EEVblog #933 - Keysight U1272A EMC Issue

[2nOrderEDO]

¿Is Keysight replacing out-of-warranty U127x multimeters then?

I have two of them and I am planing on exchange one for the revised version and the other one for the U1282A.

[Tom45]

¿Is Keysight replacing out-of-warranty U127x multimeters then?

I have two of them and I am planing on exchange one for the revised version and the other one for the U1282A.

Yes, out of warranty is not a problem. My U1272a was an Agilent, not a Keysight. As posted recently, I got a replacement with the EMC fixes.

[hayatepilot]

Could someone post a high resolution picture of the pcb of the revised U1272A?
I'm really curious what they changed. 

Greetings

[bitseeker]

It looks like the U1252B may also have an issue. Just got an email from Keysight with a link to the following service note:

http://literature.cdn.keysight.com/litweb/pdf/U1252B-08.pdf

[richnormand]

Just checked the Keysight website in Montreal and they do have a service note about the U1252B....
Contacted them by email. Let's see what the response is.

Thanks for the heads-up bitseeker.

Wonder if it is related to the U1272A or a completely different issue.

Edit: Feb 06.
Just got a RMA from Keysight today to send the unit.
No fuss, no issues.

[bitseeker]

From the service note, it sounds similar, but at a much higher frequency (80-120 MHz) than what affected the U1270 series and potentially triggering an error message. When Dave's video, episode 933, came out, I only tested my U1252B at 10 MHz and there was no issue. I never tried it at 10x that frequency.

I talked with an engineer at Keysight, but the service note is so new, he didn't have specifics on the issue nor the fix. He said I just have to send my meter in and they'll take care of it.

[2nOrderEDO]

¿Is Keysight replacing out-of-warranty U127x multimeters then?

I have two of them and I am planing on exchange one for the revised version and the other one for the U1282A.

Yes, out of warranty is not a problem. My U1272a was an Agilent, not a Keysight. As posted recently, I got a replacement with the EMC fixes.

Just for the record:

Hi
 
Thank you for your response and for choosing the option to receive a higher performance U1280 handheld digital multi-meter and redesigned U1270 series handheld digital multimeter. We will process your request base on the information you provided as listed below.
 
Your Name:                XXXXXX
Your mailing address: XXXXXXX
Your phone number:   +XXXXXXXXX
Product Model:           U1272A x 2
Serial Number:            MYXXXXXX & MYXXXXXX
 
A Keysight representative will contact you for the delivery arrangement in two weeks. The delivery time may vary depending on the destination country. Please contact hh_promo@keysight.com if you have any questions.
 
 
Thanks and regards,
 
XXXXXXXXXX
Business Process Coordinator
Keysight Technologies

I was thinking on buying the Rohde & Schwarz HMO12xx oscilloscope, but after this excellent service from Keysight I am starting to reconsider my brand choice. Such customer care is top notch.

[Tom45]

Could someone post a high resolution picture of the pcb of the revised U1272A?
I'm really curious what they changed. 

Greetings

Here you go:

Rev 6: https://s15.postimg.cc/kp16tfnkb/220_A1205_stitch.jpg

Rev 5: https://s15.postimg.cc/6vcu4dfjf/220_A1197_stitch.jpg

Due to the size, I'm just posting links.

Note that both boards say Agilent. But the Rev 5 vs. 6 gives it away. I have higher res. versions but even PostImage wasn't happy about  the size.

Happy sleuthing!

[bernroth]

BD4 and BD5 are 0 Ohm resistors on Rev 6 whereas on Rev 5 thes look like caps or inductors - interesting.
Looks like a hand solder job.

These are located right of the 11A fuse.

[bernroth]

And yes, two 0 Ohm resistors seem to fix the issue 

BD4 and BD5 are each 1.5µH (measured @ 100kHz) 0805 inductors.
I took one of the broken units, replaced these inductors by a wire and tested them with the signal-generator.

Side-by-side comparison: the one with the wire shows no problems, the other original U1272A (rev 5) the strange values.

@Tom45: Thank you very much for your hi-res pictures!

[TheSteve]

If it is that simple of a fix I would have thought they would put out service note detailing the fix. They have done such things before with other products(such as swapping an inductor/cap in the DSOX2000/3000A series scopes).

[bernroth]

Well, it's all about security, certification etc.

If they would tell the owners "fix this and that to make it work", they could be made liable if something goes wrong (even if unrelated to the fix).

[hayatepilot]

Here you go:

Rev 6: https://postimg.org/image/4jctxkjaf/

Rev 5: https://postimg.org/image/5l8v2zban/

Due to the size, I'm just posting links.

Note that both boards say Agilent. But the Rev 5 vs. 6 gives it away. I have higher res. versions but even PostImage wasn't happy about  the size.

Happy sleuthing!

Fantastic! Thank you very much. 

And yes, two 0 Ohm resistors seem to fix the issue 

BD4 and BD5 are each 1.5µH (measured @ 100kHz) 0805 inductors.
I took one of the broken units, replaced these inductors by a wire and tested them with the signal-generator.

Side-by-side comparison: the one with the wire shows no problems, the other original U1272A (rev 5) the strange values.

Time to mod my U1272A. 

Greetings

[HighVoltage]

Wow, I had expected a much more difficult fix.
Thanks for the great pictures.
Now I can update my U1272A

[Zbig]

There's also some flux residue near C52. Adjacent resistor (R55?) seems to differ in value between two revisions. I've annotated the low-res image with approximate location. Please refer to Tom45's high-res originals for pixel-peeping.

EDIT:
Also, any chances there are some modifications on the other side of the PCB as well? As per Dave's original teardown photos: https://www.flickr.com/photos/eevblog/sets/72157626761928610/, there's some input protection stuff on the other side.

[bernroth]

@ZBig: Yes, indeed! Good eyes

The original Rev 5 has a R55 value of 3300 (330?)

The new Rev 6 has a R55 value of 2401 (2.4k?)

Maybe an improved low-pass filter?

[Zbig]

@ZBig: Yes, indeed! Good eyes

The original Rev 5 has a R55 value of 3300 (330?)

The new Rev 6 has a R55 value of 2401 (2.4k?)

Maybe an improved low-pass filter?

Revised LPF would be my guess as well.

[Tom45]

So it seems that the replacement is a modded old meter rather than a new run. Paying people to replace 3 parts in existing inventory is more cost effective than tossing the existing inventory.

Wonder how soon they will exhaust the existing modded meters and start shipping the new design. I seem to be the first here to get a replacement U1272a, and it arrived earlier than they had predicted. Maybe the new run will begin sometime later in February.

I also wonder how many people other than EEVBlog members heard about this issue and applied for replacement meters.

Tom

[Zbig]

And yes, two 0 Ohm resistors seem to fix the issue 

BD4 and BD5 are each 1.5µH (measured @ 100kHz) 0805 inductors.
I took one of the broken units, replaced these inductors by a wire and tested them with the signal-generator.

Side-by-side comparison: the one with the wire shows no problems, the other original U1272A (rev 5) the strange values.

@Tom45: Thank you very much for your hi-res pictures!

How about calibration? Doesn't your mod seem to have affected it in any way?

[bernroth]

I made a quick test, chaining some new and old fixed (0Ohm + 2k4 resistor) meters together and they all display the same current (tested from 0.5 to 4A)
At the beginning the fixed meter was 2mA off but after 1 minute it was bang on.

I will do some more testing with my DMM-Check plus soon.

[wraper]

Wonder how soon they will exhaust the existing modded meters and start shipping the new design.

Dunno if there will be any new design. Quiet likely there will be the same boards with 3 different parts assembled from beginning.

[Tom45]

Wonder how soon they will exhaust the existing modded meters and start shipping the new design.

Dunno if there will be any new design. Quiet likely there will be the same boards with 3 different parts assembled from beginning.

I considered the changed parts to be the new design. As you say, the board is the same. At some point all boards already stuffed with the old parts will have been modified. Then all new boards will be stuffed with the changed parts.

Thanks to the efforts of Bernroth and ZBig, we now know that it is just a part change rather than a new board design. So the Rev 5 vs Rev 6 of my two meters apparently has nothing to do with the EMC issue. My original U1272a was purchased before Agilent became Keysight. If my original U1272a had been a recent purchase, it quite likely would have had a Rev 6 board too.


Tom

[Dwaine]

This incident has a happy ending for everyone.   Keysight has a very simple fix; instead of having to redesign a brand new product.   We get the replacement and can fix the original meter that was broken.

Nice....  Not too often you see that.

Dwaine

[FrankBuss]

Well, it's all about security, certification etc.

If they would tell the owners "fix this and that to make it work", they could be made liable if something goes wrong (even if unrelated to the fix).

Right, but someone with a video blog could fix it and record a video how to do it

[testmode]

Got my replacement U1282A this week!   Thanks Keysight!



IMO, it is pretty big as compared to the U1272A.  Its size actually made me realize I don't have a compact DMM on my bench yet.  I like the removable holster though.