• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

tubes, killawatt, current

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I was testing a kill-a-watt device to see how accurate vs. my meter. I had the killawatt plugged into my variac. And then I had a line splitter plugged into the kilowatt so I can monitor with my clamp meter at the same time.

When powering tube amplfiers there is difference between current readings between the killawatt and my clamp meter. However, when powering a light bulb or a heatgun, the current measurements are the same. also when powering a solid state amplifier the current measurements are the same between both devices.

What is it about the tube amps that is causing the different readings?
 
Administrator
Joined 2007
Paid Member

PRR

Member
Joined 2003
Paid Member
Aside from reactive "apparent power", most DC power supplies work by taking huge spikes of current only 10% of the time. This will read different with an averaging meter, a peak meter, or an RMS meter.

Which is "right" depends on why you want to know. If you wish to predict your Electric Bill, again you must ask how the power company meters your house. I suspect my mechanical meter (with digital data-link) averages-out sub-cycle spikes. Newer meters may read sub-cycle spikes several different ways.

If you simply wish to compare the clamp-meter to the KillAWatt (remember the saying about a man with two watches), and do not want to think about reactance or wave-forms, "simple resistors" should give the same number on both, no matter what they read. This will be incandescent lamps, heaters, toasters, etc. (Well, actually I have a heater which turns-down by dropping part-cycles; run such a thing on max heat.)
 
If your clamp meter uses a traditional meter movement with a needle, then it almost certainly is an average-responding unit. They typically use a resistor, rectifier, and d'Arsonval movement meter.
Those current meters are calibrated using a sine wave to read out in RMS current, but it will only do so for pure sine wave current. Any other wave shape will give an incorrect reading (will not read out the correct RMS value).

As PRR stated, DC power supplies do not present a linear load to the power sine wave voltage. It distorts the sine wave. That makes the average-responding clamp meter read incorrectly.

My Kilowatt meter reads out RMS current and RMS volts. And it reads out true Watts and also reads out Volt*Amps (VA). When the phase of voltage and current is 0 degrees, then Watts = VA.
But when the phase of voltage and current is not 0 degrees, then Watts is less than VA.
W < VA.
The Kilowatt meter also reads out Power Factor (0 - 1.0). My tube amps have power factors from about 0.65 to 0.9).
For example, if the VA is 100, and the Watts is 90, then the power factor reads 0.9.
 
Thanks for pointing me in the right direction....

do the readings on schematics assume true rms?

I think my clamp meter is true rms (ideal 400). So the killawatt (p3) is certainly just a rms device for current? I read conflicting information about this.
 
Last edited:
Member
Joined 2015
Paid Member
From the Ideal 400AAC instruction manual, the model number 61-732 is average sensing, while only the model number 61-736 is true RMS. According to my experience, most low cost multimeters aren't RMS or have a severe bandwith limit and give wrong AC values on audio measurements. I am about to upgrade my multimeter to a real one because I am sick of the guesswork. I once had a very old analog bench multimeter and it was better on AC than the cheap brand new digital multimeter I am using now.
 
Thanks for pointing me in the right direction....

do the readings on schematics assume true rms?

I think my clamp meter is true rms (ideal 400). So the killawatt (p3) is certainly just a rms device for current? I read conflicting information about this.
Don't assume that the power measured by the product of rms voltage and rms current will be active (average) power like the one measured by your killawatt or a normal domestic meter: imagine the load is a pure capacitor, drawing 1A.
The clamp-meter will measure 1Arms, the voltmeter 120Vrms, but the active power will be zero, not 120W.
True power meters make a real-time vector product of I and V and average the result.
There is no way to achieve such a result with separate, independent instruments.
It is also the reason why overunity fans can claim positive (though false) results....
 
My clamp meter is NOT true RMS and the Kill-a-watt is.


So when tinkering with tube amps and the non linear load from their dc power supply, I should probably invest in a true RMS clamp meter. Should I be concerned that my voltages in my amps are way off? With the current for example, the difference didnt appear subtle, one amp was .84 vs .99 and another was like 1.10 vs. 1.50.... or maybe that +/- is subtle!

With tube amplfiers, when is it essential to use a true rms meter?

Thanks again, lots to learn.....
 
Last edited:
Residential customers in the US have been traditionally charged for fundamental 60Hz watts, with harmonics being ignored. The old time electromechanical watthour stators were responsive to the fundamental. Electronic meters therefore extracted the fundamental 60 Hz sine wave through any number of possible analog and digital filters. Similar argument applies to reactive power, though normally only primary rate customers are penalized (or credited) for reactive conditions.

Point being, rms current is not the best way to arrive at watt consumption if the concern is $$.

If you consider the rationale for paying for power, fundamental makes perfect sense. Utility generators are strict 60 Hz positive sequence devices. They are only able to supply fundamental watts, so it is this component that the utility should be able to get reimbursed for. Harmonic quantities do not get sourced from the generator; they are absorbed by it. Harmonics are generated by the load and forced back into the positive sequence system as an undesirable circulating current which heats the supply system.
 
Actually, the power company can be 'at fault' too. They use transformers (wow!). And some of them are pushed to the limit of saturation (not just because of customer's electronic loads that put large transient currents at and near the crest of the sine wave).

They are supposed to deliver power with 3% or less THD (the 3rd harmonic at 180Hz is the major harmonic energy contributor).
My friend has power line distortion that is sometimes at 5% or 6%.
It starts to look like it is filling in (headed just a little way toward a square wave).
The RMS voltage can be correct, but the peak voltage is low.
With the peaks of the sine wave quite badly squashed down then the Rectified DC B+ is low (with either capacitor input filters, or choke input filters).

At the initial testing of my amps I use the kilowatt to measure rms V and rms A.
When I do my schematics I measure my B+ that I get with the AC power I get from the power company.
When I check for what fuse value is best, I use the amp's operating rms A to calculate what value of slow blow fuse I use.
But I also measure the worst case inrush current at multiple cold turn-ons to set the value of the fast blow fuse I use. I can measure with a sense resistor in the neutral line, differential probing, and a scope.
Caution: you can not use a single ended probe and neutral sense resistor to measure this (Neutral is only at ground potential (0 Volts) all the way back at the power mains input at the building power delivery.
I use fast blow and slow blow fuses in series to get the best amplifier fault protection (i.e. 0.5A slow and 1.0A fast).
 
Last edited:
That can be true, but doesn't change the fact that you pay for fundamental only. Regardless of the source of harmonic being within the residence or in the distribution line, they furnish real power from one location only, and that is a generator. Any losses are either 'eaten' by them, or more generally incorporated into the rate structure.

The marketed purpose of the killawatt is to understand where your power consumption is, and how to potentially save money on your utility bill. They even claim to measure power the same as your utility. If that's true, it should be fundamental quantities only. I find no claim in their literature that harmonics are relevant, measured, or considered. Which is appropriate. Put down the true rms meter; it's not supposed to agree with the killawatt.

Last comment that is just as important; the presence of harmonics due to nonlinear loading presents a predominantly reactive load to the system. This could be in a lagging or leading mode, but the critical item is that the loading is not in real power, it is in reactive power. So the degree of harmonics will certainly change rms current readings, slightly alter rms voltage readings, and therefore significantly change VA readings. But the effect on watt (real power) readings should not change. The VAR (reactive power) readings will change with harmonics. Further understanding to explain why fundamental watt readings are appropriate.
 
Last edited:

PRR

Member
Joined 2003
Paid Member
...With tube amplfiers, when is it essential to use a true rms meter?.....

Uh, "never"?

I grew-up with two meters: needle (averaging) meters, and VTVMs (often peak on the AC ranges). Tube work is "never" so exact that we should be confused.

A special case: put both Average and Peak meters on a power amp output and turn up the level. At clipping, the Peak meter will stop rising, the Average keeps going up (because the Sine is going toward a Square with the same peak but more area over the waveform).

Mostly we "know" our wave-shape and, if unusual, can correct the reading.

The first time I saw "true RMS" was for calibrating theater dimmers. SCR/Triac dimmers make VERY ugly part-waves. One of the few waves so ugly that an Averaging meter didn't come close. (The Peak meter read 100% for anything over 50%.)
 
Outside the house:
I had never heard or thought that the electric company charges us only according to 60Hz power usage with no regard to harmonics (intended or realized). Whether it takes into account the phase or not is also interesting.

About 3 years ago, the power company replaced my ‘all electro-mechanical spinning wheel and
mechanical spinning numbers power meter’ with an ‘all digital meter’. If the new one only measures 60Hz power. It must have a couple of A/Ds. In addition it either needs: either 2 very effective low pass filters and a phase detector; or it has to run 2 FFTs, and phase detection.
Without phase detection, it certainly does not measure the true power dissipated.

Because the power company does not always deliver the same voltage, and because the customer load current varies, both the old electro-mechanical meter and the new digital meter must perform some kind of multiplication of voltage and current.

Inside the house:
The phase of current versus voltage varies according to the load(s), and current and voltage
may not be the same wave shape either.

In order to produce a Kill A Watt so that it only measures the 60Hz voltage and 60Hz current would also require either 2 very effective low pass filters, or it would require an FFT; and phase detection of the 60Hz (and only 60Hz) current and voltage (just like the electric power company’s meter above, if it does that).

My P4400.01 Kill A Watt cost about $20 or $25. I doubt that it measures it that way (FFTs or low pass filters).
It only displays one quantity at a time. But it must measure multiple factors at once to measure
Watts and Power Factor.
I am believing that it digitizes the waveforms, and does cycle-time integral of V, cycle-time integral of I, and does cycle-time integral of V*I (Watt).

The P4400 manual says: RMS Amps, RMS Volts, and Watts active power;
and Power Factor: Watts/(Vrms*Arms).
I do not see any claims that it only measures 60Hz, and not the harmonic contribution too.

The home Variac is an auto transformer, it may saturate, so it also may be a source of the harmonic distortion.

A power transformer with a high voltage secondary that allows choke input filter to get the desired B+ voltage, is what I use in my tube amps. It does not have nearly the current spikes of a cap input filter.

I can measure the real cycle rms of the current that the tube amp draws using the scope,
a differential probe setup, and a current sense resistor in the Neutral line. This allows me
to determine what value of slow blow fuse to use in the primary of the tube amp transformer.
This requires true rms measurement, a really bad current wave shape will fool any other measurement device (peak responding or average responding) from making an accurate determination of the amp rating of the slow blow fuse.

I can also use a True RMS voltmeter across the current sense resistor, and verify the scope reading.
In any case, the slow blow fuse responds according to true RMS current (and time versus overload value).
A DC current source and stopwatch will allow you to test fuses, and they are not always
very exact, they vary quite a bit. But once you test one, it is no good any more (get a box and check at least a couple).

I like PRR's comment that tube work is "never" so exact that we should be confused.
 
I don't know the details of the "smart meters" being deployed (and they are not all the same). I had assumed they only allowed simpler reading of power use (simpler to the Utility).

The old "rotor" meters to residential single-phase power cannot measure reactive elements; only 60 Hz power.* I suppose it's possible the newer smart meters could, but I have my doubts as it's still a single phase connection to the utility, and it seems to me you need three phase power to measure true power (power factor).

* This means that the Utility has no means to bill residential customers for the full power used via devices with a poor Power Factor, such as CFL bulbs. Recent increases in Utility rates are partly due to a widespread change to CFL bulbs in the home, as the Utility has no other means of accounting for the actual power used in single phase connections.
 
Last edited:
Maybe a simpler generalized statement would help:

Real power in a real world harmonic rich system is largely (almost completely) 60 Hz only.

Reactive power in a real world harmonic rich system consists of the fundamental component and all harmonics.

This paper is a fun little read about how to measure VARs by various methods.
http://www.radianresearch.com/white_papers/harmonicmetercal.pdf

Single phase vs three phase measurements really don't make any difference in measurement of power or power factor. The old time electromechanical meters used single phase stators on a common shaft. Sum of the shaft torques moved a register. A three phase meter was a combination of single phase devices. Read up on Blondel's theorem. The utility has no challenges with the measurement of power factor in single phase circuits; they simply by tradition did not charge residential customers for poor power factor (see post #9). Measuring VARs was simply phase shifting a quantity and using the same stator design. But all single phase.

I think the most interesting factoid coming out of the Radian paper is that measurement of VARs should perhaps be deprecated, and we should consider measuring real power (60Hz only is sufficient) and rms-VA (including all harmonics). Other desired values one might want to use for billing (power factor, VAR hours) would be calculated from those two quantities.
 
The 3 attached scope screen captures tell the story of the power at my house.

Plot # 15: 125Vrms 60.0076 Hz
The 60Hz fundamental is slightly less than 125V, (the wave shape includes the harmonics).

Plot # 17: The cursors show the relative amplitudes of the 60Hz and 180Hz.
The 180Hz 3rd harmonic is 36dB less than the 60Hz fundamental.
-36dB = 1.58% You can also see the 5th, 7th, and 9th harmonic.
All the other harmonics, odd and even are insignificant (too small to be of much effect).

I put the cursors just below the peaks of the spectral lines. That gives the same differential
result of 36 dB, but it allows you to see the peak of each special line. If I did not put them exactly the same distance below the peak, you can forgive me for a 1 dB error or so.
You will see the actual difference in the auto-crosshairs in screen shot #19

From this screen shot, I estimate the total harmonic distortion to be about 3%, maybe 4%.
But I am not going to measure all the harmonics. Then re-exponentiate them into voltages, and
calculate the THD from the root sum squares of the harmonic voltages; divided by the re-exponentiated fundamental voltage; and multiplied by 100 to get % THD.

As an example for you, suppose that there was only the 3rd harmonic, and the 5th harmonic, and that they were each -40dB. That means each harmonic is 1/100 of the fundamental voltage. 10^(-40dB/20) = 0.01
That is 1% 3rd, and 1% 5th.
But according to the (correct) root sum square method, that is 1.4%THD (not 2% THD).
Root ((1 squared) + (1 squared)) = Root (2) = 1.414….. 1.4%
A little rule of thumb works here for me, nobody pays me to make measurements anymore.

Plot # 19: Another set of cursors shows the frequency as well as the amplitudes (automatic crosshairs) of the 60Hz and 180Hz.
You will see the auto-crosshair on the 60Hz measures 41.4dB (that is 41.4dBV).
That calculates to 117.5Vrms of the 60Hz fundamental.

How does 117.5Vrms compare to the 125Vrms in screen shot #15?
The screen shots #15, versus screen shot #17 / #19, were captured at slightly different times.
The on and off of my electric heater, my water heater, the neighbor off the street transformer changed his power draw, etc., may all have affected this. These all can make the voltage (and distortion) jump up and down.
 

Attachments

  • TEK0015.BMP
    76.1 KB · Views: 43
  • TEK0017.BMP
    76.1 KB · Views: 42
  • TEK0019.BMP
    76.1 KB · Views: 33
Johnny2Bad,

Thanks for mentioning CFLs.
As I understand it, the purposes were to:
1. Reduce the use of Tungsten (expensive).
2. Increase the use of Mercury (toxic). Oh, was that on purpose, or just overlooked?
3. All in the name of saving energy.
Question: How much energy does it take to properly dispose of a single CFL?
 

PRR

Member
Joined 2003
Paid Member
...Because the power company does not always deliver the same voltage, and because the customer load current varies, .... meter must perform some kind of multiplication of voltage and current.....

Suggested Read: Electric Power Metering, Knowlton, 1934, $13 reprint (but I have a 6th printing from WWII)

Gives a short history of attempts to meter electric power. The 1934 AC Watt-Hour meters are recognizably same as the ones I grew up with. And same as the one I have now except behind the gear-train is electronics which beam data back to the company (no meter-reader). Full treatment of that, and much more about PF, demand, and polyphase metering.

The classic rotating disk meter has a Voltage coil and a Current coil. Their magnetic fluxes combine as a Multiplication. Torque on disk == V*I. Permanent magnets provide a retarding effect proportional to speed. The result is that RPM == V*I. The number of revolutions in an hour, or a month, is proportional to Watt-Hours.

The V*I multiplication is essentially "instant". If you hung a pure capacitor on the meter, you would have Current at zero Voltage, and Voltage at zero Current. These four points would correctly read zero actual Power. Thinking now, midway between V and I zeros there is finite V with finite I. There may be some reading.(*)

I think the thinking is that NO customer would hang a PURE reactance on the line. It does the customer no good. Customer wants some work done. Customer needs Power. Some common Power-sucking loads also have reactance, Power Factor. The big one is large motors which run with significant lagging current. This heats the lines and also upsets system balance. Large factories etc submit to Power Factor metering, penalties for high PF, and may opt to add opposite reactance (capacitors for motors) to null their PF surcharge.

This was never done for residential customers. The early loads were filament lamps, unity PF. The bonus loads, electric cooking, ironing, heating, are also unity PF. By this time the domestic power industry was neck-deep in Practices and Tariffs, so changing the rules became very difficult. The next big domestic load was air-conditioning, now a major part of the load in some areas. A/C pump motors have significant PF. Apparently the electric companies just wrap it up in the relatively high rates they charge small distributed customers (residences). Residential work also requires significant over-investment to cover a few high-use days a year, line and transformer capacity which sits 90% idle 360 days a year. Compare to the rates a large factory working steady and hard can negotiate.

And no, you do not need 3-phase to have or meter PF. Just that we overlook PF for small loads, enforce it for large loads, typically large enough to justify the costs of 3-phase delivery. 3-ph is sorta a discriminatory line between little customers and big customers; we play on different rules.

I'm working at another house with an "opaque" meter. It has an LCD behind glass but no rotating disk. Such things could work any way they want. Even pull random numbers from a hat. (Some customer got a $800,000 bill, but was offered a 10-month payment plan with it. She thought she had connected her holiday lights wrong. A human at the company admitted it must be a mistake, her bill was $200.) We would hope some approximation of the classic disk meter. Even to the point that a disk versus an opaque-meter would give the same reading for any likely residential load (+/-$1).

(*)Knowlton says there is torque but both + and -, so the result is correct: zero.

Remember that frequency DOES vary. In modern gridded systems, not much; but this was considered in 1934. Knowlton shows that with unity PF a meter will read 0.5% high at 50Hz, 0% at 60Hz, and 1.6% low at 70Hz. Extrapolating unwisely, it may read 30% low at 120Hz. The coil reactance is not totally swamped and begins to matter at higher frequencies.
 

PRR

Member
Joined 2003
Paid Member
... CFLs. As I understand it, the purposes were to:...

No. Just efficiency. Incandescent is terrible efficiency. Fluorescent is better, but does not scale to small size (excess surface/volume ratio).

Mid-1970s, Ed Hammer at G.E. worked on small fluorescents, dimensions and gas mixes, phosphors, and devised the spiral tube. GE liked it but it would need a whole new factory. They passed on that but considered licensing. Meanwhile the idea leaked out. Philips did a magnetic ballast version. Osram did an electronic ballast CFL. Asian glassblowers huffed and puffed, along with their electronic ballast designers.

The cost of tungsten has never been an issue.

In retrospect, the savings from CFLs may have been in vain. We would not pay for good ones, and the cheap ones crapped-out too fast. CFLs in every room invites Mercury poisoning; also the tubes make far more shards of glass. H-D does not charge me to take-back CFLs, but safe disposal can't be free, and may not be good for the workers. I have been working to remove ALL the CFLs from my home; it is amazing how many we acquired in the decade they were dominant.
 
PRR,

Very informative about power measurement, thanks!

My comments about CFLs was more tongue in cheek. I never liked CFLs, and have been replacing with LEDs. I am thinking they may be less environmentally friendly (production to grave) than the old incandescents.
I believe my oven light will always be an incandescent, or the oven will just be dark.

When the cold and wind dies down, I need to take another walk around the block. I think I saw a big capacitor on a power pole in a residential area.

I use choke input filters on my B+ whenever possible. It does keep the B+ secondary cooler (and probably the primary as well), for the same current output.
The disadvantages are DC = 0.9 * rms, and magnetic spray from the choke.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.