Boot Speed Differential

[Quitsa]

One of my TZT 3 16 displays boots up much faster than the other (30 seconds or so). I tried a few things to see if it made a difference:

-- switched chart master to the unit that booted faster
-- did full factory rest on slower unit

Neither made any difference in the speed at which the two displays boot and one remains faster. The other thing I did not try was to reinstall the unit software, which perhaps will help?

 

[Quitsa]

Being trapped in by the nearby hurricane, I went ahead and reinstalled the unit software and did another factory reset. The slower unit boots a little faster but is still noticeably slower than the other one.

 

[Johnny Electron]

There are a lot of factors that can cause one to be slower than the other such as chart indexing time and if it was the first unit turned on that day making it the NMEA 2000 data master. It normally wouldn't be anything to be concerned with. The data master manages the 2000 communications for the entire system. This unit is the first one being powered on.

 

[hiccup]

They should have the same charts on their cards, so the indexing should be the same. You've already tested the impact of which is the chart master. Checking the impact of being the n2k data master is easy too as you just turn one on, time its boot sequence, turn it off, and repeat for the other. I would temporarily remove all peripherals from both MFDs during the test as well.

My point is that it would be easy to demonstrate that your issue is MFD specific, which is what it sounds like to me.

 

[Quitsa]

Actually that's a good idea -- I am going to disconnect both units from the NMEA 2000 backbone and ethernet and time the boot for each. That should reveal if it is unit-specific.

 

[Quitsa]

I did my experiment of disconnecting the ethernet and NMEA 2000 cables. Previously one unit took 60 seconds to boot and the other took 110 seconds. Without the cables, both boot up in about 60 seconds. Guess there isn't a problem with the slower unit, just a network load issue.

 

[hiccup]

Good detective work. Now the question is whether it was the ethernet or N2K connections. If it is the N2K, the question would be why in the world would it take so much longer for the data master unit to boot? Even entry level MFD's with low-level CPUs are able to handle N2K data streams without problems and boot much faster than these units with much more powerful processors.

 

[Quitsa]

I will try and do the experiment leaving the NMEA 2000 cables connected but the ethernet disabled and vice versa and see what happens. I do have quite a few devices on my NMEA 2000 network and a couple on the ethernet besides the MFDs (radar and DFF-3D). Still, that not should really make much difference I would imagine.

I should also try and swap the chart cards just to see if that has anything to do with it.

 

[Bayhouse]

Following....I've noticed the same behavior on my 2 19" TZT3's.

 

[Quitsa]

I think I have definitely established that it is the connection to the NMEA 2000 network that produces the boot speed differential.

With the ethernet disconnected from both units but the NMEA 2000 connection in place, the first unit takes nearly 30 seconds longer to boot than the second.

With the ethernet connected but the NMEA 2000 network disconnected from both units, the first unit boots up within a few seconds of the second unit.

With both ethernet and NMEA 2000 connected, the first unit takes about 30 seconds longer to boot.

I would imagine that with some effort, a software developer could speed the process. It doesn't matter a whole lot normally. I did have a unit freeze up at sea the other day, however, and it was a very long two minutes waiting for it to reboot.

 

[Bayhouse]

Good info, thanks for running the tests Quitsa.

Now back to Furuno - 30 seconds just because of the N2K master seems like a lot of time for what should be only a little extra work??

 

[Quitsa]

I'm going to do one more test. There are loads of charts on the memory card that are the ones that require an unlock code. I think that deleting everything except the charts I actually use might be helpful in improving the boot time but of course it is not likely eliminate the differential.

 

[Quitsa]

Okay, final installment of my boot test program (you can tell we have been having really lousy weather).

I made new microSD cards that have only the charts for my area that I use and the tide and current data. I also deleted all of the "locked" charts from the chart list found under settings. I believe this means that only the active charts on the new SD cards are loaded into the machines's memory now.

In any case, this resulted in both machines booting up about 30 seconds faster than before. They also seem to run a little faster, which might be from not using as much memory for charts. The unit that is started first and becomes the NMEA 2000 master is still about 20 seconds slower to boot than the second machine.

 

[hiccup]

Interesting because tech support has said on many occasions that removing the unused charts from the SD card would not improve performance or decrease boot time.

 

[Quitsa]

Interesting because tech support has said on many occasions that removing the unused charts from the SD card would not improve performance or decrease boot time.

You can try yourself and see but I was pretty careful to change only one thing at a time and I saw a difference when I cleaned out the excess chart files. Another variable I am curious about is the speed of the memory card. I have a superfast card I use in my camera and might test it, although I suspect the MFD probably cannot take advantage of its higher data transfer rate.

 

[Johnny Electron]

Bootup time can not be sped up by using solid state drives for charts on the NN3D system etc. There have been talks over the years about different NavNet systems and each can be slightly different. Each support post should be taken in context. There is very little bootup time difference normally based on the chart chips as long as the same type and speed. The more charts you have to index, the longer it will take for the chart option tabs to pop-up which is a different thing than boot up. The larger impact on most individual units bootup time is whether they are the data master (not to be confused with chart master) as I explained before. The role of running data master for the system will take that unit a bit longer as it performs those additional duties. Boot up can also be affected slightly depending on what screen it was left on last. The system will attempt to put you back exactly where your left off at, on that machine. The more items going on with that particular hotpage compared to the others can also have a small impact. There are many variables and sometimes a machine will get clutter and just need a good factory reset to blow out cobwebs just as a PC needs optimization from time to time. If you find a unit that clearly running slow without reason it should be evaluated. We do find if you take the time to optimize your chart card it can save some initial time for the chart tabs to pop-up because there is less time needed to index the entire data set. Because it focuses on your current area first, much of the indexing happens in the background unnoticed. It is not to say that chart media can't have issues or the machine itself. The answer is not black and white because every system and install is different. I have taken a full hardrive of the entire world charts and plugged it in via USB and overall boot time was only a seconds different despite the slower data access speed. The indexing and accessing some of the charts away from the local area can take a bit more time to see but that isn't boot up time, that is indexing of time, related to background processing.

 

[Quitsa]

Thanks for the response. Usually the boot speed is largely irrelevant since I start up the machines well before we cast off. I have had a few mysterious screen freeze-ups out on the water that required shutting down and restarting. It doesn't happen very often (maybe two or three times a year) but the boot time seems eternal when it does. I am glad that my idea of slimming down the charts on the card to the handful that I use was probably a good step.

I also observe that it is a good idea to do a hard factory reset every so often even if there is no apparent issue. It does make a difference.