# Softstart for H-bridge controlling bidirectional motor

Joined Oct 18, 2021
24
Hi, I'm currently working on a project for controlling ball valves. Motors used for this operation are geared down 24V 775 motors. I'll be sensing limit positions with micro switches, but this will come to play in later text. The problem I'm trying to solve is the inrush current at start of the motors. Single motor draws around 500mA while running, but at the start, the current is around 1,5A (measured by an analog ammeter, about 1A by a digital multimeter). I intend to run two of these simultaneously, which would mean buying at least 3A switched supply, ideally 4A. But those are getting quite big for my application, and I think it is wasteful and overkill. I would rather buy 1,25 - 1,5A switched supply and solve the inrush current by some kind of softstart. For driving the motors in both directions, I'm planning on using L298N, but I'm open to suggestions.

The problem is that I'm not sure how to implement softstart circuit and what kind to use for it to work with a combination with H-bridge. The next thing to consider is that I need the motor to stop immediately after the voltage is cut by the H-bridge because I need the motor to stop at the limit position after hitting the microswitch. But some softstarts I've seen slowly ramp down after cutting the voltage to motor. I've considered using thermistors for limiting inrush currents, but I'm not sure if they would be able to cool down / recover to work properly again between the cycles. And in case of a faulty signal resulting in a quick startup after the last one, they would probably be of limited use. The solution to this problem might be to use a quick fuse on the output of the power supply, and in case of limited thermistor functionality, the fuse would get blown. I'll be glad for any help, thanks.

#### crutschow

Joined Mar 14, 2008
30,420
A common way for this is to momentarily connect a resistor in series with the bridge power for the time it takes the motor to start-up, and then remove (short-out the resistor with a relay or transistor) when the motor is up to speed.

You would likely need to experiment to determine the proper resistor value.
For a 24V, 500mA motor, it will probably be somewhere in the neighborhood of 5 to 10 ohms.

Alternately, if you can pulse the signal, you could generate a PWM signal of perhaps a few hundred Hertz and adjust the duty-cycle to get the soft start.
After then you just set the PWM signal to DC high.

Last edited:

Joined Oct 18, 2021
24
A common way for this is to momentarily connect a resistor in series with the bridge power for the time it takes the motor to start-up, and then remove (short-out the resistor with a relay or transistor) when the motor is up to speed.

You would likely need to experiment to determine the proper resistor value.
For a 24V, 500mA motor, it will probably be somewhere in the neighborhood of 5 to 10 ohms.

Alternately, if you can pulse, the signal, you could generate a PWM signal of perhaps a few hundred Hertz and adjust the duty-cycle to get the soft start.
After then you just set the PWM signal to DC high.
I've considered a classical soft start circuit, but it was getting more and more complex because of the switching polarity on the output of H-bridge. Also, I would prefer not to use relays and instead use solid-state technology, and I mainly wanted a different way of looking at the problem. Which you gave me. I'm not really sure why it didn't occur to me to use PWM for that purpose. For that idea, I'm thankful.

#### ag-123

Joined Apr 28, 2017
259
I'm not too sure if connecting an inductor in series may help, during the in rush, the inductor would generate back emf and limit the current surge. However, a rather large inductor may be needed.
$V_L = - L \frac{dI}{dt}$

#### Ian0

Joined Aug 7, 2020
5,819
If you have limit switches, then an extra pole on the limit switches could switch in the resistor - as soon as the motor moves away from the limit switch the resistor is switched out.

#### MisterBill2

Joined Jan 23, 2018
12,687
Given that the challenge is the current draw exceeding the supply capability during the startup time, which is short, a useful option is to add a capacitor in parallel to the supply to provide the additional current. This is common approach and it works well. The actual value of the capacitor may require some experimentation, or analysis.

#### LowQCab

Joined Nov 6, 2012
2,315
You might consider something like this ..........
.
.
.

Joined Jul 18, 2013
25,626
One popular TM manufacturer uses a novel way of soft-start is to use a series SCR on the un-filtered DC at the bridge output, the SCR is turned on gradually until full conduction.
Very effective.

#### ag-123

Joined Apr 28, 2017
259
I think crutschow has a point, I liked that PWM idea, It basically needs a microcontroller that drives the h-bridges.
And you could on top of that vary the speeds of the motor with PWM.

#### crutschow

Joined Mar 14, 2008
30,420
One popular TM manufacturer uses a novel way of soft-start is to use a series SCR on the un-filtered DC at the bridge output, the SCR is turned on gradually until full conduction.
Very effective.
Okay, I'll bite.
How does an SCR gradually control the unfiltered DC at the bridge output?
Is that unfiltered rectified AC?

Last edited:

#### MisterBill2

Joined Jan 23, 2018
12,687
Unfiltered rectified AC can be conveniently phase controlled with an SCR to deliver variable pulse with drive similar to a light dimmer. Varying the firing angle is done with a very simple RC arrangement, a lot like most light dimmers. No processor required. I thought that Crutschow lived in the USA, where phase angle triggered SCR dimmers are common.

Joined Oct 18, 2021
24
If you have limit switches, then an extra pole on the limit switches could switch in the resistor - as soon as the motor moves away from the limit switch the resistor is switched out.
There is a chance I'll be using the ball valves not just to cut off the flow but also to regulate it in the future. So there is a chance that the motor will start in between the limit switches.

Given that the challenge is the current draw exceeding the supply capability during the startup time, which is short, a useful option is to add a capacitor in parallel to the supply to provide the additional current. This is common approach and it works well. The actual value of the capacitor may require some experimentation, or analysis.
Also a nice solution, but I'll probably stick to the PWM approach. There will be a microcontroller on the board either way, so why not use it. Also, experimenting or analyzing capacitor value is extra work, and I already tested the PWM yesterday with good results. The only drawback is the noise of the motor while at a low duty cycle. But that could be expected from the square wave. Using higher frequency might also help.

I want to thank all involved. I'll proceed with the PWM solution and consider the problem solved.

#### MisterBill2

Joined Jan 23, 2018
12,687
Please let us see the circuit that you wind up using, if you can. It is always useful to see what works.

Joined Jul 18, 2013
25,626
The only drawback is the noise of the motor while at a low duty cycle. But that could be expected from the square wave. Using higher frequency might also help.
What PWM frequency are you using,? there should be no motor noise down to zero rpm .
Typically the PWM frequency for motor control starts at around 5khz - 6Khz.
In any case, the noise would be created by the current which does not really resemble a square wave.

#### crutschow

Joined Mar 14, 2008
30,420
The only drawback is the noise of the motor while at a low duty cycle
But don't you go to a 100% duty-cycle (DC) after the motor starts, so the noise would be only during the short motor start time?

#### MisterBill2

Joined Jan 23, 2018
12,687
While the motor VOLTAGE may approach a square wave, the motor inductance assures that the motor CURRENT will not be a square wave. And the magnetization is a function of the current.
It is useful to have the PWM frequency different from any motor package resonant frequencies.
So some adjustment may be required.

Joined Oct 18, 2021
24
Please let us see the circuit that you wind up using, if you can. It is always useful to see what works.
I've just tested the PWM approach on a breadboard without any final touches. I only wanted to test how well the PWM will work from my application. So I just used Arduino Mega as means of controlling the transistor. I wasn't trying to do some final tuning, just proof of concept as a reassurance.

What PWM frequency are you using,? there should be no motor noise down to zero rpm .
Typically the PWM frequency for motor control starts at around 5khz - 6Khz.
In any case, the noise would be created by the current which does not really resemble a square wave.
As it was just a proof of concept circuit, I was using the preset PWM frequency of Arduino Mega. I knew there would be space for improvements, especially if I'll be using ESP8266, where you can apparently set PWM up to 40kHz. Now that I'm looking at Mega specks, the preset PWM frequency is 490 Hz. That would explain a lot.

But don't you go to a 100% duty-cycle (DC) after the motor starts, so the noise would be only during the short motor start time?
Yes, exactly. That is what I meant by "noise of the motor while at a low duty cycle". The noise is only apparent at the start. While running, there is no noise due to the PWM. The program I used for testing was a simple ramp-up from 0-100% duty cycle. In the final design, I'll play with it more, maybe start with a non-zero duty cycle or increase the duty cycle at first slowly and then more quicly...

Joined Oct 18, 2021
24
I know that I've "closed" the thread, but I have a few questions loosely related to it.
I'm now thinking about shielding the PWM lines. At first, I was thinking that I would use a cable with seven cores (two for powering the motor and five for signaling the state of four limit switches). But now that I'm thinking about it... I could add a low pass filter to the limit switch state sensing circuit just to be sure. But that isn't the biggest problem. All around are different sensors and electronics.. so lower noise is better. What concerns me is circulation pump communication which uses 75 Hz PWM to communicate its state and another PWM line in a range of 100 Hz to 4 kHz for controlling the pump.

I was reading a bit about using shielding wires for the purpose of lowering the noise. But what I've got is a lot of conflicting messages. Some recommend ground at both ends (of a shielded cable). Others talk about ground loops if a shielded cable is grounded at both ends. I'm not even able to ground at both ends. Some say that shielded cable without any connection to the ground or negative pole of power supply will be better than nothing. Some say it will be worse than cable without shielding and acts as an antenna...

In the end, I have only a few ways of shielding.
1) Don't use any shielding.
2) Use shielded cable but leave shielding unconnected.
3) Use shielded cable and at on one end connect it to the ground (another thing is if you should use PE or negative pole of power supply used for powering the circuit generating PWM)
4) Use shielded cable and one end connect to the ground and on the next end, connect shielding with spare wire inside the shielded cable and run it back to the place of grounding the shielding and ground that wire.