EEVblog Electronics Community Forum

EEVblog => EEVblog Specific => Topic started by: EEVblog on March 01, 2017, 08:09:41 am

Title: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: EEVblog on March 01, 2017, 08:09:41 am
Teardown of the new Keysight InfiniiVision 1000 X-Series Oscilloscope leased TODAY!

https://www.youtube.com/watch?v=9KcOQsVxtoU (https://www.youtube.com/watch?v=9KcOQsVxtoU)
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: brutester on March 01, 2017, 08:38:50 am
Dave, I think that they have developed a separate compute board, because:
 1. it seems to be 6 layers (main board is 4 layers I think).
 2. it may save cost if they decide to reuse this module in other product (or in other versions of 1000X)
 3. They may manufacture the compute board in their own facility and outsource the assembly to another factory
 4. As you have pointed out - it is easier to test when they are in separate boards.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: daqq on March 01, 2017, 08:51:39 am
Thanks for the teardown.

My guess as to the separation would be that for the digital board you would use more layers, finer geometry and other things that make it a costlier material per dm^2 . Then they can get away with a simpler technology (also as far as assembly is concerned) for the giant analog board.

Integrating them into one would force the use of the same technology on the whole large board.

I've seen this done for those reasons before.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: MrBungle on March 01, 2017, 08:52:40 am
Ofcourse there's rust, that's why they were in the dumpster ;)
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: sorenkir on March 01, 2017, 08:57:33 am
For sure, separate processor board will allow design upgrade because these Spear600 processors are quite old ("NRND"):
http://www.st.com/en/microcontrollers/spear600.html (http://www.st.com/en/microcontrollers/spear600.html)
Michel.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: AndyC_772 on March 01, 2017, 08:58:20 am
Integrating them into one would force the use of the same technology on the whole large board.

I've seen this done for those reasons before.

I've done exactly the same thing for the same reasons. Escaping big BGAs, and correctly routing DDR memory, takes a lot of layers. The main PCB is probably only 4 or 6 layers, but I'd be surprised if that logic board has fewer than 10.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Avacee on March 01, 2017, 10:18:14 am
Excellent teardown as usual - learn a lot every time  :-+

Around the 24:01 mark and the LMH6552 1.5Ghz Differential Amp U39. Is it just the camera angle or are pins 1,4,5,8 using a bees dick width of solder and look like they'd snap if the unit was jostled sharply?
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: quarros on March 01, 2017, 10:36:54 am
Another possibility for the separate board is to make unlocking hacking features harder and make it easier/cheaper to provide different feature sets?
(Since one board don't have to contain all the features)


Edit: Hmm... I might be totally off mark.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: EEVblog on March 01, 2017, 11:02:01 am
For sure, separate processor board will allow design upgrade because these Spear600 processors are quite old ("NRND"):
http://www.st.com/en/microcontrollers/spear600.html (http://www.st.com/en/microcontrollers/spear600.html)
Michel.

Interesting.
Still selling them though, and there would be the good'ol "last buy" when the time comes.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: EEVblog on March 01, 2017, 11:04:50 am
Dave, I think that they have developed a separate compute board, because:
 1. it seems to be 6 layers (main board is 4 layers I think).
 2. it may save cost if they decide to reuse this module in other product (or in other versions of 1000X)
 3. They may manufacture the compute board in their own facility and outsource the assembly to another factory
 4. As you have pointed out - it is easier to test when they are in separate boards.

Could certainly save some assembly cost by having a higher spec assembly house for the CPU board, and an old lower end assembly house do the main board.
Actual bare board price saving will come down to panel pricing for a given manufacturing process. Could be very significant.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: capt bullshot on March 01, 2017, 01:44:37 pm
I believe Agisight has enough volume to keep ST producing the spear. Anyone else should watch out, if Keylent stops buying this chip, ST might stop producing it, causing the usual last time buy notice.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: AndyC_772 on March 01, 2017, 02:14:25 pm
NRND doesn't necessarily mean they're going to stop making it any time soon; often suppliers will continue to support existing customers for a long time.

IIRC the SPEAR range is quite an unusual CPU which can be customised for individual applications with additional hard-coded logic on the die. It may be that this customisation is a service that ST don't want to offer any longer, but if the NRE has been done already, then the particular version of the CPU used in the DSOX family will be produced for as long as there's demand.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: bji900 on March 01, 2017, 02:30:50 pm
Nice Video Dave! Sorry about changing Febuary to 28 days ;) Looking forward to the review. I find it slightly confusing that the have the two different high end ones listed almost identically on their website but the price is $150 different. You need to open it to learn that the difference is the waveform generator output.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Fungus on March 01, 2017, 04:14:58 pm
What's the CR2032 for? Does it know the time?

Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: nctnico on March 01, 2017, 05:09:14 pm
What's the CR2032 for? Does it know the time?
Ofcourse! That is a standard feature on DSOs for over 20 years so they can include time & date on a screen dump.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: coppice on March 01, 2017, 05:57:12 pm
NRND doesn't necessarily mean they're going to stop making it any time soon; often suppliers will continue to support existing customers for a long time.
Policies vary between vendors, but NRND typically means "we have something newer, that's better value for money". If they are giving up on a product type they might declare a device NRND, without having a newer option. You always need to look carefully.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: marshallh on March 01, 2017, 06:44:30 pm
Looking at the pcb design on both boards, the layout is very much like boards you see coming out of Chinese design houses. The boards were very obviously done in Altium, and the silk rectangles around resistors and ovals around the caps is a trademark giveaway. I'd suspect they contracted out the layout though they may have done the schematics themselves.

On the logic board, lack of traces on both top and bottom points to at least 6 layers and more probably 8 layers.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 01, 2017, 07:36:40 pm
What's the CR2032 for? Does it know the time?
Ofcourse! That is a standard feature on DSOs for over 20 years so they can include time & date on a screen dump.
Also for saving state - scopes have a lot of settings and people  expect them to power up in the same state as when they were powered down, ( and power down without notice) and EEProm/Flash has issues with write time and endurance.   
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: alter Ratz on March 01, 2017, 07:40:07 pm
Coated spark gaps? Does this make sense?
Anybody has some Information about this?

Best Regards,
Bernhard
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: VasinD on March 01, 2017, 07:41:11 pm
Thank you for the teardown, Dave.
But wait a minute... NRND old processor, low end old Spartan-3 FPGA, chinese capacitors on power supply... Is this is a top quality HP NEW scope or I miss something??? Come on, HP... Looks like the mountain gave birth to mouse.
The DS1054Z looks like quality/price beast comparing to this suck.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: nctnico on March 01, 2017, 08:18:19 pm
Thank you for the teardown, Dave.
But wait a minute... NRND old processor, low end old Spartan-3 FPGA, chinese capacitors on power supply... Is this is a top quality HP NEW scope or I miss
It had to compete with Tektronix' TBS1000 series so it had to be a rehash using obsolete technology  >:D
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: brutester on March 01, 2017, 08:21:04 pm
+1 I tought the same.. I think that they should not be coated.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 01, 2017, 08:37:11 pm
Spark gaps work by ionising air. Coating = no air.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Agent24 on March 01, 2017, 08:56:04 pm
Given that the PSU is also coated with a sprawl of budget capacitors, I don't think quality was at the forefront of the design there.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: IvoS on March 01, 2017, 09:06:17 pm
What a nice scope for cheap money I thought I will definitely buy one but after seeing those "lovely" electrolytic caps I lost my appetite.  :(
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: ericloewe on March 01, 2017, 09:25:35 pm
At least Rigol used CapXon. Rather disappointing selection of caps.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: rsjsouza on March 01, 2017, 09:46:35 pm
I agree as well. The overall build is fine, but the fan and the power supply were sourced through a third party and are a huge maybe. I would love to see a teardown of a TBS1000 to compare.

I wonder how soon the PSU and the fan will start giving headaches to the university lab managers. I wonder if it is more expensive to contract a maintenance plan with Keysight or treat these oscilloscopes as consumables... Only time will tell.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 01, 2017, 10:13:23 pm
What a nice scope for cheap money I thought I will definitely buy one but after seeing those "lovely" electrolytic caps I lost my appetite.  :(
What an idiotic attitude.
First, just because you don't recognise a brand doesn't mean it's bad. 
Second, caps are trivially easy to diagnose and cheap to replace if they did fail outside the 3 year warranty. Maybe an hour an ten bucks to replace all of them.
Ditto fan.
Making a purchasing decision on that alone when there are so many other factors is just stupid.


Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Hydrawerk on March 01, 2017, 10:23:40 pm
Where is an integrated circuit in the PSU? Is it in a TO-220 package??
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 01, 2017, 10:51:58 pm
Also, re. the caps, they are well spaced out and fan cooled. Heat is a far more important factor than manufacturer. If they last the warranty period ( eliminating manufacturing defects) they'll probably be good for many more years
 
I've fixed a lot of SMPSUs in my time and the vast majority of cap failures have been due to poor thermal design.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: IvoS on March 01, 2017, 10:53:53 pm
What a nice scope for cheap money I thought I will definitely buy one but after seeing those "lovely" electrolytic caps I lost my appetite.  :(
What an idiotic attitude.
First, just because you don't recognise a brand doesn't mean it's bad. 
Second, caps are trivially easy to diagnose and cheap to replace if they did fail outside the 3 year warranty. Maybe an hour an ten bucks to replace all of them.
Ditto fan.
Making a purchasing decision on that alone when there are so many other factors is just stupid.
Call me whatever you want, make yourself happy. My decision has been made.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: 128er on March 01, 2017, 11:18:04 pm
Where is an integrated circuit in the PSU? Is it in a TO-220 package??

Thought the same. No switching controller in sight. Discrete design? Would be somewhat strange.

EDIT:
Ok, at second glance, there are two tiny SOT23-6 devices U1, U6 on the bottom side of the PSU.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 01, 2017, 11:31:17 pm
Where is an integrated circuit in the PSU? Is it in a TO-220 package??

Thought the same. No switching controller in sight. Discrete design? Would be somewhat strange.

EDIT:
Ok, at second glance, there are two tiny SOT23-6 devices U1, U6 on the bottom side of the PSU.
There are some SOT23-6 devices on the primary side
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 01, 2017, 11:46:12 pm
Looking at it some more, that's one of the best engineered "cheap" SMPSUs I've seen. Plenty of input filtering, very well spaced out, conformal coated ( good where you have dusty airflow), additional LC filters on the outputs, multiple series resistors for high voltage drops.
Apart from the coating over the sparkgaps, though the coating is probably more of a plus than the less effective gaps are a negative- you rarely see them, or MOVs in PSUs at all.
Seems unlikely they'd have done all that then spoilt it with crappy caps.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: bitwelder on March 02, 2017, 06:54:19 am
Let's assume that there is no updates for this model and no other models which can benefit from the two-boards construction.
In percentage of the existing solution, what do you estimate would have been the cost of a single-board construction?
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 02, 2017, 10:13:16 am
Let's assume that there is no updates for this model and no other models which can benefit from the two-boards construction.
In percentage of the existing solution, what do you estimate would have been the cost of a single-board construction?
Hard to say as we have no idea what the total build cost - all you can do is estimate the difference in bare PCB costs, which on the figures estimated so far looks like easily covering any additional cost.
And it won't be just about the bare PCB - assembly, inspection, test will all be factors.

It also allows more flexibility for different models - the sub-board is a well-proven module that's unlikely to need changing as the design is many years old.
The mainboard may need to change for different model variants, e.g. a headless rackmount version, different bandwidth front-ends, with/without siggen etc.
The module approach reduces layout/tooling /qualification cost on a new mainboard variant.

Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: McBryce on March 02, 2017, 10:40:16 am
Could the missing BNC possibly be for a second Func. Gen. channel?

McBryce.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 02, 2017, 11:03:22 am
I notice there is an unpopulated connector on the sub-board, presumably for an MSO option
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: madires on March 02, 2017, 12:24:16 pm
Could the missing BNC possibly be for a second Func. Gen. channel?

There's another one at the side near the USB port. So it could be 4 channel and the external trigger moved to the side or a second function gen. Obviously it's kept flexible to allow variations for other models. My guess is that Keysight starts with the 2 channel version to see if they can compete at the low end and make money. If the 2 channel models are succesful I wouldn't be surpised to see 4 channel models later on.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Brumby on March 02, 2017, 12:34:16 pm
Could the missing BNC possibly be for a second Func. Gen. channel?

McBryce.

It could be for that or for something else - but I'm with Dave on it being the sort of engineering you would use for a 4ch scope.

If it were me, I would set up the tooling for a case and chassis capable of 4 channel.  Even if the 4 channel option was not a starter at the time, it gives the option of taking up the idea further down the track, with minimal additional cost.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: f4eru on March 02, 2017, 12:51:00 pm
Thought the same. No switching controller in sight. Discrete design? Would be somewhat strange.
I saw a PSU like that recently.

Really strange, as it is a SMPS for an integrated subwoofer amplifier.
+80 / -80V output, 150-250W range, full bridge design, only 4 transistors !!. The transistors are driven from aux windings, through a bunch of discretes.
Thre isn't even an optocoupler on it, so it's unregulated, and probably needs a minimum current to not have too much variation on the output.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Len on March 02, 2017, 06:21:33 pm
What's with the soldering on the IC at 23:45? Wrong footprint? or dual footprint?
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: AndyC_772 on March 02, 2017, 06:54:50 pm
It's a QFN package, and the pads have been extended. This gives somewhere to probe with a scope during development, and in production it causes a solder fillet to form at the edge of the device which can be inspected more easily with an AOI system.

If there's room, it's often good practice to extend the pads of a QFN at least a little, beyond the minimum recommended footprint in the device's data sheet.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: saturation on March 02, 2017, 07:06:13 pm
Its a very interesting blend of cost effectiveness in the parts particularly the electrolytics, the only way to know for sure if it will hold up is to test it,  it may have.  The design style reminds me of a specific PSU manufacturer in China, Fu Jia.  They began in the early 1990s, and made their own transformers.  They are now a well known OEM and the hallmarks of their designs are they do burned ins before leaving the factory, they get 3rd party safety certification [ UL, CSA, TUV etc.,] for their products as well as parts, and they have very good labels for parts and test points.  They commonly label their smps designs FJ-SWxxx which you can see all over the 1000x photos.  I was surprised to see the preferred abbreviations for DC volts, Vdc, with the current in lower case, which is an IEEE preferred label that few ever use.  So if it is Fu Jia you can read about their reputation and build quality via google.


Looking at it some more, that's one of the best engineered "cheap" SMPSUs I've seen. Plenty of input filtering, very well spaced out, conformal coated ( good where you have dusty airflow), additional LC filters on the outputs, multiple series resistors for high voltage drops.
Apart from the coating over the sparkgaps, though the coating is probably more of a plus than the less effective gaps are a negative- you rarely see them, or MOVs in PSUs at all.
Seems unlikely they'd have done all that then spoilt it with crappy caps.

Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: rrinker on March 02, 2017, 07:52:59 pm
 I was rather impressed by the very clearly labeled test points and so forth all over the unit, not only just on the power supply.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: thm_w on March 02, 2017, 08:47:14 pm
Excellent teardown as usual - learn a lot every time  :-+

Around the 24:01 mark and the LMH6552 1.5Ghz Differential Amp U39. Is it just the camera angle or are pins 1,4,5,8 using a bees dick width of solder and look like they'd snap if the unit was jostled sharply?

They won't snap, its a dual footprint and there is a pad underneath.

It's a QFN package, and the pads have been extended. This gives somewhere to probe with a scope during development, and in production it causes a solder fillet to form at the edge of the device which can be inspected more easily with an AOI system.

If there's room, it's often good practice to extend the pads of a QFN at least a little, beyond the minimum recommended footprint in the device's data sheet.

Its an alternate footprint to allow WSON/SOIC-8 packages, you can see from the datasheet: http://www.ti.com/lit/ds/symlink/lmh6552.pdf (http://www.ti.com/lit/ds/symlink/lmh6552.pdf)
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mmagin on March 02, 2017, 11:30:01 pm
So is the 'Siglent Rust' just because in these cheap metal stampings, the plating happens to whole sheets prior to stamping the parts and they're manufactured in a very humid climate?

It certainly seems like nothing to get upset about on sub-$1000 products.  I'd be more critical of the thickness of the metal or other substantial mechanical aspects.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: matthewpang on March 02, 2017, 11:52:57 pm
Does anybody know how they are supporting the 3 wire SPI interface on a 2 channel scope ( http://literature.cdn.keysight.com/litweb/pdf/5992-1965EN.pdf (http://literature.cdn.keysight.com/litweb/pdf/5992-1965EN.pdf) ) ? Do you have to pick either MISO or MOSI or are they doing something interesting like routing the clock through the external trigger port ?
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: nfmax on March 03, 2017, 11:07:42 am
Some comments:
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: tautech on March 03, 2017, 11:33:40 am
So is the 'Siglent Rust' just because in these cheap metal stampings, the plating happens to whole sheets prior to stamping the parts and they're manufactured in a very humid climate?

It certainly seems like nothing to get upset about on sub-$1000 products.  I'd be more critical of the thickness of the metal or other substantial mechanical aspects.
Now known as KeySig rust.  ;)

Yes it is very humid and hot in southern China.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: tautech on March 03, 2017, 11:36:11 am

  • Actually, this is an MSO - just an MSO with only one digital channel. We know the MZ4 ASIC has MSO built in, so the hardest part was probably just coming up with the idea of munging an external trigger and a digital input together onto one BNC. A good bit of lateral thinking!

Easily done in firmware.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: nfmax on March 03, 2017, 11:37:43 am
That's my point - easily done, but not so easily thought of in the first place.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Cervisia on March 03, 2017, 11:45:56 am
Does anybody know how they are supporting the 3 wire SPI interface on a 2 channel scope?

The manual download page (http://www.keysight.com/main/techSupport.jspx?searchT=dsox1102g&id=2766207&pid=2766207) shows only the German manual for me; it says in table 28 (translated):
Quote
SPI Signal Configuration
[…] This menu has separate soft keys and sub menus to set source channels and threshold voltages for clock, MOSI/MISO, and CS (chip select).
NOTE: The two-channel oscilloscopes of the DSOX1000 series support 3-wire SPI. The MOSI and MISO signal settings are identical by default; you can basically sample one of them.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Jay_Diddy_B on March 03, 2017, 11:49:37 am
So is the 'Siglent Rust' just because in these cheap metal stampings, the plating happens to whole sheets prior to stamping the parts and they're manufactured in a very humid climate?

It certainly seems like nothing to get upset about on sub-$1000 products.  I'd be more critical of the thickness of the metal or other substantial mechanical aspects.

The material is called 'Satin coat steel'. It is essential zinc coated steel. There is no protection from corrosion on cut edges.

Regards,

Jay_Diddy_B
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: rf-loop on March 03, 2017, 11:49:50 am
So is the 'Siglent Rust' just because in these cheap metal stampings, the plating happens to whole sheets prior to stamping the parts and they're manufactured in a very humid climate?

It certainly seems like nothing to get upset about on sub-$1000 products.  I'd be more critical of the thickness of the metal or other substantial mechanical aspects.
Now known as KeySig rust.  ;)

Yes it is very humid and hot in southern China.

Is it better say "wet" because many times dev point is very near current temp, quite nice when temp is 35 celsius and humidity 98%. . But, this is not only problem.
Take rain water or water condensed to some surface - its not like water in Fnland at all. Well this south chine rain water need sell in chemistry and drug store. (but better than some years ago)
Example we have galvanized  metal rain roof over balcony. Few years all rusted through in. Next time we change rest and use hard plastic.  Some times this rain is so acidic that it is best to go shower and rinse clothes  after come inside from rain. (but far away what is in example Beijing. (but there is not sub tropic climate as in south)
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: tautech on March 03, 2017, 11:56:20 am
So is the 'Siglent Rust' just because in these cheap metal stampings, the plating happens to whole sheets prior to stamping the parts and they're manufactured in a very humid climate?

It certainly seems like nothing to get upset about on sub-$1000 products.  I'd be more critical of the thickness of the metal or other substantial mechanical aspects.

The material is called 'Satin coat steel'. It is essential zinc coated steel. There is no protection from corrosion on cut edges.

Regards,

Jay_Diddy_B
:bullshit:
Engineers know otherwise.  :P
https://www.astm.org/SNEWS/APRIL_2006/dallynside_apr06.html (https://www.astm.org/SNEWS/APRIL_2006/dallynside_apr06.html)

Read the Galvanic Protection paragraphs.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: rrinker on March 03, 2017, 01:47:04 pm
 Reading the paragraph under the box seems to say that the flat sides are safe - cutting the edge won't cause the whole plate to rust, nor will any corrosion on the edge undercut the coating on the flat side. To me that is saying that while the cut edge itself may acquire a small amount of rust, it will not go deep or in any way harm the rest of the sheet. That rust on the cut edges is very shallow surface rust only and thus is not of any real concern (other than cosmetic).

Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Jay_Diddy_B on March 03, 2017, 02:12:39 pm
Reading the paragraph under the box seems to say that the flat sides are safe - cutting the edge won't cause the whole plate to rust, nor will any corrosion on the edge undercut the coating on the flat side. To me that is saying that while the cut edge itself may acquire a small amount of rust, it will not go deep or in any way harm the rest of the sheet. That rust on the cut edges is very shallow surface rust only and thus is not of any real concern (other than cosmetic).



1+

I have used satin coat in products. My experience was consistent, a small amount of surface rust on the exposed edges.

Regards,

Jay_Diddy_B
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Paul Moir on March 04, 2017, 06:17:16 am
:bullshit:
Engineers know otherwise.  :P

So I bolt a hunk of zinc to my car:   that means I'm all right then?
Title: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Dubbie on March 04, 2017, 08:10:22 am
That's literally what they do on boats Paul. It's called a sacrificial anode.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: McBryce on March 04, 2017, 11:14:25 am
:bullshit:
Engineers know otherwise.  [emoji14]

So I bolt a hunk of zinc to my car:   that means I'm all right then?
Car bodywork is already zinc coated, has several layers of paint and underbody sealant. Cars rarely rust these days unless you really try.

McBryce.

Gesendet von meinem Motorola DynaTAC 8000X mit Tapatalk.

Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 04, 2017, 12:30:56 pm
Back on topic, the sub-board is 10 layers.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 04, 2017, 12:44:28 pm
The anti-reflection display coating on the display is actually glass, so  should be pretty durable.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: JPortici on March 04, 2017, 03:50:29 pm
Does anybody know how they are supporting the 3 wire SPI interface on a 2 channel scope ( http://literature.cdn.keysight.com/litweb/pdf/5992-1965EN.pdf (http://literature.cdn.keysight.com/litweb/pdf/5992-1965EN.pdf) ) ? Do you have to pick either MISO or MOSI or are they doing something interesting like routing the clock through the external trigger port ?
in the main thread, mike said it does (he tried on his unit)
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Fungus on March 04, 2017, 04:03:31 pm
Does anybody know how they are supporting the 3 wire SPI interface on a 2 channel scope ( http://literature.cdn.keysight.com/litweb/pdf/5992-1965EN.pdf (http://literature.cdn.keysight.com/litweb/pdf/5992-1965EN.pdf) ) ? Do you have to pick either MISO or MOSI or are they doing something interesting like routing the clock through the external trigger port ?
in the main thread, mike said it does (he tried on his unit)

The trigger input acts like a single digital channel. It actually shows a trace on screen for that connector and you can use it as part of the serial decode.

So really it's a 2-and-a-half channel 'scope.

(...which is a big feature IMHO. I'm not sure why they aren't shouting about it in the brochures)
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 04, 2017, 05:16:27 pm
It is exactly like a 1 digital channel mso. It does not need to be the trigger source.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: nctnico on March 04, 2017, 05:33:45 pm
That's literally what they do on boats Paul. It's called a sacrificial anode.
That is something totally different. It works because of the electric conductivity of water.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: nctnico on March 04, 2017, 05:36:29 pm
So really it's a 2-and-a-half channel 'scope.

(...which is a big feature IMHO. I'm not sure why they aren't shouting about it in the brochures)
Or use it in comparisons :palm:
Having a digital third channel is a rare but useful feature so to me it would make sense to put much more emphasis on it.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: KNSSoftware on March 04, 2017, 05:50:14 pm
Other than some cost saving, is there any reason why it could not be 3 channels, with the ability two switch the 'third' to a trigger only?  To be the honest my own question is better posed the other way around... Is there a reason why a four channel scope, where there is no room for the trigger, can't be switched down to three, when the ex trigger is needed?
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 04, 2017, 07:20:55 pm
I just noticed there is no DAC for the function gen - looks like it uses PWM from the FPGA.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: PA0PBZ on March 04, 2017, 07:21:07 pm
Other than some cost saving, is there any reason why it could not be 3 channels, with the ability two switch the 'third' to a trigger only?  To be the honest my own question is better posed the other way around... Is there a reason why a four channel scope, where there is no room for the trigger, can't be switched down to three, when the ex trigger is needed?

There is no need to switch a channel when you want to use it for triggering, you don't need to display it.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: PA0PBZ on March 04, 2017, 07:23:05 pm
I just noticed there is no DAC for the function gen - looks like it uses PWM from the FPGA.

Or maybe R/2R from the FPGA?
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 04, 2017, 08:00:25 pm
I just noticed there is no DAC for the function gen - looks like it uses PWM from the FPGA.

Or maybe R/2R from the FPGA?
Don't see a benefit to that as it needs precision Rs and very stable outputs. The FPGA could be doing PWM ( or some other modulation ) at a few hundred MHz, so 20MHz ( max the wavegen goes to) would be easily doable.  There is a multistage LC filter in there.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: JPortici on March 04, 2017, 08:30:26 pm
Does anybody know how they are supporting the 3 wire SPI interface on a 2 channel scope ( http://literature.cdn.keysight.com/litweb/pdf/5992-1965EN.pdf (http://literature.cdn.keysight.com/litweb/pdf/5992-1965EN.pdf) ) ? Do you have to pick either MISO or MOSI or are they doing something interesting like routing the clock through the external trigger port ?
in the main thread, mike said it does (he tried on his unit)

The trigger input acts like a single digital channel. It actually shows a trace on screen for that connector and you can use it as part of the serial decode.

So really it's a 2-and-a-half channel 'scope.

(...which is a big feature IMHO. I'm not sure why they aren't shouting about it in the brochures)
Exactly. Kinda ends the 2 chan vs 4 chan debate... kinda.

At least in the decoding

I think i won't stop saying it anytime soon: finally. And why is this scope the only one at the moment doing this? (apparently)
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: EEVblog on March 04, 2017, 10:00:00 pm
Does anybody know how they are supporting the 3 wire SPI interface on a 2 channel scope ( http://literature.cdn.keysight.com/litweb/pdf/5992-1965EN.pdf (http://literature.cdn.keysight.com/litweb/pdf/5992-1965EN.pdf) ) ? Do you have to pick either MISO or MOSI or are they doing something interesting like routing the clock through the external trigger port ?
in the main thread, mike said it does (he tried on his unit)

The trigger input acts like a single digital channel. It actually shows a trace on screen for that connector and you can use it as part of the serial decode.
So really it's a 2-and-a-half channel 'scope.
(...which is a big feature IMHO. I'm not sure why they aren't shouting about it in the brochures)

Because you CAN'T display that trace.
You can use it as a serial data input and show the decoded serial signal, but you can't show an analog waveform. i.e. it does not have "trigger view" which was common back in the analog scope days to get an extra input, albeit with limited vertical functionality.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: EEVblog on March 04, 2017, 10:06:27 pm
I just noticed there is no DAC for the function gen - looks like it uses PWM from the FPGA.

Or maybe R/2R from the FPGA?
Don't see a benefit to that as it needs precision Rs and very stable outputs. The FPGA could be doing PWM ( or some other modulation ) at a few hundred MHz, so 20MHz ( max the wavegen goes to) would be easily doable.  There is a multistage LC filter in there.

I noticed that filter.
(https://c1.staticflickr.com/1/603/33181257995_fdf5aeae9d_b.jpg) (https://flic.kr/p/Sy7BZ6)Keysight 1000 X-Series Oscilloscope Teardown (https://flic.kr/p/Sy7BZ6) by Dave Jones (https://www.flickr.com/photos/eevblog/), on Flickr

Did they use a DAC on the 2000/3000X? I don't recall.
If so then another obvious cost saving measure.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Brumby on March 05, 2017, 01:46:37 am
Does the test point labelled "PWM5V" give any hints?
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 05, 2017, 01:56:52 am
I just noticed there is no DAC for the function gen - looks like it uses PWM from the FPGA.

Or maybe R/2R from the FPGA?
Don't see a benefit to that as it needs precision Rs and very stable outputs. The FPGA could be doing PWM ( or some other modulation ) at a few hundred MHz, so 20MHz ( max the wavegen goes to) would be easily doable.  There is a multistage LC filter in there.

I noticed that filter.
(https://c1.staticflickr.com/1/603/33181257995_fdf5aeae9d_b.jpg) (https://flic.kr/p/Sy7BZ6)Keysight 1000 X-Series Oscilloscope Teardown (https://flic.kr/p/Sy7BZ6) by Dave Jones (https://www.flickr.com/photos/eevblog/), on Flickr

Did they use a DAC on the 2000/3000X? I don't recall.
If so then another obvious cost saving measure.
Can't find your pics - had a quick skim of the vids & didn't see a similar filter
incidentally your teardown stills are is something I've had problems finding in the past - maybe you should add an index of the hi-res pics on the eevblog site, AFAICS they're only linked from (some of) the videos
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 05, 2017, 02:08:07 am
Does the test point labelled "PWM5V" give any hints?
Maybe - I was playing with it on a desk that too far away from my main a scope, and it was too fast for the 1000x to probe itself!
 Will look into it some more soon.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: MrBungle on March 05, 2017, 05:05:15 am
Does the test point labelled "PWM5V" give any hints?
I figured that would've just been a power rail, there is one labeled "PWM-FG" though
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Deridex on March 05, 2017, 07:58:40 am
I would not be so hard against the powersupply. I had to fix the powersupply of rarely used Tektronix TDS2012 a few times and i don't think that the tek-powersupply had a better quality. If i remember correctly the problems there were mostly broken solder points
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: tautech on March 05, 2017, 09:49:23 am
I would not be so hard against the powersupply. I had to fix the powersupply of rarely used Tektronix TDS2012 a few times and i don't think that the tek-powersupply had a netter quality. If i remember correctly the problems there were mostly broken solder points
:-//
One I have here (2012B) at least has caps that I've heard of before, mostly Rubycon but still with that gawd awful RTV plastered everywhere.  :--
Why would one want to lock heat into caps ? Just dumb IMHO.  :scared:
PSU is labeled ARTESYN.

And the metalwork has rust on the cut edges too.
TekKeySig rust.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: nfmax on March 05, 2017, 11:16:23 am
I would not be so hard against the powersupply. I had to fix the powersupply of rarely used Tektronix TDS2012 a few times and i don't think that the tek-powersupply had a netter quality. If i remember correctly the problems there were mostly broken solder points
... with that gawd awful RTV plastered everywhere.  :--
Why would one want to lock heat into caps ? Just dumb IMHO.

The RTV is to provide mechanical support for the capacitors, so the solder joints don't break (under shock & vibration). Capacitor mounting clips would be better, but cost a lot more.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Fungus on March 05, 2017, 11:52:53 am
Because you CAN'T display that trace.
You can use it as a serial data input and show the decoded serial signal, but you can't show an analog waveform. i.e. it does not have "trigger view" which was common back in the analog scope days to get an extra input, albeit with limited vertical functionality.

It doesn't show an analog waveform but this screenshot from the other thread shows it being used as a digital input.

(https://www.eevblog.com/forum/testgear/new-keysight-scope-1st-march-2017/?action=dlattach;attach=296451;image)

https://www.eevblog.com/forum/testgear/new-keysight-scope-1st-march-2017/msg1150668/#msg1150668 (https://www.eevblog.com/forum/testgear/new-keysight-scope-1st-march-2017/msg1150668/#msg1150668)

If that screenshot's correct then it's a big feature IMHO. It almost kills the 2 vs. 4 channel debate.

If I was Keysight I'd be shouting about it.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 05, 2017, 12:23:23 pm
To repeat - it IS a completely seperate digital channel in all respects.
It is NOT just trigger view, and can be used as a source for trigger, decode, measurements and display just like a digital channel on an MSO. It has 2 ranges, +/-1.6 and +/-8v with variable threshold - it even lets you specify a probe ratio to show the threshold correctly
Keysight have really missed a trick by not making this clear in their publicity.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Fungus on March 05, 2017, 12:38:14 pm
Keysight have really missed a trick by not making this clear in their publicity.

Agree.

They should be calling it a "2+1 channel oscilloscope" or something.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 05, 2017, 12:51:14 pm
Looking through the datasheet :
It does list "1 digital channel" in the datasheet, but they then confuse things by lumping it in with bus mode
Quote
External trigger can be used as a 3rd channel and displayed on-screen to create a bus-type display
The second part adds confusion. They should really have a paragraph fairly early on, with a screenshot showing 3 different waveforms. Plenty of space on pages 2/3 for this.
The only screenshot showing the 3rd trace is combined with the marginally uiseful bus mode, which most prople will miss.

Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 05, 2017, 02:39:53 pm
Does the test point labelled "PWM5V" give any hints?
I figured that would've just been a power rail, there is one labeled "PWM-FG" though
I did some more poking around....
PWM-FG is a 244KHz PWM  signal used to set the offset.
I traced the analogue signal back to a pin on the Meagzoom ASIC, so looks like they integrated a DAC into that. Seems a liitle odd to put an analogue function into what I'd have thought would be an all digital chip.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: SilverSolder on March 05, 2017, 02:48:20 pm

On the subject of "good" vs. "unknown" quality capacitors like we have seen in this product -  is there a non destructive way to measure or test electrolytic capacitors to get an indication of how reliable they are likely to be?
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 05, 2017, 03:08:35 pm

On the subject of "good" vs. "unknown" quality capacitors like we have seen in this product -  is there a non destructive way to measure or test electrolytic capacitors to get an indication of how reliable they are likely to be?
No, because the only thing that matters is how they perform at the temperature, voltage  & ripple current they're running at. You might find they degrade differently to known high-quality caps at higher temps or ripple current but that won't tell you much about lifetime in a specific application. 
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 05, 2017, 03:20:54 pm
Found a UART - boot text below.
Nothing too interesting, apart from "Bandwidth 200MHz" perhaps...
BLT_MODULE is the plug-in module, presumably CANINE_BOARD is the main PCB.
Looks like they are using ADC+resistors for variant detection (e.g."BLT_PRODUCT_CONFIG_0, 1.246v, ID4")
I wonder if the thing about  "USB PHY workaround" is connected with the unpopulated chip near the USB connector. This is bypassed with a couple of 0R links on the underside. Hard to think what this might be, as it's in the device, not host connection - latter could be a hub for touchscreen etc.
Only other connections I could see ( including by x-ray) are power and possibly one line to the module.
I wonder if it was something like a hub chip to overcome an errata issue on the SPEAR chip.
 
Code: [Select]
<00>
U-Boot 2010.03 (Oct 18 2011 - 14:28:06)Agilent P500
CPU:   SPEAr600
DRAM:  128 MiB
Flash: 512 KiB
NAND:  internal ecc 128 MiB

Debug serial initialized ........OK
RTC: 2024-17-3   1:95:1.27 UTC
Microsoft Windows CE Bootloader Common Library Version 1.4 Built May  7 2015 01:38:03
Microsoft Windows CE 6.0 Ethernet Bootloader for the Agilent P500 board
Adaptation performed by Agilent Technologies (c) 2008
PHY not found.
System ready!
Preparing for download...
RTC: 2024-17-3   1:95:1.27 UTC
 Loading image 1 from memory at 0xD0600000
O
BL_IMAGE_TYPE_BIN
X
XXXXOOOOXXOOOOOOOOXOXOOOOOOOOXOOOXOOOOXXOOOOOOOOOXOOOOXOXXOXOXXOXOXOXOXXXXOOXXXOOOOOOXXOXXOXXXXXXOOOXXXOXXOOOXXXOXXOOOOXOOXXOOOXOOOOXOXOOOOOXOOOXOOXOXXOXOXXXXXXOXXXXOOOXOOOXOXOOOOXOOOOXOXOXOOOOOOXX
OOOXOOXOOOOXOOOOXOOXXOOXOOOOOOOOOXOOOOXOOOOOOXOXOOOOXOXOOOOOOOXXOOXOOXOXOOOXOOOXOOXXOXOXOOOXOXXXXXOXOXXXOXXXXOXOXXOOOXXXXOXXXXOXXXXXXXOXXXXXXOXXOXXOXXOOXXOXXXOXXXXOOOXXX
OOOXXXOXXOOXOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOXXOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOXXXXXOOOXOXOOXOOXXXXXXXXXXXXXrom_offset=0x0.
XXImageStart = 0x80361000, ImageLength = 0x1A80C40, LaunchAddr = 0x80362000
Completed file(s):
-------------------------------------------------------------------------------
[0]: Address=0x80361000  Length=0x1A80C40  Name="" Target=RAM
 Loading image 1 succeeded.
ROMHDR at Address 80361044h
Preparing launch...
RTC: 2024-17-3   1:95:1.30 UTC
Launching windows CE image by jumping at address 0x  362000
Windows CE Kernel for ARM (Thumb Enabled) Built on Mar  8 2013 at 17:05:33
Setting up for a Cold Reboot
Done Setting up for a Cold Reboot
Windows CE Firmware Init
BSP 1.0.0 for the SPEARHEAD600AB board (built Sep 28 2016)
Adaptation performed by ADENEO (c) 2005
+OALIntrInit
-OALIntrInit(rc = 1)
Initialize driver globals Zeros area...
pDrvGlobalArea 0xa0060000  size 0x800 (0xa0060800 -0xa0060000)
Initialize driver globals Zeros area...done
 OALKitlStart
Firmware Init Done.
OALIoctlHalEnterI2cCriticalSection init i2c cs
++SER_Init: context Drivers\Active\14
SER_Init, dwIndex:2
SER2 got sysintr:0x00000017
SER2 Serial Port, new baud rate:0x1c200  (UARTCLK:48000000 IBRD:0x1a FBRD:0x2)
OHCI\system.c, GCFG_USBH1_SW_RST
OHCI\system.c, GCFG_USBH2_SW_RST
LAN PHY NOT detected.
DeleteP500EnetRegistry:
   \Comm\GMAC 0x0
   \Comm\GMAC1 0x0
   \Comm\Tcpip\Linkage 0x0
   \Drivers\Virtual 0x0
   \Drivers\BuiltIn\LIN 0x5
LIN: Data Valid
BALDWIN_DDI: cBaldwinHwIf::Init: Initializing...
BALDWIN_DDI: cBaldwinHwIf::Init: Scope successfully identified.
BALDWIN_DDI: cBaldwinHwIf::Init: Success!
Device load time:
   NANDFLASH: 0 ms
   SNANDFLASH: 0 ms
SHIM DLL, LoadRealDll [PalIO.dll] for [AgilentPalIO.dll]
SHIM [AgilentPalIO.dll] Get Process Addresses
LaunchInfiniiVision:
=========================================
BLT Product Config 24
   Bandwidth   : 200MHz
   #Channel    : 2
   Board Rev   : FPR
   Clk Gating  : Baldwin
   Sample Rate : 4GSa
   LAN PHY     : No
BLT Module Config 02
   Rev         : LP3
   Sample Rate : 5GSa/s
=========================================
BLT_PRODUCT_CONFIG_0, 1.246v, ID4
BLT_PRODUCT_CONFIG_1, 0.694v, ID2
BLT_MODULE_CONFIG_0, 0.687v, ID2
BLT_MODULE_CONFIG_1, 0.007v, ID0
CANINE_BOARD_REV, 0.000v, ID0
CANINE_MODEL_NAME: MARSUPIAL, 1.733v, ID6, MARSUPIAL
CANINE_EXTMODULE, 2.485v, ID8, SWID8
CANINE_MSO_REV, 0.628v, ID2, SWID2
SHIM DLL, LoadRealDll [PalSStorage.dll] for [AgilentPalSStorage.dll]
SHIM [AgilentPalSStorage.dll] Get Process Addresses
Released build, Sep 28 2016, 00:17:51
Initializing FPGA...
************************************
FPGA Type: Marsupial
Ver: 1.067 Released
Build Time: Tue Jun 14 17:13:42 2016
Build Machine: 2UA5461ZWH
************************************
cMarsupialCalMgr::cMarsupialUserCalFactors::cMarsupialUserCalFactors size 146412
cMarsupialCalMgr::cMarsupialServiceCalFactors::cMarsupialServiceCalFactors size 704
cMarsupialCalMgr::cMarsupialFactoryCalFactors::cMarsupialFactoryCalFactors size 896
Calibration mode User
Recall \Secure\cal\FactoryCal2.dat - ok
Recall \Secure\cal\ServiceCal1.dat - ok
Recall \Secure\cal\UserCal8.dat - ok
Cal Date Tue Sep 27 16:54:24 2016
will do USB phy workaround: CheckCRC
Startup sequence is complete.
System has been running 17.270830 seconds
Start Up Sequence 7.534334
Memory Load 52%
   System Physical Memory 37.898 / 73.465 MB
   Process Virtual Memory 46.938 / 1024.000 MB
-----> InfiniiVision is running <-----
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: amitchell on March 05, 2017, 04:25:42 pm
Are these measured voltages or versions or something else?
Quote
BLT_PRODUCT_CONFIG_0, 1.246v, ID4
BLT_PRODUCT_CONFIG_1, 0.694v, ID2
BLT_MODULE_CONFIG_0, 0.687v, ID2
BLT_MODULE_CONFIG_1, 0.007v, ID0
CANINE_BOARD_REV, 0.000v, ID0
CANINE_MODEL_NAME: MARSUPIAL, 1.733v, ID6, MARSUPIAL
CANINE_EXTMODULE, 2.485v, ID8, SWID8
CANINE_MSO_REV, 0.628v, ID2, SWID2

Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 05, 2017, 04:39:06 pm
Are these measured voltages or versions or something else?
Quote
BLT_PRODUCT_CONFIG_0, 1.246v, ID4
BLT_PRODUCT_CONFIG_1, 0.694v, ID2
BLT_MODULE_CONFIG_0, 0.687v, ID2
BLT_MODULE_CONFIG_1, 0.007v, ID0
CANINE_BOARD_REV, 0.000v, ID0
CANINE_MODEL_NAME: MARSUPIAL, 1.733v, ID6, MARSUPIAL
CANINE_EXTMODULE, 2.485v, ID8, SWID8
CANINE_MSO_REV, 0.628v, ID2, SWID2
Looks like measured voltages from option resistors, about 0.3V per ID number
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: amitchell on March 05, 2017, 05:03:48 pm
So potential to change values and set different options/configs? I guess we can probe around and search for those numbers, I don't have my scope yet though.  I assume that the daughter board has hardware that can do a lot more than what it is configured as.


Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: TheSteve on March 05, 2017, 05:15:21 pm
So at boot it appears to indicate what it is capable of. Swapping jumpers might just bump it to a 200 MHz model with 4 GS/s.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: amitchell on March 05, 2017, 05:19:21 pm
So at boot it appears to indicate what it is capable of. Swapping jumpers might just bump it to a 200 MHz model with 4 GS/s.

It looks like it may do 5 GSa/s Do we have info on potential memory unlock?

Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: RGB255_0_0 on March 05, 2017, 05:19:25 pm
Need a firmware dump  :-+
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: nctnico on March 05, 2017, 05:45:11 pm
Am I crazy in thinking did they send one to Mike for the purpose of hacking it?  >:D
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: floobydust on March 05, 2017, 05:55:38 pm
The firmware is so buggy on these Keysight scopes that I avoid them for any troubleshooting work.
Sexy LCD and fast CPU but I had to laugh when I kept getting a blank trace and putting a Tektronix to the same probe point gave me a trace, just as I expected.
I especially love setting it to AC coupled and turning the trigger level negative -0.2V really makes sense  :palm:
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: tautech on March 05, 2017, 06:28:26 pm
I would not be so hard against the powersupply. I had to fix the powersupply of rarely used Tektronix TDS2012 a few times and i don't think that the tek-powersupply had a netter quality. If i remember correctly the problems there were mostly broken solder points
... with that gawd awful RTV plastered everywhere.  :--
Why would one want to lock heat into caps ? Just dumb IMHO.

The RTV is to provide mechanical support for the capacitors, so the solder joints don't break (under shock & vibration). Capacitor mounting clips would be better, but cost a lot more.
Of course, but needed on short secondary side caps.....not.
In this SMPS in this Tek the caps are only 22mm high.  ::)

Maybe this the new way to protect caps from thermal degradation...smother them in RTV.  :-DD
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 05, 2017, 07:08:49 pm
So at boot it appears to indicate what it is capable of. Swapping jumpers might just bump it to a 200 MHz model with 4 GS/s.
I doubt it's that simple - the application software will need to also co-operate, and I think the functionality, layout etc. is sufficiently different that it's likely to be different builds, not universal firmware that decides funcitonality based on what hardware it thinks it's running on. Also the flash may be smaller to reduce cost.
I could be wrong-  I'll leave it to others to investigate-too much other stuff to play with.

Not too implausible that some extra funcitonality might be possible. I did look at the SPEAR datasheet to see if adding a magjack would be possible, but looks like it needs an external PHY as well. The UART pins are on that unpopulated connector and I'd put money on the PHY connections also being on there. The SPEAR BGA is fully via'd so it would be possible to trace it out. Would be a lot of work only to maybe find the firmware didn't have networking built in.

It may be that there are some hidden options in the firmware, but can't imagine there's a link to turn it into a 3000X anywhere.


Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: TheSteve on March 05, 2017, 07:43:26 pm
Well it is probably that easy to fool it into believing it has those spec's. That doesn't mean the analog front end etc will cooperate.

The jumper matrix could be just above and to the right of the 25P40, unless it reads the settings from the analog board through a connector.

I don't think we can blame Mike for not experimenting further, there are now some areas to play with if one falls into the right hands.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 06, 2017, 12:45:44 am
And don't forget that many features will be controlled by license keys etc., though I suspect there will be people working on that at some point.

Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: EEVblog on March 06, 2017, 12:57:10 am
And don't forget that many features will be controlled by license keys etc., though I suspect there will be people working on that at some point.

Yes, you can forget about hardware hacks being useful if you don't hack the license keys as well.
In theory it should only require a license key hack on the EDU G model.
The bandwidth, serial decode, and wavegen are all software license key options. So even if you got the non-G version and added the signal gen hardware (if the footprints are there), then you still need the license key to enable it.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: amitchell on March 06, 2017, 01:55:36 am
And don't forget that many features will be controlled by license keys etc., though I suspect there will be people working on that at some point.

Yes, you can forget about hardware hacks being useful if you don't hack the license keys as well.
In theory it should only require a license key hack on the EDU G model.
The bandwidth, serial decode, and wavegen are all software license key options. So even if you got the non-G version and added the signal gen hardware (if the footprints are there), then you still need the license key to enable it.

Are the prize oscilloscopes fully loaded @EEVblog?
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: EEVblog on March 06, 2017, 02:22:56 am
Only other connections I could see ( including by x-ray) are power and possibly one line to the module.

Did the x-ray show you how many layers the processor board is?
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: MrBungle on March 06, 2017, 02:46:41 am
Only other connections I could see ( including by x-ray) are power and possibly one line to the module.

Did the x-ray show you how many layers the processor board is?

A visual did....

Back on topic, the sub-board is 10 layers.

Edit: bah! Photo didn't copy across. Maybe this will work....
(https://www.eevblog.com/forum/blog/eevblog-976-keysight-1000-x-series-oscilloscope-teardown/?action=dlattach;attach=296623;image)
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: EEVblog on March 06, 2017, 03:11:07 am
A visual did....
Edit: bah! Photo didn't copy across. Maybe this will work....
(https://www.eevblog.com/forum/blog/eevblog-976-keysight-1000-x-series-oscilloscope-teardown/?action=dlattach;attach=296623;image)

Thanks. Getting medieaval was going to be my next step  ;D
It did look like at least 10
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: EEVblog on March 06, 2017, 04:21:16 am
Found a UART - boot text below.

Where did you find the UART?
I found a UART port on a via near the unused board-board header operating at 57600 but it seems to get gibberish out on boot. Unless I'm not decoding it right.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Fungus on March 06, 2017, 07:28:03 am
The firmware is so buggy on these Keysight scopes that I avoid them for any troubleshooting work.

Weird. I was assured by several forum members that only Rigol 'scopes had bugs.

Sexy LCD and fast CPU but I had to laugh when I kept getting a blank trace and putting a Tektronix to the same probe point gave me a trace, just as I expected.
I especially love setting it to AC coupled and turning the trigger level negative -0.2V really makes sense  :palm:

 :-//

Anywhere that's "on the signal" makes perfect sense.

"Off the signal" makes sense too - you might be looking for rare signal glitches.

If your Tek can't do it then it's a problem with the Tek, not the Keysight.

Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 06, 2017, 09:08:11 am
Found a UART - boot text below.

Where did you find the UART?
I found a UART port on a via near the unused board-board header operating at 57600 but it seems to get gibberish out on boot. Unless I'm not decoding it right.
On the unpopulated board-board connector on the processor board - looks like the only one of the 2 SPEAR600 UARTS connected to anything.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: PA0PBZ on March 06, 2017, 09:48:35 am
Found a UART - boot text below.

Where did you find the UART?
I found a UART port on a via near the unused board-board header operating at 57600 but it seems to get gibberish out on boot. Unless I'm not decoding it right.
On the unpopulated board-board connector on the processor board - looks like the only one of the 2 SPEAR600 UARTS connected to anything.

There is at least one more port, because half way into the boot text:

SER_Init: context Drivers\Active\14
SER_Init, dwIndex:2
SER2 got sysintr:0x00000017
SER2 Serial Port, new baud rate:0x1c200  (UARTCLK:48000000 IBRD:0x1a FBRD:0x2)


1C200 = 115.200
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 06, 2017, 10:41:12 am
Found a UART - boot text below.

Where did you find the UART?
I found a UART port on a via near the unused board-board header operating at 57600 but it seems to get gibberish out on boot. Unless I'm not decoding it right.
On the unpopulated board-board connector on the processor board - looks like the only one of the 2 SPEAR600 UARTS connected to anything.

There is at least one more port, because half way into the boot text:

SER_Init: context Drivers\Active\14
SER_Init, dwIndex:2
SER2 got sysintr:0x00000017
SER2 Serial Port, new baud rate:0x1c200  (UARTCLK:48000000 IBRD:0x1a FBRD:0x2)


1C200 = 115.200
As far as I could tell only one of the 2 UARTs (UART2) on the SPEAR600 is connected to anything. Didn't test comprehensively - just a continuity swipe over all the connector pins and anything that looked like a test pad.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: EEVblog on March 06, 2017, 11:04:19 am
My terminal dump:
(Should be the same as Mike's, I haven't compared)

Code: [Select]
U-Boot 2010.03 (Oct 18 2011 - 14:28:06)Agilent P500

CPU:   SPEAr600
DRAM:  128 MiB
Flash: 512 KiB
NAND:  internal ecc 128 MiB

Debug serial initialized ........OK
RTC: 2024-16-10   7:86:38.44 UTC

Microsoft Windows CE Bootloader Common Library Version 1.4 Built May  7 2015 01:38:03
Microsoft Windows CE 6.0 Ethernet Bootloader for the Agilent P500 board
Adaptation performed by Agilent Technologies (c) 2008

PHY not found.

System ready!
Preparing for download...
RTC: 2024-16-10   7:86:38.44 UTC
 Loading image 1 from memory at 0xD0600000
O
BL_IMAGE_TYPE_BIN

X
XXXXOOOOXXOOOOOOOOXOXOOOOOOOOXOOOXOOOOXXOOOOOOOOOXOOOOXOXXOXOXXOXOXOXOXXXXOOXXXOOOOOOXXOXXOXXXXXXOOOXXXOXXOOOXXXOXXOOOOXOOXXOOOXOOOOXOXOOOOOXOOOXOOXOXXOXOXXXXXXOXXXXOOOXOOOXOXOOOOXOOOOXOXOXOOOOOOXX
OOOXOOXOOOOXOOOOXOOXXOOXOOOOOOOOOXOOOOXOOOOOOXOXOOOOXOXOOOOOOOXXOOXOOXOXOOOXOOOXOOXXOXOXOOOXOXXXXXOXOXXXOXXXXOXOXXOOOXXXXOXXXXOXXXXXXXOXXXXXXOXXOXXOXXOOXXOXXXOXXXXOOOXXX
OOOXXXOXXOOXOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOXXOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOXXXXXOOOXOXOOXOOXXXXXXXXXXXXXrom_offset=0x0.
XXImageStart = 0x80361000, ImageLength = 0x1A80C40, LaunchAddr = 0x80362000

Completed file(s):
-------------------------------------------------------------------------------
[0]: Address=0x80361000  Length=0x1A80C40  Name="" Target=RAM
 Loading image 1 succeeded.
ROMHDR at Address 80361044h
Preparing launch...
RTC: 2024-16-10   7:86:38.47 UTC
Launching windows CE image by jumping at address 0x  362000

Windows CE Kernel for ARM (Thumb Enabled) Built on Mar  8 2013 at 17:05:33
Setting up for a Cold Reboot
Done Setting up for a Cold Reboot
Windows CE Firmware Init
BSP 1.0.0 for the SPEARHEAD600AB board (built Sep 28 2016)
Adaptation performed by ADENEO (c) 2005
+OALIntrInit
-OALIntrInit(rc = 1)
Initialize driver globals Zeros area...
pDrvGlobalArea 0xa0060000  size 0x800 (0xa0060800 -0xa0060000)
Initialize driver globals Zeros area...done
 OALKitlStart
Firmware Init Done.
OALIoctlHalEnterI2cCriticalSection init i2c cs
++SER_Init: context Drivers\Active\14
SER_Init, dwIndex:2
SER2 got sysintr:0x00000017
SER2 Serial Port, new baud rate:0x1c200  (UARTCLK:48000000 IBRD:0x1a FBRD:0x2)
OHCI\system.c, GCFG_USBH1_SW_RST
OHCI\system.c, GCFG_USBH2_SW_RST
LAN PHY NOT detected.
DeleteP500EnetRegistry:
   \Comm\GMAC 0x0
   \Comm\GMAC1 0x0
   \Comm\Tcpip\Linkage 0x0
   \Drivers\Virtual 0x0
   \Drivers\BuiltIn\LIN 0x5
LIN: Data Valid
BALDWIN_DDI: cBaldwinHwIf::Init: Initializing...
BALDWIN_DDI: cBaldwinHwIf::Init: Scope successfully identified.
BALDWIN_DDI: cBaldwinHwIf::Init: Success!
Device load time:
   NANDFLASH: 1 ms
   SNANDFLASH: 1 ms
SHIM DLL, LoadRealDll [PalIO.dll] for [AgilentPalIO.dll]
SHIM [AgilentPalIO.dll] Get Process Addresses
LaunchInfiniiVision:
=========================================
BLT Product Config 24
   Bandwidth   : 200MHz
   #Channel    : 2
   Board Rev   : FPR
   Clk Gating  : Baldwin
   Sample Rate : 4GSa
   LAN PHY     : No
BLT Module Config 02
   Rev         : LP3
   Sample Rate : 5GSa/s
=========================================
BLT_PRODUCT_CONFIG_0, 1.251v, ID4
BLT_PRODUCT_CONFIG_1, 0.692v, ID2
BLT_MODULE_CONFIG_0, 0.687v, ID2
BLT_MODULE_CONFIG_1, 0.005v, ID0
CANINE_BOARD_REV, 0.002v, ID0
CANINE_MODEL_NAME: MARSUPIAL, 1.738v, ID6, MARSUPIAL
CANINE_EXTMODULE, 2.488v, ID8, SWID8
CANINE_MSO_REV, 0.628v, ID2, SWID2
SHIM DLL, LoadRealDll [PalSStorage.dll] for [AgilentPalSStorage.dll]
SHIM [AgilentPalSStorage.dll] Get Process Addresses
Released build, Sep 28 2016, 00:17:51
Initializing FPGA...
************************************
FPGA Type: Marsupial
Ver: 1.067 Released
Build Time: Tue Jun 14 17:13:42 2016
Build Machine: 2UA5461ZWH
************************************
cMarsupialCalMgr::cMarsupialUserCalFactors::cMarsupialUserCalFactors size 146412
cMarsupialCalMgr::cMarsupialServiceCalFactors::cMarsupialServiceCalFactors size 704
cMarsupialCalMgr::cMarsupialFactoryCalFactors::cMarsupialFactoryCalFactors size 896
Calibration mode User
Recall \Secure\cal\FactoryCal2.dat - ok
Recall \Secure\cal\ServiceCal1.dat - ok
Recall \Secure\cal\UserCal8.dat - ok
Cal Date Sun Sep 25 15:11:58 2016
will do USB phy workaround: CheckCRC
Startup sequence is complete.
System has been running 16.841095 seconds
Start Up Sequence 7.470958
Memory Load 50%
   System Physical Memory 36.441 / 73.465 MB
   Process Virtual Memory 46.938 / 1024.000 MB
-----> InfiniiVision is running <-----
[00]
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: EEVblog on March 06, 2017, 11:07:46 am
As far as I could tell only one of the 2 UARTs (UART2) on the SPEAR600 is connected to anything. Didn't test comprehensively - just a continuity swipe over all the connector pins and anything that looked like a test pad.

I can't find another port either.
If it's there and we have probed it then maybe it doesn't output anything at boot?

I might have identified one of the config resistors, no time right now to play though unfortunately.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: EEVblog on March 06, 2017, 11:27:14 am
I also have a prototype unit and here is the dump:
Some slight differences including resistor set voltages and modes.
Code: [Select]


U-Boot 2010.03 (Oct 18 2011 - 14:28:06)Agilent P500

CPU:   SPEAr600
DRAM:  128 MiB
Flash: 512 KiB
NAND:  internal ecc 128 MiB

Debug serial initialized ........OK
RTC: 2024-17-3   2:85:23.27 UTC

Microsoft Windows CE Bootloader Common Library Version 1.4 Built May  7 2015 01:38:03
Microsoft Windows CE 6.0 Ethernet Bootloader for the Agilent P500 board
Adaptation performed by Agilent Technologies (c) 2008

PHY not found.

System ready!
Preparing for download...
RTC: 2024-17-3   2:85:23.27 UTC
 Loading image 1 from memory at 0xD0600000
O
BL_IMAGE_TYPE_BIN

X
XXXXOOOOXXOOOOOOOOXOXOOOOOOOOXOOOXOOOOXXOOOOOOOOOXOOOOXOXXOXOXXOXOXOXOXXXXOOXXXOOOOOOXXOXXOXXXXXXOOOXXXOOXXOOXXXOXXOOOOXOOXXOOOXOOOOXOXOOOOOXOOOXOOOXXXOXOXXXXXXOXXXXOOOXOOOXOXOOOOXOOOOXOXOXOOOOOOXOOOX
OOXOOOOXOOOOXOOXXOOXOOOOOOOOOXOOOOXOOOORewrite recommended, internal ECC corrected data at 0x166f
OOXOXOOOOOXXOOOOOOOXOXOXOOXOXOOOXOOOXOOXOXXOXOOOXOXXXXXOXOXXOXXXXXOXOXOXOOXXOXXXXXXOXXXXXXXOXXXXXXOXXOXXOXXOOOXXXXXOXXXXOOOXOXXOOX
XOXXXOOXOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOXXOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOXXXXOXOOOXOXOOXOOXXXXXXXXXXXXXrom_offset=0x0.
XXImageStart = 0x80361000, ImageLength = 0x1AEEE00, LaunchAddr = 0x80362000

Completed file(s):
-------------------------------------------------------------------------------
[0]: Address=0x80361000  Length=0x1AEEE00  Name="" Target=RAM
 Loading image 1 succeeded.
ROMHDR at Address 80361044h
Preparing launch...
RTC: 2024-17-3   2:85:23.42 UTC
Launching windows CE image by jumping at address 0x  362000

Windows CE Kernel for ARM (Thumb Enabled) Built on Mar  8 2013 at 17:05:33
Setting up for a Cold Reboot
Done Setting up for a Cold Reboot
Windows CE Firmware Init
BSP 1.0.0 for the SPEARHEAD600AB board (built Aug 30 2016)
Adaptation performed by ADENEO (c) 2005
+OALIntrInit
-OALIntrInit(rc = 1)
Initialize driver globals Zeros area...
pDrvGlobalArea 0xa0060000  size 0x800 (0xa0060800 -0xa0060000)
Initialize driver globals Zeros area...done
 OALKitlStart
Firmware Init Done.
OALIoctlHalEnterI2cCriticalSection init i2c cs
++SER_Init: context Drivers\Active\14
SER_Init, dwIndex:2
SER2 got sysintr:0x00000017
SER2 Serial Port, new baud rate:0x1c200  (UARTCLK:48000000 IBRD:0x1a FBRD:0x2)
ERROR: c:\WINCE600\PLATFORM\COMMON\SRC\SOC\STM\COMMON\DRIVERS\NandFlash\.\stm_NandFlash.c line 1460: RewriteEbootBlocks from 0x166f to 0x166f
OHCI\system.c, GCFG_USBH1_SW_RST
OHCI\system.c, GCFG_USBH2_SW_RST
LAN PHY NOT detected.
DeleteP500EnetRegistry:
   \Comm\GMAC 0x0
   \Comm\GMAC1 0x0
   \Comm\Tcpip\Linkage 0x0
   \Drivers\Virtual 0x0
   \Drivers\BuiltIn\LIN 0x5
LIN: Data Valid
BALDWIN_DDI: cBaldwinHwIf::Init: Initializing...
BALDWIN_DDI: cBaldwinHwIf::Init: Scope successfully identified.
BALDWIN_DDI: cBaldwinHwIf::Init: Success!
Device load time:
   NANDFLASH: 0 ms
   SNANDFLASH: 0 ms
SHIM DLL, LoadRealDll [PalIO.dll] for [AgilentPalIO.dll]
SHIM [AgilentPalIO.dll] Get Process Addresses
LaunchInfiniiVision:
=========================================
BLT Product Config 25
   Bandwidth   : 200MHz
   #Channel    : 2
   Board Rev   : PP
   Clk Gating  : Baldwin
   Sample Rate : 4GSa
   LAN PHY     : No
BLT Module Config 02
   Rev         : LP3
   Sample Rate : 5GSa/s
=========================================
BLT_PRODUCT_CONFIG_0, 1.515v, ID5
BLT_PRODUCT_CONFIG_1, 0.689v, ID2
BLT_MODULE_CONFIG_0, 0.694v, ID2
BLT_MODULE_CONFIG_1, 0.015v, ID0
CANINE_BOARD_REV, 0.002v, ID0
CANINE_MODEL_NAME: MARSUPIAL, 1.740v, ID6, MARSUPIAL
CANINE_EXTMODULE, 2.490v, ID8, SWID8
CANINE_MSO_REV, 0.648v, ID2, SWID2
SHIM DLL, LoadRealDll [PalSStorage.dll] for [AgilentPalSStorage.dll]
SHIM [AgilentPalSStorage.dll] Get Process Addresses
QA build, Aug 30 2016, 06:12:04
Initializing FPGA...
************************************
FPGA Type: Marsupial
Ver: 1.067 Released
Build Time: Tue Jun 14 17:13:42 2016
Build Machine: 2UA5461ZWH
************************************
cMarsupialCalMgr::cMarsupialUserCalFactors::cMarsupialUserCalFactors size 146412
cMarsupialCalMgr::cMarsupialServiceCalFactors::cMarsupialServiceCalFactors size 704
cMarsupialCalMgr::cMarsupialFactoryCalFactors::cMarsupialFactoryCalFactors size 896
Calibration mode User

cCalMgr: ************************************************
cCalMgr: >>>>> Verifying ASCII cal factors. <<<<<
cCalMgr: User cal mode - ok
cCalMgr: Service cal mode - ok
cCalMgr: Factory cal mode - ok
cCalMgr: ************************************************

Recall \Secure\cal\FactoryCal2.dat - ok
Recall \Secure\cal\ServiceCal1.dat - ok
Recall \Secure\cal\UserCal8.dat - ok
Cal Date Wed Aug 31 12:54:45 2016
will do USB phy workaround: CheckCRC
Startup sequence is complete.

CEO: **************************************************
CEO: >>>>> Validating control system. <<<<<
CEO: Total groups = 38, Total groups validated = 38.
CEO: Total ctrls = 1464, Total ctrls validated = 1464.
CEO: OK, errors = 0, warnings = 0, info = 0
CEO: **************************************************

System has been running 19.383053 seconds
Start Up Sequence 10.295580
Memory Load 55%
   System Physical Memory 39.547 / 73.184 MB
   Process Virtual Memory 50.250 / 1024.000 MB
-----> InfiniiVision is running <-----
[00]
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 06, 2017, 11:38:14 am
Might be interesting compare messages and ID values with a 2/3000X unit.

Here are the points on the 1000X module
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Fungus on March 06, 2017, 01:02:54 pm
Microsoft Windows CE

 :scared:
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 06, 2017, 01:20:15 pm
Microsoft Windows CE

 :scared:
Nothing wrong with that - it gets the job done.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mmagin on March 06, 2017, 04:20:33 pm
My terminal dump:
(Should be the same as Mike's, I haven't compared)

Code: [Select]
U-Boot 2010.03 (Oct 18 2011 - 14:28:06)Agilent P500

Wow, is there anything that doesn't have U-Boot in it?
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: rrinker on March 06, 2017, 07:56:23 pm
 The German Navy is everywhere!

Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Keysight DanielBogdanoff on March 06, 2017, 08:45:01 pm
The firmware is so buggy on these Keysight scopes that I avoid them for any troubleshooting work.
Sexy LCD and fast CPU but I had to laugh when I kept getting a blank trace and putting a Tektronix to the same probe point gave me a trace, just as I expected.
I especially love setting it to AC coupled and turning the trigger level negative -0.2V really makes sense  :palm:

What scope are you using? Overall I'd expect that it's more of a use-model thing than a buggy firmware thing...
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 06, 2017, 11:36:30 pm
The firmware is so buggy on these Keysight scopes that I avoid them for any troubleshooting work.

Weird. I was assured by several forum members that only Rigol 'scopes had bugs.

Sexy LCD and fast CPU but I had to laugh when I kept getting a blank trace and putting a Tektronix to the same probe point gave me a trace, just as I expected.
I especially love setting it to AC coupled and turning the trigger level negative -0.2V really makes sense  :palm:

 :-//

Anywhere that's "on the signal" makes perfect sense.

"Off the signal" makes sense too - you might be looking for rare signal glitches.

If your Tek can't do it then it's a problem with the Tek, not the Keysight.
I've been using HP/Agilent/Keysight scopes daily for 25+years and have yet to see a bug of any significance.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 06, 2017, 11:38:09 pm
Am I crazy in thinking did they send one to Mike for the purpose of hacking it?  >:D
If that was the intention, they'd have sent the lowest end version ;D
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: SilverSolder on March 07, 2017, 12:22:57 am
The firmware is so buggy on these Keysight scopes that I avoid them for any troubleshooting work.
Sexy LCD and fast CPU but I had to laugh when I kept getting a blank trace and putting a Tektronix to the same probe point gave me a trace, just as I expected.
I especially love setting it to AC coupled and turning the trigger level negative -0.2V really makes sense  :palm:

What scope are you using? Overall I'd expect that it's more of a use-model thing than a buggy firmware thing...

Is it not possible for an AC signal to drop below zero volts?  You should be able to set the trigger at any level in the AC wave form - positive or negative - no?  That seems to be how my scope works.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 07, 2017, 12:29:14 am
The firmware is so buggy on these Keysight scopes that I avoid them for any troubleshooting work.
Sexy LCD and fast CPU but I had to laugh when I kept getting a blank trace and putting a Tektronix to the same probe point gave me a trace, just as I expected.
I especially love setting it to AC coupled and turning the trigger level negative -0.2V really makes sense  :palm:

What scope are you using? Overall I'd expect that it's more of a use-model thing than a buggy firmware thing...

Is it not possible for an AC signal to drop below zero volts?  You should be able to set the trigger at any level in the AC wave form - positive or negative - no?  That seems to be how my scope works.
Yes, just need to bear in mind that the level would be relative to the avarage level of the signal, not absolute voltage to ground
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: AndyC_772 on March 07, 2017, 07:12:03 am
I've been using HP/Agilent/Keysight scopes daily for 25+years and have yet to see a bug of any significance.

Quoting this because it's probably the single most important specification the scope could have.

My experience has been similar.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: EEVblog on March 08, 2017, 10:09:10 am
https://www.youtube.com/watch?v=YC6JCVHk80c (https://www.youtube.com/watch?v=YC6JCVHk80c)
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: nfmax on March 08, 2017, 12:16:36 pm
Dave - I think you are wrong when you say the 1000X does not have an MSO capability - it is an MSO, with just one digital input. This input seems to be able to do everything a digital input can do on the 2,3,4,6-X series MSO's (provided that is supported by the installed decode options), it's just it's shared with the external trigger function.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 08, 2017, 01:01:02 pm
Dave - I think you are wrong when you say the 1000X does not have an MSO capability - it is an MSO, with just one digital input. This input seems to be able to do everything a digital input can do on the 2,3,4,6-X series MSO's (provided that is supported by the installed decode options), it's just it's shared with the external trigger function.
It isn't really shared with the ext trig function - it is an independent digital input, just like an MSO. You can trigger on any source and still display, decode or measure the digital input
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Jay_Diddy_B on March 08, 2017, 02:55:44 pm
Hi,

I was watching Dave's latest video, are these the configuration resistors:

(https://www.eevblog.com/forum/blog/eevblog-976-keysight-1000-x-series-oscilloscope-teardown/?action=dlattach;attach=297497;image)


They look suspicious. May be try continuity to the ADC inputs on the SPEAR chip?

Regards,

Jay_Diddy_B
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: David Hess on March 08, 2017, 05:36:39 pm
Dave - I think you are wrong when you say the 1000X does not have an MSO capability - it is an MSO, with just one digital input. This input seems to be able to do everything a digital input can do on the 2,3,4,6-X series MSO's (provided that is supported by the installed decode options), it's just it's shared with the external trigger function.

It isn't really shared with the ext trig function - it is an independent digital input, just like an MSO. You can trigger on any source and still display, decode or measure the digital input

An external trigger is more than just a digital MSO input.  Usually the external trigger is sampled at a rate much lower than the digitizer rate so the trigger to sample time has to be interpolated to prevent jitter in the waveform display.

I searched through the available DSOX1000 documentation and did not find any specifications for when the external trigger input is used as a 1-bit digital input.  Maybe they sample it at the vertical sample rate so no interpolation is required?  Can it even be used outside of protocol decoding?
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Someone on March 08, 2017, 10:03:17 pm
I was watching Dave's latest video, are these the configuration resistors:
[picture of processor card]
They look suspicious. May be try continuity to the ADC inputs on the SPEAR chip?
As mike said previously the separate PCB for the processor and ASICs is likely tested independently of the scope (configured to limit signals to 200MHz) and then once its put into the scope assumes its full identity. I'd be looking to the main board for resistor straps or flash memory to identify what its plugged into, unless someone wants to get their hands on the lower end models to compare against.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 09, 2017, 12:33:33 am
Can it even be used outside of protocol decoding?
YES, for the umpteenth time. You can display it, trigger on it, measure from it and decode from it, in any permutation. You can set the threshold over 2 ranges +/-1.6v and +/-8V, and apply probe scaling factors.
It works JUST LIKE AN MSO INPUT.
The only limitation I've noticed is when you are displaying it, you can't select some acquisition modes like avaraging.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: bitseeker on March 09, 2017, 01:59:40 am
I was watching Dave's latest video, are these the configuration resistors:

Were you referring to video #978? Some of those do affect the configuration voltages.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Jay_Diddy_B on March 09, 2017, 02:13:03 pm
I was watching Dave's latest video, are these the configuration resistors:

Were you referring to video #978? Some of those do affect the configuration voltages.

I was referring to the video. I had just scanned the video when I posted the message. Last night I watched the video properly and I realized my question was answered in the video.

I would try and map the known resistive dividers to the corresponding ADC pins on the Spear600, listed in this table:

(https://www.eevblog.com/forum/blog/eevblog-976-keysight-1000-x-series-oscilloscope-teardown/?action=dlattach;attach=297796;image)

Then try and track the other ADC channels to the configuration dividers that control the other functions. It is these other configurations that are the most interesting.

Regards,

Jay_Diddy_B
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: David Hess on March 10, 2017, 03:09:55 am
YES, for the umpteenth time. You can display it, trigger on it, measure from it and decode from it, in any permutation. You can set the threshold over 2 ranges +/-1.6v and +/-8V, and apply probe scaling factors.
It works JUST LIKE AN MSO INPUT.
The only limitation I've noticed is when you are displaying it, you can't select some acquisition modes like avaraging.

You sure cannot tell from the documentation which also leaves out a whole bunch of other questions I had.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: amitchell on March 10, 2017, 03:42:28 am
YES, for the umpteenth time. You can display it, trigger on it, measure from it and decode from it, in any permutation. You can set the threshold over 2 ranges +/-1.6v and +/-8V, and apply probe scaling factors.
It works JUST LIKE AN MSO INPUT.
The only limitation I've noticed is when you are displaying it, you can't select some acquisition modes like avaraging.

You sure cannot tell from the documentation which also leaves out a whole bunch of other questions I had.

The documentation that also uses the word "cheap" to describe the product. I had a laugh at that, not exactly how you want to describe a product from a marketing perspective. It could use a revision to say the least.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 10, 2017, 09:49:34 am
YES, for the umpteenth time. You can display it, trigger on it, measure from it and decode from it, in any permutation. You can set the threshold over 2 ranges +/-1.6v and +/-8V, and apply probe scaling factors.
It works JUST LIKE AN MSO INPUT.
The only limitation I've noticed is when you are displaying it, you can't select some acquisition modes like avaraging.

You sure cannot tell from the documentation which also leaves out a whole bunch of other questions I had.
Yes, the datasheet in particular is very lacking in this respect
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Fungus on March 10, 2017, 09:55:00 am
YES, for the umpteenth time. You can display it, trigger on it, measure from it and decode from it, in any permutation. You can set the threshold over 2 ranges +/-1.6v and +/-8V, and apply probe scaling factors.
It works JUST LIKE AN MSO INPUT.
The only limitation I've noticed is when you are displaying it, you can't select some acquisition modes like avaraging.

You sure cannot tell from the documentation which also leaves out a whole bunch of other questions I had.

Yep. This is one of the 'scopes biggest features IMHO but Keysight doesn't think it's important.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Nobody2 on March 10, 2017, 12:39:02 pm
Considering the separate processor board:
Maybe it has to do with layer count? e.g. they need 4 or 6 layers to route the zillion BGA pins, while the rest is pretty simple and only needs 2. I haven't noticed any layer markings in the video though.

Sorry if this was already mentioned, but I don't have the time right now to check the previous 6 pages.
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 10, 2017, 01:44:41 pm
Considering the separate processor board:
Maybe it has to do with layer count? e.g. they need 4 or 6 layers to route the zillion BGA pins, while the rest is pretty simple and only needs 2. I haven't noticed any layer markings in the video though.

Sorry if this was already mentioned, but I don't have the time right now to check the previous 6 pages.
The sub-board is 10 layers - there's a pic earlier in this thread
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: JPortici on March 10, 2017, 02:51:03 pm
mike, there is an unpopulated QFP + others components near the area where a lan connector would be (according to the mechanical assembly)
what could it be?
could it be perhaps of an ethernet phy? even though there is no jack footprint on the board...
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: mikeselectricstuff on March 10, 2017, 03:09:50 pm
mike, there is an unpopulated QFP + others components near the area where a lan connector would be (according to the mechanical assembly)
what could it be?
could it be perhaps of an ethernet phy? even though there is no jack footprint on the board...
It is in line with the USB device port,and bypassed with 0R resistors  on the underside. I think it may have been something like a hub as a workaround for silicon errata on the SPEAR - the boot text references a USB PHY workaround.
It is definitely not an ethernet PHY.
I haven't checked but I'd be surprised if the unpopulated connector on the sub-board didn't have all the PHY signals. The SPEAR needs an external PHY, so it's not just a case of lashing on a magjack. And it probably isn't supported by the software anyway.

Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: Nemezy on April 07, 2017, 10:26:18 pm
Hi, I decided to replace the fan in my DSOX1102G. Now it is much quieter and less frequent. The next step is to replace the capacitors   ;D
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: amitchell on April 07, 2017, 11:40:37 pm
Hi, I decided to replace the fan in my DSOX1102G. Now it is much quieter and less frequent. The next step is to replace the capacitors   ;D

Nice I bet the rubber mounting grommets help, I am going to order one for myself.   :-+
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: jusaca on April 21, 2017, 11:35:56 am
Is there already some information about the review? Still waiting for it :D
Title: Re: EEVblog #976 - Keysight 1000 X-Series Oscilloscope Teardown
Post by: markus-k on June 30, 2017, 06:07:53 am
I replaced the fan on my 1000X with a Noctua fan too a few weeks ago, the stock one was incredibly loud and annoying. A lot better now, I'm not sure yet if I should silence it further with one of the included low-noise adapters. It's mostly airflow noise now. The heatsinks on the board are getting quite hot, I might have to make some measurements.

Then I got a second 1000X yesterday. To my surprise, it is way quieter than the first one when I turned it on. No need to replace the fan. Now I get why people are saying the scope is silent enough, but it's odd there is such a difference between them. I don't have anything to measure the noise level, but I'd guess it's half as loud and an order of magnitude less annoying.

Maybe there is a quality problem with the fan? The second scope's serial number is lower than the first one, so I guess it's a bit older.