Wondering if anyone might have experienced a problem that I'm having. My ArduPilot has been working fine, but yesterday, the servo outputs suddenly quit working. After doing some troubleshooting, I think I've narrowed the problem down the the multiplexer (74ACT157). It gets power, and gets a proper signal from the failsafe, and I found a 50 Hz signal on the inputs. However, the outputs are always low, at 0V.I've been using a 4-cell, 4.8V receiver battery to power my setup. When fully charged, it gets up to 5.3V, which should be safe for the multiplexer. Maybe I shorted out one of the outputs, or zapped it with static? Anyone else had a similar problem?
I also fried the 74ACT157, I replaced it with a 74HCT157 (cmos), now my servo's run directly to the corner. What could be wrong? Perhaps the gate stays down or something like that. I try a 10k pullup resistor but no positive result.
I have never fried anything with ESD and I have taken very little effort in my 5 years of tinkering. I would suspect bGatti's claims and serch out any scenario where you put the MUX out of spec that or it simply went bad. Really a 50cent part went bad.... Dang it gonna break the bank. Enjoy the search!
@Damon,
Now we're getting closer.
You're elevons don't have a ground wire?, They are floating until you plug something else in. If the next thing has a ground first plug, or it's put in sideways etc, you can definitely push 12x into the board.
Ben
Ben,
Thanks for the tip. I'll try that out when I receive my replacement AP. I tested the broken one, and the voltages are exactly the same, but that doesn't mean much since it is already fried.
The humidity varies a lot here, but during summer its probably humid more often than not. So static is not usually a problem, but the foam on this plane does seem to generate some static. I know, I don't really have an quantifiable evidence that points to static.
The only situation I can think of where the AP could receive a signal when not powered is from the elevon signals. However, in that case, ONLY the signals would have been connected, and the ground and VCC of the
AP would be floating. If those inputs were at different levels, maybe this could cause a problem, but I'm not familiar enough with CMOS logic to know.
the test is easy.
disconnect the receiver from the AP,
measure the AC voltage of the servo signals. reconnect the AP and remeasure the AC of the servo signals. The change is being absorbed by the MUX.
Ben
@Damon,
What's the humidity like where you live?
- I take you're point on the battery set-up.
Essentially you're ruling out the idea of the receiver being powered up and connected to the AP without the ap also being powered up. That would take the 12x overvoltage out of the equation.
I would take a second look at how you're doing that and confirm that there isn't any combination of connections which pushes a receiver signal at the AP without it being powered.
Next we're ack to the issue of the regulator pulling down the VDD on the mux. This is a small overvoltage, but might as well rule that out. Explaining everything with ESD, is about like explaining everything with GOD in my book, neither one of them is a persuasive answer.
Since the schematic (ArduPilot-V15) for the Ardupilot on the Sparkfun website does not include the ATTiny nor the Multiplexer, I going to assume that all of the ICs on the Ardupilot board are powered by the receiver throught the servo inputs. The only time that the onboard regulator is used is when the board is powered through the battery input and the onboard jumper is moved from the servo to the battery position.
Most RC manufactures are explicit about powering the receiver with only 4 rechargeable batteries and not 4 AA cells. The NMHi batteries ususally charge up to ~1.3 v each resulting in a 4 pack voltage of around ~5.2 v which is well within spec. If the receiver is powered by the 4 pack and the Ardupilot is powered by the receiver, there should be no voltage differential between the receiver and the Ardupilot. If the board is powered by the onboard regulator from a separate battery and the receiver voltage is greater than say 5.5 v, which is possible with present day receivers, there definitely could be an issue.
I should clarify how I have my ArduPilot setup. I am not powering the AP separately. I have a 3-pin female to female lead to connect the receiver to the CTRL input on the AP. This connects the power buses on the receiver and AP. I then use jumpers to connect the 2 servo signals for the elevons. I connect my 4.8 V receiver battery to either an extra servo input on the AP, or to the revceiver. Either way accomplishes the same thing, as the power buses are connected together. So, because of how I connected my system together, I do not believe that I violated the specs on the MUX. I guess that it could be an issue when the board the battery is being connected or disconnected, but it seems unlikely to be the problem, as I have been working with it like this for several weeks.
The voltage from my battery varies from maybe 4.7 to 5.3 volts. This is within the spec of all the devices on the AP. I am assuming that the servo signals from the receiver are not boosted above the supply voltage to the receiver. My setup is almost identical to that of someone using a BEC, with the exception that my battery is not regulated to 5 V. A typical BEC can source 2 amps, while my battery can probably source more if it is shorted. However, 2 amps is way more current than it would take to fry the MUX, so the current limit from the BEC is not protecting the MUX.
I believe that the MUX was damaged due to ESD. I have not be very careful about ESD protection, and have been playing with the board on the bench for several weeks, and it has worked just fine. However, I recently installed the board in a Zagi (foam) and brought it outside for some ground testing, after which the multiplexer was fried. The combination of a foam plane and wind outside leads me to believe that ESD was probably the cause.
If you are using what Chris Anderson calls the typical setup: Battery -> ESC/BEC -> Receiver -> ArduPilot, I believe that you should not have to worry about over voltage problems with the MUX. You certainly should take care not to short the outputs, as that could easily damage the MUX.
@Ryan,
This would only effect people who use separate power, with separate switches or connections, and who turn on the power in the unlucky order. Presumably the time to failure would be a factor, then general unluckiness, receiver source current, and a few other factors, but when you are 12x the published factors.
What I have tried to do here is to determine the most likely cause of the problem on the main post. I have identified a 12x over voltage which fits with the facts as known. No one else has posited an explanation which could reproduce a sustained 12x overrating. Chris is suggesting a 0x overvoltage explanation - which I reject. This is the better explanation, until a better one comes along - IMO.
5.1 should be fine - although it technically depends on the droppyness of the Ardupilot regulator. Frankly, I think the power situation is a mess. You can't stack regulators, ie you can't have a 5v reg powered by a 5v reg.
Presumably, the Ardupilot is a customer of a regulated power supply. It ought to behave as one (ie and operate at the same voltage level.) Instead, it drops the voltage through a regulator, and then violates the specs on every pin. That's a mess. It might do better to clamp the input voltage, use good conditioning, include a fuseable part to protect the clamping diode, but in general to operate at the system voltage - particularly with respect to the mux which has all of its pins exposed, and is flight-supercritical.
Basically - if you are powering the Ardupilot from a 5.1 v BES (or ESC/BEC), the voltages and startup timing are close to spec (presuming the dropout on the Ardupilot reg is less than .4v)
The real problem occurs when using a separate power supply for the Ardupilot. Here, you'll want to power-on the AP before the receiver.
Comments
Now we're getting closer.
You're elevons don't have a ground wire?, They are floating until you plug something else in. If the next thing has a ground first plug, or it's put in sideways etc, you can definitely push 12x into the board.
Ben
Thanks for the tip. I'll try that out when I receive my replacement AP. I tested the broken one, and the voltages are exactly the same, but that doesn't mean much since it is already fried.
The humidity varies a lot here, but during summer its probably humid more often than not. So static is not usually a problem, but the foam on this plane does seem to generate some static. I know, I don't really have an quantifiable evidence that points to static.
The only situation I can think of where the AP could receive a signal when not powered is from the elevon signals. However, in that case, ONLY the signals would have been connected, and the ground and VCC of the
AP would be floating. If those inputs were at different levels, maybe this could cause a problem, but I'm not familiar enough with CMOS logic to know.
disconnect the receiver from the AP,
measure the AC voltage of the servo signals. reconnect the AP and remeasure the AC of the servo signals. The change is being absorbed by the MUX.
Ben
What's the humidity like where you live?
- I take you're point on the battery set-up.
Essentially you're ruling out the idea of the receiver being powered up and connected to the AP without the ap also being powered up. That would take the 12x overvoltage out of the equation.
I would take a second look at how you're doing that and confirm that there isn't any combination of connections which pushes a receiver signal at the AP without it being powered.
Next we're ack to the issue of the regulator pulling down the VDD on the mux. This is a small overvoltage, but might as well rule that out. Explaining everything with ESD, is about like explaining everything with GOD in my book, neither one of them is a persuasive answer.
Since the schematic (ArduPilot-V15) for the Ardupilot on the Sparkfun website does not include the ATTiny nor the Multiplexer, I going to assume that all of the ICs on the Ardupilot board are powered by the receiver throught the servo inputs. The only time that the onboard regulator is used is when the board is powered through the battery input and the onboard jumper is moved from the servo to the battery position.
Most RC manufactures are explicit about powering the receiver with only 4 rechargeable batteries and not 4 AA cells. The NMHi batteries ususally charge up to ~1.3 v each resulting in a 4 pack voltage of around ~5.2 v which is well within spec. If the receiver is powered by the 4 pack and the Ardupilot is powered by the receiver, there should be no voltage differential between the receiver and the Ardupilot. If the board is powered by the onboard regulator from a separate battery and the receiver voltage is greater than say 5.5 v, which is possible with present day receivers, there definitely could be an issue.
Just a thought.
Regards,
TCIII
The voltage from my battery varies from maybe 4.7 to 5.3 volts. This is within the spec of all the devices on the AP. I am assuming that the servo signals from the receiver are not boosted above the supply voltage to the receiver. My setup is almost identical to that of someone using a BEC, with the exception that my battery is not regulated to 5 V. A typical BEC can source 2 amps, while my battery can probably source more if it is shorted. However, 2 amps is way more current than it would take to fry the MUX, so the current limit from the BEC is not protecting the MUX.
I believe that the MUX was damaged due to ESD. I have not be very careful about ESD protection, and have been playing with the board on the bench for several weeks, and it has worked just fine. However, I recently installed the board in a Zagi (foam) and brought it outside for some ground testing, after which the multiplexer was fried. The combination of a foam plane and wind outside leads me to believe that ESD was probably the cause.
If you are using what Chris Anderson calls the typical setup: Battery -> ESC/BEC -> Receiver -> ArduPilot, I believe that you should not have to worry about over voltage problems with the MUX. You certainly should take care not to short the outputs, as that could easily damage the MUX.
This would only effect people who use separate power, with separate switches or connections, and who turn on the power in the unlucky order. Presumably the time to failure would be a factor, then general unluckiness, receiver source current, and a few other factors, but when you are 12x the published factors.
What I have tried to do here is to determine the most likely cause of the problem on the main post. I have identified a 12x over voltage which fits with the facts as known. No one else has posited an explanation which could reproduce a sustained 12x overrating. Chris is suggesting a 0x overvoltage explanation - which I reject. This is the better explanation, until a better one comes along - IMO.
Presumably, the Ardupilot is a customer of a regulated power supply. It ought to behave as one (ie and operate at the same voltage level.) Instead, it drops the voltage through a regulator, and then violates the specs on every pin. That's a mess. It might do better to clamp the input voltage, use good conditioning, include a fuseable part to protect the clamping diode, but in general to operate at the system voltage - particularly with respect to the mux which has all of its pins exposed, and is flight-supercritical.
Basically - if you are powering the Ardupilot from a 5.1 v BES (or ESC/BEC), the voltages and startup timing are close to spec (presuming the dropout on the Ardupilot reg is less than .4v)
The real problem occurs when using a separate power supply for the Ardupilot. Here, you'll want to power-on the AP before the receiver.
-
1
-
2
-
3
of 3 Next