http://opalkelly.com/

Author Topic: EEVblog #972 - Operating Chips Outside Their Spec  (Read 6483 times)

0 Members and 1 Guest are viewing this topic.

Offline oldway

  • Super Contributor
  • ***
  • Posts: 1443
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #25 on: February 17, 2017, 10:49:52 PM »
...
Concerning chips out site specs, it can be  OK,  if the risks have been properly rated, and the outside spec validated as thoroughly as possible. A eol test is a good bonus. I seen it done at a  big Corp for million automotive products. It was done properly.
It may be acceptable in the United States and China but not in Europe .... In Europe, it is necessary to respect the specifications of the manufacturer, it is not negotiable.
It is probably a difference of mentalities.
« Last Edit: February 17, 2017, 10:55:33 PM by oldway »
 

Offline f4eru

  • Frequent Contributor
  • **
  • Posts: 270
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #26 on: February 17, 2017, 10:58:53 PM »
As an example: an ads 7850 as stand alone fast adc is already
 cheaper than the rms to DC converter, adc nonwisstanding......
 

Online joeqsmith

  • Super Contributor
  • ***
  • Posts: 3048
  • Country: us
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #27 on: February 17, 2017, 11:50:16 PM »
But... I've been madly googling images and for the life of me I cant work out the manufacturer.

Apple lists the software as well now. 
https://itunes.apple.com/us/app/eevblog-121gw/id1173156224?mt=8

The meter looks good from the back. 
How electrically robust is your meter?? http://www.youtube.com/channel/UCsK99WXk9VhcghnAauTBsbg
 

Online mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 9681
  • Country: gb
    • Mike's Electric Stuff
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #28 on: February 17, 2017, 11:59:41 PM »
Operating chips outside their spec is a bad practice, I don't agree with this.

There is no such thing as "bad practice". Only more or less appropriate for a given situation.

In this case, for example,  the minimum supply voltage will be guaranteed over the entire temperature range, which is not required in a handheld DMM.

Obviously there is some risk, e.g. they could chnage their process in the future, but that's one of things that need to be taken into account.

Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 

Online mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 9681
  • Country: gb
    • Mike's Electric Stuff
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #29 on: February 18, 2017, 12:03:08 AM »
...
Concerning chips out site specs, it can be  OK,  if the risks have been properly rated, and the outside spec validated as thoroughly as possible. A eol test is a good bonus. I seen it done at a  big Corp for million automotive products. It was done properly.
It may be acceptable in the United States and China but not in Europe .... In Europe, it is necessary to respect the specifications of the manufacturer, it is not negotiable.
It is probably a difference of mentalities.
That is just a ridiculous generalisation. You do not speak for all of Europe.
The multimeter either meets its published specs or it doesn't. How it achieves that internally is irrelevant.
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 

Offline Zbig

  • Frequent Contributor
  • **
  • Posts: 744
  • Country: pl
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #30 on: February 18, 2017, 12:21:39 AM »
But... I've been madly googling images and for the life of me I cant work out the manufacturer.

Apple lists the software as well now. 
https://itunes.apple.com/us/app/eevblog-121gw/id1173156224?mt=8

The meter looks good from the back.

Dave, I hate to be that guy, but I honestly think the mobile app as it looks today, is a disservice to the image of your soon to be released product. I know that you've stated more than once that it is just a kind of a proof of concept, to verify meter's basic functionality. I say it'd be better to start with no app at all than this. It just screams amateur personal late 90s Geocities home pages, the UI looks truly awful. It's not always that anything is better than nothing. The app will be associated with your product, the look and feel of this abomination will affect how the actual meter is perceived, IMO. How would you like a color scheme, fonts, drop shadows, etc. like this on the actual meter's faceplate?

EDIT:
Has anyone clicked the "EEVBlog 121GW Support" link on the iTunes listing? It leads to http://www.ueitest.com/ so now is seems we know who the OEM is :-)
« Last Edit: February 18, 2017, 12:27:40 AM by Zbig »
 

Offline madires

  • Super Contributor
  • ***
  • Posts: 3422
  • Country: de
  • A qualified hobbyist ;-)
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #31 on: February 18, 2017, 12:47:46 AM »
Operating chips outside their spec is a bad practice, I don't agree with this.
Ask to the manufacturer to change their specifications....

The statement " the manufacturer (Analog Device) says it can work at lower voltage than specified" is not a profissional statement....It means that good working in this condition is not garanteed by the manufacturer of the chip.

Not quite. The answers could be:
- No!
- Yes, but at your own risk.
- Yes, it will be stated in the next revision of the datasheet.
- Yes, we'll send you a confirmation written on paper if you wish.
- Yes, but not officially.
- any other shade of grey :)

Unless you get an official confirmation or an updated or custom datasheet, you have to perform a risk analysis besides running tests. The  most prevalent risk in this case is that the measured value could have a larger tolerance. But no risk of electric shock, exploding DMM or other showstoppers.

« Last Edit: February 18, 2017, 12:49:21 AM by madires »
 

Offline NANDBlog

  • Super Contributor
  • ***
  • Posts: 3060
  • Country: be
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #32 on: February 18, 2017, 12:49:07 AM »
Expensive? Noooo, it's not! There are plenty of us today with integrated adc 12 bits or higher. Take one that is fast enough. The cost of the Micro is already included
If we calculate with a conservative (fluke 87) 20 Khz bandwidth for the TrueRMS core, you need to take samples at around 200KHz, otherwise it is not really true rms. An ADC with those specs there is the ADS8339 which is 3.5 USD from TI. You need to buffer and amplify the signal. I will assume the AD8436 is getting a 1V signal, so the amplification is only 5. Assume 0.5% max error, you need an opamp doing less than 1mV offset, 2MHz GBW. Something like a OPA197 can do it, 60 cents. You will need to dedicate an SPI port of your micro to it. Also coding needs to be done, verifying that the RMS calculations are correct. Luckily, there is only one "root calculation" in the code. I think with the usual big company impotence, a few months should be enough to develop this.
Compare this with the price of the AD8436, which is ~3USD. You gain nothing, but more chance of error for more money, and more code. While you could just solve it with an IC. #analogisbetter

At this particular chip, I would suspect the following:
At low power supply voltage, over the temperature range, it might have linearity errors. Maybe they had problems with the output buffer going only to supply voltage -1V. Sometimes they just dont test the part with a certain scenario, because they haven't thought about that.
 

Offline coppice

  • Super Contributor
  • ***
  • Posts: 2489
  • Country: hk
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #33 on: February 18, 2017, 01:05:18 AM »
Is that a commercial chamber in the video, or something else repurposed? If it's a real chamber, who makes it?
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 23749
  • Country: au
    • EEVblog
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #34 on: February 18, 2017, 01:12:22 AM »
Dave, I hate to be that guy, but I honestly think the mobile app as it looks today, is a disservice to the image of your soon to be released product.

Meh.
 

Offline Brumby

  • Supporter
  • ****
  • Posts: 4668
  • Country: au
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #35 on: February 18, 2017, 01:25:29 AM »
The App available on release doesn't worry me in the slightest.

I think Dave has said the protocol will be open - so it's only a matter of time before we have someone with an EE bent and some App writing skills sit down and put together something far more useful.
 

Offline Zbig

  • Frequent Contributor
  • **
  • Posts: 744
  • Country: pl
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #36 on: February 18, 2017, 02:08:34 AM »
Is that a commercial chamber in the video, or something else repurposed? If it's a real chamber, who makes it?

I believe it's an incubator for pet reptiles.
 

Offline coppice

  • Super Contributor
  • ***
  • Posts: 2489
  • Country: hk
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #37 on: February 18, 2017, 02:58:30 AM »
Is that a commercial chamber in the video, or something else repurposed? If it's a real chamber, who makes it?

I believe it's an incubator for pet reptiles.
Ah, right. It seems to be this one https://www.amazon.com/Exo-Terra-PT2499-Incubator-Unit/dp/B004AJLREA . It seems to get a lot of negative feedback from customers over reliability.
 

Offline NivagSwerdna

  • Frequent Contributor
  • **
  • Posts: 748
  • Country: gb
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #38 on: February 18, 2017, 03:26:56 AM »
...it's clearly going to be either a rebadged existing model or as has been suggested a rebadged exisiting model with maybe a tweak or two.

Nope, and nope.

Quote
It's likely that this model will have the same glass as an existing or planned model since an EEVBlog run will only form part of a larger commercial run..

Nope again.
:-DD

Fair enough!  Looks like I wasn't warm then.  :) Looking forward to seeing the final version!
 

Offline NivagSwerdna

  • Frequent Contributor
  • **
  • Posts: 748
  • Country: gb
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #39 on: February 18, 2017, 03:29:07 AM »
Is that a commercial chamber in the video, or something else repurposed? If it's a real chamber, who makes it?

I believe it's an incubator for pet reptiles.
Ah, right. It seems to be this one https://www.amazon.com/Exo-Terra-PT2499-Incubator-Unit/dp/B004AJLREA . It seems to get a lot of negative feedback from customers over reliability.
As an owner of some tortoises, whom have just come out of the fridge, I would agree with that.  I think this might be a peltier device... it would struggle to keep a stable 5C required for wildlife hibernation.
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 1895
  • Country: us
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #40 on: February 18, 2017, 04:57:26 AM »
Use of a device outside of spec has dangers that vary with the product.  The most significant ones come for very high volume or very long production life products.  In my career the products I associated with were usually produced for a decade or two.  Over that interval common parts can see not just process variation but complete process changes.  The manufacturer will validate their new process to the spec sheet, no more, and anything not defined on the spec sheet is at risk.  Large volume producers may purchase from multiple suppliers of the same device.  Some of which will work while others don't.  A famous example of this came in the WWII VT fuse which was based on a single vacuum tube (valve for those of you speaking Old English ;)).  As I recall there were eight makers of this type tube, only one worked in the application.

For a product like the EEVBlog branded meter, which will likely be produced for a short time, maybe in a single production run, the risks are much smaller.
 

Offline oldway

  • Super Contributor
  • ***
  • Posts: 1443
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #41 on: February 18, 2017, 05:10:46 AM »
....But no risk of electric shock, exploding DMM or other showstoppers.
No risk of electric shock ? You adjust the low battery sensor at 3.6V but the Analog Device rms converter chip is only garanteed to work down to 4.8V....and if this chip stops to work at 3.8V (no low battery alarm) and you measure an 600V ac voltage and it indicates no voltage at all....??? :scared:

You think it safe, but it isn't ... :--

How can you garantee something if you use parts out of specifications ?
« Last Edit: February 18, 2017, 05:16:24 AM by oldway »
 

Offline Len

  • Frequent Contributor
  • **
  • Posts: 504
  • Country: ca
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #42 on: February 18, 2017, 05:37:32 AM »
Is that a commercial chamber in the video, or something else repurposed? If it's a real chamber, who makes it?


 

Online mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 9681
  • Country: gb
    • Mike's Electric Stuff
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #43 on: February 18, 2017, 05:45:51 AM »
How can you garantee something if you use parts out of specifications ?
Production test.
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 

Offline madires

  • Super Contributor
  • ***
  • Posts: 3422
  • Country: de
  • A qualified hobbyist ;-)
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #44 on: February 18, 2017, 05:46:00 AM »
No risk of electric shock ? You adjust the low battery sensor at 3.6V but the Analog Device rms converter chip is only garanteed to work down to 4.8V....and if this chip stops to work at 3.8V (no low battery alarm) and you measure an 600V ac voltage and it indicates no voltage at all....??? :scared:

You think it safe, but it isn't ... :--

How can you garantee something if you use parts out of specifications ?

If the RMS converter only works down to 3.8V you would adjust the low battery warning accordingly. And if rechargeable batteries would be a problem you would put a "non-rechargable batteries only" sticker on the DMM. And we shouldn't forget that any DMM could fail for whatever reason and show the wrong voltage.
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 2972
  • Country: de
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #45 on: February 18, 2017, 06:03:07 AM »
The datasheet for the RMS converter chip already shows a few things that happen at low supply voltage: there is an increased error expected for large signal amplitudes. Depending one the ADC following this might not be a problem, or only at a high crest factor. AFAIR the high crest factor tests in the video where done at low amplitude (fixed peak-peak value). There is no real risk of showing grossly wrong voltage due to the low supply, more like maybe a few percent off for extreme cases. This can be an acceptable compromise between power consumption and accuracy for an handheld DMM. For a very high crest factor signal, the RMS value might not be that good to judge if a voltage is dangerous anyway.

It might be a good idea to include tests of the RMS function in the QC, just in case.

When using the digital method, there is no absolute need to have such a fast ADC. Digital RMS even works with sub-sampling. So even the 15 kSPS inside an AVR µC could be good for software RMS up to about 40 KHz (BW limit of the ADC when sub-sampling). The SW method would be mainly an option with an µC internal ADC. The downside is that is does not work so well with small amplitudes, so one might need some extra auto-range steps (e.g. another 1 or 2 gain stages). However it also has an advantage: settling time can be lower and is not slowing down on low amplitudes.
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 14031
  • Country: za
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #46 on: February 18, 2017, 07:16:11 AM »
Use of a device outside of spec has dangers that vary with the product.  The most significant ones come for very high volume or very long production life products.  In my career the products I associated with were usually produced for a decade or two.  Over that interval common parts can see not just process variation but complete process changes.  The manufacturer will validate their new process to the spec sheet, no more, and anything not defined on the spec sheet is at risk.  Large volume producers may purchase from multiple suppliers of the same device.  Some of which will work while others don't.  A famous example of this came in the WWII VT fuse which was based on a single vacuum tube (valve for those of you speaking Old English ;)).  As I recall there were eight makers of this type tube, only one worked in the application.

For a product like the EEVBlog branded meter, which will likely be produced for a short time, maybe in a single production run, the risks are much smaller.

Don't worry, most of those specs for a chip are "typical" and can vary in some cases widely, with only very broad minimum and maximum values. There are a lot of parameters that are "typical" and which are only tested on a sample basis, and those often are not guaranteed, even if they are actually an important reason for choosing one chip over another in selection. Things like noise levels, actual open loop gain, true bandwidth and such are only sampled, not tested on each lot, they might only be done on a few chips from each run of wafers.

A lot of the lower figures like supply voltages are often "we never tested below xx", and similar for things like gain, bias current ( and some can be nasty, with bias current that depends on the input voltage applied to BOTH inputs, and the supply voltage and temperature, and some can swap from source to sink at some point) and saturated output voltages.
 

Offline oldway

  • Super Contributor
  • ***
  • Posts: 1443
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #47 on: February 18, 2017, 07:48:34 AM »
4.8V is min supply voltage of AD8436, not typical supply voltage.
It has to be respected.
There some reason why Analog Device has specified such a voltage : below this voltage, working of the chip is no more garanteed.

http://www.analog.com/media/en/technical-documentation/data-sheets/AD8436.pdf
 

Offline f4eru

  • Frequent Contributor
  • **
  • Posts: 270
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #48 on: February 18, 2017, 08:55:12 AM »
Expensive? Noooo, it's not! There are plenty of us today with integrated adc 12 bits or higher. Take one that is fast enough. The cost of the Micro is already included
If we calculate with a conservative (fluke 87) 20 Khz bandwidth for the TrueRMS core, you need to take samples at around 200KHz, otherwise it is not really true rms. An ADC with those specs there is the ADS8339 which is 3.5 USD from TI. You need to buffer and amplify the signal. I will assume the AD8436 is getting a 1V signal, so the amplification is only 5. Assume 0.5% max error, you need an opamp doing less than 1mV offset, 2MHz GBW. Something like a OPA197 can do it, 60 cents. You will need to dedicate an SPI port of your micro to it. Also coding needs to be done, verifying that the RMS calculations are correct. Luckily, there is only one "root calculation" in the code. I think with the usual big company impotence, a few months should be enough to develop this.
Compare this with the price of the AD8436, which is ~3USD. You gain nothing, but more chance of error for more money, and more code. While you could just solve it with an IC. #analogisbetter
Wrong price comparison. You already need a front end and an ADC anyway.
You got many advantages: a good power meter. Flexibility. Low cost. More precision on measurements (thanks to oversampling). Simpler hardware...

Nope. Analog is  not always  better, especially when dealing with lowish frequencies.

Personally, I did this exact implementation  of an rms and power calculation, at only 10 ksps, with acuracy better than 0.05% on an EFM8BB1. A 32 cent micro. With integrated 12 bit ADC!!!!!! Used in an industrial product.

I think the actual state of digital uCs and ADC allows us to make much better cheap multimeters.
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 23749
  • Country: au
    • EEVblog
Re: EEVblog #972 - Operating Chips Outside Their Spec
« Reply #49 on: February 18, 2017, 10:25:26 AM »
How can you garantee something if you use parts out of specifications ?
Production test.

Yep. Chip performance doesn't magically change at any time, it's fully characterisable.
Every unit (like any quality meter) is fully tested and calibrated at the factory. It has to be fully calibrated on every range, that's how modern meters work, the cal values are stored in the EEPROM.
And in the this case it's tested at different frequencies and signal levels on all AC ranges.
There is no risk here for the consumer, the risk of production parts changing falls upon the manufacturer.
« Last Edit: February 18, 2017, 10:27:17 AM by EEVblog »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf