jonnyGURU Forums
Home Site Search Reviews Articles Contest Links PSU FAQs  


Go Back   jonnyGURU Forums > Site and Forum Topics > Testing Methodology Discussion

Testing Methodology Discussion Questions and comments regarding the testing methodologies used on jonnyguru.com

Reply
 
Thread Tools Display Modes
  #31  
Old 10-16-2016
Behemot Behemot is offline
Banned
 
Join Date: Aug 2013
Location: 504'53.169"N, 1423'30.699"E
Posts: 1,578
Thanks: 90
Thanked 249 Times in 183 Posts
Default

Quote:
Originally Posted by -The_Mask- View Post
He doesn't say that. He only stated that it's not repeatable by another person. Simple because a homemade load tester is one of a kind. Sunmoon or Chroma load testers are quite common. If you get some strange results with those load testers another person with the same load tester can verify you results. With a homemade tester that's not possible. Aris knows because his first load tester was a homemade load tester.
That's kinda difficult if it is "programmable" and, naturally, nobody gives his program settings to somebody else. Besides, this discussion was already here, according to somebody (don't remember) them ppl using Chroma way too often have wrong settings.

And I still believe there may be systematic error with the fact they would all use the same equipment. As each PC configuration is different it is also good to use different loading machines to test the PSUs in slightly different scenarios. If everybody had the same equipment we may get the same results. But same results do not mean right results and definitelly do not mean the tested unit works OK under all scenarios.

In understand the point is that others may not reproduce the same behavior on different equipment. Well, yes. They very well may not even on the same equipment. But I fail to see the reason, who wants to reproduce anything anyway? Each reviewer does the thing his way over and over no matter what the others do. Very rarely they go elsewhere to see how a unit behaved there. One of them is often me, and quite often I find a flawed unit behaving similarly bad on totaly different equipment as well. What else do you need to consider such unit Pos?
Quote:
Originally Posted by ITelektro View Post
I do not see how you will menage to get a slew rate of 1 A/s with resistor based design.
If it has to be so precise, than use precise switch. Like SSR for example.

BTW as for the panel meters, it has been suggested at EEV things should be fine as long as the measured voltage stays within the supply voltage margins. Also that the measured ground should be common for all the meters. Well kinda difficult to use ground when measuring voltage across a shunt…

Last edited by Behemot; 10-16-2016 at 03:36 PM.
Reply With Quote
  #32  
Old 10-16-2016
-The_Mask-'s Avatar
-The_Mask- -The_Mask- is online now
ssssssmokin'
 
Join Date: Jun 2010
Location: The Netherlands
Posts: 4,928
Thanks: 350
Thanked 1,427 Times in 937 Posts
Default

Quote:
Originally Posted by ITelektro View Post
Hint : 1A example.
Hint:
Quote:
Originally Posted by -The_Mask- View Post
There is a difference between 1A load with one load tester and 1A with another load tester, even if they're both exactly 1A.
Power supplies act different with different loads. One uses resistors for loading the PSU another uses mosfets, another uses something else. And the PSU is gonna act differently with different loads. Not in efficiency, that's probably gonna be the same. Difference between that is because different load patterns and used measurement equipment. The differences are gonna occur in the performance measurements. There is gonna be a difference in voltage regulation, ripple measurement and transient response. Doesn't mean the measurements are wrong, but no one can verify them because they don't have the same load tester. Just look at the difference between Aris and OklahomaWolf/Tazz in voltage regulation. Load testing with the Chroma always gives you worse voltage regulation compared with a Sunmoon. Doesn't mean that someone is wrong, could be just a difference in equipment. But that can be verified because other people have a Sunmoon or Chroma, with a custom made load tester that's not the case.

Quote:
Originally Posted by ITelektro View Post
Aris home made dc load is not working as it shoud, obviously. It is not personal attack on Aris, it is fact that Aris did not made dc load that work as it shoud.
The conclusion, therefore, that none dc load is not working as it shoud is wrong and baseless, it is logical mistake.
Your just saying that Aris is stupid but he should take it personally?

I'm quite sure his custom load tester works as it should, I probably have read all his reviews with the Faganas ATE, even the ones in Greek. All were great and I've never seen strange results.

This is the load tester by the way http://www.jonnyguru.com/forums/showthread.php?t=6661
Reply With Quote
  #33  
Old 10-16-2016
ITelektro ITelektro is offline
500W User
 
Join Date: Jul 2015
Location: Croatia
Posts: 70
Thanks: 12
Thanked 9 Times in 8 Posts
Default

Quote:
Your just saying that Aris is stupid but he should take it personally?
Quote exactly where I wrote something like that.
If we can discuss about test methodology, dc load and general technical stuff, ok, but if you're going debate down to the personal level implying these I will not participate in it.
Reply With Quote
  #34  
Old 10-16-2016
ITelektro ITelektro is offline
500W User
 
Join Date: Jul 2015
Location: Croatia
Posts: 70
Thanks: 12
Thanked 9 Times in 8 Posts
Default

Quote:
Originally Posted by Behemot View Post
If it has to be so precise, than use precise switch. Like SSR for example
Specified by Intel ..... if I'm not mistaken.
To my knowledge there are only two states, on and off.(for SSR)
There will be some kind of slew rate but you will not have the ability to control it.

Quote:
BTW as for the panel meters, it has been suggested at EEV things should be fine as long as the measured voltage stays within the supply voltage margins. Also that the measured ground should be common for all the meters. Well kinda difficult to use ground when measuring voltage across a shunt…
As I recall, there is an option to separate the power supply(intrument) of the measuring points, then contact on the shunt will be "in the air". On my a cheap was a small bridge.
Reply With Quote
  #35  
Old 10-16-2016
crmaris's Avatar
crmaris crmaris is offline
Cybenetics CTE
 
Join Date: Jan 2010
Location: Greece
Posts: 1,520
Thanks: 173
Thanked 487 Times in 232 Posts
Default

I wanted to write a heavy reply but I won't, since I respect this forum and above all myself. However I cannot pass the fact that I am sorry about where this forum is going by a bunch of guys who try to sell themselves as first-class reviewers and highly educated individuals, by insulting other members. Personally I would prefer if they actually offered something that would enlighten all of us instead of trolling.

The post in eevblog with the scope's upgrade should be an example for all of us. This is how this forum also used to be but now we had to make an internal forum in order to be able to talk without having any trolls messing with our posts.

Let me first begin why I wrote this stuff about custom made loads.

In the academia world or to be more specific in the scientific word if you make an experiment which cannot be reproduced in another lab using exactly the same equipment, the results are debatable and anyone can argue on them, so your paper which will be based on those results won't be valid. How I know this? Because I happen to have a PhD and over a dozen of published papers on high impact magazines. The cherry on top? I work for 22 years now as a telecommunications engineer, so I know a thing or two on electronics. And guess what. All of our gear has to be calibrated frequently else we aren't allowed to use it.

I have nothing with custom loaders since I have built a couple of them years ago. Those loaders tested over 200 PSUs and work just fine even today. The only problem is that they are custom made so nobody else has them, in order to reproduce the results. BTW my loaders were fully controlled by PC through a custom made program and could put to shame many Chinese made loads. But apparently some know better and can say for sure that my custom made loads weren't working properly. Well this is good to know even now, ha ha. Before I decide to buy my first loads, I was on the works to build a highly advanced loader with Darlington arrays (I have even finished the design phase) but eventually I abandoned this project because even if I had built the best loader, still my results would be debatable by the companies.

You say you apply 1 A load. How are you sure about it? You have calibrated machines and by whom?? Is there a certificate or is there a calibration report for the equipment that you have? Or for you 1 A is the same with 1.1 A or 0.9 A? In any case if you don't have calibrated equipment no manufacturer will take you seriously when you tell him that your PSU isn't 80 Plus Platinum or Gold etc. And from the moment they won't know if your equipment works properly (the calibration certifies this as well) they won't trust all of your results. How I know this? Because I have been already there. This is why I bust my ass and spend my own money to get the Chromas I use now.

FET loaders need a very fast and highly reliable feedback circuit in order not to have a significant drift in the applied load. Also in order to measure this drift and of course the applied load you have to use high precision shunts and a good quality logger, able to cooperate well with the mVs of the shunts. Needless to say that any noise in the circuit will mess with your circuit's readings so you say goodbye to accuracy. And even in the best scenario you will have noise. If you need a quality multi-channel logger, which is essential to a custom made loader, you need to spend some bucks but this is a one way road since you have to take into account multiple readings (both Volts and Amps) and use them in sync, in order to have a usable loader. But you also needs a piece of software to fully utilize the logger, even a plain one. As for the shunt/panel combo from ebay I also tried them back in the day and threw them from day one. They were highly unreliable and not accurate at all. The only thing that saved my day was a High Voltage LabJack, which I still use but for other, much simpler purposes.

There are so many things to mention in this topic, but I don't feel ok to go down to it, after all this mess around my name. Actually this is my last post in this thread which was totally derailed.


ps. I am not afraid of any competition, on the contrary I do want competition since this is what makes us pursue perfection and get even better.
Reply With Quote
The Following 5 Users Say Thank You to crmaris For This Useful Post:
Microflop (10-17-2016), Nuckles_56 (10-17-2016), Philipus II (10-16-2016), quest for silence (10-17-2016), Tator Tot (10-16-2016)
  #36  
Old 10-16-2016
ITelektro ITelektro is offline
500W User
 
Join Date: Jul 2015
Location: Croatia
Posts: 70
Thanks: 12
Thanked 9 Times in 8 Posts
Default

......
EDIT;
Aris pulled argument calibration equipment, he's right. In all the results of the measurements we must have information of accuracy of measuring instruments and shown test methodology so results could be replicate in other labs.
If someone is using DIY device or or any device that is not calibrated and if a device does not meet the specifications required by the intel in the description of the methodology should write that as a possible disadvantage that can cause inaccurate results.

Aris;
If I did anything offended accept my apology, was not the intention.
Sadly, if somebody can not make critique and discuss about problem without being imply that someone has something against a person, or even better to imply that someone is someone said that stupid then I do not feel comfortable to participate a discussion.

Last edited by ITelektro; 10-17-2016 at 03:11 AM.
Reply With Quote
  #37  
Old 10-17-2016
quest for silence's Avatar
quest for silence quest for silence is offline
Silencer & heathen relic of a human ( by OW)
 
Join Date: Sep 2013
Location: Italy
Posts: 3,353
Thanks: 1,037
Thanked 486 Times in 421 Posts
Default

Quote:
Originally Posted by ITelektro View Post
Aris;
If I did anything offended accept my apology, was not the intention.
Sadly, if somebody can not make critique and discuss about problem without being imply that someone has something against a person, or even better to imply that someone is someone said that stupid then I do not feel comfortable to participate a discussion.
Nobody put me in charge here, so pick my following rant as a very personal opinion.

You're just too much confident in yourself and in your ability to vehiculate your ideas correctly in english: in other words, to be kind you often didn't take into account language and cultural barriers (we're american, croatian, dutch, greek, aussie, polish, italian, etc.), and I won't write down the "not kind" version.

Rephrasing what our co-forumer sith'ari once wrote, we're not here to make friends.

Well, that's not exactly true: personally I like so much people like Douwe (The Mask), Tyho (McSteel), Philip (Philipus II), Charles (Tator), Aris (crmaris), and many others (to many degrees, even Pavel/behemot, despite his temper: I don't include JG, Tazz and Wolfie because they're our hosts and I don't want to be branded of flattery), so I stay here even because they're here.

However, that's just a personal note: the default scenario is that even in a small, specialized community we're not friends, and I've learned that the hard way (ending arguing with people I didn't really want to attack).

Given that, you (royal you) must always take into account the chance of being misunderstood, and I promptly warned you:

Quote:
Originally Posted by quest for silence View Post
Quote:
Originally Posted by ITelektro View Post
Okay, that tells us that he was not able to make a dc load.
That tells us now you're acting needlessly polemically (working off some private grudge, perhaps?): and eventually that may turn out as acting stupid, so be careful.
Instead you just replied:

Quote:
Originally Posted by ITelektro View Post
Aris home made dc load is not working as it shoud, obviously.
Well, mind you, that "obviously" is probably one of the most offensive adverb I read here: said differently, if you say about an engineer he's not able to do his work, you're calling him as a stupid engineer, and that was clearly evident to any of us but you.

Didn't want to? Always glad to hear that, just take this lesson always into consideration, please.

This is my last post about that here, ITelektro, so if you ever wanted to reply me, then use PMs only.
__________________
Best, Luca
Reply With Quote
  #38  
Old 10-17-2016
rafal_iB_PL rafal_iB_PL is offline
Flux Capacitor User
 
Join Date: Aug 2010
Posts: 2,688
Thanks: 113
Thanked 176 Times in 156 Posts
Default

Quote:
In the academia world or to be more specific in the scientific word if you make an experiment which cannot be reproduced in another lab using exactly the same equipment, the results are debatable and anyone can argue on them, so your paper which will be based on those results won't be valid.
Quote:
The only problem is that they are custom made so nobody else has them, in order to reproduce the results.
By that very logic, there is actually no inherent problem with custom load testers, as long as technical data needed to create copy of that custom load can be shared on request. By then any party that wants to replicate results and has enough resources (including spare time), can make said copy and see for themselves.

The only problem is that usually nobody will bother. Might have sth to do with that whole "enough resources" thing.
Reply With Quote
  #39  
Old 10-17-2016
Behemot Behemot is offline
Banned
 
Join Date: Aug 2013
Location: 504'53.169"N, 1423'30.699"E
Posts: 1,578
Thanks: 90
Thanked 249 Times in 183 Posts
Default

Quote:
Originally Posted by crmaris View Post
FET loaders need a very fast and highly reliable feedback circuit in order not to have a significant drift in the applied load. Also in order to measure this drift and of course the applied load you have to use high precision shunts and a good quality logger, able to cooperate well with the mVs of the shunts. Needless to say that any noise in the circuit will mess with your circuit's readings so you say goodbye to accuracy. And even in the best scenario you will have noise. If you need a quality multi-channel logger, which is essential to a custom made loader, you need to spend some bucks but this is a one way road since you have to take into account multiple readings (both Volts and Amps) and use them in sync, in order to have a usable loader. But you also needs a piece of software to fully utilize the logger, even a plain one. As for the shunt/panel combo from ebay I also tried them back in the day and threw them from day one. They were highly unreliable and not accurate at all. The only thing that saved my day was a High Voltage LabJack, which I still use but for other, much simpler purposes.
That's the main reason I do not want to go that road, for some time at least. IMO FETs could bring much more troubles than gains.

As for the panel meters, for the record, I do not intend to buy any of those combos. Just the meters themselves, and than separate highly precise shunts from branded companies elsewhere.

Quote:
Originally Posted by rafal_iB_PL View Post
By that very logic, there is actually no inherent problem with custom load testers, as long as technical data needed to create copy of that custom load can be shared on request. By then any party that wants to replicate results and has enough resources (including spare time), can make said copy and see for themselves.

The only problem is that usually nobody will bother. Might have sth to do with that whole "enough resources" thing.
This.

Well, if somebody asked they could most likely even have such second device made by its author. But guess what…if a manufacturer does not like the results, his intention is usually not to find the cause and if there indeed is a flaw in their design, to correct it. No. They just have a nice reason to state "your results are not reproducable so we are going to ignore it". Their choice of course, for me it's a sign that something stinks in that company.

When the results are good, their behaviour is quite the opposite, they are so happy all of sudden and OFC also use your reward for the unit for marketing purposes.

My point is, reasonable company does not play that "unreproducible" card and they actually investigate. The rest, well…

Last edited by Behemot; 10-17-2016 at 06:57 AM.
Reply With Quote
  #40  
Old 10-17-2016
Tazz's Avatar
Tazz Tazz is offline
Site Administrator
 
Join Date: Oct 2006
Location: Modders-ville
Posts: 1,281
Thanks: 47
Thanked 189 Times in 85 Posts
Default

I get that I'm a few days behind here. Had a bad weekend myself.

So, this goes out to everyone.

Keep the personal shit out of here.

We are supposed to respect one another. Unfortunately, there is no way to stop someone from mistaking something someone else says, especially when you run into the language barriers and so forth. I don't think this started as a personal attack, but I see where it went and I believe there are a few that owe crmaris an apology for their conduct.

There's a difference between discussing ones work properly and being an ass about it.
__________________
Reply With Quote
The Following User Says Thank You to Tazz For This Useful Post:
crmaris (10-20-2016)
Reply

Tags
ds2000, fix, o-scope, oscilloscope, rigol

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -4. The time now is 06:26 PM.


Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2018, vBulletin Solutions, Inc.
Copyright © JonnyGURU.com 2000 -