5 second 'slow on' circuit

[AGCB]

I'm trying to build a soft start circuit for an electric clutch (riding mower type). So far, I can only get about 1 second or less to full power. I tried this circuit less the Zener and LED.

Voltage is 12 volts DC, load is 5 amps and using a IRFZ48N transistor.
I've increased the capacitor to 15,000 and played with resistors but nothing seems to help much.

I know I could do it with a MCU (and have) but would like to keep it simple if possible.

Would appreciate if someone could point me to a solution.

Thanks
Aaron

 

[Papabravo]

Considering all the trouble you are having; how can you claim with a straight face that this is somehow simple? We might have good cause to question your judgement.

 

[danadak]

Is clutch a bistable kind of response or does it have a linear or exponential
force transmission behavior with current ?


Regards, Dana.

 

[crutschow]

The clutch is likely operated by a solenoid, which has a bistable operation, thus will tend to suddenly snap on, even if their voltage is slowly increased.
And the problem with that simple circuit is that it still has a rapid turn-on due to the high-gain of the MOSFET.

A varying PWM signal to slowly increase the clutch current may work better:

Below is the LTspice sim of such a circuit:
The clutch current takes a little over 2 seconds to reach the maximum 5A.

Of course this circuit is some more complicated than the circuit you posted, requiring one LM339 or LM393 comparator package plus the MOSFET, diode, and a few resistors and capacitors.
The value of R7 and C2 determine how long it takes for the clutch to reach full current.

 

[danadak]

A very simple PWM control circuit :

This is block programming a Arduino UNO or Nano or ATTINY85. mBlock converts the
user block configuration into Arduino code, which you program the chip or board with.
Code mBlock generated shown on right. Note if clutch non linear response code could
easily tackle "linearizing" the control loop.

Note I do not show the catch diode for the clutch coil, you would need to add that.


Regards, Dana

 

[Karklebapper]

I have done soft start circuits using a charge pump driving an Nch MOSFET. This will provide a relatively linear output ramp comprised of many small steps.

Mainly you are charging C21 slightly with each pulse from an input pulse source such as 10V p-p. When a circuit already has a switcher running, the switchnode waveform may be useful for this (I have done that) so an added oscillator may not be needed.

The ramp rate depends mostly on the pulse frequency and voltage, and the ratio of C3 and C21 if R39 value is not too high. The gate-source voltage at Q1 does not change a lot during the duration of the ramp, but there is a delay at the beginning before the Vgs threshold is reached, and some time at the end when Vgs approaches max and Q1 Source-Drain voltage reaches its minimum.

Be careful about lossy conduction in the MOSFET, as it is possible to destroy the MOSFET due to thermal runaway in a small number of cells within the MOSFET die. This aspect was relatively new to me, but read about "linear" MOSFETs made by Infineon and some other manufacturers. Those have a better safe operating area for this type of operation. But it is entirely possible that a normal MOSFET intended for switching might work OK.

 

[crutschow]

Be careful about lossy conduction in the MOSFET

That's the advantage of a PWM ramp, which doesn't have that linear operation loss in the MOSFET during the ramp.

 

[Karklebapper]

Yes OK so long as your load is reasonably inductive, pwm'ing into it may work. The dissipative soft start can work into a variety of downstream loads which are likely to not be very inductive.

 

[danadak]

Another tradeoff using PWM is EMI and related issues.

 

[crutschow]

Another tradeoff using PWM is EMI and related issues.

Doubt that would be a problem on a riding lawnmower.

so long as your load is reasonably inductive, pwm'ing into it may work.

Doesn't necessarily have to be inductive.
The average value of the current will still ramp, even if it has a high ripple content.

 

[Karklebapper]

I think of using this sort of circuit to "hot swap" a DC load (generally with bypass capacitors) to a DC source. If you simply hot plug a low impedance DC source to this low impedance load (which starts out at 0V) you will get a large current spike which is likely to damage something, such as the switching element or bypass capacitors. Similarly, PWM switching the output to the input will not work correctly unless you have some current limiting impedance between the source and the load.

 

[danadak]

Doubt that would be a problem on a riding lawnmower.

crutschow, do electrical appliances in US have to meet FCC, ANSI, ......standards, not sure.
Ignition systems still incorporating noise suppression or is EMI a thing of the past.....

PWM are excellent noise generators, especially when involved in driving high currents
thru wires (antennas).

Of course this is a one off, and the standards police are few and far between.


Knight

 

[ZipZapOuch]

crutschow, do electrical appliances in US have to meet FCC, ANSI, ......standards, not sure.

They do but those appliances can be several feet (or less) from their neighbor's TV in an apartment complex. A lawn mower is, well, in the middle of a yard.

 

[Karklebapper]

Yes EMC rules are still a requirement so far as I know. But switching power converters are ubiquitous. Some years ago I worked for a company which made motor drive controller ICs for large (up to many horsepower) motors, driven using IGBTs and 800V input bus voltages for example. The motors are inductive and are driven directly by 10 KHz (typical) PWM output from the IGBTs. I have wondered about noise produced by the cabling and motors involved in this...

 

[Karklebapper]

Here is an LTSpice Sim demonstrating the design shown above. I am using an LM339 + inverter as a 100 KHz oscillator. Some component values have been modified to protect the guilty. The blue trace is output current.

 

[danadak]

They do but those appliances can be several feet (or less) from their neighbor's TV in an apartment complex. A lawn mower is, well, in the middle of a yard.

Except when its mowing the lawn close to the windows......

 

[crutschow]

In my circuit, the PWM EMI would only be there for the few seconds the clutch is being activated, as the clutch voltage is DC after that.

 

[Karklebapper]

Regarding EMI; the whole system is packed relatively close together within the body of the mower. So this presents less opportunity for radiation than if you had long wiring. And the startup-only aspect would probably enable any radiation to escape notice by the FCC for example.

 

[ChrisP58]

Regarding EMI; the whole system is packed relatively close together within the body of the mower. So this presents less opportunity for radiation than if you had long wiring. ...

While your statement has some validity, it's a significant over simplification.

The transmitter in your cell phone is packed very close together too, yet it's still able to transmit for miles.

 

[Karklebapper]

Yes I suppose I am simplifying, but there are some key differences. Basically I am comparing the requirement in question to some other "high frequency" power converters, such as maybe a computer power supply.

1) A cell phone operates at GHz, while most switching power converters operate below 1 MHz (low MHz is used for some low voltage power conversion.)

2) I know that cell phone cells have been getting smaller; not sure if the radius commonly exceeds 1 mile?

3) The cell phone has (what passes for an) antenna. Yeah I guess it is small but it is designed to radiate.