cdynasty001
Member
- Location
- California
Oscilloscope Problem
- Thread starter cdynasty001
- Start date
- Status
dfmischler
Senior Member
- Location
- Western NY
- Occupation
- Facilities Manager
I assume you are using the oscilloscope's "calibrator" output.
As in any other troubleshooting scenario you need to be able to make a change and see the effect. This could include trying:
- A known-working signal source to verify that the probe and scope are working (do not exceed input specs).
- A known-working oscilloscope to verify that the calibrator output is OK.
- A known-working probe to verify that everything else is OK.
Is there an operator's manual? Make sure you follow the probe compensation instructions to the letter before contacting the vendor for an RMA.
As in any other troubleshooting scenario you need to be able to make a change and see the effect. This could include trying:
- A known-working signal source to verify that the probe and scope are working (do not exceed input specs).
- A known-working oscilloscope to verify that the calibrator output is OK.
- A known-working probe to verify that everything else is OK.
Is there an operator's manual? Make sure you follow the probe compensation instructions to the letter before contacting the vendor for an RMA.
junkhound
Senior Member
- Location
- Renton, WA
- Occupation
- EE, power electronics specialty
At 1 kHz the probe compensation has little effect. Your best bet is to look at an external waveform generator and then go back to the owners manual.
Like df says, look at your owners manual, you likely have simply pulled up that particular waveform from the memory.
Do not have a siglent scope myself, but curious, so
look at page 17, http://www.siglentamerica.com/USA_w...UserManual/SHS800_UserManual_UM03008-E05A.pdf;
Have you tried a different probe? Say a x1?
I think it's the scope's own test output.
cdynasty001
Member
- Location
- California
Thanks for your responses. I followed the manual. I did the compensation test and adjusted the probe. It's a little plug you put on the USB port and you put the leads on to get a square wave. I was unsure if I had to match the oscilloscope attenuation ratio with the probe's but I did that as well. There was very little change in the test wave. I changed between 1X and 10X - the voltage changed but not the shape of the wave. And I am switching between 3 different probes. I noticed the test wave would get more distorted over time as if heat was causing it or something else.
I do have a function generator. Before I wasn't getting a solid square wave. But now I am! I guess that's all that matters here right? I just wanted to know if my scope was working properly
I do have a function generator. Before I wasn't getting a solid square wave. But now I am! I guess that's all that matters here right? I just wanted to know if my scope was working properly
gar
Senior Member
- Location
- Ann Arbor, Michigan
- Occupation
- EE
161026-2122 EDT
cdynasty001:
How you got the waveform is important.
Immediate observation of the waveform leads one to suspect this is the output of a simple series connection of a single resistor and capacitor fed from a balanced square wave source with the voltage measurement across the resistor. In this type of circuit an exponential decay of the voltage with time occurs. Your waveform looks like this.
From your measurement the RC time constant is around 500 microseconds. A combination of 5.6 k and 0.1 mfd is close to this. Or you could use 560 k and 0.001 mfd. If I experimentally setup a 1 kHz square wave with 5.6 k and 0.1 ufd I can approximate your waveform. If you lower the frequency to 0.2 kHz the exponential curve is more apparent.
RC values like these would not be expected in your probe or your source. A typical scope input is 1 M shunted by possibly 100 pfd. This means a compensated probe would contain a 9 M resistor shunted with about 9 pfd (that 9 pfd is the adjustable compensating capacitor). If the probe resistor opened, then the equivalent source circuit is approximately a square wave of about 1/10 its original value with an internal impedance of about 110 pfd. The time constant of this in combination with the scope 1 M resistance would be about 110 microseconds. Not the approximate 500 microsecond value indicated from the scope waveform, but may point in the direction of the problem. Likely the probe.
I haven't proofread, but consider the concept.
.
cdynasty001:
How you got the waveform is important.
Immediate observation of the waveform leads one to suspect this is the output of a simple series connection of a single resistor and capacitor fed from a balanced square wave source with the voltage measurement across the resistor. In this type of circuit an exponential decay of the voltage with time occurs. Your waveform looks like this.
From your measurement the RC time constant is around 500 microseconds. A combination of 5.6 k and 0.1 mfd is close to this. Or you could use 560 k and 0.001 mfd. If I experimentally setup a 1 kHz square wave with 5.6 k and 0.1 ufd I can approximate your waveform. If you lower the frequency to 0.2 kHz the exponential curve is more apparent.
RC values like these would not be expected in your probe or your source. A typical scope input is 1 M shunted by possibly 100 pfd. This means a compensated probe would contain a 9 M resistor shunted with about 9 pfd (that 9 pfd is the adjustable compensating capacitor). If the probe resistor opened, then the equivalent source circuit is approximately a square wave of about 1/10 its original value with an internal impedance of about 110 pfd. The time constant of this in combination with the scope 1 M resistance would be about 110 microseconds. Not the approximate 500 microsecond value indicated from the scope waveform, but may point in the direction of the problem. Likely the probe.
I haven't proofread, but consider the concept.
.
winnie
Senior Member
- Location
- Springfield, MA, USA
- Occupation
- Electric motor research
That waveform looks exactly like what you get with a low frequency square wave and AC coupling...however according to the manual the cut-off for AC coupling is 10Hz, so a 1kHz signal should not have a problem.
The manual shows that the _trigger_ input can have a 'LF reject' coupling mode which filters below 7kHz, so if by some chance the 'LF reject' coupling mode were active for the channel trace (rather than for trigger) then you might see this sort of output.
I'd say it is worth trying different coupling settings, just to see if you have the scope in a strange mode.
-Jon
The manual shows that the _trigger_ input can have a 'LF reject' coupling mode which filters below 7kHz, so if by some chance the 'LF reject' coupling mode were active for the channel trace (rather than for trigger) then you might see this sort of output.
I'd say it is worth trying different coupling settings, just to see if you have the scope in a strange mode.
-Jon
gadfly56
Senior Member
- Location
- New Jersey
- Occupation
- Professional Engineer, Fire & Life Safety
gar
Senior Member
- Location
- Ann Arbor, Michigan
- Occupation
- EE
161027-1329 EDT
gadfly56:
What manual?
Any of the many scope probes I have, when over-compensated, look
like an apparent time constant of about 50 microseconds. Nothing close to what cdynasty001 showed. Also cdynasty001 was not able to adjust his probe compensation to obtain a square wave.
My 10x Tektronix scope probe input capacitance is 6.5 pfd when compensated, and 14 pfd when fully over-compensated. As measured with a Tektronix LC meter. The scope input is 28 pfd and the probe cable 91 pfd. The combination of all these does not quite correlate, but are approximately as expected.
A single stage RC high pass filter using 0.1 ufd and 5 k has a 3 db frequency of about 300 Hz.
We need more information from cdynasty001.
.
gadfly56:
What manual?
Any of the many scope probes I have, when over-compensated, look
like an apparent time constant of about 50 microseconds. Nothing close to what cdynasty001 showed. Also cdynasty001 was not able to adjust his probe compensation to obtain a square wave.
My 10x Tektronix scope probe input capacitance is 6.5 pfd when compensated, and 14 pfd when fully over-compensated. As measured with a Tektronix LC meter. The scope input is 28 pfd and the probe cable 91 pfd. The combination of all these does not quite correlate, but are approximately as expected.
A single stage RC high pass filter using 0.1 ufd and 5 k has a 3 db frequency of about 300 Hz.
We need more information from cdynasty001.
.
gadfly56
Senior Member
- Location
- New Jersey
- Occupation
- Professional Engineer, Fire & Life Safety
gar
Senior Member
- Location
- Ann Arbor, Michigan
- Occupation
- EE
161027-1457 EDT
If I simply use a 10 pfd capacitor from the square wave source to the scope input, then the time constant is about (10 + 28) * 1 = 38 microseconds, and the waveform apppears to have about a 40 microsecond curve.
With expected components associated with the probe we don't get close to a 500 microsecond time constant. Thus, it is likely there was simple high pass RC filter in the circuit.
.
If I simply use a 10 pfd capacitor from the square wave source to the scope input, then the time constant is about (10 + 28) * 1 = 38 microseconds, and the waveform apppears to have about a 40 microsecond curve.
With expected components associated with the probe we don't get close to a 500 microsecond time constant. Thus, it is likely there was simple high pass RC filter in the circuit.
.
gar
Senior Member
- Location
- Ann Arbor, Michigan
- Occupation
- EE
1621027-2204 EDT
gadfly56:
The manual in post #3 was suggested by junkhound. No time base defined.
Attached are two actual plots from my scope with a 1 kHz square wave and a fully over-compensated Tektronix probe. Time constant is ballpark 100 microseconds. One plot is the full waveform, and the other is the same waveform with DC 0 shifted off the screen, and the steady state level of the transient set at the screen mid-point.
.
.
.
The plot in the original post looks like the transient decay is toward 0 V. This would imply no resistance around a coupling capacitor. Note in my first plot the decay is toward the voltage defined by the resistance ratio of the probe resistance and the scope input resistance.
.
gadfly56:
The manual in post #3 was suggested by junkhound. No time base defined.
Attached are two actual plots from my scope with a 1 kHz square wave and a fully over-compensated Tektronix probe. Time constant is ballpark 100 microseconds. One plot is the full waveform, and the other is the same waveform with DC 0 shifted off the screen, and the steady state level of the transient set at the screen mid-point.
.
.
.
The plot in the original post looks like the transient decay is toward 0 V. This would imply no resistance around a coupling capacitor. Note in my first plot the decay is toward the voltage defined by the resistance ratio of the probe resistance and the scope input resistance.
.
Last edited:
gar
Senior Member
- Location
- Ann Arbor, Michigan
- Occupation
- EE
161028-1257 EDT
MY guess is that cdynasty001 had his scope Y-axis set to AC coupling and not DC. This means the scope input looks like a capacitor and resistor in series, with the capacitor connectee to the square wave source, and the scope input connected across a 1 megohm resistance.
When you look at the waveform in post #1 it appears the curve is approaching 0 V. The square wave input frequency is too high for the decay to reach 0.
The following plots illustrate this on my scope. My scope has much better low frequency response in AC coupling mode. It is about an 0.1 mfd capacitor for isolation into a 1 megohm resistance.
This time I inverted color so the plots are more readable.
*****.
DC coupling 3.5 V/div, step change 4.4 div, 5 Hz sq-wave
.
*****.
AC coupling 3.5 V/div, note step change 4.4 div, 5 Hz sq-wave.
63% drop occurs somewhat over 100 milliseconds
..
*****.
AC coupling 5 V/div, note step change 3 div, 0.5 Hz sq-wave.
Same voltage change as before just different scale factor.
63% change about 140 milliseconds. This plot clearly shows approach to 0 V.
.
.
MY guess is that cdynasty001 had his scope Y-axis set to AC coupling and not DC. This means the scope input looks like a capacitor and resistor in series, with the capacitor connectee to the square wave source, and the scope input connected across a 1 megohm resistance.
When you look at the waveform in post #1 it appears the curve is approaching 0 V. The square wave input frequency is too high for the decay to reach 0.
The following plots illustrate this on my scope. My scope has much better low frequency response in AC coupling mode. It is about an 0.1 mfd capacitor for isolation into a 1 megohm resistance.
This time I inverted color so the plots are more readable.
*****.
DC coupling 3.5 V/div, step change 4.4 div, 5 Hz sq-wave
.
*****.
AC coupling 3.5 V/div, note step change 4.4 div, 5 Hz sq-wave.
63% drop occurs somewhat over 100 milliseconds
..
*****.
AC coupling 5 V/div, note step change 3 div, 0.5 Hz sq-wave.
Same voltage change as before just different scale factor.
63% change about 140 milliseconds. This plot clearly shows approach to 0 V.
.
.
gar
Senior Member
- Location
- Ann Arbor, Michigan
- Occupation
- EE
161029-0900 EDT
In the plot of post #1 note the inflection point in the curve at about 250 microseconds. The waveform is not a clean exponential decay of a single RC stage, but rather looks like two separate time constants. The first is probably the RC from an overcompensated probe, and the second from the AC coupling at the scope input.
These two can be sorted out by more experiments with the scope of the original post.
If one does not have an adjustable square wave generator, then such a device can be made easily with a 555 timer chip. The CMOS version is probably the better one to use. The datsheet shows the wiring for a simple near square wave generator.
More experiments from cdynasty001 would be useful.
.
In the plot of post #1 note the inflection point in the curve at about 250 microseconds. The waveform is not a clean exponential decay of a single RC stage, but rather looks like two separate time constants. The first is probably the RC from an overcompensated probe, and the second from the AC coupling at the scope input.
These two can be sorted out by more experiments with the scope of the original post.
If one does not have an adjustable square wave generator, then such a device can be made easily with a 555 timer chip. The CMOS version is probably the better one to use. The datsheet shows the wiring for a simple near square wave generator.
More experiments from cdynasty001 would be useful.
.
gar
Senior Member
- Location
- Ann Arbor, Michigan
- Occupation
- EE
161029-1051 EDT
Actual waveforms are better than just talk. Following are some plots to illustrate my above points. I used some convenient "cherry picked" values.
The scope is direct input without a probe. Thus, input is about 1 megohm shunted with about 30 pfd. There are three plots. The first is the square wave directly to the scope input. The second is thru a series 0.001 mfd capacitor. This filters out any steady DC signal. And third 200 pfd shunted by 1 megohm is added in series with the 0.001 mfd capacitor.
.******
DC coupled square wave into scope. Peak-to-peak change is 4 major divisions. This won't change in the subsequent plots.
.******
Square wave coupled into scope thru a series 0.001 mfd capacitor. Step change is 4 major divisions. Very smooth exponential curve.
.******
Square wave coupled into scope thru a series 0.001 mfd capacitor in series with a 200 pfd in parallel with 1 megohm. Step change is 4 major divisions, but not readily visible in this plot. Here you see the effect of compounding the two different RC time constants. Inflection point is obvious.
.
Actual waveforms are better than just talk. Following are some plots to illustrate my above points. I used some convenient "cherry picked" values.
The scope is direct input without a probe. Thus, input is about 1 megohm shunted with about 30 pfd. There are three plots. The first is the square wave directly to the scope input. The second is thru a series 0.001 mfd capacitor. This filters out any steady DC signal. And third 200 pfd shunted by 1 megohm is added in series with the 0.001 mfd capacitor.
.******
DC coupled square wave into scope. Peak-to-peak change is 4 major divisions. This won't change in the subsequent plots.
.******
Square wave coupled into scope thru a series 0.001 mfd capacitor. Step change is 4 major divisions. Very smooth exponential curve.
.******
Square wave coupled into scope thru a series 0.001 mfd capacitor in series with a 200 pfd in parallel with 1 megohm. Step change is 4 major divisions, but not readily visible in this plot. Here you see the effect of compounding the two different RC time constants. Inflection point is obvious.
.
- Status