Much, MUCH better!!!

 

Much, MUCH better!!!

Good! back to the drawing board then...

 

I think I'm changing my mind about using the MIC4428ZN for this application. For starts, I'm not sure how I'm supposed to connect it to a bipolar supply, since I don't fully understand its functional diagram:

​

But this I understand:

​

I'll be using the ZVNL120A and the ZVP2106A for this purpose, and it looks like they can be driven directly from my MCU's TTL outputs.

 

I would say go with what you're comfortable with, then; I don't see any big difference between the two approaches.

 

I would say go with what you're comfortable with, then; I don't see any big difference between the two approaches.

Thanks... my big difference is that I understand the latter, but not the former.
I mean, to which pin in the MIC4428ZN would I have to connect the -5V source to? The ground pin? Wouldn't that interfere with the input logic?

 

That would be my assumption, yes. Which would also mean the signal source driving the chip would have to be operated from -5V & ground rather than from ground and +5V.

 

So I've tried to drive the positive-negative switching logic for the mosfets, and I can't make the negative part switch the way I want to. I've already googled my mind off, and stared at the screen for hours to no avail...

I want to be able to switch a p-mosfet connected between -5V and ground through a 350 ohm load, the way it's shown in the circuit at the right of my sim.

I've already simmed switching an n-mosfet with a positive voltage load... no problem.
I've already simmed switching a p-mosfet with a negative voltage load... but I had to use a negative pulse into the gate to make it work.
Now I want to do the same thing with the pFet, but using a 5V - 0V signal (just like TTL from my MCU) to switch it on and off... and I can't make heads from tails of it... the circuit at the right is my last failed attempt before I decided to show my face here again. I'm sure the answer is quite simple, but I've never worked with negative voltages before.

 

I'm sure the answer is quite simple, but I've never worked with negative voltages before.

Negative voltages are just like positive voltages, except they're negative.

Sorry, couldn't resist...

The problem isn't that you're working with negative voltages, it's that you need to perform a level shift from your 0V/+5V MCU logic levels to the -5V/0V drive levels required by M1.

Try this:

1. Disconnect V5 from the bases of Q1 and Q2.
2. Connect a resistor (around 1K - 5K) between the Q1/Q2 bases and -5V. Call it R5.
3. Add a 2N3906 transistor to the circuit, call it Q3.
4. Connect the collector of Q3 to the bases of Q1/Q2.
5. Connect the emitter of Q3 to V5.
6. Add a resistor, around 10K, to the circuit. Call it R6. Connect a 33pF capacitor across R6. Call it C1.
7. Connect one end of R6/C1 to ground, and the other end to the base of Q3.

Voila. There you have a non-inverting level shifter.

(C1 is there mainly to reduce the turn-off time of Q3, although it will speed up turn-on a bit as well.)

 

Negative voltages are just like positive voltages, except they're negative.

Gee... thanks for your positive input... couldn't resist either...

Thanks! I'll do as you say.

 

OH... MY... GOD...

It worked! Have you any idea of how long it would've taken me to figure that out by myself ????
Now I'll try to integrate that circuit into the half h driver I'm trying to build, see if I can do it on my own.

Thanks!

 

Yep, that's what I had in mind, except I had R5 connected to -5V rather than -15V. Should work either way, though.

 

Yep, that's what I had in mind, except I had R5 connected to -5V rather than -15V. Should work either way, though.

Would you recommend I connected it to -5V? the -15V come from a switching power supply. What I'm trying to accomplish here is the most stable and accurate AC square wave I can get, that's why I worked on the current-boosted references first.

 

Just check the specs for your MOSFET and see what the absolute maximum rated Vgs is. So long as it's rated for +/- 10V or more, you should be OK with R5 connected to -15V; otherwise, connect it to -5V.

 

Just check the specs for your MOSFET and see what the absolute maximum rated Vgs is. So long as it's rated for +/- 10V or more, you should be OK with R5 connected to -15V; otherwise, connect it to -5V.

Thanks, I'm sure it will work both ways (max V at the gate is ±20V, min trigger is 2.5V). My question is more about which of the two sources will effect more noise at the output.

 

Little difference, I suspect.

 

Aaaaaanddddd... it works! ... Now on to finishing the ±5V current-boosted reference.

I'm going to take a close look at how the R6-C1 affects timing, see if it actually helps preventing both mosfets from being on at the same time. The mosfets take more than twice as long to turn off than to turn on.

 

In LTSpice, see if you can display the drain current of M1 & M2, to check for shoot-through.

 

Ok, here's the ±5V current-boosted reference.
I placed a 2n3906 transistor to boost the positive stage... no problem-o there.
But I was forced to place another OP07 in series with the previous one, to be able to connect the LT1010 in follower configuration just as it says on its datasheet. Hope that's not a problem in real life.
Both stages are now easily delivering the 100mA that I want



Now back to the AC square wave mosfet circuit.

 

Why do you need to operate U1/U2 as a follower? Seems to me you could just get rid of U4, C3, R4 & R5, ground the non-inverting input of U2, and connect R3 between the inverting input of U2 and +7V.

 

Why do you need to operate U1/U2 as a follower? Seems to me you could just get rid of U4, C3, R4 & R5, ground the non-inverting input of U2, and connect R3 between the inverting input of U2 and +7V.

Couldn't make it work that way... the negative output is giving me -10V instead