Oscilliscope reading 1/2 voltage

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Echo North

Well-known member
Joined
Jan 21, 2006
Messages
4,452
Location
Seattle, WA
Hi,

Let me start by saying I am totally new to oscilloscopes. I just got my first scope (free!) a Kikusui COS5020. I have all the knobs off or zero'd I think...

When take try to take the voltage of a nine volt battery with the voltage units set to 1v the line jumps 4.5 lines or 4.5 volts.

When I attach the probe to the calibration lead which is .5 (Vp-p) I get a square wave that reads .25 v peak to peak.

Everything seems to be reading 1/2 voltage...am I missing something?

Thanks!

Mike
 
Check that the scope doesn't have an adjustable range trim. Some scopes have a variable scale adjustment for making specific comparisons. Sometimes this will be a dual concentric pot behind the range switch. They often have a click at full CW that should be calibrated.

If you don't have a range trim, it could be a probe or calibration problem. If you have more than one probe/channel confirm they all act the same.

JR
 
Hi,

The VAR pots on the volt/div selector are pushed in and both positions are clicked off. Same with the cal'd selector.

The line voltage selector on the back for the power supply is set to 110-120 volts. I assume this is a regional thing.

Both sets of probes and both channels behave the same. One question is - the input for the probes reads 25pf and the probes read 14pf...could this be my problem?

Here are some photos:

2268869393_ec139501bb.jpg
2269660042_8e334c86f3.jpg


Thanks,

Mike
 
If both probes act the same it looks like a calibration issue. The probe capacitance will only make a difference for overshoot on fast square wave rising or falling edges. There is typically a standard resistive input termination (1 Meg) to form a voltage divider with 10x probes.

You can eliminate the probe entirely by using a short length of wire between the calibration signal and probe input jack, but I would suspect a general calibration issue.

JR
 
If I run a short wire from the test signal to the prob input I get 2.5v which I assume since the probe is a 10:1 equals .25v. I assume the confirms no probe problems.

So where or how are oscilloscopes calibrated? Is it worth doing with a cheap scope?

Thanks again,

Mike
 
Using a wire, the scope will be reading actual voltage with no external divider. If scale is set for 10x probe it could indicate 2.5V for .25v input.

If it was just a bad probe issue you might get 5V for .5V in.

The fact that you get a similar error for 9V battery and reference suggests scope calibration is off. A factor of 2x seems a lot for adjustment range from calibration.

Is this a new scope? It appears faulty.

JR
 
It's an old scope from a lab at the university I work at. The tech seemed to think it was working fine. I'll bring it back to him and see if he has another scope.

On a side note...when tested wall voltage it read 160v when my multimeter showed 120v.

Mike
 
[quote author="Echo North"]On a side note...when tested wall voltage it read 160v when my multimeter showed 120v.[/quote]
How were you 'measuring' this, and how much do you know about RMS-versus-peak-to-peak...?

Keith
 
This is where I embarrass myself if I haven't already.

I know nothing about RMS voltage. Other then it means "Root Mean Square".

I measured the voltage by plugging a power chord into the wall. The chord had been cut and the leads where exposed. It was not a grounded chord. I attached a lead to the (+) and got approx. 160v (peak to peak). When a ran my multimeter over both leads I got 120v.

Thank you for the help.

Mike
 
So I went to an online rms calculator..and if I enter my peak voltage as 86v (which is what mine was) it tells me my P-P is 172 and my RMS is 60V. So if I double that to account for an RMS p-p that gives me 120V.

Is that coincidence...or correct?

Thanks,

Mike
 
> Everything seems to be reading 1/2 voltage...am I missing something?

Calibration errors are like 10%.

Exact-half sounds more like a push-pull stage has gone push-only.

> power chord into the wall.

"Smoke on the Water" has a power chord. Smoking technicians use a power cord.

> got approx. 160v (peak to peak). When a ran my multimeter over both leads I got 120v

The meter traditionally reads the RMS (heating value) of AC, though it may actually read Average or P-P and factor in an assumption about sine waves.

120V RMS Sine Wave is 2.828*120V= 339.4V peak-peak. So you have a near-right result except half of what it should be. (Very common for wall-line "Sines" to be flatter than a perfect Sine.)

All else appears fine. The internal square is more than square enough for poverty work. The "half reading" seems consistent. You could just re-number the knobs and get on with life.

If indeed it is a push-pull stage gone half-bad... long-term flaws are even-order distortion and DC drift. I'm old enough to have 'scopes with single-ended CRTs, and I don't think this is a major flaw. Great precision is nice, but as you get to know a 'scope's quirks they don't hurt the learning, they just look bad when you send your paper to journals.
 
I simply may just go with doubling everything! I'm going to check to see how it reads on the horizontal (time) plane. I seem to recall I was getting about one cycle per 82.5ms with wall signal. I need to figure out how to calculate this in to hertz.

Does this have anything to do with the fact that my Oscilloscope is COS5020. Is this the difference between sine and COS?

Thanks again,

Mike
 
> COS5020. Is this the difference between sine and COS?

No.

That's a mathematic distinction about what part of the cycle you call "start". SIN and COS are the same thing in a real world where we pick a start-point for arbitrary convenience. "COS" probably stands for "Cheap OscilloScope"... ah, Kikusui product numbers run like AVM13=AC voltMeter, DME1400=Digital multiMEter, PAB13= Power supply, FC01120= Frequency COunter. So maybe "COS" is "Cathode ray OscilloScope" (opposed to LCD 'scopes), whatever.

And how could they sell a box marked SIN? My purchasing dept would see "SIN69 - 1ea @ $399.98" and think it was a joke.

> one cycle per 82.5ms with wall signal. I need to figure out how to calculate this in to hertz.

60 cycles per second is 1/60 of a second per cycle. 1/60 is 0.016,666 seconds, 16.66 milliSeconds.

One half-cycle in 8.3mS is correct. Set sweep to 1mS/div and see if you get a bit more than half a complete wave.

PLEASE do this with a 120V:12V transformer, NOT right off that wall-cord.

Or just put your un-grounded finger on the probe tip. Any indoor workroom will throw a large fraction of a Volt of stray 60Hz buzz onto an un-grounded body. It will have a lot of "garbage", because stray line-spikes couple better than low 60Hz tone; but you can usually sort-out the basic 60Hz roundy wave.

5%-50% of ex-school-lab 'scopes will be defective some way, depending how sharp the students were about detecting problems (they have less practical experience than you) and how energetic the lab tech was in fixing the old 'scopes after new ones were budgeted.

For $50, the going rate for a used-working 5020, you should ask for explanation or a swop. For free, you can only throw yourself on your friend's mercy.

> the input for the probes reads 25pf and the probes read 14pf...could this be my problem?

No. A serious mis-match will round-down/spike-up high frequency shapes. But no effect on a 9V battery. And probes invariably have a range of adjustment for different inputs. That's why they give you a known-sharp square-wave. Aside from the known voltage, you know the corners are crisp. Verify with 1X probe. Then put probe in X10 mode (and bring up V/div). If the "square" is round-shoulder or spike-corner, use a fine screwdriver in the probe and you will get a different shape. When you get a best-possible square, the probe is matched to that cable and input. You won't have to re-adjust if you use that probe on the other input, but you will have to adjust if you use that probe on another 'scope, or if you do much-much finer work than a $free 'scope should be expected to do.
 
Hmmmm. The range switches run 1-2-5. Three of the four steps are 2:1 ratios.

So if the knob was installed one-off from the correct notch, and you ran readings over a wide and random range of voltages and settings, you would have a 1:2 error 75% of the time and a 1:2.5 error 25% of the time.

Fiddle that 9V battery to large and small displays and see if you can find a setting which is off by 2.5. You may need a mm ruler to be sure it is error of 2.5 and not just the known error of 2.

I think this unlikely because the inexpensive Japan 'scopes tend to have a set-screw, a flatted shaft, and an index-nub. But if it has been severely twisted by frustrated students, they could have buggered it one-off from the correct setting. You think you set 1V/div and really are at 2V/div. Then if you set "2V/div" you would really be at 5V/div, and 9V would read 1.8div instead of 2.25div as extrapolated from 4.5div at "1V/div".
 
Am I reading the time scale correctly?

I tried what you said and set the time/div to 1ms/div and got a little more then 1/2 a cycle across my screen. However I though 1ms/div meant that each little line was equal to 1ms. Since I have 10 squares each with 5 little lines that this meant my screen represented 50ms across my screen.

I'm thinking this is VERY wrong?

When I originally hooked my probe to the wall-cord I had it set 5ms/div and it about 16.6 small ticks to complete on cycle. Does this make sense?

Thank you for helping, I know my lack of basic knowledge is probably frustrating!

Mike
 
Time scale refers to each MAJOR division (square) as opposed to each sub-division ("tick").

If you have five squares either side of the center, -a total of 10 squares wide on the screen- then at "1mS per div", it should take 10mS to traverse the entire screen.

Keith
 
Oh man...I'm sorry.

So it would appear that my Time Scale is reading perfectly fine.

So just to clarify the Volts/Div is by Sub-divisions correct?
 
PRR wrote:

"So if the knob was installed one-off from the correct notch, and you ran readings over a wide and random range of voltages and settings, you would have a 1:2 error 75% of the time and a 1:2.5 error 25% of the time."

I'm not sure it could be a knob issue since both channels read the same?

Thanks again,

Mike
 
I tested the knobs and the 1-2-5 relationship seemed to be correct.

From these posts can we conclude that the Time (x-Axis) is working correctly and the Volts (y-axis) consistently reads 1/2?

If so I may take PRR's advice and move on.

Thanks!

Mike
 
Oh now I'm really confused about volts per division...

The time axis seems correct...but I have been reading the volt axis as volts per sub-division. Is this right?

My max v/div is 5v. That would mean if it was v per DIVISION my max viewable voltage would be 8 squares x 5 for 40 volts. That's why I assume it's 5v per SUB-DIVISION (small tick)?

Mike
 

Latest posts

Back
Top