How to prevent a circuit from cycling / oscillating

[Wetmelon]

How to prevent a circuit from "cycling" / oscillating

Hello everyone,

I have the following circuit I designed. Although I am open to all criticisms and suggestions, my main question is thus:

If the resistor 1.77Ω (load) becomes a dead short when the highlighted momentary switch is closed (high possibility), what's the best way to prevent this circuit from oscillating?

EDIT: I have done a bit of redesign and come up with a solution. Am I correct in thinking that if the 1.77Ω resistor is shunted to ground that the circled PFET gate will go to logic 0? Will it remain at logic 1 with the resistor in place? Does this design look sound? If not, what needs to be done to correct it?

 

[meBigGuy]

I don't see why it cycles. The switch turns on the fet, it conducts IDDsat. Not sure what you are trying to do. Explain what I am missing. Does Vdd sag? Power the switch from V+?

 

[Wetmelon]

I don't see why it cycles. The switch turns on the fet, it conducts IDDsat. Not sure what you are trying to do. Explain what I am missing. Does Vdd sag? Power the switch from V+?

When the fet turns on, it's expected that the 1.77Ω resistor explodes into a nice flame (70+ watts). If it creates an open circuit, great! If the leads create a short, it would then oscillate. See my edit for more info.

 

[meBigGuy]

Please explain exactly what you mean by it oscillates?

What are you trying to do? If I understood the point maybe I could suggest a way to avoid burning out all the FETS. What do want to have happen after the resistor opens or shorts?

 

[Wetmelon]

It oscillates during a short in the sense that the logic would turn the PFET off, then get pulled up, turn the PFET back on, turn it off, get pulled up, turn it on, etc.

Basically I'm trying to drive the 1.77Ω load, which is a "Quest Q2G2 Model Rocket Igniter". The igniter burns and hopefully creates an open circuit. In the event that it doesn't burn completely or something happens and it instead shorts together, I need to detect the short circuit and disable the high side driver. The maximum all-fire current (i.e. the igniter will always detonate) is something around 144mA, so I have to be careful about what method I use to detect a short. Hence the voltage-controlled FETs as opposed to current driven devices.

I did more designing and came up with a method to turn off the power in the case of a short, but it had an issue wherein the logic gates could switch faster than the FET (potentially) and lock out the FET, even if the load was in place.

There's an album available here if you like to see the revisions:

Imgur Album

I had one very effective version but because it too had the switching issue I didn't save it. Basically I started experimenting with logic gates instead of FETs since that's what they're designed for :P

 

[Baluncore]

What is the frequency or period of oscillation ?
What type of FETs are you using ?

 

[meBigGuy]

I don't quite understand the application yet. You say you want 144ma through a 1.77ohm igniter, but you also said it was 70W. 1.77ohms dissipates 70W at 6.2 amps (around 12V).

Put a small sense resistor is series with the igniter. Say 0.1 ohm which is 1V @ 10 amps. Use it to drive a latch that forces the igniter voltage off. When you release the trigger it should reset the latch. You can implement it with FET's.

 

[Wetmelon]

I don't quite understand the application yet. You say you want 144ma through a 1.77ohm igniter, but you also said it was 70W. 1.77Ω dissipates 70W at 6.2 amps (around 12V).

Apologies, the "max all-fire current" is the current at which the igniter/load, which is a piece of nichrome wire coated in gunpowder, will always ignite. The actual highest "detection" current that we can use is going to be around 70mA. Also, most other model rocket engine igniters are much lower resistance - the standard ones you buy in a hobby store are around .650Ω or lower. With a .1Ω resistor it'll dissipate what, about 25.6W in that configuration?

I think the key is in the latch, which I'll have to lookup how to build. It won't change until an input is pulsed, right? Or am I thinking of another design?

 

[meBigGuy]

Maybe I have some idea of what you want.

A circuit to apply 12V to a model rocket igniter through a long cable and connectors.
Cable and connector resistance is unknown (but is significant)
Igniter resistance can range from 0.65 to 1.77 ohms (?)
Up to 70ma can safely be used to verify continuity before launch.
If the igniter shorts when blown, you want the power to be removed (not sure why).
If the igniter is initially shorted before firing (faulty setup/cable), there should be an indication (I just added that)
This should run from an as yet unspecified battery.
There should be an arm switch to activate the launch button.

Anything else?

As for the latch circuit, search for "push button activated power fet" on google images and look how they use the switched-on voltage to hold the power on. That is an example of latching circuit. You want to figure out how to use that general feedback concept to allow the short circuit current detection to force the power off. (hint: you may need to stack two N channel fets to get an AND function and use resistors to get an OR function)

Personally I would tend to go with a super simple "switch the battery across the igniter with a big switch" approach. And detect continuity with an led/resistor across the switch.

 

[Baluncore]

The requirements appear identical to the box used to test and fire electric detonators.

 

[Wetmelon]

Thanks for the continued support, BigGuy, it is appreciated!

Maybe I have some idea of what you want.

A circuit to apply 12V to a model rocket igniter through a long cable and connectors.
Cable and connector resistance is unknown (but is significant)

Ah yes, I have been neglecting the cable resistances. Good catch.

Igniter resistance can range from 0.65 to 1.77 ohms (?)

Roughly, yeah.

Up to 70ma can safely be used to verify continuity before launch.

Precisely.

If the igniter shorts when blown, you want the power to be removed (not sure why).

Simply because I don't want to burn out the FETs and to help save the battery.

If the igniter is initially shorted before firing (faulty setup/cable), there should be an indication (I just added that)

This would be very helpful, yes.

This should run from an as yet unspecified battery.

Test battery is a 3 cell 11.1V (nominal) 1600mAh 20C LiPO (32 amp peak).

There should be an arm switch to activate the launch button.

Yep, that's the switch up top.

Anything else?

No, that about covers it. Preferably it would be made of passive electronics, but I'm starting to think a small PIC would be the best option. I also have a Javelin Stamp kicking around, which can theoretically program PICs but I'm not sure how to do it.

As for the latch circuit, search for "push button activated power fet" on google images and look how they use the switched-on voltage to hold the power on. That is an example of latching circuit. You want to figure out how to use that general feedback concept to allow the short circuit current detection to force the power off. (hint: you may need to stack two N channel fets to get an AND function and use resistors to get an OR function)

you may need to stack two N channel fets to get an AND function

I actually had this design at one point - two N Channels gated to different lines - one on the AND gate, one on the pull-down resistor. I think I just decided wrongly that it was getting too rediculous.

Personally I would tend to go with a super simple "switch the battery across the igniter with a big switch" approach. And detect continuity with an led/resistor across the switch.

Well I have that one already. I wanted to do this circuit as a fun project and also to miniaturize it for future "airstarts" - second stage ignition based on an altimeter.

The requirements appear identical to the box used to test and fire electric detonators.

Ematches? Yeah, a lot of rocket guys use Ematches as opposed to nichrome bridge igniters.

To prove I'm in the rocket stuff (as opposed to the explosive stuff), here's my CURRENT launch controller video (xposted to Rocketryforum.com):

Thanks again for the help, guys! :)

 

[meBigGuy]

Another thing you need is surge protection since you are driving high currents through long cables. When the igniter opens the field will cause a high voltage that could damage your components.

Ignoring the logic for a minute, detecting a short vs. the normal surge while the igniter blows may be difficult with long cables. If you add "thevenin" sense connections you can detect the voltage across the igniter more accurately. This takes a 4 conductor cable though.

 

[Wetmelon]

I.e back emf protection? No problem