How to measure B+ Ripple with a Scope?

[jjasilli]

In a nearby thread, sluckey mentioned using a scope to measure B+ ripple.  Nominal B+ may exceed the input voltage of a scope; and it may be significantly higher under signal or fault conditions. Any suggestions on how to best probe B+?


*  Is it advisable to use a 10x or 100x probe, even if the point of measurement exceeds that voltage limit is the scope's manual?


*  Is it better to build an outboard attenuator.  In an old thread it seems PRR recommended building an outboard attenuator to safely probe B+, but the file attachment is gone:  http://el34world.com/Forum/index.php?topic=3974.0  Can someone resurrect this?


There's scant reliable info on the web.

[shooter]

http://how-to.wikia.com/wiki/How_to_make_a_100X_oscilloscope_probe

the one I built used 10 resistors of equal value, and you tapped off the 9th one, stuck the sting inside a ball point pens guts with some quickset epoxy.

[jjasilli]

Ok, something like this for under $20 should be OK:  http://www.ebay.com/itm/US-STOCK-100MHz-T3100-Oscilloscope-High-Voltage-Clip-Passive-Probe-Hantek-/292146318063?_trksid=p2349526.m2548.l4275


Seems idiot-proof enough, which is what I need.  And, say, 2V of ripple > 20mV should still be readable.


Am I on the right track?

[pompeiisneaks]

Here's some documentation from a 10x probe:

http://tmi.yokogawa.com/us/products/oscilloscopes/voltage-probes/700929-101-probe/

says it can take up to 1000V Cat II with a 10Mohm resistor in it. 

Here's a 100:1 that also only does up to 1000V and has a 100Mohm resistor

http://www.tequipment.net/Hioki/9666/Oscilloscope-High-Voltage-Probes/?Source=googleshopping&gclid=Cj0KCQjwwevLBRCGARIsAKnAJvc4MMiXWcuzKi9U8STWGqn_nvDzfm8lP4-i1B_KAVe14SzcoJa8pUMaAvlHEALw_wcB 

(specs are a bit lower). 

Everything I understand is that the scope probe protects the input by guaranteeing the max voltage that gets to the input is that order of magnitude less.  1000v into a 10x is 100v, into a 100x is 10v. 

I too would like to have someone, either PRR or Sluckey I don't recall which, let me know if I'm right here, I swear it was one of them that told me this in a thread a long time ago

Edit: btw I just saw your link, I bought that exact same one, two actually, but I think they feel somewhat crappy.  I still use them at times, but my scope probes that came with my rigol seem a lot better, and at 10X I've got the 1000V protection so I use those.

~Phil

[shooter]

On my scope I can tell it what probe, 1, 10, 100, and it scales everything for me, otherwise just do the math, if it says 1v with a 10X it's 10v you're measuring.
You do loose sensitivity I believe, so trying to measure 20mV with a 100X probe might not be the best solution.

[jjasilli]

@pompeiisneaks.  Your math is right; but there's no guarantee.  If you probe a voltage that exceeds your scope's limit, there is the possibility of damage if something goes awry.  A mistake or an un-noticed fault condition is tempting Murphy's Law.


@shooter.  10x probes good for 1000V, or more, are expensive. 100x hi voltage probes can be inexpensive, but over attenuate small voltages as you say.


I'm tempted to build an outboard voltage divider in a box, and probe that, per the attached schematic.


Any comments appreciated.

[sluckey]

Quit trying to build a divider. Your scope can handle the amplitude of the AC ripple even with a 1X probe. All you gotta do is block the 400VDC that the ripple is riding on. A .1µF/630V cap in series with the probe will do that. Same applies when viewing a signal on a tube plate. So just get a cap and a couple gator clip leads. Or you could just use a RG-58 BNC cable. Cut one connector off and put a cap with gator clip on that end of the cable.

[shooter]

outboard voltage divider in a box,

that's basically all I did for mine til I found my lost 100X. 

this is my "standard" ;
virtually all DC, including ripple AC I use my fluke 87.
Huge DC >750V I use a scope, with appropriate probes

Virtually all AC, or digital tracing I use a scope.

anybody that has worked professionally where meters and scopes are must-haves will have their own "standard", and the cool thing about humans, they all get the job gone, the odd thing about humans is they want their standard, standard

[jjasilli]

Quit trying to build a divider. Your scope can handle the amplitude of the AC ripple even with a 1X probe. All you gotta do is block the 400VDC that the ripple is riding on. A .1µF/630V cap in series with the probe will do that. Same applies when viewing a signal on a tube plate. So just get a cap and a couple gator clip leads. Or you could just use a RG-58 BNC cable. Cut one connector off and put a cap with gator clip on that end of the cable.

Great, thanks!

[shooter]

All you gotta do is block the 400VDC

Like Steve says; you're measuring AC at a DC level, my scope has "ac coupled" setting with each channel.  Since I only use it for AC (mostly) it's always set to AC coupled.  That can be a "rabbit hole" if you're measuring the signal path trying to find leaking DC, if you're ac coupled it won't show the dc you're trying to find.  Time and practice works well for almost everything

[jjasilli]

Agreed.


However, I'm hammering away at a simple point which does not seem to be acknowledged:  It is not best practice to touch a probe to a point of measurement where the voltage exceeds the scope's rated input voltage.  It specifically says so in my scope's manual. I know there are ways to get around that.  But I don't want to get around it; I just don't want to do it.  Sluckey's method solves the problem: with a cap in series with the probe tip, the voltage at the point of measurement is now well within the scope's rating.

[sluckey]

It would all be so simple if the scope manufacturers would just use internal 1000V caps when AC Coupling is selected.

This was not a problem back in the days of tube scopes like Tektronix 545. 

[PRR]

> All you gotta do is block the 400VDC....

I'll disagree. The initial voltage on the 0.1u cap is zero. When one end is touched to 400V, the other end rises to 400V. Yes, this decays as 0.1uFd+1Meg... but it will be high for many milliseconds, enough to zap the probe/'scope input caps.

Idiot Assistant's figuring below. At first no connection (Mega-MegaOhm), then at 1mS connection is made. (For some reason I tole it 600V, but the results scale for any voltage.)

[PRR]

Most 'scopes read mV as well as V. A 1Meg+1K divider will divide 1000:1, making any stage-amp's voltages safe.

[jjasilli]

I'll build it!

[shooter]

like Tektronix 545

but I have a bad back

2PRR; (at PRR seems to presumptuous )
Is the right hand R222 & C73 the actual probe, scope internals or just part of the ext box?

[sluckey]

I guess the cap is not such a good idea.

[jjasilli]

@shooter:  Is the right hand R222 & C73 the actual probe, scope internals or just part of the ext box?


The Voltage Out of the attenuator cannot be part of the inner circuit of the scope or of the probe.

[shooter]

the attenuator

I was looking at the 1M 1K as the attenuation "feeding" the probe/scope internals

[jjasilli]

You are right that R223 1M & R224 1K are the main components of the voltage divider circuit.  But the circuit is more complex.  I suspect R222 & C73 are added to "stabalize" the circuit.


R222 1M is in parallel with R224; per the Rule of Tens, R222 does not materially change the value of R224. 




Note that the ground connection is essential.  Without it there is no voltage attenuation.  IMHO the attenuator ground should be connected to chassis ground of the amp under test, to avoid different ground potentials.

[PRR]

A typical 'scope probe is 1Meg+30pFd, or something like that. Those simulated parts "are the probe".

It hardly matters if the probe specs are different (except the less-common case of a 75 Ohm probe used only in RF work).

You add a 1Meg and a 1K.

[shooter]

A typical 'scope probe

Thanks PRR, that's sorta what I thought