Hey Dave, a really good way to spot differences visually is to take two identical photos and then stack them in Photoshop. Then on the top layer, change it from "Normal" to "Difference" type. Things that are the same and overlapping perfectly will show up completely black. Things that aren't the same will show up in some color and will be highly visible and easy to locate. Also, make up a "corner guide" jig to help align the boards for the photos. The jig can be something as simple as pencil marks on your bench or lengths of straight edge that you can butt the boards against. And you probably know to keep the camera fixed on a tripod between exposures . . . An easier way (the one that you're using) is to just stack them and keep them both normal layers. Then toggle the visibility on the top layer on/off while scanning the board for any visible changes. In Astronomy (and possibly other disciplines) this is called a Blink Comparator.
aerofart Way too much effort (and I don't use photoshop), and I like the method I used. Also, it's not easy to get the same alignment when the boards are still in the case with the tilting bails etc. I know there are better ways but what I got was quick and more than good enough for the comparison.
EEVblog After having tried the "difference" method myself using your photos, I have to agree, Dave. The color differences in the components and lighting throws things off too. And that's why I mentioned the Blink Comparator (before having gone through the whole video to see that that, in fact, was exactly the method you used) which worked PERFECTLY. Having now seen the complete video, I would have to say that the Blink method is perfectly suited to this sort of thing. Well done, my friend. Catch'ya Next Time!
It's always nice to see a proper design and build, no bodges just excellent quality. I wonder if in the final price of such a meter they take into account the electricity bill for the thermal chamber where they burn-in the voltage reference :-) Got me on that R324 Vishay precision resistor, I thought it was a crystal and was wondering why you are not saying anything about it cause it was so obvious, earlier in the video.
They should have separated the current return, from the voltage ground. I want to do V measurements on the high side to ground return, and current measurements on the high side. You have to buy 2 of these, to do that in an automated environment. I doubt the common mode between the V-measure ground and the I-measure return induce much error.
I wonder if the moon shaped thermal stress relief cuts are to ensure the plastic heat shield seals up properly. If they weren't moon shaped then it would have air flow going through the stress relief cuts.
What is CR1 4-pin component on main voltage reference board? Very strange. Also R231 (29:00), the bigger package is actually on older model. Also around 34:06 you can see that I_BOOT designator is still there just in different place (a little lower). The older model do not have a relay tho. 90 deg rotations of silkscreen desginators, and 180 deg rotations of resistors are just funny touch What is interesting is that the new model still have few unpoppulated footprints.
Plastic enclosure is here to prevent cooling of internally heated voltage reference with airflow form ventilator. Not to prevent heating it with with hot air coming from other components!
How about a some video about overview of classic and historical HP multimeters? There is so many different models (some of them 25 years old), and multimeters by competition that it could be good to say what is good for what and what is just obsolete now at any given price range. That would be awesome.
The Big question is are the extra features and accuracy 34470A over the worth the extra money. The “Typical Price” of the 34461A is US$ 1,173 and the “Typical Price” of the 34470A is US$ 3,215 as shown on the Keysight website. ( www.keysight.com/en/pcx-2832416/truevolt-series-65-75-digit-multimeters?nid=32051.0&cc=US&lc=eng&cmp=2835291%2C2891615%2C2891457%2C2891643 )
I absolutely believe that the Agilent engineers kick back on a Friday and have beers while watching your teardowns to see how much stuff you catch. lol Top notch vid Dave, thanks a bunch.
Could that unused footprint under the voltage reference be for a thermistor? It seems like the perfect place to put it instead of somewhere outside the plastic enclosure to read the correct temp.
What frame rate did you film this at? only reason why I'm asking is because the video doesn't look very natural. when your moving the unit around or pointing at stuff it looks weird. or it could just be my senile old man eyes. :)
i find these videos of expensive equipment thats out of reach of the hobbyist boring but i appreciate a lot of people like these vids so it's all cool :)
Interesting. If you look when he first opens it, the board has a number of 34465. yet he opened the 35570. I am guessing the only difference is the add-on board under the shield
Is the enclosure just some bog standard plastic or is it a ceramic? If it's ceramic, that's basically a miniature ovenized reference - very well done. (Pun acknowledged, but not intended.)
The half digit is the uppermost digit that has maximum value 1 and 2 possible states (hence 1 and 2 in ½ digit) in a 2½ digit display it would be between 000 and 199. A 2¾ digit display, for example, would be between 000 and 399. The full digits go from 0-9 but the ½ digit goes from 0-1 and the ¾ digit goes from 0-3, etc. Hope this helps.
Do you not have to characterize the LM399 in the same way as the LTZ1000? I only ask because if you do then you are only paying for the difference between them in price, right? what do these references cost?
Darren Jacobson plus characterization and "picking" the "good ones" in keysight manufacturing, will the little board is hocked up to an meter, that is one or two orders of magnitude better then the product for which the reference is for. This is certainly a bit of gear that will not end up as a "day to day multimeter" on someones desk, but somewhere where the specs of this device rely matter. Of you think that the premium for the extra digit is to much, then it certainly is for your purpose. But if an application needs for what ever reason this specs, then well: it has to be compared to the market/competitors and not to the BOM cost.
I really think that the "don't turn it on" part is detrimental for the show, especially for those who aren't acquainted with the device. It would have been nice to see what the device actually does. Now it's just some random device that does something... Unless you already know, it's a complete mystery what it does. May I humbly ask to not follow that gimmick? I understand it sound catchy, but it just doesn't make sense. Even a half-minute explanation and demonstration of what the device does would be marvelous.
WarpRulez This meter cost around $3K USD. Rarely people who going to spend such money for DMM don't know what is it. But often people who actually have need in these, want to know how this performance is archieved.
They didn't change their name. The (as I understand it) life sciences equipment part and the test equipment parts of Agilent split up. The life sciences part was larger and so kept the Agilent name and carried on as is. Only the test equipment part then became Keysight. Edit: Agilent still exists, they just sell life sciences equipment.
+Sheldon Holy As the new kid on the block in the life sciences world, they wanted the brand recognition of the Agilent name. It was felt that Keysight wouldn't lose any customers by the name change since the company already has such a large market share in T&M. Agilent has very little market share in life sciences, however, so any little bit helps. I'm sure Keysight was willing to give up the Agilent name as long as it meant freeing itself from the acquisition-happy and ROI-destroying succubus of a life sciences division.
Haha that unit has the Keysight DemStock sticker on it lol. You should order Keysight/Ixia 8 port 400G AresOne Test Appliance and do a tear down on that!
It is a shame that they can't use the old [hp] logo anymore, isn't it? At one time, HP, Fluke, and Tektronix were the pièce de résistance of professional test equipment. Still are, IMHO.
Shadi2 I didn't notice it, and I pay particular attention to things like that on every person that I watch. You see a lot of shaking hands from people in the UK. Their water is probably being poisoned more than the USA.
Robert Calk Jr. Well then, our water is not poisoned and neither is yours. If you're one of those anti-fluoride people you'll find only 10% of people's water here contains measurable fluoride (8% of that is natural calcium fluoride) compared to 66% in the US. If you have actual sources please share as I am yet to become paranoid of the government trying to poison my water supply.
+Hrvoje P. It's funny how the original group that does T&M stays in the same place and has to take a new name and be called the spin-off when the new company that spins-off takes the old name and goes away. HP stayed put and became Agilent whilst the new HP branched off to ruin PCs. Agilent stayed in the same original HP buildings with the same original HP people and became Keysight whilst the new life sciences company (made largely of acquisitions) took the "Agilent" name. Don't let the press releases fool you :(
That front panel looks terrible. The bolded, slightly too large model number text looks cheap. The Agilent, sorry Keysight, logo looks very plain and cheap. That USB port cutout looks crappy, and I don't like the placement. Even the bumper and kickstand look to be made of that junky plastic/rubber you'd find on a $50 eBay special. Yeah the board looks nice and I'm sure the performance is there, but jeez what a bad marketing move by Agilent. Sorry, HP. No wait.......Keysight...
The issue is will anyone actually spend that much money on such a device? Do you think that there may be a beats audio logo, or some other component advertising low analog noise for maximum audio quality from your digital audio files? Did someone fall a sleep on top of the number pad of a keyboard when coming up with the price?
there are 6 1/2 digits meters and 8 1/2 digit meters in keysights product line. I guess the guy that priced the 7 1/2 did place it in between the two and marketing checked, that there is a demand for such a product.
HP, Agilent or Keysight whatever you want to call them are a complete piece of shit, don't last forever like Fluke. Fluke is one million times better in all ways. Cheers.
I LOVE how you roast their entire branding in the first 30 seconds of the video.
I've just received a newsletter from Keysight linking to this video. Nice job, Dave!
Hey Dave, a really good way to spot differences visually is to take two identical photos and then stack them in Photoshop. Then on the top layer, change it from "Normal" to "Difference" type. Things that are the same and overlapping perfectly will show up completely black. Things that aren't the same will show up in some color and will be highly visible and easy to locate. Also, make up a "corner guide" jig to help align the boards for the photos. The jig can be something as simple as pencil marks on your bench or lengths of straight edge that you can butt the boards against. And you probably know to keep the camera fixed on a tripod between exposures . . .
An easier way (the one that you're using) is to just stack them and keep them both normal layers. Then toggle the visibility on the top layer on/off while scanning the board for any visible changes. In Astronomy (and possibly other disciplines) this is called a Blink Comparator.
You'd need to process them and warp them to match. Dave did a good job with the line up but still not good enough for a pixel perfect comparison.
aerofart Way too much effort (and I don't use photoshop), and I like the method I used. Also, it's not easy to get the same alignment when the boards are still in the case with the tilting bails etc. I know there are better ways but what I got was quick and more than good enough for the comparison.
EEVblog After having tried the "difference" method myself using your photos, I have to agree, Dave. The color differences in the components and lighting throws things off too. And that's why I mentioned the Blink Comparator (before having gone through the whole video to see that that, in fact, was exactly the method you used) which worked PERFECTLY. Having now seen the complete video, I would have to say that the Blink method is perfectly suited to this sort of thing. Well done, my friend. Catch'ya Next Time!
***** Agreed.
I sure do love to sit down to some tea to your tear downs in the morning, :)
Those slots are mainly for thermal isolation, not thermal stress.
It's always nice to see a proper design and build, no bodges just excellent quality. I wonder if in the final price of such a meter they take into account the electricity bill for the thermal chamber where they burn-in the voltage reference :-)
Got me on that R324 Vishay precision resistor, I thought it was a crystal and was wondering why you are not saying anything about it cause it was so obvious, earlier in the video.
is it just me or is the keysight logo similar to Sony's Walkman logo?
acksmawb Wow, freaky! pbs.twimg.com/media/B_9n--yUIAA8lQx.png
acksmawb awesome
You are right mate
They should have separated the current return, from the voltage ground. I want to do V measurements on the high side to ground return, and current measurements on the high side. You have to buy 2 of these, to do that in an automated environment. I doubt the common mode between the V-measure ground and the I-measure return induce much error.
I wonder if the moon shaped thermal stress relief cuts are to ensure the plastic heat shield seals up properly. If they weren't moon shaped then it would have air flow going through the stress relief cuts.
Note the old board already have the 8 holes for the pinheader under the voltage reference.
outstanding video, with awesome switching between the photos
Maybe you should turn on these multimeters and compare measurements? :)
36:55: Those two controllers on the bottom: Keysight changed them from LM3S to TM4C129.
I wonder if I-Boot stands for current bootstrap. A bootstrap circuit for the microamp scale or something.
MrBobcat975 btw the label is still there, next to the white relay.
What is CR1 4-pin component on main voltage reference board? Very strange.
Also R231 (29:00), the bigger package is actually on older model.
Also around 34:06 you can see that I_BOOT designator is still there just in different place (a little lower). The older model do not have a relay tho.
90 deg rotations of silkscreen desginators, and 180 deg rotations of resistors are just funny touch
What is interesting is that the new model still have few unpoppulated footprints.
Plastic enclosure is here to prevent cooling of internally heated voltage reference with airflow form ventilator. Not to prevent heating it with with hot air coming from other components!
How about a some video about overview of classic and historical HP multimeters? There is so many different models (some of them 25 years old), and multimeters by competition that it could be good to say what is good for what and what is just obsolete now at any given price range. That would be awesome.
The Big question is are the extra features and accuracy 34470A over the worth the extra money. The “Typical Price” of the 34461A is US$ 1,173 and the “Typical Price” of the 34470A is US$ 3,215 as shown on the Keysight website.
( www.keysight.com/en/pcx-2832416/truevolt-series-65-75-digit-multimeters?nid=32051.0&cc=US&lc=eng&cmp=2835291%2C2891615%2C2891457%2C2891643 )
Top notch presentation skills!
Is there an 11A fuse where a 3A fuse should be @ 5:01?
I absolutely believe that the Agilent engineers kick back on a Friday and have beers while watching your teardowns to see how much stuff you catch. lol
Top notch vid Dave, thanks a bunch.
Was the three-side cutout for increasing junction to ambient thermal resistance? I am not sure about the thermal stress part.
Could that unused footprint under the voltage reference be for a thermistor? It seems like the perfect place to put it instead of somewhere outside the plastic enclosure to read the correct temp.
What frame rate did you film this at? only reason why I'm asking is because the video doesn't look very natural. when your moving the unit around or pointing at stuff it looks weird. or it could just be my senile old man eyes. :)
Daniel Hughes 50fps like all my recent videos.
i find these videos of expensive equipment thats out of reach of the hobbyist boring but i appreciate a lot of people like these vids so it's all cool :)
What I don't get is the boards number and Rev are the same yet it hasn't the same layout?
What the heck is the spark gap for in a DMM and who fudge is Flynn?
the flipping of R102 and R101 is driving me crazy.
What happened to teardown Tuesday?
We need mike to do an x-ray, would like to reverse engineer the analogue section.
Question Dave.
How does the software calibration work?
Interesting. If you look when he first opens it, the board has a number of 34465. yet he opened the 35570. I am guessing the only difference is the add-on board under the shield
dcheverie Seems to be the only difference, yes. Much discussion on the forum.
I see you decided not to use the first take for the intro.
@James Fitz he tore down a digital photocopier once which features a laser printer :)
Is the enclosure just some bog standard plastic or is it a ceramic? If it's ceramic, that's basically a miniature ovenized reference - very well done. (Pun acknowledged, but not intended.)
Anyone else notice the missing screw near the fan on the old board?
énorme, merci!
Quick question... What is a half digit? (ie the .5 part of 7.5)
The half digit is the uppermost digit that has maximum value 1 and 2 possible states (hence 1 and 2 in ½ digit) in a 2½ digit display it would be between 000 and 199. A 2¾ digit display, for example, would be between 000 and 399. The full digits go from 0-9 but the ½ digit goes from 0-1 and the ¾ digit goes from 0-3, etc. Hope this helps.
Do you not have to characterize the LM399 in the same way as the LTZ1000? I only ask because if you do then you are only paying for the difference between them in price, right? what do these references cost?
Darren Jacobson
Digikey has the LTZ1000ACH for $59US and change for a single IC.
Robert Calk Jr.
Well, that's not very pricey, they must have to put a lot of work into characterizing it to justify such a price increase.
Darren Jacobson And $250 USD more for resistors which can match LTZ's performance.
Darren Jacobson plus characterization and "picking" the "good ones" in keysight manufacturing, will the little board is hocked up to an meter, that is one or two orders of magnitude better then the product for which the reference is for. This is certainly a bit of gear that will not end up as a "day to day multimeter" on someones desk, but somewhere where the specs of this device rely matter. Of you think that the premium for the extra digit is to much, then it certainly is for your purpose. But if an application needs for what ever reason this specs, then well: it has to be compared to the market/competitors and not to the BOM cost.
Price is exponential with sales volume. Quality increases linearly.
If you ask yourself how much the LTZ1000ACH costs, found it on one auction site for US $42,50 new.
$1,800.00 more seems a bit steep for basically just adding a better reference...
Yay! Ordered myself a now in stock uCurrent!
Hal Yadoin
I ordered myself one also.
with LAN builtin why do you need the GPIB option?
Backwards compatibility. A lot of labs may already have an entire GPIB setup for the whole lab.
I really think that the "don't turn it on" part is detrimental for the show, especially for those who aren't acquainted with the device. It would have been nice to see what the device actually does. Now it's just some random device that does something... Unless you already know, it's a complete mystery what it does.
May I humbly ask to not follow that gimmick? I understand it sound catchy, but it just doesn't make sense. Even a half-minute explanation and demonstration of what the device does would be marvelous.
WarpRulez This meter cost around $3K USD. Rarely people who going to spend such money for DMM don't know what is it. But often people who actually have need in these, want to know how this performance is archieved.
WarpRulez
Did you watch the video? Dave said that he will be doing a review of the meter...
If Dave did turn it on, then people would be complaining that he didn't follow his mantra. Can't please everyone...
This is a teardown. It's not a review. He'll do a review later.
I don't want a review in a teardown video. I simply want to briefly see, and be explained, what it does.
Agilent changing their name to keysight was a stupid idea
They didn't change their name. The (as I understand it) life sciences equipment part and the test equipment parts of Agilent split up. The life sciences part was larger and so kept the Agilent name and carried on as is. Only the test equipment part then became Keysight. Edit: Agilent still exists, they just sell life sciences equipment.
I was just going off what I heard
+Sheldon Holy As the new kid on the block in the life sciences world, they wanted the brand recognition of the Agilent name. It was felt that Keysight wouldn't lose any customers by the name change since the company already has such a large market share in T&M. Agilent has very little market share in life sciences, however, so any little bit helps. I'm sure Keysight was willing to give up the Agilent name as long as it meant freeing itself from the acquisition-happy and ROI-destroying succubus of a life sciences division.
Haha that unit has the Keysight DemStock sticker on it lol. You should order Keysight/Ixia 8 port 400G AresOne Test Appliance and do a tear down on that!
its teardown Thursday!!!!!
"Any stress on those leads"
It is a shame that they can't use the old [hp] logo anymore, isn't it? At one time, HP, Fluke, and Tektronix were the pièce de résistance of professional test equipment. Still are, IMHO.
Wouldn't 1ppm be represented as 0.000001? If so, the ACAL specs are 500ppm + 100ppm, NOT as Dave states 5ppm + 1ppm.
aerofart The values represent % of scale, so 1ppm = 0.0001%.
Keith MacDonald I see that now. Thanks. For a second there I thought it was another "trick up the sleeve" maneuver by manufacturers.
Interesting how we came from tubes to micro computers while reference devices still stay sort of stone age "large" oven devices.
LTZ1000 costs $36 in qtys of 200 at digikey. Wow, expensive component.
quote of the year "they are not critical parts, they are just jelly bean parts" 33:26
@Shadi2 humans do have a tendency to shake.
Dave, can you do a video on Voltcraft multimeters? Are they good?
cool.
5DX aoi passed!!! Yaaaaay!
i keep thinking it's an oscilloscope
is it just me or are his hands shaking more lately?
Shadi2 It's just you.
Shadi2
I didn't notice it, and I pay particular attention to things like that on every person that I watch. You see a lot of shaking hands from people in the UK. Their water is probably being poisoned more than the USA.
Robert Calk Jr. Not sure if trolling or actually serious... Dave's from 'stralia anyway.
Callum Aitchison
I know he is - I meant other people in the UK. And I'm always serious. I never troll.
Robert Calk Jr. Well then, our water is not poisoned and neither is yours. If you're one of those anti-fluoride people you'll find only 10% of people's water here contains measurable fluoride (8% of that is natural calcium fluoride) compared to 66% in the US. If you have actual sources please share as I am yet to become paranoid of the government trying to poison my water supply.
Hey hey hey be careful there! I work at Keysight! Great products, great people, great service, and yeah...it's a pretty boring logo...
i think you critisize at least one thing in the product no matter what. why?
JusstyteN Because no product is perfect? If only he would give up saying UN-BUH-LEEEEV-ABLE. Great video, nonetheless.
I am also disappointed with their spin-off's. Agilent sounds better.
+Hrvoje P. It's funny how the original group that does T&M stays in the same place and has to take a new name and be called the spin-off when the new company that spins-off takes the old name and goes away. HP stayed put and became Agilent whilst the new HP branched off to ruin PCs. Agilent stayed in the same original HP buildings with the same original HP people and became Keysight whilst the new life sciences company (made largely of acquisitions) took the "Agilent" name.
Don't let the press releases fool you :(
Say 34461a one more time....
Bloody Ripper!
When are you going to tear down an "E meter"?
(www.cs.cmu.edu/~dst/E-Meter/)
That front panel looks terrible. The bolded, slightly too large model number text looks cheap. The Agilent, sorry Keysight, logo looks very plain and cheap. That USB port cutout looks crappy, and I don't like the placement. Even the bumper and kickstand look to be made of that junky plastic/rubber you'd find on a $50 eBay special. Yeah the board looks nice and I'm sure the performance is there, but jeez what a bad marketing move by Agilent. Sorry, HP. No wait.......Keysight...
EEVBlog -where even outro video links are too long :P
The issue is will anyone actually spend that much money on such a device?
Do you think that there may be a beats audio logo, or some other component advertising low analog noise for maximum audio quality from your digital audio files?
Did someone fall a sleep on top of the number pad of a keyboard when coming up with the price?
there are 6 1/2 digits meters and 8 1/2 digit meters in keysights product line.
I guess the guy that priced the 7 1/2 did place it in between the two and marketing checked, that there is a demand for such a product.
:)
That logo, name and font says to me: "Chinese junk". What a dumb move by Agilent.
*Keysight
just that it's not made in China unlike Tektronix or Keithley
Not only the logo, all in Keysight equipment is a complete shit.
Hand by hand Fluke wins all the time.
Keysight IS Agilent.
HP, Agilent or Keysight whatever you want to call them are a complete piece of shit, don't last forever like Fluke.
Fluke is one million times better in all ways.
Cheers.
is it? the fluke bench meters are not graphical per se, they have fewer features