3
\$\begingroup\$

I'm building current limiter for my circuit, I saw a nice one in Maxim's document (Figure 2a): https://www.maximintegrated.com/en/app-notes/index.mvp/id/2158

enter image description here

I need current limiting for 100mA, so I modified that circuit and simulated in LTspice:
enter image description here

It works, it limits current well, but there is some glitch/spike when power is switched on (current across R1):
enter image description here

Spike in detail (power is turned at 10μs, spike starts at 10μs):
enter image description here

I tried different transistors (Q1 and M1) but it didn't help.

What causes this spike?
Is there any problem in that circuit?
Or is it just some problem in LTspice simulation?
Is it possible to get rid of this spike?

\$\endgroup\$
9
  • \$\begingroup\$ The capacitor C1 is the culprit there. It's meant to be filtering M1's voltage, to avoid slight perturbations, but, in return, at startup, it charges briefly through R1-C1-R2. If you look at the time scale, the pulse is limited in both amplitude and time (~3ns), which is minor (unless your circuit is so very sensitive). If you move C1 in parallel with R2, the behaviour will change. \$\endgroup\$
    – a concerned citizen
    Commented Apr 23, 2017 at 6:34
  • \$\begingroup\$ I believe it's a simulation bug. A 400mA current going through 1k resistor would cause a 400V drop which in practice can't happen. \$\endgroup\$
    – dirac16
    Commented Apr 23, 2017 at 6:36
  • \$\begingroup\$ @aconcernedcitizen I accused C1 too. So I removed C1, and spike was still there. \$\endgroup\$
    – Chupacabras
    Commented Apr 23, 2017 at 6:37
  • \$\begingroup\$ @Chupacabras It wasn't possible to be \$C_1\$, as there's huge \$R_2\$ in the way, despite comments otherwise. \$\endgroup\$
    – jonk
    Commented Apr 23, 2017 at 6:38
  • 1
    \$\begingroup\$ Hold on, I just discovered it: it's the input source, which delivers the current to the charging capacitors through its (machine) zero internal resistance. If you add Rser=1m Cpar=1m to V1, everything changes. \$\endgroup\$
    – a concerned citizen
    Commented Apr 23, 2017 at 6:41

1 Answer 1

Reset to default
3
\$\begingroup\$

After my failed attempts in the comments, I point the finger to V1, who has (machine) zero internal resistance and is forced to deliver the charging current for the capacitors (through the other series resistances in the circuit). Normally, you'd have both a limited resistance from the input supply, and some capacitor at the input, so simulating that by adding Rser=1m Cpar=1m to V1, seems to solve it.

\$\endgroup\$
3
  • \$\begingroup\$ Well, Rser=1m Cpar=1m helped. I did not realize that I use ideal source, and that this is the culprit. Are those values real? \$\endgroup\$
    – Chupacabras
    Commented Apr 23, 2017 at 6:49
  • \$\begingroup\$ I meant to say it, but I forgot: there are other parasitic capacitances in the circuit besides the two visible ones, and they also contribute to spikes in currents. \$\endgroup\$
    – a concerned citizen
    Commented Apr 23, 2017 at 6:49
  • \$\begingroup\$ @Chupacabras No, 1 mili-Ohm series resistance for an input source is an exceptional one, particularly for such low currents. But you will have a (dynamic) resistance, and you should place some capacitor, as well, the importance of which is clearly seen here. :-) \$\endgroup\$
    – a concerned citizen
    Commented Apr 23, 2017 at 6:51

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.