FET opto driver questions: opto current requirements and in-rush current control?

Go To Last Post
13 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

This high-side P-channel FET circuit that seems to work to switch 9V from 5V micro:

 

schematic

 

The FET is a digital-drive type.

 

Questions:

 

  • Does opto U1 (LTV-817S) require a minimum current on the output side to stay turned on?  The datasheet has a parameter Ic (collector current) with minimum value 2.5 mA, does that mean it must have at least this much current to stay on, or what?
  • Will the C1/R3 combination work ok to limit in-rush to the (capacitive) load when the FET turns on.  I looked for how to do this and found the advice to add C1 to slow the turn-on of the FET.  I guess it make sense to control the rate of change of turn-on signal for FET, but I didn't find much guidance in datasheet to determine what filter time constant to use.

 

I ask because I would like to use larger R2-R3 if possible to save current when FET is on, and also be more sure of how to control in-rush.

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The opto doesn't have a minimum current as such. Why are you using an opto in the first place? Your circuit shows a common gnd, so a simple transistor would suffice.

If your load is capacitive, you can use a pwm channel for soft start. Write your code to ramp the duty cycle from 0 to 100 over a given time. Thus not needing the r and c to give a time constant.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

Common transistor good idea, I'm not sure why not.  I think i looked at collector-base voltages for some of them and they had limits

that would cause trouble with higher voltage batteries than 9V or something.  The hope is to have it work up to 24V at least.

 

Regarding soft start via PWM, sounds really good.  Will this approach work for inductive loads also (capacitive are anticipated by

the FET has a diode and I think can handle some inductive laod at least)?

 

What frequency would you advise for this application?  Would it be possible

to use PWM on the input side of the opto (since this part is happily working I'm not sure I want to change even if simple

transistor would suffice)?

 

Thanks,

Britton

Last Edited: Wed. Aug 12, 2015 - 01:29 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

bkerin,

 

I do not see the need for your elaborate power switch.  You are simply switching power to an Arduino and a cell shield power board". Ramping up the 9 volts to either board is likely to cause more problems than not. For example... a slowly rising power rail can defeat an R/C reset circuit.

 

If your power supply cannot cope with the "in-rush" current, how do you expect it to cope with your cell's transmit current needs?

 

Ross McKenzie ValuSoft Melbourne Australia

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm not sure what transistors you were looking at, something like a bc548 is good for 45V or so.
If you want to use 24V, you'll need some extra circuitry for the mosfet gate drive - Vgs is usually around 20V.
Pwm frequency? A tricky question. 500Hz might be good. You'll want a catch diode across the load for inductive loads.
An opto really only makes sense if you want galvanic isolation.

[edit] just read Ross' comments and i agree. Obviously i didn't look too closely at the schematic.

Last Edited: Wed. Aug 12, 2015 - 02:17 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

NTC Thermistors can be used for inrush current limit.

AMWEI Thermistor Sensor

thermistors-PTC NTC Sensor/Silicon PTC Thermistor Temperature Sensor, KTY84, 83, 81 cross.

www.amwei.com, sales@amwei.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

AMWEI Thermistor wrote:

NTC Thermistors can be used for inrush current limit.

 

If your semiconductor device can withstand the time delay!

 

Last Edited: Wed. Aug 12, 2015 - 10:50 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

valusoft wrote:

bkerin,

 

I do not see the need for your elaborate power switch.  You are simply switching power to an Arduino and a cell shield power board". Ramping up the 9 volts to either board is likely to cause more problems than not. For example... a slowly rising power rail can defeat an R/C reset circuit.

 

 

Ok.  It does work fine as it is, perhaps there's reason to mess with it.  I just don't like the idea of dumping a lot of juice through the FET to charge all the caps on the Arduino or other loads, it seems like just taking a guess that it can handle it is bad, but maybe all the caps are small enough that its not an issue.

 

Regarding reset and the like I'm aware of it.  The Arduino uses the most conservative fuse settings for power rise assumption and I wouldn't come close to those times for the ramp.

 

valusoft wrote:

 

If your power supply cannot cope with the "in-rush" current, how do you expect it to cope with your cell's transmit current needs?

 

 

Well the FET spec says it can take 34A burst, and the cell shield only wants 2A max, so that part is fine.  I had hopes of making this a general purpose board that could

drive some arbitrary small load.  Admittedly I'm a bit out of my depth trying to decide what e.g. in-rush protection is required and how to characterize acceptable loads.

It just occurred to me that some caps I've dealt with have milliohm ESR and so would presumably greatly exceed that 34A limit at turn-on.  In practice the battery internal

resistance would stop it, but I also decided it would be cute to add a barrel jack and started thinking about this.

 

So I guess my question should be "how much mitigation of such problems is reasonable, and how do I determine/specify the required load characteristics?"

 

Last Edited: Wed. Aug 12, 2015 - 05:47 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:

I'm not sure what transistors you were looking at, something like a bc548 is good for 45V or so.
If you want to use 24V, you'll need some extra circuitry for the mosfet gate drive - Vgs is usually around 20V.
Pwm frequency? A tricky question. 500Hz might be good. You'll want a catch diode across the load for inductive loads.
An opto really only makes sense if you want galvanic isolation.

[edit] just read Ross' comments and i agree. Obviously i didn't look too closely at the schematic.

 

bc548 does look good.  I get it on opto issue now thx.  24 V too much also very right thx.

Catch diode thx also.

 

In looking at the IRLML5401G spec, it seems that trying to turn it only slowly is really working

against its design.  It barely has any spec for its Rds variation with Vgs, just one graph at end

of datasheet and it looks like its as sharp as possible at Vgs threshold.  So I guess a whole

different circuit is required, or simply spec that capacitive loads with tiny ESR aren't allowed?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

On the UNO R3 there's only one cap you need to worry about, and that's the one before the voltage regulator, PC1, which is 47 uF.  Every other cap on the power bus is past the regulator and will be subject to the regulator's own current limiting feature, which is max 2.2A.

 

According to the BOM for the R3, PC1 is a EMVA250ADA470MF55G, although the UNO I have in front of me is probably a EMVA160ADA470ME55G (16V rather than 25V).

 

That datasheet specifies only the dissipation factor, not the ESR.  However the dissipation factor is the ratio of ESR to reactance.

DF = ESR/Xc

or

ESR = DF*Xc.

 

The datasheet specifies a DF of 0.16 @ 120 Hz for the 25V part.

 

We can calculate:

Xc = 1/(2πfC)

   = 1/(2π(120*47E-6))

   = about 28.2 ohms

 

So the ESR of PC1 is about 0.16 * 28.2 = 4.5 ohms.

 

At 24 V, the maximum inrush current you're likely to see due to PC1 is 5.3A.  With a TC of about 200 us (4.5 ohms * 47 uF), that would drop to 500 mA after 1 ms.

 

Even when you add the 2.2A max inrush for down-bus caps, this inrush current is likely to be swallowed by your power supply's output filter cap.  I really don't think you need to worry about this at all.

"Experience is what enables you to recognise a mistake the second time you make it."

"Good judgement comes from experience.  Experience comes from bad judgement."

"Wisdom is always wont to arrive late, and to be a little approximate on first possession."

"When you hear hoofbeats, think horses, not unicorns."

"Fast.  Cheap.  Good.  Pick two."

"Read a lot.  Write a lot."

"We see a lot of arses on handlebars around here." - [J Ekdahl]

 

Last Edited: Wed. Aug 12, 2015 - 06:44 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

More information concerning Power NTC thermistor for inrush current limiting can be found in the following website:

http://www.amwei.com/views.asp?hw_id=74

http://www.amwei.com/news.asp?news_id=85

AMWEI Thermistor Sensor

thermistors-PTC NTC Sensor/Silicon PTC Thermistor Temperature Sensor, KTY84, 83, 81 cross.

www.amwei.com, sales@amwei.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It probably goes without saying, but I'll mention it anyway:

 

If your PFet is spending much time in the linear mode, i.e. not fully switched On or Off, then the Rds will be higher, and the PFet will have to dissipate more heat.

If you slow your start up down too much, with the RC configuration, you will want to decide if your PFet needs a heat sink, or not.

(Probably not, but at least think about it!)

 

One of the advantages of the PWM method is that the PFet, although switching through the linear mode many times, is mostly in the switched On or switched Off state, and isn't running in the linear mode for any length of time.

 

But, as already noted, I'm not sure your circuitry will benefit from a slow power up.

 

JC

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

What is the total load current? There are optotriacs that can handle up to 300mA directly, so you can get rid of that FET and resistors in middle.

while(!solution) {patience--;}