I have a project which uses a 5V microcontroller to produce a PWM control signal with a frequency of around 400Hz. The PWM output level however needs to be 8V. The whole thing is fed by 12V so both 5V and 8V supplies are easy to make (7805 / 7808 regs for example).
What do you think would be the best way to achieve the 8V output level? Here are a few solutions I've come up with:

1. Single npn / logic level n-Fet + pullup resistor to 8V
-Extremely simple and probably good enough for the low frequency. I expect the load to be fairly high-impedance (control input for car AC fan).

2. Mosfet driver IC connected to 8V
-Seems like a great idea since they go rail to rail. Is it OK to use the driver without any significant load on its output?

3. Rail to rail single supply opamp
-Could be fed directly from 12V but I would prefer not to since I don't want under any circumstances / failure mode to have over 8V on the output, which could damage the PWM input.

4. A different type of IC, level converter, logic gate, PWM controller or something?
-Any ideas?

 

400 Hz is slow, the plain-old NPN transistor trick should be fine, with the following caveats:

1) The PWM signal will be inverted

2) The edges will be about 200 ns delayed

 

Thanks for the reply!
The NPN would no doubt be the simplest way to do it. One thing that came to my mind though is Vce. But that would be minimal if I choose the pullup sensibly, right? (0.5-1.0 kOhm or something like that)

Would you see any problem in using a FET driver for this? (Apart from it being overkill...)

Also, afaik most voltage regs need a minimum load, although I failed to find this in the L7808 datasheet. That'd be important to know in this application.

 

Which Vce? The one where Vcc pops the transistor or the one where the transistor is conducting, VceSat?
VceSat depends on how much base current, how much collector current, and exactly which transistor part number.

Would you see any problem in using a FET driver for this? (Apart from it being overkill...)

A high current MOSFET gate driver for a high impedance load? Sure. Why use one transistor when you can use a chip with a dozen transistors in it?

voltage regs need a minimum load,

First you have to name that regulator exactly. Here is the datasheet for L7808. Not L7808C. Not LM7808
This one says Ir1=5Iq and Iq = 5ma to 8 ma depending on which model and voltage.

 

Would you see any problem in using a FET driver for this?

I believe some have an under-voltage lockout threshold greater than 8V.
I'm curious as to what demands an 8V PWM input?

 

I'll try to be more specific.
I was referring to VceSat. Seems to be 0.2V for a 2n3904 @ Ic = 10mA, which seems realistic in this application. 0.2V is probably low enough, although a n-Fet in place of the NPN would pull it a lot lower.

The voltage reg I was looking at is a L7808CV from SGS-THOMSON.

This one says Ir1=5Iq and Iq = 5ma to 8 ma depending on which model and voltage.

Did you get Ir1 from the schematic on page 17? Seems pretty high to me. Why can't they just clearly state it like with some other regs?
I was thinking of putting the 7808 and 7805 (which powers the uC) in series so that both would get atleast 10mA on the outputs.

 

Actually Io is frequently stated as 5mA-1A under "test conditions" in the ST micro datasheet so would it be safe to assume the minimum load current is 5mA?
Also, is it safe to assume the ST micro part is identical to the SGS-THOMSON part?

 

The 317LZ needs 10 ma load.

I was thinking of putting the 7808 and 7805 (which powers the uC) in series

That's a good thing.
You're not going to get much better than 0.2 Vce with a bipolar, but it's tough deciding when you don't know what you actually need!

Actually Io is frequently stated as 5mA-1A under "test conditions" in the datasheet so would it be safe to assume the minimum load current is 5mA?

No.

 

The PWM output will go to my car AC blower's PWM control input, which needs 8V and 400Hz. And yeah, I don't actually know what it detects as low level but my guess is >0.2V. The fet or fet driver might still be the safer choice.

I see the 317LZ has a clearly stated minimum Io but most others don't. So how are you supposed to know it exactly? Not that it would really matter in this case as I could easily waste 50mA by just loading it with a resistor, I'm just curious...

Edit. And apparently there are regulators with internal load resistors. They have high quiescent current (like the 7808).

 

What do you think would be the best way to achieve the 8V output level? Here are a few solutions I've come up with:

2. Mosfet driver IC connected to 8V
-Seems like a great idea since they go rail to rail. Is it OK to use the driver without any significant load on its output?

Yes, a MOSFET driver will be perfectly happy driving a light load, and that's the option I would choose; perhaps an MCP1407 or something similar. It will operate off supply voltages from 4.5V to 18V, accepts a logic-level input and gives a rail-to-rail output.

That would be my choice: simple and effective.

 

I think I actually have a single MCP1407 lying around somewhere. I used them to drive over 10 parallel fets in a welder some years ago.
Might go that route since its very simple and will achieve a nice PWM basically rail to rail.

The voltage regulator question remains...

 

So how are you supposed to know it exactly?

Spend hours deciphering datasheets.
Any datasheet that can't be deciphered is disqualified for obfuscation.

 

The voltage regulator question remains...

In a situation like this where you have 12V available and need 8V for your driver, but the 8V doesn't need to be super-precisely regulated, I'd be tempted to just put a 3.9V Zener diode (like a 1N4730 or equivalent) between the 12V supply and the Vdd pin of your driver and call it a day.

You might also need a resistor of a couple of kΩ from Vdd to ground to assure some minimum Zener current (and therefore a stable Vdd), as I don't think the MCP1407 draws much when it's not switching.

 

Not really an option in this case as the power source is a 12V car battery which can go close to 15V. Also, I think it is pretty important to precisely regulate the 8V since its going to be fed into the fan control unit of the car. I don't know how picky it's about voltage but I'd prefer not to test it.
Anyway, putting two regs on the PCB is simple enough.

 

Not really an option in this case as the power source is a 12V car battery which can go close to 15V.

Ah. I forgot about that, and you're right: my series-Zener approach would not be appropriate.

 

The 7800 series regulators don't minimum current requirement - that is reserved for the LM317/337 variable regulators.

 

Everybody has an opinion. And this is right, we are all entitled to that; as long as we respect other people's own opinions.

Having said this, my opinion is:
keep the circuit as simple and straightforward as possible. Go for the NPN solution

 

Everybody has an opinion.

That makes the hair on my neck stand up!
This is a scientific site. Sometimes we're dealing with hard facts.
Datasheets have a lot of information, and often what the numbers mean is, "equal to or less than".
Sometimes they mean, "This is the minimum the manufacturer guarantees".
Sometimes the answer is 42 and there is no room for opinions.

 

Everybody has an opinion.

And everybody has an asshole.

Everybody's lives will be much easier when you all realize I'm always right.