EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: niner_007 on June 06, 2021, 02:33:37 am
-
I’m really disappointed on the recent power products coming out of Keysight.
Their new DC electronic load in this case, it’s pretty much unusable, and at a nearly $4K price tag for the 600W dual input model.
If you send data to the instrument too fast (every 800ms is what I tried) it locks the instrument’s communication interface and remains locked, but the instrument still runs and is responsive via manual controls; that’s a really dangerous condition for a power product, and alone makes the instrument unusable.
The instrument doesn’t have a very important feature, under voltage lock out, so cannot use it for battery testing, except through external scripting and communication. So bye bye battery applications.
Their under voltage inhibit (this is not UVLO) is also broken with some weird hysteris where under voltage conditions does not always stop the load; even without the fault, I don’t understand where such a feature would be useful, as it reconnects the load when the voltage goes back up, UVLO is a more critical feature to be built in, inhibit can be programmed via the interface.
No automatic ranging, you’re either on the higher range losing resolution, but if on the lower range you cannot dial outside the range, with errors thrown at you. The auto ranging could been done in software.
The rear connectors are really nice, but the front connectors are custom, they should have selected lower current banana posts instead for the front, and limit the current through the front, this is a nearly $4K instrument after all and should have a feature like that.
No energy accumulation monitoring. So no Wh.
This is a broken instrument, they focused so much on the scope view and pulse features of the load (I don’t need these) that they got the basics completely wrong. I don’t know how good the hardware underneath is, but I think their software might require a lot if effort to fix, it’s hard to convey in words until you use it.
Sadly the new Keysight power supplies also have similar issues with the software, that’s why I will continue to use my trusted 3631A and 363xA series supplies, or go to R&S for power needs, and to BK Precision/Kikusue for DC loads, and avoid the new Keysight power products altogether. I don’t want my supply to stop responding to commands and blow the DUT up or start a fire in the lab.
I was in the market for a new DC load, and I got this Keysight DC load hopping it would do. I returned it for a refund.
-
My Korad KA102 got most of that features (altho bit of PITA to setup), that is just embarrassing
-
I use a BK 8601, but recently it got broken. I used the opportunity to look at Keysight and was very excited with their new loads, the 600W one was particularly interesting because it’s dual channel, the channels can be combined together, and high resolution. I really wish Keysight would’ve worked.
I repaired my BK, BK did an excellent job on it, and if anything, I will buy another 8601 :) Kikusui has a very nice family of DC loads, I have played with one for few days, it’s got a very nice display, and energy accumulation. Their interface is a big complex to configure, but overall for my needs, BK Precision is still the best performer. Just a rock solid instrument.
-
What did you Key-sight supplier say to your concerns regarding the instrument itself?
Would pretty surprised if they just dismissed the issue you have found :o.
The Danny Bogdanoff show is pretty active on this forum, as the brand ambassador for Keysight I am sure he could set the wheels in motion as to an investigation of your reported issue.
Just a thought?
I use Kikisui loads having tried a few others, however I have noticed on virtually all of the brands they do produce quite a bit of residue noise which is very evident especially when testing for EMC and pre compliance with spectral analysis
For general work they are great also running components as well, for more precision I built a set of dedicated specific current passive loads, the difference in the noise floor on the 9Khz to 30Mhz test suits is quiet something.
-
I did comment on Keysight youtube videos for the load, I didn’t get any replies on youtube. I needed a solution immediately, I cannot afford few months turnaround, broken I/O is a pretty big deal breaker to me. I didn't bother with support, it was just not worth it given the many issues found.
-
I don’t understand why you would think leaving YouTube comments would end up being more helpful than contacting Keysight directly.
-
I don’t understand why you would think leaving YouTube comments would end up being more helpful than contacting Keysight directly.
Agreed. You might even get a pre-release firmware version that addresses the issues (some or all). Not contacting support about the problem you have leaves the manufacturer in the dark and your issues won't be fixed for sure.
-
I checked the information of their new electronic load on the Keysight website and noticed that the values in their "Number of inputs" column are decimal numbers. Since I am not a native English speaker, I am curious that this usage is just a typo or a more precise description. :-//
-
@gamalot
I'm quite sure this is just a typesetting thing. Probably the field on the table was some decimal number by default or something like that.
@niner_007
How good/bad is the load in your opinion when operated manually via frontpanel? Is it intuitive and responsive to use?
I agree that a locking IO is pretty bad, but as far as I know the units are pretty new on the market, so some feedback from customers will probably result in a few future bug fixes ;)
-
This is a $4k instrument. When I buy it, I expect it to work as advertised, it’s not on me to report issues to Keysight for things that are fundamental, and then wait until they are addressed, that’s just not acceptable, period. This is not a DIY amateur product, or prototype, so your logic doesn’t make sense there. If it were only for the locked IO, I would have contacted support, but there were many issues. When a part of the bread is rotten, you can eat around, save what you can, but when the entire bread is rotten, there’s no eating it.
jusaca, yes for the from panel, but lack of UVLO is a non starter, specially as there is no alternative, as the IO is broken.
-
I agree, it should not happen and is annoying as hell. Unfortunatly we can see this kind of stuff happening with basicially all big players when it comes to the modern, digital driven testgear. Tens of thousends of dollars for an MSO Series 4, RTM3004 or whatever - the first firmware releases were spicked with bugs. Pressure to release new products seems to be to bug to hold back until this kind of stuff is fixed... ;/
-
I agree, it should not happen and is annoying as hell. Unfortunatly we can see this kind of stuff happening with basicially all big players when it comes to the modern, digital driven testgear. Tens of thousends of dollars for an MSO Series 4, RTM3004 or whatever - the first firmware releases were spicked with bugs. Pressure to release new products seems to be to bug to hold back until this kind of stuff is fixed... ;/
But in the end rushing out new equipment is doing nobody any good because people will be wary of new equipment and postpone purchases.
-
I agree, it should not happen and is annoying as hell. Unfortunatly we can see this kind of stuff happening with basicially all big players when it comes to the modern, digital driven testgear. Tens of thousends of dollars for an MSO Series 4, RTM3004 or whatever - the first firmware releases were spicked with bugs. Pressure to release new products seems to be to bug to hold back until this kind of stuff is fixed... ;/
But in the end rushing out new equipment is doing nobody any good because people will be wary of new equipment and postpone purchases.
New anything is getting scary. It is interesting how many bugs are found so quickly by folks outside the company. External testing could alleviate a good portion of buggy new stuff. Give Dave Jones one and he will find a lot of the bugs :-DD
-
I agree, it should not happen and is annoying as hell. Unfortunatly we can see this kind of stuff happening with basicially all big players when it comes to the modern, digital driven testgear. Tens of thousends of dollars for an MSO Series 4, RTM3004 or whatever - the first firmware releases were spicked with bugs. Pressure to release new products seems to be to bug to hold back until this kind of stuff is fixed... ;/
But in the end rushing out new equipment is doing nobody any good because people will be wary of new equipment and postpone purchases.
New anything is getting scary. It is interesting how many bugs are found so quickly by folks outside the company. External testing could alleviate a good portion of buggy new stuff. Give Dave Jones one and he will find a lot of the bugs :-DD
It is really surprising, I don’t know how they missed in testing. Could guess they released an unfinished product and knew about it. On one hand you hear marketers, and even Danny from Keysight advertise and demo the instrument as amazing. On the other hand, there is a different reality, and it’s like false advertising. Very not Keysight like. It seems they spent so much time on the Pulse and Scope View software features of the instrument software, that the rest they could not tackle, they didn’t have capacity and resources to.
-
I agree, it should not happen and is annoying as hell. Unfortunatly we can see this kind of stuff happening with basicially all big players when it comes to the modern, digital driven testgear. Tens of thousends of dollars for an MSO Series 4, RTM3004 or whatever - the first firmware releases were spicked with bugs. Pressure to release new products seems to be to bug to hold back until this kind of stuff is fixed... ;/
But in the end rushing out new equipment is doing nobody any good because people will be wary of new equipment and postpone purchases.
You previously claimed that you get something that Just Works and meets specs when you buy A brands. That they do proper QA and fundamental functional problems thus get caught before release, and that this justified their higher price. Do you still stand by that? Because it looks to me like that's becoming less true as time goes on, if this thread is anything to go by.
Maybe a better question is: at this point, what would you say are the advantages of buying an A brand over a B brand?
-
I agree, it should not happen and is annoying as hell. Unfortunatly we can see this kind of stuff happening with basicially all big players when it comes to the modern, digital driven testgear. Tens of thousends of dollars for an MSO Series 4, RTM3004 or whatever - the first firmware releases were spicked with bugs. Pressure to release new products seems to be to bug to hold back until this kind of stuff is fixed... ;/
But in the end rushing out new equipment is doing nobody any good because people will be wary of new equipment and postpone purchases.
You previously claimed that you get something that Just Works and meets specs when you buy A brands. That they do proper QA and fundamental functional problems thus get caught before release, and that this justified their higher price. Do you still stand by that? Because it looks to me like that's becoming less true as time goes on, if this thread is anything to go by.
Well, one of the lines I wanted to add to my previous posting is 'If A-brands are going to be consistent in this behaviour then people might as well buy gear from Rigol or Siglent'. It is already a fact that A-brands push out equipment ahead of schedule and need about a year to fix firmware bugs. IMHO it is still worth buying an A-brand due to pedigree; the core of the instrument usually is based on old & well proven designs / methods so it will also work in the corner cases where gear from Rigol & Siglent typically fails. Due to the latter I decided to leave the line out.
-
Have Keysight responded this at all?
Nctnico made some valid points, however the MXR last year was another case in point of being released far to early at that point in time imho
https://www.eevblog.com/forum/testgear/keysight-mxr-8-channel-scope/125/ (https://www.eevblog.com/forum/testgear/keysight-mxr-8-channel-scope/125/)
-
The Danny Bogdanoff show is pretty active on this forum, as the brand ambassador for Keysight I am sure he could set the wheels in motion as to an investigation of your reported issue.
Haha ;D
I'm passing this along to the team for feedback!
-
Well, one of the lines I wanted to add to my previous posting is 'If A-brands are going to be consistent in this behaviour then people might as well buy gear from Rigol or Siglent'. It is already a fact that A-brands push out equipment ahead of schedule and need about a year to fix firmware bugs. IMHO it is still worth buying an A-brand due to pedigree; the core of the instrument usually is based on old & well proven designs / methods so it will also work in the corner cases where gear from Rigol & Siglent typically fails. Due to the latter I decided to leave the line out.
Makes sense. Do you have some examples of where an A brand handled a corner case where the B brands failed?
I also get the sense that the A brands tend to be constructed better than the B brands, though I think the B brands have been improving their game a bit on that as well.
-
Well, one of the lines I wanted to add to my previous posting is 'If A-brands are going to be consistent in this behaviour then people might as well buy gear from Rigol or Siglent'. It is already a fact that A-brands push out equipment ahead of schedule and need about a year to fix firmware bugs. IMHO it is still worth buying an A-brand due to pedigree; the core of the instrument usually is based on old & well proven designs / methods so it will also work in the corner cases where gear from Rigol & Siglent typically fails. Due to the latter I decided to leave the line out.
Makes sense. Do you have some examples of where an A brand handled a corner case where the B brands failed?
Siglent SDG2000xyz function generator versus Tektronix AFG31000 for example. The latter stays in phase with an external reference when doing modulation. Also the Tektronix AFG31000 is a true stand-alone unit. You don't need any PC software to edit / import arbitrary waveforms.
-
The Danny Bogdanoff show is pretty active on this forum, as the brand ambassador for Keysight I am sure he could set the wheels in motion as to an investigation of your reported issue.
Haha ;D
I'm passing this along to the team for feedback!
When these are addressed, I'm willing to give EL30000 series a serious consideration.
-
Still following this terrible black design trend... ahhh ugly as hell to stare at
-
Okay most of it are features that they just didn't put in, mostly a mistake in market research there, but the remote SCPI interface locking up is a rather big deal.
I have had plenty of weird bugs and issues with remote control interfaces on test gear. Pretty much anything i had out of China has had at least some sort of quirk in the remote interface. You have to randomly stumble on them and find workarounds, usually works in the end. Yet this is Keysight we are talking about, with the reputation they are holding you would expect the remote interface to be almost fully bug free, let alone have such a serious bug as this.
-
Recent Keysight hardware is just not as great as in the old days. There are plenty of other examples discussed here:
Disappointing performance of B2962 power source (https://www.eevblog.com/forum/metrology/detailed-review-of-keysight-smupower-source-b2962a/)
Drift of 34470 DMM due to ACAL (https://www.eevblog.com/forum/metrology/keysight-34470a-calibration-and-acal-problems/)
Bugs with bursts in 33522 function generator (https://www.eevblog.com/forum/testgear/agilentkeysight-awg-fw-errors/)
-
Sorry for the delay, took a few days off and am catching up!
@niner_007 the R&D team would love to chat if you're willing, I'll send you a PM.
The SCPI lockup issue is something we haven't seen before, we've been doing nonstop / no delay commands without issues, so the team would like to take a look / chat and see if there's a bug, but we suspect it's just your load and we want to get it replaced.
Regarding battery testing, from R&D:
"
The EL30000 Series does not currently have a dedicated battery test feature, but we are considering adding it through a future firmware release. Meanwhile, you will need to use a PC to control the load and stop the discharge before the battery reaches its cutoff voltage.
"
Obviously we'll need to get the comms issue worked out for your unit.
From Bill Griffith, one of our power folks:
"- The current definition for undervoltage better is for testing power supplies. The latch mode assumes the input voltage starts low, and rises toward the undervoltage limit at which point the load turns on and stays on regardless of voltage. E.g. once live stay alive.
- I wrote a simple program to test batteries with an undervoltage cutoff: https://blogs.keysight.com/blogs/tech/bench.entry.html/2021/05/03/battery_and_celltes-HXQ1.html
- Loads potentially have two sets of ranges, one for the load and one for the measurement system. You do not want to auto-range the load as it would affect the source. It might be beneficial in some cases to auto-range the measurement system. However, it is much simpler just to tie the load and measurement system together. I personally, prefer the ease-of-use of a single setting for high, medium, and low range.
- For lower power BNC connections – I cut off a dual banana plug. You used two of them in this video: https://youtu.be/FM7dWAC6wok?t=22 I personally like the front connectors – we used them to test the 20 A output of the power supply
[attachimg=1]
"
-
Sorry for the delay, took a few days off and am catching up!
@niner_007 the R&D team would love to chat if you're willing, I'll send you a PM.
The SCPI lockup issue is something we haven't seen before, we've been doing nonstop / no delay commands without issues, so the team would like to take a look / chat and see if there's a bug, but we suspect it's just your load and we want to get it replaced.
Regarding battery testing, from R&D:
"
The EL30000 Series does not currently have a dedicated battery test feature, but we are considering adding it through a future firmware release. Meanwhile, you will need to use a PC to control the load and stop the discharge before the battery reaches its cutoff voltage.
"
Obviously we'll need to get the comms issue worked out for your unit.
From Bill Griffith, one of our power folks:
"- The current definition for undervoltage better is for testing power supplies. The latch mode assumes the input voltage starts low, and rises toward the undervoltage limit at which point the load turns on and stays on regardless of voltage. E.g. once live stay alive.
- I wrote a simple program to test batteries with an undervoltage cutoff: https://blogs.keysight.com/blogs/tech/bench.entry.html/2021/05/03/battery_and_celltes-HXQ1.html
- Loads potentially have two sets of ranges, one for the load and one for the measurement system. You do not want to auto-range the load as it would affect the source. It might be beneficial in some cases to auto-range the measurement system. However, it is much simpler just to tie the load and measurement system together. I personally, prefer the ease-of-use of a single setting for high, medium, and low range.
- For lower power BNC connections – I cut off a dual banana plug. You used two of them in this video: https://youtu.be/FM7dWAC6wok?t=22 I personally like the front connectors – we used them to test the 20 A output of the power supply
(Attachment Link)
"
Feel free to PM. I returned the unit I purchased for a refund, I don’t have access to one. If you provide a demo unit, I’d be happy to retest and report the findings here, in case it was only my unit, there might be weird interaction with network equipment, I don’t know.
Regarding battery discharge, I am not comfortable driving it via PC, too many things can go wrong in such a setup, unless you have a mode in the load where you check “keep alive” messages from the PC, and in their absence you disconnect the load.
For auto-ranging, it doesn’t need to work while the load is connected, you can restrict it only when the load is disconnected. This is a usability issue as implemented today. Can record a video and illustrate it if I had access to a load.
-
I was just bitten by this bug. Here I thought this $4,000 DC load would be better than the BK Precision and it just ignores the under voltage setting.
My understanding was this could be used to stop when discharging batteries etc.
Also no Wh reading on the display. @Daniel - I would be happy to test any updated firmware if you can provide it. My email is michael@inertialbox.com.
I purchased this through a local agent for Farnell UK as I live in Colombo, Sri Lanka so yeah I’m pretty screwed.
-
I bought E36232 and EL34143 at the same time, the experience is very bad! They both have serious anti-human logic in their basic operations. The device has knobs but noway to adjust voltage or current by bit. Only 0.001V per turn of the knob, which makes the knob useless! And even when benchvue is used, EL36232 is often disconnected, and USB has to be replugged to reconnect the instrument. Very, very bad experience. :( :( :( :( :( :(
-
To keep a running list:
Recent Keysight hardware is just not as great as in the old days. There are plenty of other examples discussed here:
Disappointing performance of B2962 power source (https://www.eevblog.com/forum/metrology/detailed-review-of-keysight-smupower-source-b2962a/)
Drift of 34470 DMM due to ACAL (https://www.eevblog.com/forum/metrology/keysight-34470a-calibration-and-acal-problems/)
Bugs with bursts in 33522 function generator (https://www.eevblog.com/forum/testgear/agilentkeysight-awg-fw-errors/)
Bug in 53230 counter implementation (https://www.eevblog.com/forum/testgear/keysight-53230-frequency-counter-review-and-bug-identification/)
-
I bought E36232 and EL34143 at the same time, the experience is very bad! They both have serious anti-human logic in their basic operations. The device has knobs but noway to adjust voltage or current by bit. Only 0.001V per turn of the knob, which makes the knob useless! And even when benchvue is used, EL36232 is often disconnected, and USB has to be replugged to reconnect the instrument. Very, very bad experience. :( :( :( :( :( :(
Yes, I ran into this same issue. The knob is practically hopeless. I've also got a R&S NGP814 - that UI is amazing in comparison.
-
I've tweeted to Keysight about this https://twitter.com/DanielBogdanoff/status/1458148062241587203?s=20 but based on their approach of sending me to blog post, and the fact that they've also seen this thread, it looks like there isn't going to be a fix for us current owners.
Sadly, I'm pulling my hair out for wasting $4,000/- and having missed this thread before shelling out |O |O |O |O
-
I'm in touch with the product teams and will give them this feedback, hoping we can work on some things here.
I put it on Twitter as well, but there is a workable option using Excel that one of our power folks put together:
https://blogs.keysight.com/blogs/tech/bench.entry.html/2021/05/03/battery_and_celltes-HXQ1.html
Personally, I'd also like to see it integrated in to the load itself.
-
Thanks Daniel. I look forward to hearing from your engineering team. Happy to assist where possible.
Cheers M
-
Hi Guys, although I have read many an article here its my first post.
I was interested in the EL34143A DC load responses as I was looking to purchase. Just arranging a demo and dont want to waste time if the units have any issues being controlled by software etc.
BK Precision being my other choice simply on accuracy spec (at the time).
Does anyone know if any of the issues were resolved by Keysight, firmware updates etc as I can see resolutions may only have started in November/December ?
Many Thanks
-
I got an email where they expect to have new firmware ready by middle of this year. That’s all I’m aware of at the moment.
Thanks M
-
Many Thanks for the quick reply. I may still get one on loan but will discuss with the sales engineer regards known issues.
I have just been looking at the BK 8601 and downloaded the software to take a peek.
Thanks
-
Any update from anyone regarding this DC load? Has the issues been fixed?
-
I was sent this firmware by Keysight a while back, and unfortunately have not had any time to test it. You can grab it here: https://drive.google.com/file/d/1IkAMjMfkPTkOfbA56MVlp3Nq419CQ1iz/view?usp=sharing (https://drive.google.com/file/d/1IkAMjMfkPTkOfbA56MVlp3Nq419CQ1iz/view?usp=sharing)
I've also verified the MD5 checksum provided by Keysight
[11:36] M1MaxMacBookPro.local:keysight_EL3000_beta_khew_13MAY2022 | md5 Jati750B_9.2.7_1.0.6_17_114_dfd9d5d1396b24671c2adaa64d85875e.dfu
MD5 (Jati750B_9.2.7_1.0.6_17_114_dfd9d5d1396b24671c2adaa64d85875e.dfu) = dfd9d5d1396b24671c2adaa64d85875e
Instructions sent from Keysight:
Please find the beta file and checksum attached.
The instruction can be found from the firmware page.
https://www.keysight.com/us/en/lib/software-detail/computer-software/el30000-series-bench-electronic-loads-firmware.html (https://www.keysight.com/us/en/lib/software-detail/computer-software/el30000-series-bench-electronic-loads-firmware.html)
Download and install the firmware utility
Connect the instrument to the PC
Load the firmware file with the utility tool and perform the firmware upgrade
Note: Please note that I am simply passing on this firmware received and anyone attempting to flash their device is doing so at their own risk.
-
As I just got one of these electronic loads myself, I'm curious what the current state of the firmware is...
My unit shipped with the following version:
1.2.5-1.0.6-17-114
The website shows this as the latest version, with a release date of 2022-01-04.
(the previous release is 1.2.3-1.0.6-17-114, listed with a release date of 2021-03-16)
The beta image file shared by the previous post shows a first number that is so far ahead of what I see officially, that I can't help but assume its special numbering for beta releases. So I'm not sure if its actually older or newer than what I have.
So in any case, I guess I'm just wondering whether a new release is around the corner, and what the current state of things is past what's been discussed so far.
-
I've just pinged the Keysight team on your behalf. Unfortunately I didn't have time to test the beta, although I've used the electronic load in a basic manner.
-
Anyone already tested the beta firmware?
I mean i would have no hesitation if there are some bugs and i can downgrade to a official firmware again.
But if the beta firmware would be broken completely, this would be kind of bad...
-
(I don't have a EL30000 - cross-shopped a couple of years ago but was turned off by the lack of Wh or Ah counting and lack of battery test features in general).
There is a new firmware on the Keysight website: 1.2.9-1.0.6-17-114
https://www.keysight.com/us/en/lib/software-detail/computer-software/el30000-series-bench-electronic-loads-firmware.pv.html (https://www.keysight.com/us/en/lib/software-detail/computer-software/el30000-series-bench-electronic-loads-firmware.pv.html)
Release notes say: "Added new battery test features."
There is no new Instruction Manual that may describe these. Anyone with an EL30000 that would know what these are?
-
Release notes say: "Added new battery test features."
I just went ahead and installed the firmware update on my EL34143A. The "Mode" menu now has a new option called "Battery Test." I've attached a screenshot of what's displayed when you select it.
-
Release notes say: "Added new battery test features."
I just went ahead and installed the firmware update on my EL34143A. The "Mode" menu now has a new option called "Battery Test." I've attached a screenshot of what's displayed when you select it.
Thank you!
Too bad it's a completely separate mode, but at least it should do the job for batteries now.
-
any ideas how that works with models that have 2 inputs? can it work in parallel mode?
-
any ideas how that works with models that have 2 inputs? can it work in parallel mode?
The battery testing function is described in the new manual, but it's not clear how or even if you can have 2 battery tests in parallel.
https://www.keysight.com/us/en/assets/9921-01404/user-manuals/EL34243-90000-User-Guide.pdf (https://www.keysight.com/us/en/assets/9921-01404/user-manuals/EL34243-90000-User-Guide.pdf)
-
Interesting that this thread has gotten some activity suddenly.
I've been looking for a good deal on the EL34243A dual channel model for the last year or so but had no luck. The added battery test feature mentioned here increased my interest recently. A few weeks back I noticed some EL34143A mid-level single channel units on the Keysight used site for right at $1k and I couldn't resist picking one up. The extra 5% off and easy purchase process sealed the deal. Shortly after I bought mine, another one showed up and I see it's now sold as well.
The battery test feature is something I'm really looking forward to, but also the additional features and accuracy compared to my Kunkin KL283 will be welcome. I had experimented with a battery capacity test setup a while back but due to my lack of time and skill never got it going.
I can see the point of Keysight wanting you to run this via their software, but I also see the point of wanting to just be able to walk up and hit a few buttons to run the battery test for example. I have a dedicated machine with the Keysight software and have not had any issues with it. Keysight is also offering PathWave BenchVue Basic for free now, which I admit I haven't tried out yet.