Storing Cookies (See : http://ec.europa.eu/ipg/basics/legal/cookies/index_en.htm ) help us to bring you our services at overunity.com . If you use this website and our services you declare yourself okay with using cookies .More Infos here:
https://overunity.com/5553/privacy-policy/
If you do not agree with storing cookies, please LEAVE this website now. From the 25th of May 2018, every existing user has to accept the GDPR agreement at first login. If a user is unwilling to accept the GDPR, he should email us and request to erase his account. Many thanks for your understanding

User Menu

Custom Search

Author Topic: Reactive power - Reactive Generator research from GotoLuc - discussion thread  (Read 362111 times)

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Hi poynt and all,

here is a video update with the scope set to DC coupling. I'm still questioning the channel 2 Inversion.
Can you please confirm which way it is and what is Watts used

Video: http://www.youtube.com/watch?v=gTJ8i5unIwQ&feature=youtu.be

Thanks

Luc

EXCELLENT VIDEO LUC!

- Glad to see the power is much closer when switching between inverted and non-inverted. They should be the same however, and hopefully a channel calibrate will take care of that problem (see next point).

- Please run a "cal" on the scope from the "Utility->DoSelfCal" and check your measurements again. Run scope for 20 min. then disconnect all probes before you run the cal. Check your measurement again, as any offset in the channels should now be compensated. It takes about 5 minutes to run its course.

- DC Coupling on both channels...all good.

- Tip1: For low frequency measurements such as these, it may be beneficial to reduce the scope bandwidth limit (BW) to 20MHz on both channels. This is what I usually do and have done for my measurements here. You will get a slightly cleaner trace and triggering is sometimes less jittery (when traces are spiky).

- Tip2: If you haven't already, set the trigger source to "line".

- I am a little confused about your current probe scaling. I am using a x1 probe, 0.1 Ohm CSR, and scaling is 100mV/A. You sure that making it x10 is correct? The 100mV/A scaling already accounts for the fact that the resistor is 0.1 Ohms, so no need to multiply by 10. Also, I noticed your MATH does not read in Watts anymore; I thought it did before.

- CH2 inverted is (still) the correct way. Sources compute to a NEGATIVE MEAN power. But I will be making a video to explain all this. So yeah, you should have about -20W (CH2 inverted and the current scale setting is correct) GRID power. And a negative GRID power (with CH2 inverted) does not mean power is going back to the grid. A positive power would however mean power is going back to the grid (with CH2 inverted).

- I'm not sure why your MEAN value did not change much when increasing the number of MATH samples on the display, but it does make a marked difference with my measurement. But it is quite dependent on the wave shape. Sometimes it makes a great difference, others not, but the bottom line is that for an accurate MEAN power measurement, you should always use at least 10 MATH cycles (5 voltage cycles) on the display.

- Your power reading on the DC Wattmeter fluctuates when you remove the batteries because the meter would be seeing a FWR signal rather than a relatively steady DC.

.99

barbosi

  • Sr. Member
  • ****
  • Posts: 261
I determined that using a 0.1 Ohm CSR, one can select current for probe CH2, and one can also set the scaling for 100mV/A so that CH2 will read directly in the correct mA.

Yes, the measurement is sufficiently accurate for these measurements at 60Hz. If that CSR was replaced with a good non-inductive resistor, we could do high frequency measurements as well. Actual Current probes are non-intrusive, but they produce a voltage output just as a voltage probe across a CSR does.

No need to lecture me, I know for many years how to use a scope.
Are you sales rep for Tek? ask them for a group discount because all should be buying Tek. Or you want to keep this as a 2 people thread with Luc which I doubt?

A propos: What happened with polarities and DVM usage and all that theory you fought for? It was just an insertion tool?
Man, in my life I met hundreds of sales reps, but you are... Don't bother to respond, I won't read, I'm out!

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
No need to lecture me, I know for many years how to use a scope.
Are you sales rep for Tek? ask them for a group discount because all should be buying Tek. Or you want to keep this as a 2 people thread with Luc which I doubt?
What's your problem fella? You asked a question, and I answered it appropriately. Sorry that you take offense so easily.

Quote
A propos: What happened with polarities and DVM usage and all that theory you fought for? It was just an insertion tool?
The video to explain the proper channel invert selections when using the common probe placement is coming. I first needed to determine why I was not getting the negative MEAN measurement Luc was, and now that I have resolved that, I can move on. But who are you to be making demands on my schedule and method of approach?

I'm not sure what you are inferring with "DVM usage".

Quote
Man, in my life I met hundreds of sales reps, but you are... Don't bother to respond, I won't read, I'm out!
Happy trails!  :D

gotoluc

  • elite_member
  • Hero Member
  • ******
  • Posts: 3096
EXCELLENT VIDEO LUC!

- Glad to see the power is much closer when switching between inverted and non-inverted. They should be the same however, and hopefully a channel calibrate will take care of that problem (see next point).

- Please run a "cal" on the scope from the "Utility->DoSelfCal" and check your measurements again. Run scope for 20 min. then disconnect all probes before you run the cal. Check your measurement again, as any offset in the channels should now be compensated. It takes about 5 minutes to run its course.

- DC Coupling on both channels...all good.

- Tip1: For low frequency measurements such as these, it may be beneficial to reduce the scope bandwidth limit (BW) to 20MHz on both channels. This is what I usually do and have done for my measurements here. You will get a slightly cleaner trace and triggering is sometimes less jittery (when traces are spiky).

- Tip2: If you haven't already, set the trigger source to "line".

- I am a little confused about your current probe scaling. I am using a x1 probe, 0.1 Ohm CSR, and scaling is 100mV/A. You sure that making it x10 is correct? The 100mV/A scaling already accounts for the fact that the resistor is 0.1 Ohms, so no need to multiply by 10. Also, I noticed your MATH does not read in Watts anymore; I thought it did before.

- CH2 inverted is (still) the correct way. Sources compute to a NEGATIVE MEAN power. But I will be making a video to explain all this. So yeah, you should have about -20W (CH2 inverted and the current scale setting is correct) GRID power. And a negative GRID power (with CH2 inverted) does not mean power is going back to the grid. A positive power would however mean power is going back to the grid (with CH2 inverted).

- I'm not sure why your MEAN value did not change much when increasing the number of MATH samples on the display, but it does make a marked difference with my measurement. But it is quite dependent on the wave shape. Sometimes it makes a great difference, others not, but the bottom line is that for an accurate measurement, you should always use at least 10 MATH cycles (5 voltage cycles) on the display.

- Your power reading on the DC  Wattmeter fluctuates when you remove the batteries because the meter would be seeing a FWR signal rather than mostly DC.

.99

Thanks for all the details. I will do them all.

How about in the Acquire menu... what should be selected there?

Luc

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Luc, to summarize:

Channel Settings:
- Coupling = DC
- BW = 20MHz
- Termination = 1M
- Offset: 0.0V, 0.0A
- Leave the vertical scaling to default for now. i.e. don't adjust the fine scale to make the traces the same on the display. This shouldn't have any effect, but just in case for now.
- Current scaling with 0.1 Ohm = 100mV/A (or 10A/V) and x1 attenuation.

Acquire Menu:
- Mode = Sample (for now)
- Record Length = 10k (increasing to 100k did not make any difference for me)
- Delay = don't care


btw, I changed the horizontal scale to see what difference the number of cycles makes for my measurements, and it was as follows:

5 MATH cycles, MEAN=2.37W
12 MATH cycles, MEAN=3.59W

So even if it doesn't seem to make a difference, please make it a habit of using 10 or so MATH cycles when making the critical MEAN power measurement. After you get the measurement, you can always decrease the horizontal to see the wave forms better if you wish.

gotoluc

  • elite_member
  • Hero Member
  • ******
  • Posts: 3096

Channel Settings:
- Termination = 1M
- Offset: 0.0V, 0.0A
- Leave the vertical scaling to default for now.

Acquire Menu:
- Record Length = 10k

Okay but all the above I don't see in menus?

Also, I think I see a problem. I tested with my 10 Ohm 1% at 10vac which should = 10 Watts and should be identical when I select Invert but see scope shots below. I added  Math Max and Min to see what's going on.

I did the scope auto calibrations with probe removed and it's clear the Math has an offset. Can this be adjusted?

Also, I have found that when I use Fine for vertical scaling and adjust each channel to fill the scope window the most that the readings are more accurate then using coarse setting and having two different size wave forms. I mostly see a difference if I make the Math as large as possible to fit the window. Try it, you should see a small change numbers.

I agree with you that it is better to check more sample to be sure the numbers are accurate or stay the same.

Luc

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Luc,

You only have about 6 MATH cycles there. Can you increase the number of cycles and see if there is better correlation between inverted and non-inverted?

Regarding those settings that you can't find, I'll check the manual tomorrow (Tue) and find them for you. They should be there somewhere.

gotoluc

  • elite_member
  • Hero Member
  • ******
  • Posts: 3096
Okay, here you go.

I'm done for the night

Luc

tinman

  • Hero Member
  • *****
  • Posts: 5365
@ Poynt

The only way i can select the math as the source,is in the cursor menu. The shots below are set up as this-ch1 x ch2. DC coupling.Voltage probe(yellow trace)set at 10x,and current probe(blue trace) set at 1x,across a .1 ohm 5 watt resistor.I just cant seem to get it to come up like Luc's shows,with it in the menu bar down the right side-math that is.All i can get is the little box of numbers in top left corner.

tinman

  • Hero Member
  • *****
  • Posts: 5365
I can also get all this junk(below),but only on ch1 and ch2,not math.
As you can see,i realy know my way around a scope.

tinman

  • Hero Member
  • *****
  • Posts: 5365
So here is what i did ,just to lead myself up the garden path a bit. I took the Crms of ch2,which is 168 mV,over our .1 ohm resistor,and come up with a current of 1.68 amp's. then multiplied 1.68 x the Crms of ch1 of 66 volt's. This comes to 110.88 watt's. My watt meter says the motor draws 48 watts without the exciter circuit conected,and 158 watts with the exciter circuit conected. 158-48=110-Oh by the way-all the above scope shots are taken on the exciter circuit-just as a demo.

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Okay, here you go.

I'm done for the night

Luc

OK, still off a bit.

I guess we'll see if any of those other settings are different.

I wonder if it could be the probe you are using. I am not even using a probe, rather a length of coax with a BNC on one end, and alligator clips on the other.

One thing you may try is changing the termination from 1M to 50 Ohm for the current probe channel.

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
I can also get all this junk(below),but only on ch1 and ch2,not math.
As you can see,i realy know my way around a scope.

LOL. Well the good news is, that you can do A x B and display it! I don't think we were able to do that several months ago.

So the challenge now is to figure out if you can apply a measurement to the MATH trace. But I suspect that is what you are struggling with at the moment, correct?

I see that you have a MEAN measurement in that mess of numbers, but you say you can't get that for the MATH trace? Hmm. Damn, having the MATH trace is almost useless if you can't apply any measurements to it!

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Channel Settings:
- Termination = 1M
- Offset: 0.0V, 0.0A
- Leave the vertical scaling to default for now.

Acquire Menu:
- Record Length = 10k


Okay but all the above I don't see in menus?

- OK, You don't have the choice of termination on your scope. 1M is the fixed setting.

- Looks like offset is not included. Good.

- Vertical scaling, I will try it with mine and see what effect if any it has.

- Record Length seems to be fixed at 2500. So no adjustment there.

So I think you have it adjusted as best we can get it for our needs. I am going to perform a similar test you just did above to see how close my measurements correlate when inverting CH2.

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Luc,

Just for curiosity sake, try running the calibration again, but this time leave the voltage and current probes connected to the scope. Disconnect them from the circuit though and run the cal. Do this once with the probes open-circuited, then remeasure to observe the balance, and once with the probe tip shorted to the gnd lead. Again remeasure to see if the inverted and non-inverted are any closer.

Hopefully one of these two methods will minimize the slight offset we are seeing. Let me know what you find.

I have done this before, and even though the instructions say to disconnect the probes, I have found that passive probes can sometimes cause some sort of offset, and calibrating with them attached removes this offset.