For me, the lack of discharge modeling/programmable load functionality is pretty devastating. I believe this sort of device should be a one stop shop. That's the point of a dedicated device. I really don't get it.
I'm sure Keithley has the in-house expertise to add a prog. load considering their SMU market share. Also, not sure why the tremendous transformer is necessary when the single channel output power is 120 watts, or 20 volts DC @ 6 amps. I wonder what the VA rating is on that toroid and if the job can be done with a smaller tx to free up some room for programmable load hardware.
Nice review! I also like the way you used it to locate the short circuit. I don't know why yours is quiet but mine makes a lot of noise. It seems from the 1st day that the fan is always operating at maximum or at least very high speed. When you power it on the FAN starts in lower rpm which still makes a noticeable noise and once it boots the FAN speed gets higher rpms producing a lot more noise. I also checked that I have the latest firmware in case there is something wrong there.
stupid equipment for simulating improbable battreies ! when you know the wide range of quality of batteries...stupid equipment , the least one to have !!!
The fact that there are a wide range of battery qualities is precisely why such an instrument is useful. It allows the the user to stress test the DUT to ensure it operates correctly for a variety of battery conditions and ESR.
Nice! I got one of these for work. First thing I noticed is the unit is super heavy, that big toroidal transformer is no joke! I think it has a current sink feature, but it is not a fully programmable electronic load too. I wish Keithley would incorporate a "discharge learning" mode where you could actually discharge an attached battery at a given rate, and learn the discharge curve/esr to the instrument, so you can then simulate the DUT you just discharged.
For battery tests I like to use an actual battery with a much cheaper, faster, and easier to use Joulescope. Battery simulators, even on very high end Keithley devices, don't capture the dynamic behavior of primary cell LiFeS2 and LTC batteries. Primary cells don't model well as a simple ESR and voltage - they behaver more like a pair of RC circuits, and have strong temperature effects. So it's nice to be able to use a real battery in circuit, and still get amazing precision and 0.5 us resolution. And it streams onto your 4k monitor in real time, where you can quickly measure power, charge and timing.
I have used this instrument for a year together with the 2450 SMU. The fact that the instrument does not have touch screen is VERY annoying. Almost every time i use the 2281 I try to use the screen. The 2450 (and other) has a really well designed GUI and touch screen inteface. And when the screen on the 2281 look the same it shall work the same. Keithly: Please come with a 2282S-20-6 with a working touch screen + some other improvements...
Thanks Shahriar for this great review. Do you have any insights as to how the terminal voltage in the battery simulator is implemented, ie purely software or hardware? Is it possible for you to show on an oscilloscope the response time of the simulated terminal voltage changing due to the ESR under a dynamic load? ie in your experiment, the cell phone consumes 0.2A and then suddenly increases to 0.3A, how long does the 2281 take for the Vt to drop accordingly due to the ESR?
One downside of this unit which Shariar doesn't mention is that the remote web interface is written with Java applets, which haven't been supported in browsers for roughly 5-7 years or so. It still advertises this as a point on the datasheet though: "Emulation of front panel display and controls via your web browser simplifies automated control and monitoring". Kind of a bummer for a very expensive power supply to have an issue like this.
I haven't look into the web interface directly, but the Kickstart software compatibility is of course independent of Java. I'll look into the web interface.
Thanks for this Shahriar, I have both the 2281S and a 2450 in the personal lab. I did run head first in to the testing limitation of the 2281 having a need for making models of coin cells. I am posting the models created with the 2450. I currently have a CR2477 burning down on the SMU this weekend. Both units are fantastic to have in the lab. Thanks for the deeper dive in to the battery simulator. Have a bunch of AA, AAA and CR cells already created I can send links if there is interest.
They should have either put in a touch screen or modified the interface to be intuitive to use with arrow keys. This middle ground seems very awkward. You can even tell that you're having to pause and hunt for the right key as you navigate the menu system.
hey what a convenient short circuit detection. quick question: i do not have / cannot afford yet any precision psu like this device... but i do have a very basic riden rd tech psu, which has a less precise voltage and current limiting. and i also have a 34401a which is a good and reliable 6.5 digits multimenter. so perhaps this will work? assuming if this cheaper psu can maintain a steady state average voltage self consistency during the whole probing and fault finding session? - then surely i must try this out!! a truly awesome and useful feature. thanks so much for demonstrating this technique makes me wonder what other new possibilities may exist with my 34401a beyond that. it cannot be the only thing
When we draw current from a battery, it causes heating inside the battery. To "test" a battery, we need to be able to draw the same current as the device we are simulating. If we draw some other current, the test result will change because the heating changes. Doesn't this render the test results from a fixed sink current useless?
It is definitely a compromise. Batteries also have memory effects (partially due to thermal effects), and you cannot capture it with a linear, lookup-table model. I have suggested this to the Keithley team.
Very nice power supply, I have the 2280S. What I am missing on that supply is a way to zero the current measurement (It has a small offset). My main problem with the supply is the fan, it is way to noisy and means I only use the supply when I need its capabilities. The 2 quadrant function is sometimes very useful and at other times a pest. I am a bit surprised about your Ri on the 18650 cell, generally they are below 0.1ohm (I have measured a lot of them), but then you 4-terminal setup is not very good.
I'm guessing that on the small PCB for the large transformer, there looks to be a relay to apply mains to the big transformer. And the small transformer is for a standby supply that enables the relay for the big one, and also powers the RTC supercaps, and whatever else, when the unit is not in use.
21:31 discharge limt = 1.25A 26:00 1.11A discharge. I thought u said whether battery is capable or not it will discharge @ 1.25 A. Please explain. I also didnt understand y u would put a resistance in series as said @ 21:45 Could u also brief on regenerative load Keysight RP7900 or alike. It looks interesting but dont know drawbacks.
Thanks Shahriar. So can battery ESR be manually calculated by (Voc - (Id * Rd))/Id where Voc = V open circuit, Rd = discharge resistor, Id = discharge current?
We have this in our lab, as you said without it you can not design any device with battery . and this is the only equipment in market that do the job , not any other company make such !
20:42 - are we seeing a clipping of the current measurement or does auto range on the current really works that fast ? say it is set to auto and before the pulse is at 10 mA range ?
Others have already mentioned this, but I also want to express the fact that I;m disappointed how Keithley fixed that discharge limit. I think that's either lazy engineering or someone being verry sloppy with the design. What's the purpose of asking and also paying so much money for an instrument with these capabilities then ? And yes, the lack of a touchscreen like on the other models is annoying and with today;s technology, it's a real misstep, just to save a few dollars.
I don't get it. I did while you were closing in on the short. But you lost me when you measured 0.86 ma at the short. How can that be when at the beginning you showed the current with the two probes touching to be 0.77 ma ? The max possible was .77, then you found .86.
The contract resistance of pressing the probes together is probably higher than the contact resistance between the sharp probe tips and the solder points.
feel like you can just do that with a power supply, a resistor and a oscilloscope which you probably have. But if the company pays for instrument, why not
A simple resistor would not be sufficient because a battery has a range of ESR across its discharge cycle. Not to mention that you cannot create a battery model with that anyway.
I have two of the 2280S models on my workbench. I bought the first one when the model initially came out and it came with a carry handle. The same model nowadays doesn't come with a carry model anymore. I wonder why they got rid of it. One other thing I noticed was that the default idle fan speed/noise seemed to increase in later firmware versions. The initial release firmware resulted in a quiet unit but the most recent version is pretty loud in contrast. I didn't see anything mentioned in the change log so I am unsure why they increased the fan speed (at least what I perceived) without mentioning why. All in all. They are pretty nice units to have given their prevision, graphing capability, and triggering functions. These units fit into the tier between general power supplies and SMUs in terms of cost and given features.
The Keysight Labs said you needed to be 100k subs. So there.. I fixed it. I'm your 100k subscriber.
For me, the lack of discharge modeling/programmable load functionality is pretty devastating. I believe this sort of device should be a one stop shop. That's the point of a dedicated device.
I really don't get it.
Much agreed, and the discontinuity/missing touchscreen from the other instruments breaks workflow, further disappointment.
In their defense, it's a power supply, not a programmable load.
I'm sure Keithley has the in-house expertise to add a prog. load considering their SMU market share. Also, not sure why the tremendous transformer is necessary when the single channel output power is 120 watts, or 20 volts DC @ 6 amps. I wonder what the VA rating is on that toroid and if the job can be done with a smaller tx to free up some room for programmable load hardware.
@@kevina.4036 I'm pretty sure the overbuilt transformer is for stabilty.
Nice review! I also like the way you used it to locate the short circuit. I don't know why yours is quiet but mine makes a lot of noise. It seems from the 1st day that the fan is always operating at maximum or at least very high speed. When you power it on the FAN starts in lower rpm which still makes a noticeable noise and once it boots the FAN speed gets higher rpms producing a lot more noise. I also checked that I have the latest firmware in case there is something wrong there.
Thank you Shahriar for this video. Could you please share the simulated battery models with me?
👍🙏
stupid equipment for simulating improbable battreies ! when you know the wide range of quality of batteries...stupid equipment , the least one to have !!!
The fact that there are a wide range of battery qualities is precisely why such an instrument is useful. It allows the the user to stress test the DUT to ensure it operates correctly for a variety of battery conditions and ESR.
Nice! I got one of these for work. First thing I noticed is the unit is super heavy, that big toroidal transformer is no joke!
I think it has a current sink feature, but it is not a fully programmable electronic load too. I wish Keithley would incorporate a "discharge learning" mode where you could actually discharge an attached battery at a given rate, and learn the discharge curve/esr to the instrument, so you can then simulate the DUT you just discharged.
Some heavy gym apparatus inside,....
Great video as always. Love the battery stuff.
For battery tests I like to use an actual battery with a much cheaper, faster, and easier to use Joulescope. Battery simulators, even on very high end Keithley devices, don't capture the dynamic behavior of primary cell LiFeS2 and LTC batteries. Primary cells don't model well as a simple ESR and voltage - they behaver more like a pair of RC circuits, and have strong temperature effects. So it's nice to be able to use a real battery in circuit, and still get amazing precision and 0.5 us resolution. And it streams onto your 4k monitor in real time, where you can quickly measure power, charge and timing.
24:08 Each of them would be 10F, you probably meant to say?
Yes.
I have used this instrument for a year together with the 2450 SMU. The fact that the instrument does not have touch screen is VERY annoying. Almost every time i use the 2281 I try to use the screen. The 2450 (and other) has a really well designed GUI and touch screen inteface. And when the screen on the 2281 look the same it shall work the same.
Keithly: Please come with a 2282S-20-6 with a working touch screen + some other improvements...
For $4,000 its odd they would have left out the touchscreen. There must have been some other reason than money.
Thanks Shahriar for this great review. Do you have any insights as to how the terminal voltage in the battery simulator is implemented, ie purely software or hardware? Is it possible for you to show on an oscilloscope the response time of the simulated terminal voltage changing due to the ESR under a dynamic load? ie in your experiment, the cell phone consumes 0.2A and then suddenly increases to 0.3A, how long does the 2281 take for the Vt to drop accordingly due to the ESR?
One downside of this unit which Shariar doesn't mention is that the remote web interface is written with Java applets, which haven't been supported in browsers for roughly 5-7 years or so. It still advertises this as a point on the datasheet though: "Emulation of front panel display and controls via your web browser simplifies automated control and monitoring". Kind of a bummer for a very expensive power supply to have an issue like this.
I haven't look into the web interface directly, but the Kickstart software compatibility is of course independent of Java. I'll look into the web interface.
This is quite annoying although the web interface does still work on internet explorer.
Thanks for this Shahriar, I have both the 2281S and a 2450 in the personal lab. I did run head first in to the testing limitation of the 2281 having a need for making models of coin cells. I am posting the models created with the 2450. I currently have a CR2477 burning down on the SMU this weekend. Both units are fantastic to have in the lab. Thanks for the deeper dive in to the battery simulator. Have a bunch of AA, AAA and CR cells already created I can send links if there is interest.
I would be interested in any models you have created for AA batteries
Really good video Shahriar! :)
Hmm nice unit , but I will wait for newer version :D with discharge current settings and touch screen.
They should have either put in a touch screen or modified the interface to be intuitive to use with arrow keys. This middle ground seems very awkward. You can even tell that you're having to pause and hunt for the right key as you navigate the menu system.
A great video.
Nice vide! But how do I earn so much money to buy as much expensive hardware as you have?
I have older generation Keithley battery sim borked, still waiting for LT1363 opamp :(
Let us know when it arrives!
(will check back in 2090)
Is the battery model CSV file format documented such that you could manually enter/edit your own, theoretical (eg discharge) models?
Yes, it is easy to edit.
hey what a convenient short circuit detection. quick question: i do not have / cannot afford yet any precision psu like this device... but i do have a very basic riden rd tech psu, which has a less precise voltage and current limiting. and i also have a 34401a which is a good and reliable 6.5 digits multimenter. so perhaps this will work? assuming if this cheaper psu can maintain a steady state average voltage self consistency during the whole probing and fault finding session? - then surely i must try this out!! a truly awesome and useful feature. thanks so much for demonstrating this technique
makes me wonder what other new possibilities may exist with my 34401a beyond that. it cannot be the only thing
Yes, works fine.
When we draw current from a battery, it causes heating inside the battery.
To "test" a battery, we need to be able to draw the same current as the device we are simulating. If we draw some other current, the test result will change because the heating changes.
Doesn't this render the test results from a fixed sink current useless?
It is definitely a compromise. Batteries also have memory effects (partially due to thermal effects), and you cannot capture it with a linear, lookup-table model. I have suggested this to the Keithley team.
Almost 100k subscribers! 😀
Very nice power supply, I have the 2280S. What I am missing on that supply is a way to zero the current measurement (It has a small offset). My main problem with the supply is the fan, it is way to noisy and means I only use the supply when I need its capabilities.
The 2 quadrant function is sometimes very useful and at other times a pest.
I am a bit surprised about your Ri on the 18650 cell, generally they are below 0.1ohm (I have measured a lot of them), but then you 4-terminal setup is not very good.
The 18650 cell you saw is very old and near its end of life. That is why its ESR is so poor.
I'm guessing that on the small PCB for the large transformer, there looks to be a relay to apply mains to the big transformer. And the small transformer is for a standby supply that enables the relay for the big one, and also powers the RTC supercaps, and whatever else, when the unit is not in use.
21:31 discharge limt = 1.25A
26:00 1.11A discharge. I thought u said whether battery is capable or not it will discharge @ 1.25 A. Please explain.
I also didnt understand y u would put a resistance in series as said @ 21:45
Could u also brief on regenerative load Keysight RP7900 or alike. It looks interesting but dont know drawbacks.
Thanks Shahriar. So can battery ESR be manually calculated by (Voc - (Id * Rd))/Id where Voc = V open circuit, Rd = discharge resistor, Id = discharge current?
thank you
We have this in our lab, as you said without it you can not design any device with battery . and this is the only equipment in market that do the job , not any other company make such !
And the wall of keithley instruments grows further 😁
Thank you for sharing.
I would bet they have a hoist on the assembly line for that transformer :D
Congratulations on your 100K subs!
Do you have any luck to find correct component for Anritsu MG3700A you were fixing? We are waiting Part 2
No, sadly still no repair on that unit.
👍👍
For those without an SMU, and the ability to create battery models, is there an online resource for battery models that the 2281S-20-6 can use?
It is a very basic CSV file for which you can create in Excel.
14:38 the short current at its highest was 0.87A. How was it possible if shorting 2 probes draws only 0.77A @ 12:57
The plate I was using to short the probes was a little too resistive.
Cheers !
20:42 - are we seeing a clipping of the current measurement or does auto range on the current really works that fast ? say it is set to auto and before the pulse is at 10 mA range ?
Yes, it works very fast!
Others have already mentioned this, but I also want to express the fact that I;m disappointed how Keithley fixed that discharge limit. I think that's either lazy engineering or someone being verry sloppy with the design. What's the purpose of asking and also paying so much money for an instrument with these capabilities then ? And yes, the lack of a touchscreen like on the other models is annoying and with today;s technology, it's a real misstep, just to save a few dollars.
I don't get it. I did while you were closing in on the short. But you lost me when you measured 0.86 ma at the short. How can that be when at the beginning you showed the current with the two probes touching to be 0.77 ma ?
The max possible was .77, then you found .86.
The contract resistance of pressing the probes together is probably higher than the contact resistance between the sharp probe tips and the solder points.
Yes, this is what happened. I think the Al plate I tested the leads on was fairly resistive.
feel like you can just do that with a power supply, a resistor and a oscilloscope which you probably have. But if the company pays for instrument, why not
A simple resistor would not be sufficient because a battery has a range of ESR across its discharge cycle. Not to mention that you cannot create a battery model with that anyway.
@@Thesignalpath thank you free education
Is this device a subset of a more expensive source meter? Or does it do things the sourcemeters don't?
I mean just its power supply features, not the battery simulation/testing.
It is a subset of an SMU, yes.
I have two of the 2280S models on my workbench. I bought the first one when the model initially came out and it came with a carry handle. The same model nowadays doesn't come with a carry model anymore. I wonder why they got rid of it.
One other thing I noticed was that the default idle fan speed/noise seemed to increase in later firmware versions. The initial release firmware resulted in a quiet unit but the most recent version is pretty loud in contrast. I didn't see anything mentioned in the change log so I am unsure why they increased the fan speed (at least what I perceived) without mentioning why.
All in all. They are pretty nice units to have given their prevision, graphing capability, and triggering functions. These units fit into the tier between general power supplies and SMUs in terms of cost and given features.
Green shrink tube seems to cover screw hole ?
That hole is not used in this chassis.
At 9:55, Shahriar said that he put that there as a bit of extra protection for the flat flex cable. That screw hole is not used in this instrument.
Unused holes ? That explains the price 😀
$4,000 ...are they INSANE!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
4:30 well played Keithly well played... who ever disassembles this is greeted by a d!ck x'D (at least from this camera angle)