4
\$\begingroup\$

The circuit is connected as drawn below:

circuit connection

MOSFET IC = MAX14919A

Valve actuator = MHE2-MS1H-3/2G-M7-K

Schottky Diode = SS16HL-TP

The pulse happens when the low-side MOSFET opens (disconnecting the valve actuator from ground) before slowly increasing to 24V from 0V.

The yellow line represents the MOSFET gate input signal (5V = on, 0V = off)

The green line represents the applied output voltage (24V)

The waveforms are shown below:

This is the pulse that I am describing

full waveform

I believe this phenomenon is not an inductive spike because if it was, the Schottky diode would have minimised or eliminate the spike. I have tried the same setup with two other MOSFET ICs, BTS3410G and Si83014BF‐IT. Both of these MOSFET are slower than the MAX14919A (with the BTS3410G being the slowest) and both of them show this pulse once the MOSFET is open.

BTS3410G:

BTS3410G waveform

Si83014BF‐IT:

Si83014BF‐IT waveform 1

Si83014BF‐IT waveform 2

\$\endgroup\$
8
  • \$\begingroup\$ What is driving the gate(pulse generator, microcontroller, etc.)? How is it connected to the gate (directly or via for example a resistor)? Is the probe of the CH1 connected directly to the gate or is it showing the output of the driver? \$\endgroup\$
    – Saadat
    Nov 30, 2023 at 8:10
  • \$\begingroup\$ Hi, i simplified the MOSFET IC drawing. The IC itself is actually being triggered using a square wave from 0V to 5V using a waveform generator. The square wave is directly connected to the signal input of the IC. CH1 is probing the square waveform signal. CH2 is probed at the drain pin of the MOSFET. \$\endgroup\$
    – KevinZ
    Nov 30, 2023 at 8:18
  • \$\begingroup\$ If there is something in between the output and the gate of the Mosfet, the problem could be from that. It's a good idea to connect the probe directly to the gate, too. If the signal on the gate is the same square pulse, then you can be sure the problem is rising from the Mosfet itself. It's a good idea to include the full schematic. It makes is easier for people to understand it and be able to help. \$\endgroup\$
    – Saadat
    Nov 30, 2023 at 8:24
  • \$\begingroup\$ I don't see anything weird. Can you point out precisely what you see as being weird. \$\endgroup\$
    – Andy aka
    Nov 30, 2023 at 8:45
  • \$\begingroup\$ I find the 24V pulse that lasts for about 2.5us weird, because when the MOSFET is off (opened) the load disconnected from ground, hence the load should rise back to 24V, however in the first waveform that I have attached, the signal went to 24V for 2.5us and then went back to 0V and slowly rise back to 24V. \$\endgroup\$
    – KevinZ
    Nov 30, 2023 at 9:18

1 Answer 1

4
\$\begingroup\$

The delayed return to the 24 V level may be due to an RC snubber circuit already within the assembly. Note that in the valve's datasheet there is this entry: "Additional functions - Spark arresting, Holding current reduction, Protective circuit". So there are likely a few other components within the assembly.

The extra pulse could be due to a long cable (as hinted by Tim W). Might there be excess cable length coiled up? On your next to last plot following the first pulse there even seems to be a bit of suppressed oscillation (ringing), potentially caused by cable inductance. If there are additional components internal to the valve perhaps these are helping to suppress some but not all of the ringing action.

If the extra pulse is causing a problem (beyond just a curiosity), a moderate size cap located near the valve may eliminate ringing due to a long cable length.

Below is a sample circuit and Vin/Vout plots showing the results with an RC snubber placed across the valve coil and a small amount of extra inductance on the in-coming cable side.

enter image description here

enter image description here

If the extra cable inductance were to increase to 3 uH the pulse (ringing) will increase considerably.

enter image description here

By adding a relatively small cap (C5 10 uF) onto the cable just ahead of the valve coil the pulse(s) can be eliminated.

enter image description here

enter image description here A final note: The RC snubber values used were not optimized at all, they were selected in an attempt to duplicate the original plots. The coil inductance, resistance, and self capacitance are also just guesses.
.

\$\endgroup\$
1
  • \$\begingroup\$ The cable length of the valve actuator is about 1.5meters long. The suppressed oscillation smoothen out when we move the valve wire around. Applying a capacitor to ground at the +24V point of the valve actuator did not affect the waveform, however, placing the capacitor to ground at the node of the MOSFET's Drain managed to reduce the spike (and caused a slower rise time). \$\endgroup\$
    – KevinZ
    Dec 4, 2023 at 1:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.