Flight controller -Electrical checking

The circuit will output the set current if possible. If it can't reach that current, it will give as much as possible. If your load resistance is low enough, then it can deliver 10 amps, but that resistance will have to be pretty low. You will lose about 1 volt over Q2 (which is a darlington transistor, not two stacked on op of each other) and 0.7 volts over R1 (0.7V is a normal Vth) so if your supply is 9V, you have a maximum of 7.3V that can be supplied to the load. So if your load resistance is 1 ohm, it will never give more than 7.3 amps, even if you set R1 for a higher value.

If you just want to push a ton of current, this isn't necessary. I've heard some people say that delivering too much current to the igniter can melt the bridgewire so fast it won't heat the explosive before breaking. I'm not sure if this is true though. Other then that, such a circuit would only be good for limiting current consumption or powering current sensitive loads (i.e. glow plugs).

Originally posted by mtwieg
...
(which is a darlington transistor, not two stacked on op of each other) [/B]

Click to expand...

I know that much

Originally posted by mtwieg
If you just want to push a ton of current, this isn't necessary. I've heard some people say that delivering too much current to the igniter can melt the bridgewire so fast it won't heat the explosive before breaking. I'm not sure if this is true though. Other then that, such a circuit would only be good for limiting current consumption or powering current sensitive loads (i.e. glow plugs). [/B]

Click to expand...

What I would like is a circuit that can provide whatever current an ematch (or other ignitor) needs to fire, but it doesn't need to push lots of current...I just dont want it to blow up when an ignitior draws 10A

So I guess its an 'as needed' type circuit rather than forcing current down the ignitors throat

Edit: Forgot to say with a non-MCU controlled failsafe.

Originally posted by basil4j

What I would like is a circuit that can provide whatever current an ematch (or other ignitor) needs to fire, but it doesn't need to push lots of current...I just dont want it to blow up when an ignitior draws 10A

So I guess its an 'as needed' type circuit rather than forcing current down the ignitors throat

Click to expand...

Well, then you have a couple options. The circuit has no way of knowing what the current requirement is, so you would need to program it somehow. I can think of some ways, but none of them are really nice. If you stuck with the circuit I showed, you could do this by replacing R1 with a few smaller resistors and shunting some of them, thus changing the total R1.

You could also do something with an op amp set up as a voltage to current converter. You could make a rough D/A converter by taking digital outputs and connecting them together through different resistances. I'll have to simulate this one to know exactly what will work.

A much simpler (I think) possibility is just having the output current increase over time. It will just stop when the pyro blows, no programming necesarry. The disadvantage would be that higher current loads would take longer to blow.

I'm working on a firing device, and I'm trying to come up with the best way to have it selectable to either limit its current to a set value or just give the maximum. That way it could work with glow plugs and low current batteries as well as demanding loads.

How about a coil inline with the load to limit the current spike? If sized right the ematch will fire while the current is ramping up.

You mean an inductor? I can't think of a reason it wouldn't work. If it does, then that might be the best solution. I imagine it would oscillate somewhat, which might interfere with the power supplies. I'll have to work it out on paper...

Originally posted by mtwieg
Well, then you have a couple options. The circuit has no way of knowing what the current requirement is, so you would need to program it somehow. I can think of some ways, but none of them are really nice. If you stuck with the circuit I showed, you could do this by replacing R1 with a few smaller resistors and shunting some of them, thus changing the total R1.

You could also do something with an op amp set up as a voltage to current converter. You could make a rough D/A converter by taking digital outputs and connecting them together through different resistances. I'll have to simulate this one to know exactly what will work.

A much simpler (I think) possibility is just having the output current increase over time. It will just stop when the pyro blows, no programming necesarry. The disadvantage would be that higher current loads would take longer to blow.

Click to expand...

Hmmmm. Lots of possibilities, I thought the firing circuit would be the easiest to implement!

How do exisiting flight controllers/timers do this?

Originally posted by jderimig
How about a coil inline with the load to limit the current spike? If sized right the ematch will fire while the current is ramping up.

Click to expand...

I don't know much about inductors but that sounds good Has this ever been done before?

Originally posted by basil4j
Hmmmm. Lots of possibilities, I thought the firing circuit would be the easiest to implement!

Click to expand...

It can be, if you want. Personally, I agree with David that reliability lies with simplicity. The best failsafe measures are things like shunts and switches, which are easy to inspect and replace and are very durable. If you just have a big switch and one FET, then most people have nothing to complain about. The more sophisticated circuits are meant for certain types of loads which not many people use. If I were you, I wouldn't worry about making something that satisfies everybody's needs, especially on your first design.

How do exisiting flight controllers/timers do this?

Click to expand...

As far as I know, most use a a setup like the one in your schematic (sans failsafe transistors). I'm sure there are exceptions, like in the G-wiz computers, which are extremely sophisticated. Good luck on trying to get anyone to give details, though.

And about using an inductor, the big question is what happen when the current stops suddenly. It will produce a very large reverse voltage spike, which will likely reek havoc on the mosfet (if not the entire circuit) if you don't deal with it somehow.

Originally posted by mtwieg


And about using an inductor, the big question is what happen when the current stops suddenly. It will produce a very large reverse voltage spike, which will likely reek havoc on the mosfet (if not the entire circuit) if you don't deal with it somehow.

Click to expand...

You put a diode across in the reverse direction.

However I agree is go simple. Reliability is more important than "features" in rocket recovery (IMHO). Find a design that minimizes component count and go with that.

Originally posted by mtwieg
It can be, if you want. Personally, I agree with David that reliability lies with simplicity. The best failsafe measures are things like shunts and switches, which are easy to inspect and replace and are very durable. If you just have a big switch and one FET, then most people have nothing to complain about. The more sophisticated circuits are meant for certain types of loads which not many people use. If I were you, I wouldn't worry about making something that satisfies everybody's needs, especially on your first design.
As far as I know, most use a a setup like the one in your schematic (sans failsafe transistors). I'm sure there are exceptions, like in the G-wiz computers, which are extremely sophisticated. Good luck on trying to get anyone to give details, though.

Click to expand...

Originally posted by jderimig
However I agree is go simple. Reliability is more important than "features" in rocket recovery (IMHO). Find a design that minimizes component count and go with that.

Click to expand...

Thats what Im beginning to think myself, I don't have the knowledge to make a one size fits all device, and I can add high current outputs with the expansion modules I am going to implement.

So im summary I can use the firing circuit attached, using FET's rather than NPN trans. (so as not to need a resistor to limit current through the NPN).

To reduce the number of components further, I could loose th 4 failsafe FETs, and just put 1 FET between 9V and the sources on the firing FET's, is this correct?

Originally posted by basil4j
To reduce the number of components further, I could loose th 4 failsafe FETs, and just put 1 FET between 9V and the sources on the firing FET's, is this correct?

Click to expand...

I don’t think this will work because all the gates of the firing FETs will then be connected together, but maybe I haven’t understood your idea. Can you post schematics of this idea?

Well, at least you can make your circuit a little bit simpler when you replace your failsave transistors with a diode and put the failsave switch between GND and the diode. This would limit Vgs to about 0.7V when the failsave switch is closed. Check back with the firing FET datasheet, to ensure its Vth is above 0.7V. Otherwise it wouldn’t work. You can use shottky diodes to reduce this voltage.

I too had some time to think about similar problems some weeks ago. My approach is a little different. The labeled contacts are connected to an ATmega88, therefore it may require some changes when used with other microcontrollers having different IO ports. This was the simplest solution that I found which offers the capabilities I was looking for: ARM/SAFE not affected by defective parts (especially the FETs), ARM/SAFE independent from software, ability to detect ARM/SAFE, continuity check in both states (armed, safe), no fire during reset and no big currents into the controller in case of software errors. The circuit is only suitable for ematches and other low power igniters that can be ignited with some 12.5mJ (1000µF@5V). Of course its quite easy to modify this part.
The 1k resistors (R26-R28) aren’t really necessary. But they prevent the µC from damage, that may be caused by software errors and they eliminate single points of failure (regarding accidental firing of igniters) . For example, a shorted R2 may otherwise cause the igniter to fire, even with flawlessly working software and the arming switch in “Safe” Position.

Reinhard

Forgot the attachment...

Originally posted by jderimig
You put a diode across in the reverse direction.

Click to expand...

You mean in parallel with the inductor? That's a good idea. It would have to be a pretty big diode though.

Originally posted by basil4j

To reduce the number of components further, I could loose th 4 failsafe FETs, and just put 1 FET between 9V and the sources on the firing FET's, is this correct?

Click to expand...

Um, the source is at ground, so doing that would effectively short the entire thing.

Originally posted by Reinhard

Well, at least you can make your circuit a little bit simpler when you replace your failsave transistors with a diode and put the failsave switch between GND and the diode. This would limit Vgs to about 0.7V when the failsave switch is closed. Check back with the firing FET datasheet, to ensure its Vth is above 0.7V. Otherwise it wouldn’t work. You can use shottky diodes to reduce this voltage.

Click to expand...

Since he's firing off of 3 volts, he needs a low Vth mosfet. His is rated at 0.4 - 0.7 V, so even schottkys are cutting it close. The current is really small, so it should probably be below 0.4, but applying even a couple tenths of a volt makes me nervous.

The 1k resistors (R26-R28) aren’t really necessary. But they prevent the µC from damage, that may be caused by software errors and they eliminate single points of failure (regarding accidental firing of igniters) . For example, a shorted R2 may otherwise cause the igniter to fire, even with flawlessly working software and the arming switch in “Safe” Position.
Reinhard

Click to expand...

That's pretty much exactly what I'm using for my dual event timer, except I don't have an arm/safe contact, or the gate pull down resistors. For designs like mine and Alec's where the battery voltage is higher than the MCU voltage, R26 and R27 are necessary to divide the voltage down to a level that the MCU can tolerate.

What do you use for the safe/arm contact? I want to have one, but I want it to be small and mountable on a pcb. I was thinking a simple shunt and a couple header pins, but I don't think they're meant to take such high currents.

Originally posted by mtwieg
Since he's firing off of 3 volts, he needs a low Vth mosfet. His is rated at 0.4 - 0.7 V, so even schottkys are cutting it close. The current is really small, so it should probably be below 0.4, but applying even a couple tenths of a volt makes me nervous.

Click to expand...

You’re right. I should have looked into the datasheet. I guess its quite likely, that the current wouldn’t be enough to fire the igniter, but even then it would be a bad design, because It draws more current than necessary.

That's pretty much exactly what I'm using for my dual event timer, except I don't have an arm/safe contact, or the gate pull down resistors

Click to expand...

Maybe I’m paranoid. The gate pull down resistors are only there, to ensure that a controller reset, when the IOs go tristate, can not cause the igniter to fire.

What do you use for the safe/arm contact? I want to have one, but I want it to be small and mountable on a pcb. I was thinking a simple shunt and a couple header pins, but I don't think they're meant to take such high currents.

Click to expand...

I will use terminal blocks. Something in the 3.5mm range. They don’t need to much space, and the solution is quite flexible. On the downside, this means an external switch or whatever else is needed.

Reinhard

Forget what I said about the single FET, Im at work and don't have Eagle to play with...my brain saw it differently to reality

That schematic looks good and simple And if it is used by both of you and proven to work then I might as well go with it!

Only question, without the 1000uf cap, will this still function? I would like to connect directly to the battery so as not to have a charge time. If not, how long does the cap take to charge to capacity? (I know the formula, but its not coming to me right now)

I assume using a larger cap would allow it to fire higher current devices to operate.

If I was to replace the FET's with the FDS6064N7's I need to use (low Vth), I take it this would be the only change needed? I.e, no resistor changes?

Regarding the fail safe connectors, I too am using 3.5 (or 3.81mm) PCB terminals commonly used in industry. They are rated at 10A and they are nice and small. I work at a connector company so can get these free, but they only cost under $1NZ anyway

Originally posted by basil4j
Only question, without the 1000uf cap, will this still function? I would like to connect directly to the battery so as not to have a charge time. If not, how long does the cap take to charge to capacity? (I know the formula, but its not coming to me right now)

Click to expand...

You can do it without the cap, but you will also need to remove the 220 ohm resistor (R1). Capacitor discharge is usually a good idea when you use a battery that can't give lots of current. My timer uses an A23 (12 volt, very small size, low current) which couldn't hope to fire igniters normally, but can charge a large capacitor enough to fire off seven ematches.

If I was to replace the FET's with the FDS6064N7's I need to use (low Vth), I take it this would be the only change needed? I.e, no resistor changes?

Click to expand...

Aren't the FETs in the picture already FDS606n7's?

Aside from that, not really. I would use higher resistances for the resistors from the MCU to the gates (at least 1K).

Originally posted by mtwieg
You can do it without the cap, but you will also need to remove the 220 ohm resistor (R1). Capacitor discharge is usually a good idea when you use a battery that can't give lots of current. My timer uses an A23 (12 volt, very small size, low current) which couldn't hope to fire igniters normally, but can charge a large capacitor enough to fire off seven ematches.

Click to expand...

7 ematches should be enough for most people!
I am planning on using a 9V battery. If this will provide enough current on its own then ill go with 9V, otherwise I was going to use the A23 with a big cap due to its light weight.

Originally posted by mtwieg
Aren't the FETs in the picture already FDS606n7's?

Click to expand...

Ummm, are you refering to the schematic I posted or Reinhard's?
I was refering to Reinhard's

Originally posted by basil4j
7 ematches should be enough for most people!
I am planning on using a 9V battery. If this will provide enough current on its own then ill go with 9V, otherwise I was going to use the A23 with a big cap due to its light weight.

Click to expand...

I would assume that people who invest in a full fledged flight computer are going to want to use there own batteries, and they're going to want them to be high current. An A23 would be very unsuited for a flight computer not only because of its low current, but its short lifespan. My timer uses about 2.5ma, and the battery operates it for around 20 hours. Your design is going to use much more current (probably over 20ma), so it won't last nearly long enough. Just give it a terminal block and circuitry that can accommodate a wide range of battery types.

Ummm, are you refering to the schematic I posted or Reinhard's?
I was refering to Reinhard's

Click to expand...

In that case, you will also need to adjust the resistors for the continuity check so the voltage into the MCU is below 3V when there is continuity (note, his design will not work with resistance checks in that exact configuration). So R26 and R2 as well as R27 and R4 form dividers (1/4 is a good ratio).

Then I’d suggest to calculate the voltage divider for the lowest expected battery voltage and add a z-diode (parallel to R2 and R4) to limit the voltage to protect the controller.

Reinhard

Alec, where do you plan on buying the adxl78 from?

To my horror, the accelerometer I'm using just went obsolete (easn't rohs compliant) and apparently no distributors carry it yet. I'm thinking of switching to Analog devices accelerometers, but I can't find a place to buy those either.

EDIT: Digikey has the adxl321, which is okay for me, but not much else. I think you had better rethink your accelerometer choice, since no distributors appear to sell it in reasonable quantities.

The analog devices ones are more expensive, but some of their specs are nicer... adjustable bandwidth, smaller size, large ratiometric voltage supply range, lower quiescent current, and larger acceleration range (not necesarrily good...). Maybe the increase in cost will be worth it.

The ADXL78 is distributed by Avnet here in New Zealand, I think they are global though so you could try them. There are MOQ's on the device buts its pretty low (around 20 pieces from memory). It was around $20NZ, which I think is about $12USD at the moment.

Edit: Post deleted

Originally posted by mtwieg
In that case, you will also need to adjust the resistors for the continuity check so the voltage into the MCU is below 3V when there is continuity (note, his design will not work with resistance checks in that exact configuration). So R26 and R2 as well as R27 and R4 form dividers (1/4 is a good ratio).

Click to expand...

Oh ok. Resistors R26 & R2, and R27 & R4, would they happen to have the same purpose as the ones ive circled in the attached drawing which you posted earlier? I see a remarkable similarity, (except that the R26/27 resistor is on the ADC side of R2/R4 rather than the FET).

Originally posted by Reinhard
Then I’d suggest to calculate the voltage divider for the lowest expected battery voltage and add a z-diode (parallel to R2 and R4) to limit the voltage to protect the controller.

Click to expand...

Ive quickly added the diode into the drawing also

EDIT: Original file was corrupt, please see post below for fixed one.

I think the pictures are corrupt. They're mostly gray area.

Originally posted by mtwieg
I think the pictures are corrupt. They're mostly gray area.

Click to expand...

Right then, lets try again

Originally posted by basil4j
Right then, lets try again

Click to expand...

Wait, what? The diode was for Reinhard's design, which only gives a continuity check, not a resistance check. If you really want the resistance check, you don't need the diode, since the op amp effectively isolates the firing circuit from the MCU. Everything we've dicussed in the last several posts have been aimed at using continuity testing, which it seems you don't want.

Originally posted by mtwieg
Wait, what? The diode was for Reinhard's design, which only gives a continuity check, not a resistance check. If you really want the resistance check, you don't need the diode, since the op amp effectively isolates the firing circuit from the MCU. Everything we've dicussed in the last several posts have been aimed at using continuity testing, which it seems you don't want.

Click to expand...

Oh lol. I didn't pick up on that sorry!

Do you think the resistance check is really needed? Or am I wasting all of your time? (you've been very helpful by the way!)

I mean, how often will a simple continuity test provide a false result?

I don't mind putting it in, but if most altimeters that are currently in use only use a simple con't test, then why change what isn't broken?

Ok, I have attached a schematic which I HOPE will give me resistance checking and failsafe. Its kind of a merged version of yours and Reinhard's circuits. Please feel free to pull it down or whatever

(Boy you guys have alot of patience!! )

I'd like to trust people to take the trouble to inspect their pyro clips/contacts or whatever they use for crud and corrosion. I'd like to think a simple continuity test is adequate. Most people are only going to expect continuity testing, but a few may be impressed enough by a resistance check to be turned towards your device. I think it just comes down to how much they have to pay for it in the end. I can't really say just yet, but I imagine just that feature is going to add around $7 (US) to the price of parts.

As for the schematic, you've got the gist of it, except:
1) the op amp configuration is wrong (I think I got it wrong in the one I showed since I forgot it was inverting). They need to be operated on the 9V line, and therefore the outputs need to be divided down (another thing I missed. When there is no continuity, the amp would be saturated at 9V, which would kill the MCU). Also, you need to switch the inputs to the op amp.

2) You don't need the divider pairs (R12 + R6, R13 + R8, R14 + R8).

Here's a fixd one:

Keep in mind that all resistor values are rough estimates. You'll need to adjust them yourself. Also, the OP07s will not work. They need dual supplies and I don't think they are rail to rail. The OP7xx series is very similar and does not have these limitations, but it is more expensive (3.50 USD for a dual package). Either these or the CS3014 I mentioned would probably be your best bet.

Also, about what I said earlier about putting filtering on the ADXL output, scratch that. You need a single capacitor to ground, with no resistor. This sets the bandwidth and resolution of the sensor. The datasheet wxplains it well.