How To Amplify a Signal?

Hey

Possibly a stupid question but here it goes...

I am looking to amplify a voltage of about 4.8V to 48V

I have the 4.8V, I have +48V and obviously a ground.

If I use an op amp can i put Vp as the +48 and Vn as 0V then use an appropriate resistor ratio to amplify?

Any help or suggestions would be great. Thanks

Yes in principle you can, using a single supply op-amp setting the gain to 10. However even using a rail-to-rail amplifier you will not reach an output of 48V, but some 200 mV less.

I'm not sure you will find a rail-rail opamp that will run off 48V. Presumably you have a 0V to 4.8V input range? You say it is DC so presumably any changes will be very slow?

Keith

An OPA445 opamp has a max allowed supply of 80V and would amplify 4.8V to 48V easily if its supply is about 50V or more and its load current is very low.

The OPA445 opamp is made in lots of about 2000 every 9 months by Texas Instruments. They sell out in one week. Today, nobody has one for sale.

Audioguru said:

The OPA445 opamp is made in lots of about 2000 every 9 months by Texas Instruments. They sell out in one week. Today, nobody has one for sale.

Click to expand...

can you just clarify this,

The OPA445 is in stock at loads of places according to TI's site. I just checked Farnell and they have stock.

I don't think it really helps though - it's not rail-rail. I think a solution with discrete devices on the output of an opamp to add a high voltage output stage to a normal opamp may work.

Keith

Hey Guys

This is basically what I am trying attached as a quick JPEG please excuse the quality just pulled together quickly.

I basically want to make the msp430 output pin go high to the base of a MOSFET to act as a switch to allow the 48V through. I cannot just use a MOSFET straight on the line as the Vgs is too high for standard MOSFETS (biggest I found was Vgs plus or minus 30).

I know there wont be quite 48V on the ouput but something smaller. will this circuit work and would an op amp such as this one be useful:

https://uk.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=6610981#header

Thanks guys!

So you are just wanting to switch the 48V rather than progressively control it? What current?

Keith.

in a sense yeah.

I am using the 48v to power an IP camera. the power consumption of the camera is about 3 W max so a current of 62 mA would be needed I guess...

cheers

I have to go out, but I can draw something properly later, but all you need is something like this



or something similar with MOSFETs. You can easily get round the Vgs limit by adding an extra resistor to the MOSFET version. There are also neat dual packages with N+PMOS or NPN+PNP devices in them. Rohm may even do one with the resistors built in.

Anyway I must rush so will post later unless you have already sorted it out.

Keith.

Added after 1 hours 9 minutes:

This may do the job:

**broken link removed**

All you need to do is put a potential divider from the NMOS drain to the PMOS gate so the gate voltage doesn't exceed 20V, pretty much as I drew with the bipolar transistors, but R3 isn't needed with the MOSFETs. Otherwise you could use any number of discrete bipolar transistors or MOSFETs in the same configuration.

Keith.

To switch 60mA @ 48V on and off you can use a simple micro-relay with a 5V coil ..
The relay can be driven directly from 5V or through a transistor, depending on the coil current .. see attached picture ..

Rgds,
IanP

hey guys

Thanks for the input. I think I will go for keiths idea as I am not sure I could power a relay. I forgot to mention the device causing the switching is remote so current consumption is a big issue as it needs to stay switched for as long as possible.

Keith, so i can use a similar circuit to that in the diagram, miss out R3 as I am using mosfets and R1 to R2 is just a suitable pot divider relation so Vgs isnt over the 20. basically Vg of what you have called Q2 needs to be very minimum 24V for the 48V source. Is that right?

Cheers
Thomas

Yes, leave out R3, NMOS for Q1, PMOS for Q2. Your PMOS is likely to only need something like 5V to 10V to be fully turned on, so maybe use a resistor ratio such as 10k for R1, 27k for R2. That will give you around 13V maximum on Q2 gate. You can scale the resistors up to save power with the only risk being that they will take longer to turn on due to gate capacitance. During that switch on time the power dissipation in Q2 can be high. So, picking a transistor that is not too oversized will keep the capacitance low and the switching speed high, even with large resistors.

Keith.

Hey

As you may have guessed analog circuitry isn't my best area. Why would you use the PMOS and NMOS as in when would you know to use each one?

I have this N - Channel MOSFET

http://www.datasheetcatalog.org/datasheet/fairchild/HUF75337P3.pdf

Would this be suitable as a P for my application.

http://www.farnell.com/datasheets/46053.pdf

If my voltage at Q2 gate is 13V, and teh source is 48V then will Vgs not be greater than the maximum allowed of 20V ? Or is that not how Vgs works?

Don't use that NMOS one. It is nearly 1000 times bigger than you need! Also, the current will be controlled by Q2 which will need to be PMOS anyway.

NMOS and PMOS are simply being used as switches in your case. An NMOS will switch current to ground when you aplly a positive voltage to the gate (with respect to ground). A PMOS would normally be connected to a positive supply and will switch the load to the positive supply when you apply a voltage on its gate which is lower than the positive supply. I hope that makes sense. There is a threshold voltage with them both i.e. a voltage on the gate below which nothing will happen.

So, in your case you need an NMOS and PMOS (you could actually switch with 2 NMOS but then you would drop a few volts through the NMOS, so it isn't worth it).

So, I suggest you buy a small PMOS & NMOS, maybe 1A or 2A for the PMOS and anything tiny for the NMOS (it only switches a few milliamps at most). I had a quick look at Farnel and something like the 2SJ632, SI2309CDS, MCH3316, CPH3324 would do for the PMOS (all 1-2A, 60V).

Then something like 2N7002, PMBF170, TP0610K, VN10LFTA, ZVN3306FTA for the NMOS - all <1A 60V.

Keith.

Added after 3 minutes:

I forgot to look at the PMOS link you showed - that one looks a bit weedy - 140mA at 70C and you want >60mA. OK, but a bit more margin would be nice.

Keith.

Added after 3 minutes:

thaitam said:

If my voltage at Q2 gate is 13V, and teh source is 48V then will Vgs not be greater than the maximum allowed of 20V ? Or is that not how Vgs works?

Click to expand...

Be careful when working with PMOS & negative currents and voltages, but yes, that is correct.

You can often only put 20V on the gate of a MOSFET relative to the source, but the drain could be at 600V. The large MOSFET transistors are not symmetrical even though small devices usually are (i.e. drain & source are interchangeable on low voltage devices within an IC).

Keith.

Hey

I am going to use MCH3316 as a PMOS for Q2. How exactly is it connected as in where is the drain connected to and where is the source connected to? Is the source the 48V?
If I use R1 as 10K and R2 as 27K then this will give me a Vgs of about 13V for Q2 if you ignore voltage drop over Q1?

PMBF170 as an N channel. This will be connected so that 3.3V will be at the gate and 0V at the source? Thus the MOSFET will be on. ie Vgs > 2 for this MOSFET.

Cheers again
Thomas

Yes, for the PMOS, source to 48V, drain to the load and the gate to the resistors. R1 as 10K and R2 as 27K is fine and 13V Vgs as you say.

For NMOS, source to 0V, input to the gate and drain to R2.

Your transistor choices are OK, although the threshold of the NMOS could be 3.5V at -55C. As long as you are not working at extreme temperatures you should be OK. Otherwise something with a lower threshold voltage would be required.

Keith.

Hey

I ordered these MOSFETs however I wasn't aware they were so small :S

My bad! I don't really know much about this sort of stuff!

I thought they would be big enough to be hand soldered? I am struggling to find MOSFETs that I can solder myself and test on a protoboard.

How is best to use these MOSFETS if I cant get an alternative?

Any help?

Thanks
Thomas

You need to look for components in through hole packages if you find SOT23 types too small, although it is perfectly practical to hand solder them provided your soldering iron tip isn't too large. You can buy 0.05" pitch prototyping board for it as well. Mind you it is getting more difficult to find the newer devices in anything other than surface mount. With some types of components you have to use QFN packages.

Keith.

Don't look at the threshold voltage rating for a Mosfet because that is where it barely turns on (or barely turns off) with a current of only 0.25mA.

Instead, look at the gate voltage for its On-Resistance rating.