0
\$\begingroup\$

I am attempting to control the current through a generator rotor by adjusting PWM input to a DC to DC solid state relay (SSR), specifically a TWTADE SSR-40DD.

The power supply for the output side of the SSR is 20VDC. The input side of the SSR is an ESP32 microcontroller. I have verified that the output voltage (with the rotor disconnected) on the SSR is what is expected according to PWM input.

With the rotor connected to the output side of the SSR, when I set the SSR input to 100% duty cycle, the output through the rotor goes to 2A which is expected and great. However at anything less than 100% duty cycle, the current passed through the rotor drops precipitously. I would like to be able to smoothly control the time-averaged rotor current from 0A to 2A, but instead the current goes from 5mA gradually to about 30mA then jumps to 2A at about 100% duty cycle.

The rotor tests at 10 Ohms and 2.2 H inductance. I have varied the PWM input pulse frequency over a wide range (60 Hz to 6000Hz) and the results on the output are only slightly different.

When I was using an L298N motor controller instead of the SSR, I was able to modulate the current through the rotor quite well. This is also PWM output. However the L298N does not have as safe a power output margin as the SSR, so that is why I tried to replace it with the SSR.

Why can't the SSR modulate a time-averaged current through the rotor with PWM input?

\$\endgroup\$
4
  • 1
    \$\begingroup\$ This is expected. SSRs are not made to switch quickly. Read what barely qualifies as a product datasheet next time. It lists 10ms transition time. If you saw no datasheet to begin with from which to gleam this info, do not buy. Never buy without a datasheet. A real datasheet, not a random list of specs on a web page that could have been typed up by anybody. \$\endgroup\$
    – DKNguyen
    May 25, 2022 at 1:49
  • \$\begingroup\$ I significantly increased the pulse width (to 200ms) as a test. This is fairly large relative to the quoted "<10ms" response time of the SSR, but I still see very large nonlinearity in the current response as I gradually increase duty cycle of the PWM. Not sure what to make of this. \$\endgroup\$
    – user156904
    May 25, 2022 at 2:49
  • 1
    \$\begingroup\$ Pulse width is not frequency. Besides, the time constant of a 2.2H, 10 Ohm rotor is 220ms so any pulse width that is 200ms is inherently far too low a frequency. Free-wheeling current paths are also required. Also, I'm guessing you aren't measuring current with a scope otherwise you would just be able to see all this. \$\endgroup\$
    – DKNguyen
    May 25, 2022 at 3:23
  • \$\begingroup\$ You don't mention a flyback (freewheeling) diode. Are you using one? \$\endgroup\$ May 25, 2022 at 11:31

1 Answer 1

0
\$\begingroup\$

The L298 is more suited to high V, low current for efficiency as the Vce(sat) is too large thus self-heating is a problem.

@ 1A Vce(sat) = 1.8 ~ 3.2 V i.e. Rce= 1.8 to 3.2 Ohms for the pair.
@ 2A Vce(sat) = 4.9 V max or 2.45 ohms for the pair.

With thousands of FETS in the milliohm range or a half-bridge IC is a far better solution. Get something rate for >> 2x your max current to keep cool and/or use an appropriate heatsink. It will need a suitable dead-time for commutation, Otherwise a single switch and diode with same or more current rating.

\$\endgroup\$
2
  • \$\begingroup\$ This is good information but not directly related to my question: why the SSR was not working as expected. As I mentioned in my question, the L298 did work well from a control standpoint, but I was concerned about power capacity in this application, about which you also are concerned. \$\endgroup\$
    – user156904
    May 25, 2022 at 2:56
  • \$\begingroup\$ It is relevant because your solutions won't work \$\endgroup\$ May 25, 2022 at 3:18

Your Answer

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

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