Clock not keeping time

V

Vern Little(Guest)

Guest
My D10A (SN 3034) does not keep time.

Originally, I installed it with the keep-alive power connected, but it was draining my main battery, so I disconnected it. I have the internal backup battery option, which is currently showing 15.1 Volts.

The EFIS is not keeping time correctly. If I turn it off for short periods and check again it seems ok, but if I leave it overnight, it loses several hours.

Vern Little
 

dynonsupport

Dynon Technical Support
Staff member
Joined
Mar 23, 2005
Messages
13,226
15.1 Volts represents a partially, but not fully charged internal battery, though even a low battery should power the clock indefinitely.

A few questions:

Are you sure the EFIS is losing time, or is it perhaps resetting to 12:00 at some point?
How much time is the EFIS gaining/losing?
Is the error repeatable or does it vary?
Is the battery always show 15.1 Volts, or does it creep up when powered externally?
 
V

Vern Little(Guest)

Guest
Thanks for the quick response.

I'm not sure about the resetting. Yesterday, when I checked after overnight, it had lost seven hours and some minutes. Today, when I checked it has lost a total of 10 hours and some minutes (could be more than that and rolled over).

I'll run an experiment on repeatability.

I can't power it up externally just yet because the panel is out of the aircraft. I have noticed that the battery does charge up well over 16 volts when power is applied.

Vern Little
 
V

Vern Little(Guest)

Guest
I have not yet been able to replicate the several hours of errors, but I have found what looks like a temperature instability in the real time clock.

I plotted the clock error vs. time and found about a +2 second per hour error during the evening and morning with the house thermostat set at 70F, but very little error overnight when the house thermostat was set back to 62F.

Dynon%20Clock%20Error.jpg


This indicates a temperature instability that could lead to errors of several minutes per week or more, depending on ambient temperatures. It may also indicate that the oscillator is unstable... which could explain why I get several (-) hours of errors from time to time (oscillator stopping?).

The graph starts at 1750 (5:50 pm). You can see the overnight stability compared to the evening/morning errors.

Is this helpful?
 

dynonsupport

Dynon Technical Support
Staff member
Joined
Mar 23, 2005
Messages
13,226
Vern,

We definitely expect better performance than what you've graphed and actually test each EFIS's clock performance during burn-in and calibration. It sounds like there is something wrong with your unit. If you give us a call at 425-402-0553 we can arrange for you to send it in for repair under warranty.
 
V

Vern Little(Guest)

Guest
OK, I returned my unit for service and received it back yesterday. I had asked for repair and calibration.

I ran some tests, and discovered that the unit is still running fast, however, I have found the problem.

Every time I power the unit on, it gains one or two seconds in time. That explains my previous graph, where the clock did not gain much time overnight, but would gain time during the day when I was powering it up to check. I think that this is the Heisenberg uncertainty principal at work-- the process of measuring it changes it :)

I can reproduce this problem reliably now. It sounds like a software problem to me, at least I hope it can be fixed in software.

I look forward to a confirmation and a planned bug fix.

Vern Little
 

dynonsupport

Dynon Technical Support
Staff member
Joined
Mar 23, 2005
Messages
13,226
Hi Vern,

We have done similar testing here and have noticed the same behavior: power cycling the unit will cause the clock to gain a small amount of time, generally between 0-1 second as far as we can tell. We know it has to do with the transition between power states, but we are less clear on whether it is something that can be remedied in software. We will look into it.
 
V

Vern Little(Guest)

Guest
Any update on the RTC bug fix? I've got 30 hours on my aircraft, and my D10A has gained 15 minutes in 2 months.

Does this prove the adage that all aircraft clocks are broken?

If we assume two power up/down cycle per flight hour, plus power cycles for maintenance, this would give me more than 60 cycles since clock reset.

Normally, I'd expect this to be about one to two minutes of error (based on the known bug), not 15.

I have the internal backup battery, with no clock keep-alive power.

Vern
 

dynonsupport

Dynon Technical Support
Staff member
Joined
Mar 23, 2005
Messages
13,226
The past couple months have kept us busy releasing new products, and we haven't had the time to devote to this issue. However, it is now higher up the list of things to work on and we will devote some time towards fixing this by Oshkosh.

We apologize for the inconvenience, but hope to have a solution here soon.
 
V

Vern Little(Guest)

Guest
More information on the real time clock. I reset the clock at 30 hours, and checked at 35 hours (flight time). The clock was slow by 15 minutes.

It looks like that power off mode is the problem. This may be in addition to the power off/on errors that I'm seeing. The internal battery is fully charged.

Vern Little
 
Top