3458A - black edition

[Kleinstein]

A problem with those high end DMMs is that there is a rather limited market. For most cases one does not need a super low noise meter as the voltage source one is looking at will have much higher noise. One usually only sees the combined noise of the source and meter. So it is only with those very low noise source, that may be used in metrology that one really gets a real advantage from a very low noise meter.

The 3458 also has one big asset - it's reputation and history. So the minimal changes are understandable.
Because many of the old meters still get maintenance and repairs, it does help if the boards are interchangeable  - this also limits possible changes.

It seems that they did not change the ground referenced software (at least not much). This is at least the easier part.
Updates to the software an GND references part would be a area that looks possibly, as it does not need slow and expensive new testing of the analog part. I can understand that the priority is to meet RoHS so one is sure one has the hardware, before one invests in SW updates or extra features.

[niner_007]

with Helium coming in thru the multimeter? that wouldn’t be feasible size wise, if anything, I want the DMMs to be smaller, not larger, and not very wise system engineering; there’s also the other components, like resistors, capacitors, and ultimately the PCB itself, the reference is just a small part

on the other hand, you could have the reference external as an option, in the same way some instruments can use a more stable external clock

[ramon]

Not sure everything was done so well-planned and deliberately.

Pobably more like this:

1) Sales: Shit. We cannot sell our 3458A anymore to Europe because of this bloody ROHS stuff.
2) Managemeent to engineers: Whats the cheapest way to get around that ?
3) Engineers to management: We can leave all circuits as they are, but we need to make new PCBs and buy lead-free components
4) Management: OK, do that. And paint it black so it at least looks different and new.
5) Engineers: done.

Management: Is there anything that needs to be improved or fixed?

Engineers: Yes, there is a problem with a custom asic in one board that makes the device drift out of specs, it has happened even on new units. The customer either needs to pay us a few thousands (if not covered by warranty) or then they need to sell the unit on the second hand market. This makes the second hand market full of faulty units that don't meet spec that cannot be recovered into service again without '$,$$$ help' from us.

Management: Ok ... Understood. Just make the ROHS change, nothing else. And be sure that this custom asic works in the same way for this new version.

6) Management to Marketing: Make a campaign about the "new" 3458A and explain why "no change" is a lot better than improved specs, a better user interface or long missing interfaces
7) Markting: Groovy ! Almost no investment, no improvement, still looking as new. Lets write a few journal articles that make us look smart.

[niner_007]

The 30 years of compatibility is a strengths though, no matter what the underlying reasons for the smaller or larger update might have been on the Keysight part, we should be grateful for it, even if it was fueled by reduced investments and cost cuts. For the ASIC, unless you have at large failure data, it's a moot point, for all we know the 8588A has similar failure points We might never know, we don't have access to the data, and likely the Fluke DMM will be obsoleted 10 years from now

Touchscreen, LAN, WiFi, USB, you name it, can absolutely be done, and for little effort on the Keysight side, but why, such UX improvements would be pretty useless, that's not how (and if, it should not) these multi-meters are used. Keithley did the touch screen treatment in DMM7510, and even has some pretty advanced triggering and state machine construction directly in the multimeter, but as long as you have to do something more elaborate, it all gets to down to scripting or remote control, and it all falls down. Never mind the insane number of bugs causing all sort of crashes still lurking around in the DMM7510. You hit one of these crashes, you say goodbye to your memory buffers.

Unless it's analog improvements, you too should not be interested in it And if improvements are made, I very much would prefer them to be iterative and maintain compatibility, that's a hell of a feature to have for over 30 years, the only other system with such compatibility might be Windows lol

[bsdphk]

"Touchscreen, LAN, WiFi, USB, you name it, can absolutely be done, and for little effort on the Keysight side"

I will dispute the "for little effort" part of that claim.

As much as we drool over the gorgeous hardware in the 3458A, more than half the magic is in the software.

While there are things you can do to improve that firmware, adding "Touchscreen, LAN, WiFi, USB" with "little effort" is not one of them.

Just moving the firmware to a CPU which does not execute mc68000 instructions would be a major undertaking, due to the way command-dispatch is implemented.

Ensuring that the new firmware measures *exactly* the same way as the current firmware would take a lot of effort over a couple of years.

Of course one could design a UX unit with a touch-screen, which pretends to be the existing screen & keyboard electrically, but it would be intolerably slow.

Likewise, one could stick a small embedded computer on the GPIB to offer USB/LAN/WIFI interfaces, but that would increase the command latency.

I'm pretty sure that if you poll the people who pay list-price for new 3458A's, they will either tell you that they press a few buttons to do ACAL and to select what to measure and have nothing hooked up to GPIB, or they have something hooked up to GPIB which they *really* do not want to have to change in any way, because it is part of the certification of their manufacturing line.

In other words:  Yes, everybody would absolutely expect better UX and interfaces on a new design, but to the paying 3458A customers adding that would be a good reason to look for a different solution instead.

[SilverSolder]

"Touchscreen, LAN, WiFi, USB, you name it, can absolutely be done, and for little effort on the Keysight side"

I will dispute the "for little effort" part of that claim.

As much as we drool over the gorgeous hardware in the 3458A, more than half the magic is in the software.

While there are things you can do to improve that firmware, adding "Touchscreen, LAN, WiFi, USB" with "little effort" is not one of them.

Just moving the firmware to a CPU which does not execute mc68000 instructions would be a major undertaking, due to the way command-dispatch is implemented.

Ensuring that the new firmware measures *exactly* the same way as the current firmware would take a lot of effort over a couple of years.

Of course one could design a UX unit with a touch-screen, which pretends to be the existing screen & keyboard electrically, but it would be intolerably slow.

Likewise, one could stick a small embedded computer on the GPIB to offer USB/LAN/WIFI interfaces, but that would increase the command latency.

I'm pretty sure that if you poll the people who pay list-price for new 3458A's, they will either tell you that they press a few buttons to do ACAL and to select what to measure and have nothing hooked up to GPIB, or they have something hooked up to GPIB which they *really* do not want to have to change in any way, because it is part of the certification of their manufacturing line.

In other words:  Yes, everybody would absolutely expect better UX and interfaces on a new design, but to the paying 3458A customers adding that would be a good reason to look for a different solution instead.

This is why it might make more sense to develop an entirely new DMM with few backward compatibility concerns as an upgrade path, while leaving the 3458A as it is.

[Andreas]

"Touchscreen, LAN, WiFi, USB, you name it, can absolutely be done, and for little effort on the Keysight side"

How do you keep your Fluke 732 (specced 0.18V/m max field strength) away from such a EMI-Source?
No real volt nut wants avoidable RF near his measurements.

With best regards

Andreas

[rodpp]

Some interesting box showed up today with Keysight Technologies label on it. Wonder what that could be...

Ah, right....

Hi @TiN, could you already share some data about the black 3458A, better yet, compared with the old 3458a model?

Thanks!

[Edwin G. Pettis]

Guys this is from the horse's mouth (Keysight), the specifications of the 'black' 3458A is the same as the beige 3458A, yes there were changes made.  It seems that the folks over in the EU was making a real stink over the 3458A having a handle on it so they declared that it did not meet the exemption requirements as it had previously even with the handle being on there all along.  The main reason for putting a bunch of SMT in it was because some of the old leaded components were becoming difficult to obtain as was the MC68000 CPU system chips, so where necessary they had to substitute where possible for components except where it was prudent to stick with leaded components.  The software was a real challenge to keep it backwards compatible with the original system, it basically emulates the MC68000 in software.  The LTZ sub assembly was left alone as it was a proven design.  They did nothing to disturb the original design, it works just like always.

While there were some small difference such as a little decrease in system power for instance, they said the specifications will remain the same, if they were to change them then the unit would be called the 3458B.

I agree with Dr. Frank, the speculation was nothing more than that.

[rodpp]

Yes, I believe that changes were minimal.

But nothing better than data to show that the performance remains the same.

[Dr. Frank]

Guys this is from the horse's mouth (Keysight), the specifications of the 'black' 3458A is the same as the beige 3458A, yes there were changes made.  It seems that the folks over in the EU was making a real stink over the 3458A having a handle on it so they declared that it did not meet the exemption requirements as it had previously even with the handle being on there all along.  The main reason for putting a bunch of SMT in it was because some of the old leaded components were becoming difficult to obtain as was the MC68000 CPU system chips, so where necessary they had to substitute where possible for components except where it was prudent to stick with leaded components.  The software was a real challenge to keep it backwards compatible with the original system, it basically emulates the MC68000 in software.  The LTZ sub assembly was left alone as it was a proven design.  They did nothing to disturb the original design, it works just like always.

While there were some small difference such as a little decrease in system power for instance, they said the specifications will remain the same, if they were to change them then the unit would be called the 3458B.

I agree with Dr. Frank, the speculation was nothing more than that.

Sorry Mr. Pettis,
the rumors you have replicated here, do not match the truth. At our Metrology Meeting in Stuttgart last year, we have received reliable details from KS, and that the redesign has been done here. So there's no speculation from my side, either.
The 3458A was redesigned, because many components were really obsolete, like the infamous ELANTEC comparators, and many others.
It's obvious, that the engineers then chose up-to-date RoHS compliant SMD components, wherever possible and useful. That has nothing to do with the EU, FLUKE has done the same on their DMMs also.

There were several improved parameters, although this is not reflected in the specification, due to compliance requirements.
The 68000 processor is definitely NOT  emulated in software, and the firmware is not changed at all.
Please remember, that there exist cores from several FPGA suppliers and projects for 68000, 8051, so it's quite straightforward to merge most of the  logic of the controller and the A/D board into single ICs.

Frank

[RandallMcRee]

Guys this is from the horse's mouth (Keysight), the specifications of the 'black' 3458A is the same as the beige 3458A, yes there were changes made.  It seems that the folks over in the EU was making a real stink over the 3458A having a handle on it so they declared that it did not meet the exemption requirements as it had previously even with the handle being on there all along.  The main reason for putting a bunch of SMT in it was because some of the old leaded components were becoming difficult to obtain as was the MC68000 CPU system chips, so where necessary they had to substitute where possible for components except where it was prudent to stick with leaded components.  The software was a real challenge to keep it backwards compatible with the original system, it basically emulates the MC68000 in software.  The LTZ sub assembly was left alone as it was a proven design.  They did nothing to disturb the original design, it works just like always.

While there were some small difference such as a little decrease in system power for instance, they said the specifications will remain the same, if they were to change them then the unit would be called the 3458B.

I agree with Dr. Frank, the speculation was nothing more than that.

Sorry Mr. Pettis,
the rumors you have replicated here, do not match the truth. At our Metrology Meeting in Stuttgart last year, we have received reliable details from KS, and that the redesign has been done here. So there's no speculation from my side, either.
The 3458A was redesigned, because many components were really obsolete, like the infamous ELANTEC comparators, and many others.
It's obvious, that the engineers then chose up-to-date RoHS compliant SMD components, wherever possible and useful. That has nothing to do with the EU, FLUKE has done the same on their DMMs also.

There were several improved parameters, although this is not reflected in the specification, due to compliance requirements.
The 68000 processor is definitely NOT  emulated in software, and the firmware is not changed at all.
Please remember, that there exist cores from several FPGA suppliers and projects for 68000, 8051, so it's quite straightforward to merge most of the  logic of the controller and the A/D board into single ICs.

Frank

Hmmm, seems like a tempest in a teapot, Dr. Frank. Actually, everything I read from Mr. Pettis, above, seems to have a basis in everything you say, too!  . Not sure why there is the beginning vehement statement from you about unsubstantiated rumors? Could not the opinions of a Keysight engineer in the US, for example, differ from one in Germany? Lets not go to personal attacks.

[Dr. Frank]

Hmmm, seems like a tempest in a teapot, Dr. Frank. Actually, everything I read from Mr. Pettis, above, seems to have a basis in everything you say, too!  . Not sure why there is the beginning vehement statement from you about unsubstantiated rumors? Could not the opinions of a Keysight engineer in the US, for example, differ from one in Germany? Lets not go to personal attacks.

Definitely not.
I just don't like to be cited as to spread speculations.
I either publish my  personal opinions or observations I've made, or I publish facts-
If you or others read carefully, both cases can be distinguished easily.
Frank

[Edwin G. Pettis]

Well Frankie, take a pill, calm down and let the adults speak, according to Keysight USA, you know the home office, the 3458A has been and always will be 100% designed, developed, and the redesign was 100% done in the the USA.  The main manufacturing facilities are in Malaysia with a special facility on occasion in Singapore, they will also manufacture the 3458A wholly in the USA if the customer requires it (such as the military).  While thee are R&D facilities in Belgium and Ireland, there are 5 facilities in the US in including one here in Colorado where I reside.  Outside of the R&D facilities, the other Keysight facilities in Europe only deal with local issues in applications and software, no design work.

They said the redesign of the 3458A was completely done here in the USA as was any other changes, the firmware is 100% compatible with previous units.

As part of the agreement when Keysight was split off from Agilent, Keysight agreed to change the color of all of their instruments to this black (marketing picked it) you are seeing, all of the newer designs are already in black and most of the legacy instruments have been changed over, the 3458A being one of the last ones to change.....that is all the color change is, nothing more.

Perhaps there was a misunderstanding in Stuttgart but this is what the home office is saying and I have no reason to believe what they are saying is false.  You are welcome to call Keysight here yourself.

[MiDi]

Independent from Stuttgart meeting I got the information, that at least A3 board was developed in Germany.
Maybe @Henrik_V wants to contribute.

[Dr. Frank]

Well Frankie, take a pill, calm down and let the adults speak, according to Keysight USA, you know the home office, the 3458A has been and always will be 100% designed, developed, and the redesign was 100% done in the the USA.  The main manufacturing facilities are in Malaysia with a special facility on occasion in Singapore, they will also manufacture the 3458A wholly in the USA if the customer requires it (such as the military).  While thee are R&D facilities in Belgium and Ireland, there are 5 facilities in the US in including one here in Colorado where I reside.  Outside of the R&D facilities, the other Keysight facilities in Europe only deal with local issues in applications and software, no design work.

They said the redesign of the 3458A was completely done here in the USA as was any other changes, the firmware is 100% compatible with previous units.

As part of the agreement when Keysight was split off from Agilent, Keysight agreed to change the color of all of their instruments to this black (marketing picked it) you are seeing, all of the newer designs are already in black and most of the legacy instruments have been changed over, the 3458A being one of the last ones to change.....that is all the color change is, nothing more.

Perhaps there was a misunderstanding in Stuttgart but this is what the home office is saying and I have no reason to believe what they are saying is false.  You are welcome to call Keysight here yourself.

Hello Edwin,

that history of hp and agilent and Keysight, I know all that, also that the 3458A was initially designed and tested in Loveland, before 1989.
The 344xxA DMM platform was the last redesign project by several engineers of the old staff from that era, but obviously they all are gone, or being retired.
These instruments are all hp-white, so what's that fuss over the black color?

Sorry, there is absolutely no misunderstanding (and no speculation), all the new boards have been redesigned outside the KS organisation, by an Engineering agency, located here in Germany.

I don't have to call Keysight, because this information came first-hand from Keysight Germany.

Frank 

[Zucca]

Like in aviation: If only we could see what's inside the black box...

[branadic]

I've visited the german company that redesigned the A3 board myself last year and saw some different design stages of the A3 board as well as all the parts they needed to build it, including multiple U180 and entire 3458A in the different design stages, even one that was just painted black to match the new "design", the lab, some test environments and the like.
It's no disgrace to except that Keysight needed help from outside their country to get that challenge on 3458A done. But if it is easier for you to live with, believe whatever you want to believe.

-branadic-

[Wolfgang]

... whoever made this remarkable achievement to put a 30 year old 68000 processor into an FPGA and integrate some glue logic,
still nothing to be very proud of. I think that at least a USB interface plus a reasonable display could have been integrated as well,
with no effect on backward compatibility. The major innovation here is black paint.

[Zucca]

The major innovation here is black paint.

KS did not want innovation in this project.
KS had two pure main marketing (not innovation) goals in this project:

- RoHS
- absolutely backwards compatible to the white 3458A

They not even change the name for a reason. It is not a 3458B.
It this way KS can continue to sell this first place golden standard DMM to everybody again as before, no requalification needed.
Best bang for the bucks, well done KS!

[notfaded1]

I'm just glad I finally got one and I don't even care that it's white!!!  I see no reason to argue too much about US vs German engineering.  Most companies spread engineering out across all their offices these days.  With teleworking we have engineers in multiple states and multiple countries all working on the same piece of a project often.  I'd be willing to bet that was case here as well.  We aren't limited to doing one project in one place anymore.

Bill

[iMo]

The origin of the re-engineering effort depends on the effort's result, afaik from my previous work with large multinational corporations:
Success - HQ
Failure - Engineering Company off-shore
 

[Edwin G. Pettis]

In the long run it really doesn't matter where what was done or by whom, only that the 3458A is still the same machine to the user, it is identical henceforth there was no model change, it is still the reliable 3458A.

There is really no need to get irate about the details.

To some degree, yes they could have added more bells and whistles but it is apparent that the customer base is not interested and adding those bells and whistles would contribute to more EMI inside the unit and also require re-certification and I don't think that was a viable option.

[ramon]

... yet nobody has shown any proof ...  I am with rodpp   !

[ScoobyDoo]

Gentlemen - as long as my direct Keysight info is correct - Keysight Böblingen (Germany) commissioned the facelift of the A3 (ADC) board of the 3458A to a small engineering bureau near Munich in collaboration with PTB - PTB assisted with the verification (testing) of the INL of the A/D converter board.
The early 3458 B.E. units had a  dark blue-ish colour as one of the engineers repainted the panels in some sort of local garage (paintshop) and some photos - shot during a PTB meeting - were spread "by accident".
As we had seen those PTB pictures and discussed it among engineers and Metrology Voltnuts we asked Keysight (USA & Germany) to send a knowledgeable speaker to divulge more info of the 3458A B.E. during first Metrology seminar - organised late May in 2019 ( https://www.eevblog.com/forum/metrology/volt-nut-meeting-2019-in-stuttgartgermany/ ). This year the event is planned in beginning of June (info: BRANADIC - Dr. Frank - ... ) and again we invited knowledgeable speakers (from that german engineering bureau and Keysight Germany) to unveil more details of this remarkable project on next event early June 2020. I would say mark your diaries and be there ... - the organization is in hands of Hahn & Schickard (info: BRANADIC) and resides on and near the campus of the University of Stuttgart - a very nice place for Voltnuts to visit ...



Our major concern:
We have observed that many spare boards (A3 ADC, A9 03458-66529 2ppm ... ) have been withdrawn from public sale - and new 4ppm and 8ppm VREF A9 boards were introduced at much lower cost than before - I strongly "guess" (not info from controlled source)  Keysight is encountering a yield problem as we observed prices of the new A3 board sky rocketing before boards were withdrawn - I hope this is only temporarily until they control better the yield of the boards in the factory in Penang - Malaysia. A lot of the success of the old 3458A was contributed by the fact that the unit could be repaired endlessly - and I hope that Keysight continues to follow that former strategy in making available all spare parts for the 3458A owners community as they did in the past.

Herzlichen Grüßen/Meilleures salutations/Best regards
ScoobyDoo