In the simulation you can clearly see a predicted response curve.....which is partially represented in the real wold....BUT have a look at the HF energy which the simulation missed.
This problem is compounded with a dynamic load (I will simulate that on the next experiment).
I purposely use crap power supplies for these tests....to see how input phase/ripple changes the dynamic responses of the POL under test. It's great to have a collection of crap SMPS and linear PSU's. Feeding precise, filtered signal into the POL front end is a terrible way to test these devices. In the real world we have to assume that our hardware is being connected to crap wall-warts or similar. I can also run this experiment using a battery, with both high and low margining. You will see just how far off the predictions are from the real world responseIMHO that is due to the limited model you are using for the simulation. One of the things I like to simulate are worst case conditions which often reveal hidden problems. A 'crappy SMPS' is not very scientific
While not true in all cases, my experience in ME was that by the time I created a suitable model for a simulation that was to be believable - I could have built and destructive tested it a few times already which gives me the ultimate answer. Even after a good simulation has been built and analyzed, you still have to built and delicately test the circuit to confirm and look for anything unexpected. If there is any truth in that - why/when bother with simulation?
I too test with "crappy SMPS" sources to learn how my circuit deals with a variety of non-ideal situations. It seems that simulating and scientific analysis [guessing] would not be nearly as good as throwing the real world at it, right? I have not spent much time understanding my own definition of 'crappy' simply because my customers wont or cannot adhere to that - if my device says '12v IN' and their random SMPS says '12V OUT' - they will simply plug it in and expect it to work. I can setup a test with a variety of DC sources and examine the resulting performance impact on my circuit.
Is this a logical fallacy, or a reasonable way to go about it? Maybe use a simulation to set a baseline to measure from?
In the simulation you can clearly see a predicted response curve.....which is partially represented in the real wold....BUT have a look at the HF energy which the simulation missed.
This problem is compounded with a dynamic load (I will simulate that on the next experiment).
I purposely use crap power supplies for these tests....to see how input phase/ripple changes the dynamic responses of the POL under test. It's great to have a collection of crap SMPS and linear PSU's. Feeding precise, filtered signal into the POL front end is a terrible way to test these devices. In the real world we have to assume that our hardware is being connected to crap wall-warts or similar. I can also run this experiment using a battery, with both high and low margining. You will see just how far off the predictions are from the real world responseIMHO that is due to the limited model you are using for the simulation. One of the things I like to simulate are worst case conditions which often reveal hidden problems. A 'crappy SMPS' is not very scientific
Once again, you are missing the boat here.
Number one, I didn't include screen shots of the entire simulation....only it's predicted output. i would think if you had simulation experience that you would assume the simulation is complete? Do you need screen-shots of the other 5 pages of parametric simulation? Isn't the predicted response enough?
In the simulation you can clearly see a predicted response curve.....which is partially represented in the real wold....BUT have a look at the HF energy which the simulation missed.
This problem is compounded with a dynamic load (I will simulate that on the next experiment).
I purposely use crap power supplies for these tests....to see how input phase/ripple changes the dynamic responses of the POL under test. It's great to have a collection of crap SMPS and linear PSU's. Feeding precise, filtered signal into the POL front end is a terrible way to test these devices. In the real world we have to assume that our hardware is being connected to crap wall-warts or similar. I can also run this experiment using a battery, with both high and low margining. You will see just how far off the predictions are from the real world responseIMHO that is due to the limited model you are using for the simulation. One of the things I like to simulate are worst case conditions which often reveal hidden problems. A 'crappy SMPS' is not very scientific
Once again, you are missing the boat here.
Number one, I didn't include screen shots of the entire simulation....only it's predicted output. i would think if you had simulation experience that you would assume the simulation is complete? Do you need screen-shots of the other 5 pages of parametric simulation? Isn't the predicted response enough?You keep dumping loads of incomplete information without getting your point across. I'm pretty sure you know what you are doing but to me you seem like a professor who is exited about his area of expertise but is just mumbling and writing incomplete equations on the black board. You are writing about a 50mV overshoot; that can be relevant for certain loads but all the equipment in your pictures doesn't show that overshoot hence there is no evidence in your pictures to justify the test setup you are showing. Let's keep the discussion more generic instead of focussing on one niche area of circuit design.
Of course...I see the story now.
the HF spike would never be seen at 100Mhz and you would think all is well with the world. BTW, what timebase was the scope on? Is it 1us/div?
It would be such a frustrating scenario when I am experiencing issues on the load, but the scope trace looks perfect at 100Mhz. The cost difference from 100Mhz to 500Mhz is not trivial, but clearly needs some consideration. I am putting financing together for a Pick & Place system, maybe adding a no excuses scope could keep me out of trouble. With all that is going into this effort, it would be criminal to let a glitch sneak by creating some mysterious problem that may never be found.
There are Agilent MSOX3000's in 4ch 500Mhz and 1Ghz nearly 50% off list at the Keysight eBay store. At those prices, I would be foolish to not go big.
and there would of course be my consultation fee
Quoteand there would of course be my consultation feeYour've earned one with your contribution.
There will be many that find this thread and be grateful.
reading mr tunersandwich input is very eye opening. and i agree about the extra bandwdth. but for someone going beyond starter scopes, there is just so much to consider
i have been looking at ... many used scopes options. but when i fall on lecroy, i saw some that are year 2006 types ... still with 1.44 floppies ... would you (or anyone) care to give some hints about its ... usability? iirc, its also a 10G/sa. for under USD4k. i am guessing just for the bandwidth, there is no way i can get anything at over 1Ghz nearly as new or usable at under USD4k ... but on the other hand ... these are really old units
well i guess i maybe side tracking the issue of high end scopes, but it wouldnt hurt to ... kinda of compare old lecroys ... to say the current problem the rigol is having : jitter issues? ... im not familiar with anything higher than my 20Mhz analogue. but i sure is a fan of high sample rates (or BW). do high end scopes suffer any kind of bugs?
1 of the things i noticed about lecroy is, it is actually just a XP PC running very dedicated hardware to grab A to D ... so comparatively ... even if i do get something like say an agilent 2xxx series or anything up to a USD 20k unit ... a lecroy 1.5Ghz 10G/sa is still worth so much more @ USD4k right? (and it is 4 channels) ... that "juice" is in its pristine AD and DSP engine ... if its broken in its knobs or whats nots ... those are ... sort of generally cosmetic? yes?
and when you say x stream engine, that would be in their scopes models that are after 2008? (im totally unfamiliar with their tech milestones, but from what im seeing ... it seems they have left the Mhz game a long long long long time ago)
edit : see what i mean?
http://www.ebay.com/itm/Lecroy-WavePro-960-WP960-4-Channel-2-GHz-Digital-Oscilloscope-/271697521072?pt=BI_Oscilloscopes&hash=item3f426f29b0
4kUSD i think will only get a agilent 300Mhz .... *expletives* ... this 1 is 2GHZ !!!
1 of the things i noticed about lecroy is, it is actually just a XP PC running very dedicated hardware to grab A to D
... so comparatively ... even if i do get something like say an agilent 2xxx series or anything up to a USD 20k unit ... a lecroy 1.5Ghz 10G/sa is still worth so much more @ USD4k right? (and it is 4 channels) ... that "juice" is in its pristine AD and DSP engine ... if its broken in its knobs or whats nots ... those are ... sort of generally cosmetic? yes?
and when you say x stream engine, that would be in their scopes models that are after 2008? (im totally unfamiliar with their tech milestones, but from what im seeing ... it seems they have left the Mhz game a long long long long time ago)
edit : see what i mean?
http://www.ebay.com/itm/Lecroy-WavePro-960-WP960-4-Channel-2-GHz-Digital-Oscilloscope-/271697521072?pt=BI_Oscilloscopes&hash=item3f426f29b0
4kUSD i think will only get a agilent 300Mhz .... *expletives* ... this 1 is 2GHZ !!!
this seller is near me, 14 year old 584, but what i dont know is, how much functions depend on the floppy, and if that FDD could take the "chingchongchang" FDD to SD converter lol
http://www.ebay.com/itm/LECROY-LC584AL-1GHZ-4CH-DIGITAL-OSCILLOSCOPE-2-8-GS-s-/321583104404?pt=BI_Oscilloscopes&hash=item4adfd8c194
The downside is the lower waveform update rate
The downside is the lower waveform update rate
The (M)Xi Series is spec'd for a max wfm rate of 1.25M waveforms/sec, but I don't know in what configuration this is.
It gets really dodgy under heavy processing and long time bases. I think that is more an issue of the GPU/OS side of things though....not sure how that waveform rate is calculated (since it's dependent on an SoC GPU host)....
I have been running my MXi with the outer plastic case off the back.....that dropped temps by orders of magnitude....which lowered the CPU load (goddamn thermal provisioning/stepping). It still needs a better cooling solution. Quite disappointing in that regard.
On a side note I scored an un-used LeCroy motherboard/cpu/ram etc assembly (MX855LC)....for $50!!! I might just whack that into a case and do some experimenting with it.
Hoe much more specific do you need this to get? Just because you don't understand the results....doesn't mean the problem is on MY end....like I said you are throwing generalizations out, with NO data to back them up.....on a VERY specific subject....here is your overshoot
In this picture you can clearly see the circled are of the FFT (to the left) which the simulator predicted.....(have a look at the response charts again if you missed it). On the right you can clearly see HF content that was not predicted. In the waveform you can clearly see that the HF content is causing the control loop (vsense coupling) to cause a switching overshoot, that measures ABOVE the predicted range in the simulation...the ripple shown, which is more frequent (brighter part of the trace) is well within the spec of the system......the overshoot event is NOT....and if I didn't loosen up the over provisioning specs in the controller, it would have caused the entire system to turn the POL off....![]()
So then why does our scope pick-up 5.55044 V peak? Despite the POL margining the output above 9 amps?
I go with what I see and that is that your multimeters are showing 5.50V like you expected. I also see an oscilloscope image with C3 set to 20mV/div and a relatively clean trace with some spikes but they all stay within 20mVpp. The offset is adjusted to -5.29V in order to see the trace at 20mV/div.QuoteSo then why does our scope pick-up 5.55044 V peak? Despite the POL margining the output above 9 amps?There is a simple answer to that: If I look up the DC accuracy specs(*) of the 64MXi I see that it has a DC accuracy of 1.5%. At 5.5V that is already an uncertainty of +/-82mV. As a rule of thumb you can't use an oscilloscope for really accurate DC or amplitude measurements. For such measurements a DC or AC voltmeter are much better tools. IMHO Lecroy shows too many meaningless digits in the result.
BTW: IMHO the HF content could be GSM interference. It matches some GSM frequency bands perfectly. The same goes for the HF on the left. It matches the frequencies used for FM radio too perfect to be a coincidence.
* http://www.testequipmenthq.com/datasheets/LECROY-WAVERUNNER%2064MXI-Datasheet.pdf
It gets really dodgy under heavy processing and long time bases. I think that is more an issue of the GPU/OS side of things though....not sure how that waveform rate is calculated (since it's dependent on an SoC GPU host)....
I have been running my MXi with the outer plastic case off the back.....that dropped temps by orders of magnitude....which lowered the CPU load (goddamn thermal provisioning/stepping). It still needs a better cooling solution. Quite disappointing in that regard.
As said before I think there's something dodgy with your particular setup. My 64Xi with 1.8Ghz processor gets barely hand warm, and I know that the cooling is perfectly adequate for the 2GHz processor LeCroy offers as an upgrade for these scopes.QuoteOn a side note I scored an un-used LeCroy motherboard/cpu/ram etc assembly (MX855LC)....for $50!!! I might just whack that into a case and do some experimenting with it.
Nice catch!