eyes to see, unfortunately.

 

bug13: Oh okay, sorry about that misunderstanding; I was thrown off by ronv's response. I still can't use a PWM though because then I would see the slow pulses when I just want the lights to be very dim. Otherwise, you would be right about it being more efficient.

It's all about your PWM frequency. I don't think normal human eyes cannot see the slow pulses even it's very dim, if your PWM frequency is 100Hz or above. The brightness is control by the duty cycle, nothing to do with frequency.

Unless you see it through a video cam.

 

Ahh. Lots of people helped me with this IC to make it simple.
This shows the difference between PWM and what you are doing.
The blue trace is the output from the 386. The red - the way you have it and the green PWM.
As you can see some low amplitude notes don't turn on the LED and if it is high enough to turn it on.. well it just turns on. While with the PWM the lower amplitude notes appear dimmer and the loud ones brighter. The frequency of the PWM is high enough that you eye perceives it as average light.

 

Bug13: Hm that's interesting. Good point. I forgot about varying the duty cycle haha. Would I use a 555 to do the PWM? Either way.. How would I do it? I'd like to see what I can get.

Ronv: Ah, yes, I can see your model proves exactly what you have been saying. Perhaps I can link you a video.. my light brightnesses are varying just as well as they did with the BJT before (contradicting your model for some reason..). Anyway, I appreciate you pointing this out. I'll keep an eye out for FETs having this 'easy-to-just-blink' characteristic

 

If your happy, I'm happy.

 

NE555 Clock Generator and PWM adjustable circuit.

 

So I've just done a lot of reading on parallel FET configurations optimized for the linear region. However, I still need help with figuring out what value resistors to put in series with each of the gates. I think I've learned the important concepts, but now I just really need the actual number. (Or is this something I need to find with trial and error?). I've seen recommendations ranging from
.050 to 500 ohms and would like to pinpoint what would be the best for me. Again, I am paralleling 6 IRF540N's.

ScottWang: Thanks for the link, I will look into it! Btw, regarding your post on 9/30, I am trying method 3 and will experiment to see whether I can have a source resistor without making my LED strips noticeably dimmer.

 

So I've just done a lot of reading on parallel FET configurations optimized for the linear region. However, I still need help with figuring out what value resistors to put in series with each of the gates. I think I've learned the important concepts, but now I just really need the actual number. (Or is this something I need to find with trial and error?). I've seen recommendations ranging from
.050 to 500 ohms and would like to pinpoint what would be the best for me. Again, I am paralleling 6 IRF540N's.

ScottWang: Thanks for the link, I will look into it! Btw, regarding your post on 9/30, I am trying method 3 and will experiment to see whether I can have a source resistor without making my LED strips noticeably dimmer.

If you wish to use pwm to adjust duty cycle then the Rg can't be too high, there are two reasons will be affecting the frequency, one is the Cgs(internal capacitor), another is Rg, if you don't know what value of resistor is the best then you can use 100Ω to be a start and adjust it according to the frequency response of Vgs, you can using the O'scope to measure it and to see the waveform of rising and falling edge.