BNC Cable as 1X Probe

[alank2]

The low frequency compensation is wrong.  This is how 5 volt logic circuits appear to put out 10 volts.

Does this mean that the 7.5pF cap I am using is too big?  A smaller cap would yield a larger reactance which would yield more attenuation, do I have that right?

[David Hess]

The low frequency compensation is wrong.  This is how 5 volt logic circuits appear to put out 10 volts.

Does this mean that the 7.5pF cap I am using is too big?  A smaller cap would yield a larger reactance which would yield more attenuation, do I have that right?

Yes.  The low frequency compensation is done with a low frequency square wave, like 1 kHz, and if it is off, then high frequency square or sine or other waves will have the wrong amplitude but otherwise could look good.

[alank2]

Yes.  The low frequency compensation is done with a low frequency square wave, like 1 kHz, and if it is off, then high frequency square or sine or other waves will have the wrong amplitude but otherwise could look good.

Rigol RP1300 compensation voltage (amp) - 2.997V
My probe compensation voltage (amp) - 2.974V

It is a bit too small at 1kHz, but too large at 40MHz...

[alank2]

So, for a probe that has an adjustment in the BNC connector, is there ANY benefit to having an adjustment at the probe tip as well?  Will having both allow for more compensation/calibration that just using a fixed one at the probe tip?

[T3sl4co1l]

If it's too large (or small) past a certain frequency, simply add another R+C or R||C network to adjust it.

Don't be surprised if, for three digits accuracy, you need a huge pile of them...

Tim

[David Hess]

So, for a probe that has an adjustment in the BNC connector, is there ANY benefit to having an adjustment at the probe tip as well?  Will having both allow for more compensation/calibration that just using a fixed one at the probe tip?

In practice it is better to minimize the probe tip capacitance and move the compensation to the BNC connector but both designs work.  I have never seen one with low frequency compensation at both ends but it is common to have the low frequency compensation at the probe and the high frequency compensation at the BNC.

[Phaedrus]

So we got our new cables with 50 ohm terminators. Just some preliminary testing without the 450 ohm series resistor gave us this result (Ch1 is old 75ohm cable, Ch2 is new cable)





So we've gotten rid of a lot of the high frequency crap that was throwing off our measurements before, probably just from eliminating reflection. The result overall looks very good! The only thing strange is I was expecting more DC attenuation from the 50 ohm terminator; but aside from the high frequency noise the signal amplitude looks pretty much the same. Is the 450 ohm series resistor strictly necessary? I'm going to have to attenuate the +12V signal regardless, due to the terminator's 2W power rating, but if I could skip it on our +5V and +3.3V signals that would be great.

EDIT: Actually I'm going to have to attenuate all of them anyway or the power consumption will throw off our efficiency measurements. Still would like to know for the record anyway.

EDIT 2: Also, if it isn't clear we're using a 50 ohm pass thru terminator, then 1M termination on the scope. The reasoning being that if the technician accidentally applies a large DC voltage, or if DUT isolation breaks down, we'd rather lose a $25 terminator than a channel on our $3000 scope.


EDIT 3:

Whipped up a perfboard prototype of the test point with 450 ohm resistor. Results exceed expectations, though I'm not sure if my amplitude is strictly accurate.

Ch1 = 50 ohm terminated cable with 450 ohm series resistor, X10 attenuation
Ch2 = 50 ohm terminated cable, X1 attenuation
Ch3 = 75 ohm unterminated cable, X1 attenuation

[T3sl4co1l]

EDIT 2: Also, if it isn't clear we're using a 50 ohm pass thru terminator, then 1M termination on the scope. The reasoning being that if the technician accidentally applies a large DC voltage, or if DUT isolation breaks down, we'd rather lose a $25 terminator than a channel on our $3000 scope.

A few weeks ago, I hooked up, what was it, I think it was a current limited 150V supply, essentially -- the probed signal was at supply voltage -- to my terminated scope input.  [Beeee], popup on the screen reads: the terminator was disconnected because it was overheating.     Well, sometimes I'm dumb, but at least Tek still had some good ideas in that age (TDS460 -- one of the first series scopes, during their "good instrument / bad interface / no service" early decline).

Did they drop that feature?  Does anyone know if Rigols monitor this?

Tim

[Phaedrus]

There's two things I never want to stake the safety of my life or property on: firearm mechanical safeties, and equipment thermal overload protection.

[alank2]

So my tiny pcb came in, it is a 4mm x 20mm little adapter that has the coax at one side and pigtails at the other.  It has the 9M/7.5pF cap in parallel and the 100R in series with them.

I could not compensate properly with the scope compensation with only the 7.5pF so I took it off and put an adjustable 5-20pF cap on it.  I can compensate at the 3kHz, but at higher frequencies it is not attenuated enough.  I am looking at a 50MHz square wave at 5V and my Rigol probe shows an amplitude of about 4.6V.  My home built probe with adapter is about 5.7V.  I can adjust the 5-20pF cap on it and the cap in the bnc connector, but they do not adjust for the voltage being high.  Do I need to adjust the 100R?

Thanks,

Alan

[David Hess]

A few weeks ago, I hooked up, what was it, I think it was a current limited 150V supply, essentially -- the probed signal was at supply voltage -- to my terminated scope input.  [Beeee], popup on the screen reads: the terminator was disconnected because it was overheating.     Well, sometimes I'm dumb, but at least Tek still had some good ideas in that age (TDS460 -- one of the first series scopes, during their "good instrument / bad interface / no service" early decline).

Did they drop that feature?  Does anyone know if Rigols monitor this?

This feature exists in many of their oscilloscopes predating the TDS460 and as far as I know, they have kept it in every oscilloscope made since then.  It is trivial to implement in a DSO.

There is also a way to implement instantaneous high bandwidth protection against overload using a current driven diode bridge which Tektronix used in the 7A29 vertical amplifier.

I do not see anything in the Rigol documentation about overload protection of the 50 ohm termination.

[c4757p]

Wait, some scopes don't have overload protection on the termination? What the fuck?

My only scopes with internal termination were Tek 2400-series... Not only do they have it, they make you test it as part of the calibration procedure - something like "at W voltage, must not trip for X seconds, at Y voltage, must trip within Z seconds".

[David Hess]

Wait, some scopes don't have overload protection on the termination? What the fuck?

Maybe it is considered a feature by the manufacturer in the sense of forcing payed replacements or repairs.  Occam's Razor implies that they are just being cheap although it only requires firmware on a DSO to implement it.  I would be happy enough if they just documented which part to replace and made it easy to do.

My only scopes with internal termination were Tek 2400-series... Not only do they have it, they make you test it as part of the calibration procedure - something like "at W voltage, must not trip for X seconds, at Y voltage, must trip within Z seconds".

On old Tektronix analog oscilloscopes and DSOs which support 50 ohm internal terminations, I have always been surprised at how much extra effort they went to in protecting them from overload.  They used several different custom ASICs for just this purpose.