SevenTeam ST-1200E-AF (now the official: Kill-A-Watt means squat thread)

[jonnyGURU]

I won't have time to open it or load it up until the weekday, but I wanted to take pictures of this thing right away in case it got damaged because of the hair-brained packaging used to ship this thing from Taiwan to US via UPS. :(

[jonnyGURU]

Now that I'm passed the funky packaging... more pics....

[Chilly]

Now that I'm passed the funky packaging... more pics....

This is the same seventeam design that silverstones use in their DA1200 and OP1200 correct?

[jonnyGURU]

Yup

[Chilly]

Yup

Wicked, can't wait to see your test results :)

[Per Hansson]

Well, nice to see the PSU in that protective foam

Usually it's just laying there loose in it's cardboard box :P

I'm guessing the PSU itself was not damaged by the very rough handling by the post companies?

[jonnyGURU]

No. It looks fine. Just the big dent in the corner. :D

[jonnyGURU]

SevenTeam efficiency test - part 1

http://youtube.com/watch?v=EgB9F0rc0qE

SevenTeam efficiency test - part 2

http://youtube.com/watch?v=fou5ciblF64

SevenTeam efficiency test - part 3 (The Home Game)

http://youtube.com/watch?v=G0s7-rRUSwU

[Oklahoma Wolf]

I'm going to be up all night trying to figure out how you got those numbers out of that thing :D

[jonnyGURU]

I don't know how I got those numbers either. It's weird. But it also goes to show that you not only can't trust a Kill-A-Watt... you can't trust the expensive stuff either! The draw from the wall is actually probably in the neighborhood of 900W in those first two videos.

I'm going to use a clamp meter next and see what it tells me.

[Spectre]

Its the calculator.



F***ing CVS. Damn you!

[Bbq]

The answer is pretty obvious.

The psu really is 97.5% efficient. Duh.

[Per Hansson]

Hmm, this is quite odd

Could it be the APFC circuit fooling the meters?

[Krohling]

A PSU with cheats?

[jonnyGURU]

Using clamping ammeter yielded no different results:

http://www.youtube.com/watch?v=DlNeBnn4gWc

:(

Resorting to getting Chroma test report from factory. :(

[Spectre]

Using clamping ammeter yielded no different results:

http://www.youtube.com/watch?v=DlNeBnn4gWc

:(

That doesn't sound good, I wonder what is barfing back through the AC side that is the problem. I assume you asked SevenTeam, did they have a clue?

[Per Hansson]

Have you made sure that your incoming line power is reported correctly with something else also?

Could also make sure that the line frequency is correct

Also, do you get the same efficiency no matter how you cross load the thing?

Maybe something in the Chroma is malfunctioning and sending current to earth, put the clamp meter or better a digital MM inline with the Earth/Neutral cable

[jonnyGURU]

Have you made sure that your incoming line power is reported correctly with something else also?

Yep. Any other PSU, this does not happen.

Could also make sure that the line frequency is correct

Yep. You can actually see in some of the pictures that the LED's are reading 60Hz.

Also, do you get the same efficiency no matter how you cross load the thing?

Only when I'm loading this unit at 50% or higher do the numbers jump up like that. At 600W and lower, the efficiency is actually quite low (77%, 78%, etc.)

Maybe something in the Chroma is malfunctioning and sending current to earth, put the clamp meter or better a digital MM inline with the Earth/Neutral cable

Doubt it. Only happens with this one unit and only at high loads.

[HKPolice]

LOL PSU efficiency cheating

[Per Hansson]

Ok, then my ironic guess is that the PSU has 2 transformers

One with a normal regulated circuit, providing up the 600w

The other just a Transfomer going directly to output with no regulation :P

[Spectre]

Using clamping ammeter yielded no different results:

http://www.youtube.com/watch?v=DlNeBnn4gWc

:(

Resorting to getting Chroma test report from factory. :(

I got lucky today with one:

http://img.photobucket.com/albums/v478/S_p_e_c_t_r_e/100_2710.jpg

Input AC is 120v

[Per Hansson]

Yea, over 100% efficiency must be acceptable, wonder how the PSU generates the extra power, built in nuclear power plant?

But hadn't you already found out that the Kill A Watt was unreliable?
Get the same reading with a ampere meter?

[Spectre]

Yea, over 100% efficiency must be acceptable, wonder how the PSU generates the extra power, built in nuclear power plant?

But hadn't you already found out that the Kill A Watt was unreliable?
Get the same reading with a ampere meter?

The current transducer data is being displayed on the digital power meter next to the variac/kill-a-watt. I figured I would try and catch the Kill-A-Watt and or the transducer being fooled while the other was giving the correct data since Jon brought the subject up again. The power meter next to the Kill-A-Watt cycles so it is hard catching all three together at once since the SM-8800 also cycles. I just got really lucky to get that much data in one photo.

[Spectre]

Jonny did you get the report yet?

[jonnyGURU]

Not yet. :(

Chinese National Day Week Holiday. :(

[Spectre]

Not yet. :(

Chinese National Day Week Holiday. :(

Bummer. I got SuperFlower to get back to me though an issue pretty quick over the holiday.

Anyway, so far I have been able to correctly read all the units with the Brand maybe I can give one a go and see what happens.

[Chilly]

Captin, shes already at 97% effency, she was only designed for 85%, I don't know how much more of this she can take!

I think Scotty has snuck into your SevenTeam PSU!

[jonnyGURU]

Apparently newer PSU's are giving me this issue more than ever.

Check this one out:

http://www.jonnyguru.com/PSU/AD-MT8/efficiency_goof.jpg

357.8 DC / 359.7 AC = 99.5%!!! That's an Andyson AD-MT8 doing that. I have a Topower built HP Blackbird 1100W PSU doing the same thing. :(

[Super Nade]

Reactive loads confuse any AC power measurement that senses RMS values. An old skool analog magnetometer type device or a Hall-effect based sensor is a better bet.

I just sat down and did all the Math that shows how wrong a simple RMS sensing device can be. Another point to note is that the accuracy of the kAW is proportional to the number of harmonics in your AC line. The key is that it measures only those components where V and I have no phase lag between them. Everything else (power) is "lost" to measurement, although the current drawn is real.

I don't know how else to put this, but maybe a pic or two of my behind-the-napkin calculations will make it more clear.

[Per Hansson]

If the meters are not correcting for power factor then you will get incorrect readings

But that should be in the other direction IMO, i.e. showing too much?

[jonnyGURU]

The thing is, all three of those meters up on that shelf DO correct for power factor. I've never had this problem before, at least this frequently, with measuring AC.

Also, it only happens at certain loads. If I drop that 357.8 load down to < 200W, I get an efficiency of 82% (which is correct for that unit.)

[Super Nade]

The KAW site is an information blackhole. Do you have a circuit schematic or a rough idea on how they work. If its an A->D system, there can be some issues that are a function of the complexity of the device (read price).

[Per Hansson]

Info on PFC and why cheap meters are incorrect:

http://www.microconsultants.com/tips/pwrfact/pfarticl.htm

Edit:
It boils down to cheap meters relying on that the sine wave and frequency is perfect
A Fluke 43B seems like a good PF measuring device
Cost is of course $2000 for the basic uncalibrated version

A tad more than the $20 for the Kill A Watt no? :P

If anyone have some recommendations on a device that can measure the power factor taking all variables into it's calculation and it does not cost $1000+ I'm all ears...

[jonnyGURU]

The KAW site is an information blackhole. Do you have a circuit schematic or a rough idea on how they work. If its an A->D system, there can be some issues that are a function of the complexity of the device (read price).

Maybe this will help: http://www.jonnyguru.net/forums/showthread.php?t=2695

Info on PFC and why cheap meters are incorrect:

http://www.microconsultants.com/tips/pwrfact/pfarticl.htm

What's sad is neither the Weibo or iDRC are "cheap." :(

[Super Nade]

O.k from that thread, the kAW reads only the in-phase I/V components to get results. The sampling etc happens only when V and I are in phase. The sampling takes care of spikes that are IN PHASE. With that schematic, I don't see how phase lag/lead introduced by reactive loads is going to be measured?

The variation is not dependent on the load by by the NATURE of the load. If the load is more inductive or capacitive (leading to phase lead/lag between V and I), the meter will report inaccurate results.

A simple test would be to connect a motor or some such inductive device and see what the kAW measures, comparing it to the known power dissipation of the device with a clean sine wave. Sometimes even room lighting can have an effect on such measurements by polluting the input line with harmonics.

This is what I mean:

Dirty Line-->kAW (misses out of phase components)--> PSU+PFC (performs phase correction so that the load looks purely resistive to it and uses all the available power).

[LordEC911]

Interesting.
(Wish I could understand the last handful of posts..)

[Spectre]

O.k from that thread, the kAW reads only the in-phase I/V components to get results. The sampling etc happens only when V and I are in phase. The sampling takes care of spikes that are IN PHASE. With that schematic, I don't see how phase lag/lead introduced by reactive loads is going to be measured?

The variation is not dependent on the load by by the NATURE of the load. If the load is more inductive or capacitive (leading to phase lead/lag between V and I), the meter will report inaccurate results.

A simple test would be to connect a motor or some such inductive device and see what the kAW measures, comparing it to the known power dissipation of the device with a clean sine wave. Sometimes even room lighting can have an effect on such measurements by polluting the input line with harmonics.

This is what I mean:

Dirty Line-->kAW (misses out of phase components)--> PSU+PFC (performs phase correction so that the load looks purely resistive to it and uses all the available power).

But what if the line conditions aren't changing?

[Per Hansson]

The line conditions are changing when you do the PFC corrections

Remember that in a passive (and active) PFC design you are actually returning current onto the line also. To make the load look linear/not dynamic
I might have the wrong idea tho but that is how I understand it

[Super Nade]

But what if the line conditions aren't changing?

How do you know that? Unless you use a spectrum analyzer, it is going to be hard to determine. Even if you directly view the line voltage on a fast scope and use the scope's built-in FFT, you may miss quite a bit of info. Per Hansson, that is correct but the thing is, your line is always dirty because of other loads. I know this for sure because of indirect ultra-sensitive effects I can detect in my experiments when the damn building maintenance guys turn on their big ass pumping device. I'm not gonna talk about that here because it would mean spamming Jon's thread, but I'll let you know the why's and wherefore's by e-mail. :)

[Spectre]

How do you know that? Unless you use a spectrum analyzer, it is going to be hard to determine. Even if you directly view the line voltage on a fast scope and use the scope's built-in FFT, you may miss quite a bit of info. Per Hansson, that is correct but the thing is, your line is always dirty because of other loads.

Well obviously I won't know but what I have been able to do is set up two units at once and fast switch back and forth with the kill-a-watt reading one unit incorrectly on one unit and correctly on a second. Sure it takes me ~10 seconds or so to pop off the ATX connector and pop on the second but no matter how many tiems I go back and forth the units that read wrong still read wrong the ones that read right read right. It would seem that the line conditions minus those induced by the power supply would be fairly similar.

[Super Nade]

So you setup 2 units with 2 killAWatts simultaneously and obtained crazy readings on one but normal readings on the other? Hmm...that is puzzling.

Sorrry, I misunderstood! The only other reason I can think of is that that particular PSU+Load is showing up as a reactive load for the kAW.
The original assumption I made is still valid though with one small changei.e ignoring line condition and reasoning that PSU+Load is seen by the kAW as reactive (inductive/capacitive)

(Dirty or not) Line-->kAW (misses out-of-phase components)--> PSU+PFC (performs phase correction so that the load looks purely resistive to it and uses all the available power).

In short, this is what I think:

KAW--> PSU-->LOAD

Load looks resistive to the PSU as it conditions the phase between V and I before DC-DC conversion, so actual power usage ought to be determined on the DC side.
PSU+Load looks inductive/capacitive to the kAW causing it to miss out a lot of information.
Does all PSU/Load combo's do this? Perhaps not as you found out. Is this a PSU only issue or a PSU+Load combo issue? Hard to say.

[Spectre]

So you setup 2 units with 2 killAWatts simultaneously and obtained crazy readings on one but normal readings on the other? Hmm...that is puzzling.

Yeah I did it once when I had two working Kill-a-watts. It was a bit tight fitting them on the AVR but if line them up just right they fit. MAkes for a hard pciture though because they are then on the back of the AVR with the laod tester on teh front. Once I get settled out next week I might give it another shot on video for you.

[Super Nade]

Roger that. Meanwhile, would you like to blow up a vastly overrated HEC? :D

[Spectre]

Does all PSU/Load combo's do this? Perhaps not as you found out. Is this a PSU only issue or a PSU+Load combo issue? Hard to say.

Personally I think it is a PSU only thing as I have not seen a unit go from not be read correctly to be read correctly. Now I have seen one be read correctly at low load but then go incorrect at high loads. For me at least to say it is a combo load and PSU I would have to see it go both ways.

[Super Nade]

The thing is, there is no way to reliably isolate the load from the PSU. O.K you may say the load side does not matter because it is on the DC side, but the kAW may see the entire setup as an LC or RC lump and I'm not sure if you can completely isolate the load as the R and the PSU as the L/C. That would be my guess. Of course, you have a better handle on this stuff that I do, so you may be correct. :)

[cypherpunks]

You've measured the wall power three ways. With the clamp-on, you're measuring amps and thus VA (does its amps agree with the other equipment?), and you have Watts out / VA in ridiculously high.

The one piece of equipment you haven't corroborated is the SunMoon. what if IT is making mistakes at certain loads? It's a known good" piece of equipment with a long history, it for a long time, but how long have you been asking it to pull 800 W?

One way you could check up on it without additional equipment would be to measure the voltage drops along the various power supply leads. They should be exactly proportional to current draw. (And close to the values for the wire gauge at http://www.epanorama.net/documents/wiring/wire_resistance.html)

If you get sane figures at a 400W load, but not when all the currents are doubled, check if the wiring voltage drops also double. If they don't, call up SunMoon and complain.

You could also try sharing the load differently between the SunMoon's various inputs. If the load from the wall changes when you change the way the total +12V load is shared across various connectors, then you can suspect the SunMoon isn't measuring current properly.

[jonnyGURU]

You've measured the wall power three ways. With the clamp-on, you're measuring amps and thus VA (does its amps agree with the other equipment?), and you have Watts out / VA in ridiculously high.

The one piece of equipment you haven't corroborated is the SunMoon. what if IT is making mistakes at certain loads? It's a known good" piece of equipment with a long history, it for a long time, but how long have you been asking it to pull 800 W?

One way you could check up on it without additional equipment would be to measure the voltage drops along the various power supply leads. They should be exactly proportional to current draw. (And close to the values for the wire gauge at http://www.epanorama.net/documents/wiring/wire_resistance.html)

If you get sane figures at a 400W load, but not when all the currents are doubled, check if the wiring voltage drops also double. If they don't, call up SunMoon and complain.

You could also try sharing the load differently between the SunMoon's various inputs. If the load from the wall changes when you change the way the total +12V load is shared across various connectors, then you can suspect the SunMoon isn't measuring current properly.

I have it putting 800W loads on PSU's all of the time (besides, the load in the last photo is only 357.8.) And it doesn't have this problem w/ non-PFC PSU's. And my math shows that the total wattage number is correct for the loads I've programmed and voltages being read. It is a logical deduction that the problem is the SunMoon, but I just don't see how it could be.

That said, I do have a second SunMoon coming Tuesday. I'm going to try this again then. This will be in a different room too, so if the results vary, I'm going to swap the SunMoon locations as well.

[Super Nade]

Freaking A man..you rich b@s#$#%...2 SM's!:eek::D

[jonnyGURU]

No. Work is paying for it. I convinced them they needed it to QC the returns that come back.

About 80% of RMA's are good. Then again, BestBuy is the biggest client and we know how brilliant their customers are.

[Super Nade]

It is only going to get worse Jon. BB are now axing their Geek squad strength big time. They have this automated diagnostic booth called "Johnny Utah" (:D) which is really useless. It will be replacing the techs. Ofcourse, the sales guys will now be the "techs". We all know how brilliant those guys are...

[signmeuptoo]

And US Corporate idiocy continues...

[B3Nut]

It is only going to get worse Jon. BB are now axing their Geek squad strength big time. They have this automated diagnostic booth called "Johnny Utah" (:D) which is really useless. It will be replacing the techs. Ofcourse, the sales guys will now be the "techs". We all know how brilliant those guys are...

I would say "Unbelievable!" here, but being that it's Best Buy we're talking about I can't honestly say that. They took a rather good tech-support company, gobbled it up, mediocritized it, and now are slowly killing it off completely. Ah well...nobody could ever accuse Best Buy of having good customer service, it seems they take great pains to ensure a decent minimum level of patheticness.

The race to the bottom continues apace...

Todd in Cheesecurdistan

[Spectre]

The one piece of equipment you haven't corroborated is the SunMoon. what if IT is making mistakes at certain loads? It's a known good" piece of equipment with a long history, it for a long time, but how long have you been asking it to pull 800 W?


The SunMoon is correct.

[jonnyGURU]

Yeah.. not sure how the SunMoon could be wrong, but you can't rule anything out.

So I got an SM-268, put it out on the RMA line in the warehouse, programmed the same load, plugged in the same PSU units, and got the exact same results.

So I'm at a loss.

[Per Hansson]

Well, haven't we already semily assumed that it is the ampere meters that get fooled by the APFC circuit?

Tried borrowing some stupidly expensive fluke ampere meter or the likes?

[jonnyGURU]

Yes. Did that already. And get the same sort of results.

[Per Hansson]

What are the PSU manufacturers saying?

[HKPolice]

Still no answer to this conundrum? Hipro5 seems to be poking fun at you now :wtf:

[Super Nade]

I seriously doubt that HK. You may it see it differently because his English is a bit hard to understand, being that he is not a native speaker of the language. ;)

[jonnyGURU]

Still no answer to this conundrum? Hipro5 seems to be poking fun at you now :wtf:

Haha... He's not poking fun at me. He recognizes the humor of the conundrum. ;)

[HKPolice]

Haha... He's not poking fun at me. He recognizes the humor of the conundrum. ;)

So what's the final answer? Is it really 97% efficient? :confused:

[Oklahoma Wolf]

As someone who just took one of these apart - it's impossible for it to hit 97% efficiency. Should be around 85% at best.

The PFC circuit's a little odd on this one - two coils, two transistors, two diodes. The controller is one I've never seen before too... a Fairchild FAN4800.

[jonnyGURU]

So what's the final answer? Is it really 97% efficient? :confused:

According to the Chroma report I obtained from SevenTeam, it's barely 80%. 77 to 79% typical.

[Marios]

Well, I don't understand.
Aren't these efficiency values (http://www.jonnyguru.com/review_details.php?id=137&page_num=2) yours?

[Oklahoma Wolf]

Those efficiency values are mine ;)

It took a while to find a meter that worked, as I mentioned in the review thread.

[Marios]

Sure ... Thank you.

[jonnyGURU]

Discussed here: http://www.jonnyguru.net/forums/showthread.php?t=3387

[jonnyGURU]

FYI: I installed a Tripp-Lite line conditioner in between the mains and the load and now my efficiency numbers are exactly where they should be.

Strange that even the large transformer that is my Variac could not filter enough of whatever non-linear load was creating my wacky efficiency numbers, but the Tripp-Lite definitely seems to do the trick.

[Spectre]

FYI: I installed a Tripp-Lite line conditioner in between the mains and the load and now my efficiency numbers are exactly where they should be.

Strange that even the large transformer that is my Variac could not filter enough of whatever non-linear load was creating my wacky efficiency numbers, but the Tripp-Lite definitely seems to do the trick.

Oh yeah? I wonder.......

So it goes wall --->line conditioner--->load tester? Or wall---->line conditioner---->power supply

Oh and which line conditioner?

[jonnyGURU]

Wall ---> Line Conditioner ---> Load Tester ---> Power Supply.

I'm using the load tester's AC analysis right now (shift F5 gives you AC stats). That's not what solved the problem because I did this prior to adding the line conditioner and was still getting 99%-type efficiency figures. The only reason I wasn't doing this before is because I couldn't use the Variac after the load tester and get good efficiency results because there's a good deal of loss at the Variac. It sucks that the load tester only has one AC input and it's the AC input for the load tester itself AND the pass through for the PSU. I may have to open the unit up and re-wire that.

The line conditioner is the same one you have: A Tripp-Lite LCR-2400. Probably best value for a line conditioner out there.

[Spectre]

Wall ---> Line Conditioner ---> Load Tester ---> Power Supply.

I'm using the load tester's AC analysis right now (shift F5 gives you AC stats). That's not what solved the problem because I did this prior to adding the line conditioner and was still getting 99%-type efficiency figures. The only reason I wasn't doing this before is because I couldn't use the Variac after the load tester and get good efficiency results because there's a good deal of loss at the Variac. It sucks that the load tester only has one AC input and it's the AC input for the load tester itself AND the pass through for the PSU. I may have to open the unit up and re-wire that.

The line conditioner is the same one you have: A Tripp-Lite LCR-2400. Probably best value for a line conditioner out there.

I just tried it with one of my units on a kill-a-watt and it didn't seem to help.

I ran it:

wall---> line conditioner-->sm-8800--->psu with the PSU also plugged into the line conditioner.

Hmmm....looks like at least in my case here the kill-a-watt is still in the spare bin.

[jonnyGURU]

Do a SHIFT I5 on the SunMoon. Does the SunMoon give you different numbers than the Kill-A-Watt?

[Spectre]

Do a SHIFT I5 on the SunMoon. Does the SunMoon give you different numbers than the Kill-A-Watt?

I tried pluggin the unit into the SunMoon but it won't turn on when I do.

[jonnyGURU]

Make sure the ON ESC button is on as well as the PS ON. One activates power to the plug while the other is your GREEN to BLACK turn on switch.

[Spectre]

Make sure the ON ESC button is on as well as the PS ON. One activates power to the plug while the other is your GREEN to BLACK turn on switch.

Whoops! Lets try that.

Edit: Got nothing.

Edit 2: I threw the little red switch and poof it works.

Edit 3: Must still have connection path wrong since it is giving me bum numbers still with that set up.

[jonnyGURU]

Looks like the Kill-A-Watt swings both ways!

The unit on the top of my SunMoon is just a display extension. Essentially it gives me the AC input data I normally get on the SunMoon when I press SHIFT F5, but on a display I can view at all times. In these pictures, we can see that the Kill-A-Watt is reporting AC power consumption GREATER than the SunMoon. In other words, where the Kill-A-Watt normally gives us efficiency numbers of 99%.... 150%... ;) etcetera, we're now seeing lower number. DESPITE THIS, the numbers are still very realistic 80% to 85% numbers and have been ever since I isolated the PSU with the Tripp-Lite line conditioner.

[jonnyGURU]

Zap had an idea....

I was just putting an ES-800 on the load tester, 800W load, and the Weibo is telling me that it's pulling 530W from the wall.

EH?!?!

That's WITH the Tripp-Lite in place.

So... Zap suggested using an online UPS between the wall and the load. Has anyone tried this?

[Spectre]

Zap had an idea....

I was just putting an ES-800 on the load tester, 800W load, and the Weibo is telling me that it's pulling 530W from the wall.

EH?!?!

That's WITH the Tripp-Lite in place.

So... Zap suggested using an online UPS between the wall and the load. Has anyone tried this?

No I haven't but the larger units would overdraw my online UPS I have at home. I don't think I have used the Kill-A-Watt in months now.

[Micutzu]

I just saw this topic, thanks to a discussion on another forum. The PFC construction alters the current waveform, you do not have pure sine wave for the current, it has a larger fill factor. Your measuring tool assumes it's sine and gives you lower RMS readings, thus lower power consumption. Add a 0.1 ohm resistor in series on the AC line and put an oscilloscope on it.

[Spectre]

I just saw this topic, thanks to a discussion on another forum. The PFC construction alters the current waveform, you do not have pure sine wave for the current, it has a larger fill factor. Your measuring tool assumes it's sine and gives you lower RMS readings, thus lower power consumption. Add a 0.1 ohm resistor in series on the AC line and put an oscilloscope on it.

Well the funny thing is it is only some APFC units that fool some equipment some of the time. I solved my issue with it by changing the power meter I was using and since then have always found one of mine to be accurate.

[Oklahoma Wolf]

Yeah - I actually haven't had a unit fool my Kill-A-Watt yet. Including the Seventeam that started this whole discussion.

[Micutzu]

To be honest i don't know Kill-A-Watt's inner workings so i'm speculating, but I can bet the current waveform it's making it go haywire, even if if it's not repeatable with any KAW rises some question marks.

[OingoBoingo]

And US Corporate idiocy continues...

shhhhh Beray..