Best way to enable an H-Bridge

Thread Starter

akistler

Joined Jan 8, 2018
9
I am working on a project to use a single output to change the state of a latching solenoid (magnetic retention one direction, spring retention opposite, no holding power required)

So far I have the 3 elements of the circuit breadboarded, but I am running into an issue for the final enable. At this stage, all transistors are BJT.

1. Direction is controlled by a sequential switching bistable multivibrator circuit. See http://www.electronics-tutorials.ws/waveforms/bistable.html (The current orientation is irrelevant, it is verified through another means an can alternate until the desired result is achieved)

2. Timing is controlled by a (2) sequence 555 timer. First timer triggers off the same input controlling the multivibrator and is on long enough to ensure everything is ready to energize the motor (For test, 5 seconds). Upon completion of the first timer, the second timer triggers an H-bridge to energize the solenoid for a pre-defined length of time (For test, 5 seconds). See http://www.gadgetronicx.com/sequential-process-control-ic-555/

3. An H-bridge to energize the solenoid in the direction controlled by the multivibrator, for the duration from the timer. See https://electrosome.com/dc-motor-driving-using-h-bridge/

TL;DR
I am running into an issue "enabling" the H-bridge. My initial thought was to have a high-side NPN transistor acting as a switch to the H-bridge, but it doesn't seem to work. I suppose I could AND the signal from the direction controller and the timer on the base of the H-bridge, but that would substantially increase chip count. When I get near a scanner, I can add some circuit sketches, but the links were the starting point.
 

tcmtech

Joined Nov 4, 2013
2,867
If you are having trouble getting the high side N-channel switching devices to turn on by sending current/voltage to them try reversing that and use P-channel devices that turn on by pulling current/voltage off of them.
 

Thread Starter

akistler

Joined Jan 8, 2018
9
I drew it out in circuitlab (a little easier to read than hand drawn sketch) and might have found the issue. I had my "Enable" transistor hooked up in common collector instead of common emitter orientation. I'll try swapping out the NPN with a PNP.
 

Attachments

tcmtech

Joined Nov 4, 2013
2,867
As you have it drawn you will effectively destroy your switching devices on both sides being you have them wired in a self activating shoot through short circuit.
 

tcmtech

Joined Nov 4, 2013
2,867
To rephrase, Take a look at your Q1,Q3 side of the H-Bridge and think the current flows through for a bit. Then do the same for the Q2, Q4 side.

See a big problem? o_O
 

Thread Starter

akistler

Joined Jan 8, 2018
9
AnalogKid,

Skewing timing would be difficult as the input is coming from a bistable multivibrator. It seems that shoot-though current would be avoided if the h bridge had an enable transistor, unless I am misunderstanding. Would you agree, if so, do you think a low-side switch or high side switch would be better?
 

AnalogKid

Joined Aug 1, 2013
10,986
Without careful design to create a deadband when no transistors are conducting, there is a time during the transition from one phase to the other when at least 2 or possibly all 4 transistors are on at the same time.

One way to address this is to put a zener diode or some other voltage offset between the two bases on each side of the bridge, so that as the input signal transitions from one drive level to the other, there is a time when neither transistor has enough Vbe to begin or maintain conduction.

Another way to do it is to insert an edge-dependent time delay so that which ever transistor is turning off gets the edge transition immediately, while the one turning on gets a slightly delayed edge so it doesn't come on until the other one has had time to turn off. If you are driving all 4 transistors with separate outputs from a microcontroller you can adjust the delays in firmware. In hardware it takes some form of an R-C delay stage with a diode in parallel with the R to "short it out" in one edge direction but not the other.

If it makes you feel any better, this is a very common design issue. The bipolar 555 draws about 5-10 mA of static current, but spikes to over 300 mA with every output transition, independent of the actual output current.

ak
 

crutschow

Joined Mar 14, 2008
34,280
Skewing timing would be difficult as the input is coming from a bistable multivibrator. It seems that shoot-though current would be avoided if the h bridge had an enable transistor, unless I am misunderstanding.
Yes, if you have a separate PNP enable transistor controlling the bridge power, then you can change the bridge direction when the bridge is not powered, to avoid shoot-through currents.

Below is the LTspice simulation of a simple H-bridge circuit with two separate direction inputs and no shoot-through as long as both inputs are not on at the same time.

upload_2018-1-10_12-16-50.png
 

Thread Starter

akistler

Joined Jan 8, 2018
9
I got the bridge working with the inputs, timers and h bridge, as long as I'm using a dummy load (LEDs in opposite direction, A turns on with voltage potential one direction, B turns on with voltage potential other direction). If I use a DMM, I get ~11V with first trigger, ~-11 with second trigger. However, when I put my solenoid in place with the DMM in parallel, the voltage is about .5V. I've attached the circuit as built.
 

Attachments

Top