Poll

What is an acceptable boot up time for a general purpose bench multimeter

<1 second
4 (4.5%)
up to 5 seconds
23 (26.1%)
up to 10 seconds
18 (20.5%)
up to 20 seconds
10 (11.4%)
up to 30 seconds
7 (8%)
up to 45 seconds
1 (1.1%)
up to 60 seconds
2 (2.3%)
It doesn't matter
19 (21.6%)
It Depends....
4 (4.5%)

Total Members Voted: 88

Author Topic: [Poll] Benchtop DMM Acceptable Power On Time?  (Read 4058 times)

0 Members and 1 Guest are viewing this topic.

Offline Jester

  • Frequent Contributor
  • **
  • Posts: 898
  • Country: ca
Re: [Poll] Benchtop DMM Acceptable Power On Time?
« Reply #25 on: November 28, 2022, 05:15:16 pm »
What I wonder is, what the hell is the software doing that takes 30s?

Many many years ago I got out voted on the OS for a new portable product, the consensus was to use Windows CE I think it was called. Product turned out to be the biggest looser of anything our company ever designed. Boot time was in the order of 5 minutes. We had 90% return rate.

Next time round we used Linux, far more complicated product, boot time was 18s
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 15148
  • Country: de
Re: [Poll] Benchtop DMM Acceptable Power On Time?
« Reply #26 on: November 28, 2022, 06:43:09 pm »
One part the OS is usually doing is some kind of hardware test and looking for possible present hardware (especially if you don't turn that part off explicitely). Parts of the OS can be encrypted / compressed in some way and may need to be precessed first. There can be some kind of check-disk running to make sure the file systems are OK. Some installations copy hundreds of files to a RAM disk.  Linux does a lot of configurations from scrits that need to run through. The newer version are often made for a more up to date hardware, like GHz and more CPU speed - so the developers often don't see the long boot times with a slow system.

I don't know about WinCE, but with Linux there is a tendency for many distributions to start and install a lot of features / expensions / compatibility hacks that are often not really needed. For linux this may require to compile your own, hardware specific kernal.  Things already get a lot easier by skipping multi CPU support, often not needed with small systems. Reducing this to a bare minimum can speed up the boot time a lot and reduce the required memory.

Windows typical would be trying to get a network connection (including waiting for time-outs) in case there is an update available or a time server available.
 

Offline colorado.rob

  • Frequent Contributor
  • **
  • Posts: 420
  • Country: us
Re: [Poll] Benchtop DMM Acceptable Power On Time?
« Reply #27 on: November 28, 2022, 07:45:49 pm »
What's it doing?

Booting OS, programming FPGAs, negotiating IP via DHCP, calibration and self test.

Depending on what "standby" means, it may need to do everything but boot the OS. The start time could go from 45 seconds to 25 seconds. Programming FPGAs and DHCP are certainly not instantaneous operations.

You could get rid of almost all but DHCP by using an embedded system, analog calibration, and ASICs. But that's high-risk and expensive. Time to market for Linux/Android systems with established UI conventions, FPGAs which can be reconfigured, and digital calibration/self-test to ensure long-term accuracy/stability has huge advantages for manufacturers (cost/risk) and for consumers (price/features).
 

Offline sokoloff

  • Super Contributor
  • ***
  • Posts: 1799
  • Country: us
Re: [Poll] Benchtop DMM Acceptable Power On Time?
« Reply #28 on: November 28, 2022, 08:10:40 pm »
Many many years ago I got out voted on the OS for a new portable product, the consensus was to use Windows CE I think it was called. Product turned out to be the biggest looser of anything our company ever designed. Boot time was in the order of 5 minutes. We had 90% return rate.

Next time round we used Linux, far more complicated product, boot time was 18s
Democracy is a terrible structure for product design and engineering decision-making.
 

Offline LeeimaTopic starter

  • Contributor
  • Posts: 31
  • Country: gb
Re: [Poll] Benchtop DMM Acceptable Power On Time?
« Reply #29 on: November 29, 2022, 12:54:53 pm »
The startup has 2 parts: one is booting the OS, that modern DMMs, especially those with graphics or LAN interface use. The parts is doing some self test on power one - for certain instruments new regulation require a self test. The problem there is that a really stringent self test also needs the analog hardware to be reasonable settled a possibly even thermally reasonable stable (e.g. for a heated reference) before the more sensitive parts of the self test are done.

With a standby mode the boot time should be relatively short - no real excuse for more than some 10 s there.

For high accuracy measurments one needs the warm-up time anyway, but quite often one also uses the meter (maybe not so much a 8 digit one) for tasks that don't need the full performance and just want a fast result. With 4 wire ohms a handheld meter is not really an option there.

I wasn't aware of specific regulations related to self-test, what you've said is all valid.
It's now quicker for me to switch on my 500 MHz scope than use the DMM  :-DD
 

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 7462
  • Country: hr
Re: [Poll] Benchtop DMM Acceptable Power On Time?
« Reply #30 on: November 29, 2022, 01:20:55 pm »
There are few confusions here..

First standby doesn't necessarily mean OS standby. It might mean some parts of device stay powered (OCXO, voltage reference) but system still boots full every time.

Also I'w seen WinCe devices that booted in 8-10 sec. With full GUI.
There were ways to optimize which parts of CE  get loaded, and which you don't need at all.
Same thing is happening with Linux today: if you use "lean" distribution tailored for the hardware you can get super fast boot. If not, it boots much more slowly..
"Just hard work is not enough - it must be applied sensibly."
Dr. Richard W. Hamming
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf