# Driving motor with op amp output - but how?

Discussion in 'General Electronics Chat' started by summersab, May 13, 2016.

1. ### summersab Thread Starter Active Member

Apr 8, 2010
134
0
So, I had another thread on this same topic, and while I hate to start another, my situation has changed, and that thread turned into a bit of a confusing learning exercise (much to my benefit in the end - sincere thanks to everyone). I simplified my circuit from that thread, no longer have multiple voltage sources, and am just powering everything with 5V. Here's what I've got. I found the following circuit on EDN:

http://www.edn.com/design/analog/4314544/Integrator-ramps-up-down-holds-output-level (article)
http://m.eet.com/media/1126960/12071-figure.pdf (actual circuit image)

Using a 5V source, I want to take a 5V motor (0,5A nominal, 2A stall) from off to full power using the ramp output signal of this circuit. I built the circuit using the LMC6484 op amp suggested by the article (omitting IC1c since I want to use the ramp signal to automatically ramp the motor's speed and not control the speed directly with a PWM). The circuit works beautifully and ramps up linearly from 0V to 5V in about 10-15s. However, the problem I'm having is that I don't know what to connect to the output of IC1b to power the motor. Here is where my mind has gone:

Problem: LMC6484 can only supply ~10-15mA with a 5V supply. This isn't really enough to drive a power transistor.
Solution: Pick a different rail-to-rail op amp with a higher output current.
...but:
Problem: The circuit is really designed to provide a voltage ramp, not exactly a current ramp. Maybe I'm going down the wrong path.
Solution: Perhaps this is a job for a MOSFET instead of a BJT?

Problem: 5V isn't enough to power most MOSFETs unless I get something logic-level/really low threshold.
Solution: Get a MOSFET with a really low threshold.

Problem: I'm not finding many MOSFETs that have a <1V threshold.
Solution: Use a logic-level MOSFET and supply and pre-bias the gate voltage to the threshold.

Problem 1: Despite sounding like I know what I'm talking about, I have zero idea how to properly pre-bias a MOSFET and then input the ramp voltage to the gate.
Problem 2: The output of a MOSFET is is rather logarithmic, not linear, correct? Is this really going to do what I'm wanting, or is there something else that would better accomplish what I'm looking for?

Thank you all again for your help. If you've been following my inane posts, I hope it's clear that you've managed to teach me a thing or two, at least...

2. ### summersab Thread Starter Active Member

Apr 8, 2010
134
0
Crap - I meant to post this in the projects forum. Is there any way to move it, or does it not really matter?

3. ### crutschow Expert

Mar 14, 2008
13,465
3,353
You could use a switched-capacitor charge-pump to get the desired voltage bias for the MOSFET as here and here.

Nov 12, 2008
3,392
1,447

5. ### crutschow Expert

Mar 14, 2008
13,465
3,353
He wants to power it with 5V.

6. ### summersab Thread Starter Active Member

Apr 8, 2010
134
0
@ronv, I actually loved that circuit from a certain perspective (it worked exactly as intended - thank you!), but I had a few limitations. To go a bit into detail
• 3.7V wasn't enough to really get the motor vibrating enough. Since it's rated up to 6V, I got a high-power step-up converter to supply 5V (though I will probably use 6V in the end).
• Since I already had a 12V step-up converter to power the timer I'm using (not discussed in this thread - trying to keep things simple), I wanted to see if I could power everything with the single step-up. After a little reverse-engineering of the timer, I eliminated part of its control board and the 12V converter.
So, I improved and simplified my circuit. However, I also got myself back to where I was initially with the other thread, only worse - now, I don't have that second 12V voltage source (I suppose I do have the battery and the 5V step-up, but it's a different configuration, this time).

@crutschow, I'll look at those links. I was also seeing some suggestions of providing a constant voltage to the gate a the threshold and driving the supply voltage to the gate with a capacitor. Not sure if that's a good way to do it, though.

In general . . . is there a proper way to use the circuit that I found in a way that is "by design?" I wanted to reach out to the author who published the circuit to ask for his suggestions, but he unfortunately passed away in January. Hey, it led me to learn things, I suppose . . .

7. ### crutschow Expert

Mar 14, 2008
13,465
3,353
I'm sorry but after all the discussion about the all the ways you changed the circuit I'm rather confused on what you actually have.

You can boost the voltage to the gate with a bootstrap circuit if the signal is AC.
How long is the motor on at any one time?

8. ### ronv AAC Fanatic!

Nov 12, 2008
3,392
1,447
If you can pick the voltage , chose one that is a couple of volts above the desired motor voltage. Then a darlington will work.

9. ### ronv AAC Fanatic!

Nov 12, 2008
3,392
1,447
The circuit is a slow programable ramp that can hold at a selected voltage.
The problem is he is trying to drive a single transistor with a low poer op amp to get 1+ amps to the motor.

10. ### summersab Thread Starter Active Member

Apr 8, 2010
134
0
@crutschow Yeah, you're right - here's what I'm working with:
http://goo.gl/K9ZhUQ

The motor should go to full power (5V @ 0.5A nominal, 2A stall) at the end of the ramp and then stay on indefinitely (or until the battery dies).

@ronv Yes, I can pick the voltage - I got some cheap adjustable boost circuits. So, a Darlington will work even with such a small current from the op amps? Also, will using a Darlington and a higher supply voltage to get the motor to the desired voltage end up putting stress on the boost circuit due to the decreased efficiency? In other words, since the Darlington is going to need >5V at the collector in order to provide 5V at the emitter, some power will be lost as heat. How much power - enough to make it a concern to the overall system?

I'll give the Darlington a go. I guess I'm just trying to figure out if there's an ideal way make use of this "cute" circuit (as @ronv previously called it ).

11. ### ronv AAC Fanatic!

Nov 12, 2008
3,392
1,447
Yes, the darlington will have much higher gain than just the 3055. Maybe this one: http://www.mouser.com/ds/2/389/CD00000911-249927.pdf
Yes there will be heat generated. About 2 volt X the current. So maybe 1 watt.
The first circuit I posted (PWM) is more efficient, but more complex.

12. ### summersab Thread Starter Active Member

Apr 8, 2010
134
0
I wondered about that first circuit and thought about pursuing that before posting (again . . . sorry). However, I noticed that it was still making use of that 12V rail, and I wasn't sure if it needed that or not, especially since it is using a MOSFET. To make it work, could I use the same 5V supply for both the ramp/PWM and the MOSFET (substituting a logic-level power MOSFET)? I may just pursue the Darlington in the short term since I have the parts, but a MOSFET seems like the better application (per my extremely limited understanding).

13. ### crutschow Expert

Mar 14, 2008
13,465
3,353
If you power the op amp with a 10-12V rail and connect a logic-level N-MOSFET as a source follower with the drain going to 5V then you could drive the motor at the source with a full 5V signal.

14. ### ronv AAC Fanatic!

Nov 12, 2008
3,392
1,447
Yes, you could run it off of 5 volts with a logic level FET. Please note the source is ground in this case (like the schematic)
In the other case(hooked up like the transistor) it is a source follower so the voltage on gate has to be higher than the source by a few volts.

15. ### summersab Thread Starter Active Member

Apr 8, 2010
134
0
Somehow, I don't think there's a way around having multiple voltages in this system, is there?

I tried using a TIP120 Darlington (I know - hardly the most efficient, but Radio Shack and all . . . ) If I use 10V to power the circuit, the motor will get 4V, but the transistor was hitting 60° C, and I don't want to go any higher. I'm trying to avoid high temperatures and heatsinks as much as possible in the interest of space. While a more efficient Darlington may work, perhaps it's not the best solution. Perhaps if I got a different op amp like one of these? http://www.digikey.com/short/3pj017 Still, this circuit is intended as a voltage ramp and not really a current ramp, right?

I don't have a logic level FET at the moment, but I can pick one up on Monday. Seems like the right way to go unless you have any other thoughts.

Thanks for all the help/suggestions/education.

16. ### crutschow Expert

Mar 14, 2008
13,465
3,353
If you want a full 5V delivered to the motor from a 5V supply then I think that's correct.

You could generate the higher voltage to drive a MOSFET gate using a 555 astable capacitive voltage multiplier circuit if that would be better for you then one of the dedicated charge-pump circuits.

Last edited: May 14, 2016
17. ### ronv AAC Fanatic!

Nov 12, 2008
3,392
1,447
Did you power both the circuit and the darlington with the higher voltage? The drop across the transistor seems high. However: no matter what you use in the follower configuration it will generate a watt of wasted power. 1 watt in a TO220 package increases it's temperature 62 degrees C.

18. ### summersab Thread Starter Active Member

Apr 8, 2010
134
0
Yes, I was using the same voltage to power the circuit and the Darlington. I'm taking it that isn't wise/correct?

19. ### crutschow Expert

Mar 14, 2008
13,465
3,353
You can't get the full 5V out using a single 5V source.

Here's an LTspice simulation of how to get the full 5V (minus the drop due to the ON resistance of the MOSFET) to the motor using two supplies, one rail-rail op amp, and a logic-level MOSFET.
V3 could be derived from V2 using a charge-pump voltage doubler circuit.

20. ### ronv AAC Fanatic!

Nov 12, 2008
3,392
1,447
No, that is correct. You should be able to get 6 volts across the motor with about an 8 volt supply. But it will still get warm.
A measurement at base emitter and collector might be in order.