Question about 50 ohm scope inputs

[allanw]

Can someone explain this statement?

Page 8 of http://s.eeweb.com/articles/2010/4/17/power-supply-noise-measurement-1290016135.pdf

"Most power supplies will easily drive a 50-ohm load. A 50-ohm load will only draw 3 mA from a 1.5-V power supply. This allows you to connect a 50-ohm coax cable directly to your power supply and use a scope with a 50-ohm input for maximum sensitivity, instead of using a 10:1 probe."

Is the 3mA a typo or is there a 10:1 divider somewhere? Doesn't make much sense...

[bfritz]

Yes, there is a divider... the 10:1 probe!

[amspire]

It is a mistake. It should be 30mA, not 3mA.

To me, is sounds like a really stupid idea. The 50 ohms inputs on scopes is for low voltage - perhaps 5V maximum. It is really easy to damage.

If you have a 10:1 probe, use it, unless there is a reason you cannot use it.

Richard.

[allanw]

Yes, most scopes spec it to be 5Vrms. Some Tek scopes say +/- 24V peak.

Anyway, I think I will be inserting a DC blocking capacitor in series when doing my power supply measurement in order to not load it down. My measurements will probably be in the range 1 MHz - a few GHz so I think I should be able to find a BNC adapter that includes one.

[amspire]

Yes, most scopes spec it to be 5Vrms. Some Tek scopes say +/- 24V peak.

Anyway, I think I will be inserting a DC blocking capacitor in series when doing my power supply measurement in order to not load it down. My measurements will probably be in the range 1 MHz - a few GHz so I think I should be able to find a BNC adapter that includes one.

The blocking capacitor sounds reasonable as long as there is no chance of the power supplies becoming unstable resulting in a large amplitude AC. If I was doing this test regularly and I didn't need the scope's full sensitivity, I would add a 450 ohm SMD at the start of the coax to make a 10:1 divider.

[vk6zgo]

Can someone explain this statement?

Page 8 of http://s.eeweb.com/articles/2010/4/17/power-supply-noise-measurement-1290016135.pdf

"Most power supplies will easily drive a 50-ohm load. A 50-ohm load will only draw 3 mA from a 1.5-V power supply. This allows you to connect a 50-ohm coax cable directly to your power supply and use a scope with a 50-ohm input for maximum sensitivity, instead of using a 10:1 probe."

Is the 3mA a typo or is there a 10:1 divider somewhere? Doesn't make much sense...

The 3mA quoted is either a typo or a miscalculation.In any case,it looks like Agilent are "losing it".

Things aren't that bad,though!
Even with 30mA,  I^2 R only gives us 45mW,which should be well within the capabilities of a 50 Ohm termination,internal or external.

In the case of external terminations,many of these are rated at around  2 watts which is around 200mA through,or 10volts across 50 Ohms.

The danger in the quoted values of 1.5v & 3mA is that people might extrapolate from that to think that a 15v supply would still be useable.

Why would anyone use 1.5 v as an example,anyway?

[allanw]

The danger in the quoted value of 3mA is that people might extrapolate from that to think that a 15v supply would still be useable.

Why would anyone use 1.5 v as an example,anyway?

You'd only use 50 ohm probing for high frequency signals above 500MHz, and this is in the context of measuring power supply noise. Only modern high-speed devices will care about power supply noise at those frequencies, and they'll have supply voltages in the 1V range.

Anyway... sent off a technical support request to Agilent asking for clarification. I'll see what they say.

[vk6zgo]

Using your Oscilloscope high Z input with a BNC tee,you could use any size of 50 Ohm RF test load.
The only limit then is how high a voltage you can place on your 'scope's input,& the current carrying capacity of your coax!

Of course,the adaptors  & things would affect the  response at 500MHz!

I'm a bit dubious about the significance of 50 Ohms in this application,as it assumes that the source impedance of the noise component
is 50 Ohms,or at least,very low.
The latter may be the case with the DC output,but is it the same for the noise?

If we can assume a matched 50 Ohm system for the noise,a better scheme would be to use an RF coupler of known characteristics & a Spectrum Analyser.
In this case,the termination could be of any appropriate power rating.

Hell,what do I know? I'm only an RF Technician!

[amspire]

I think this article was describing how to use an an Agilent DSO80000 series scope (40GS/s) to measure noise on the power rails of a high frequency PCB circuit.

This oscilloscope family has 50 ohms/5V inputs only, and I think the DC offset range is +/- 1.5 volts which is probably why they picked the example of a 1.5V supply. No AC coupling built in the scope.

These are very expensive scopes, and it still seems stupid to be feeding power supply voltages straight into the input. I suppose that the Agilent service department wouldn't complain with the practice. The odd $3000+ repair job everytime the probes touch the wrong points on the board is a very nice earner.

If you have a 1Mohm input, then I agree with vk6zgo - use it with a BNC tee and an external 50 ohm load.

Richard

[vk6zgo]

If it is only to be used in a 1.5v circuit,I would think that the  internal termination would have no problems with 45mW,which is about
+16.5 dBm.(OK,it's not AC,but dBms are calculated on RMS,so it is still a fair comparison).
The internal terminations of most Spectrum analysers can handle +30dBm (1 watt),so probably it would be similar on the 'scope.

Where the danger lies,is,as Richard implied,when there are multiple supplies of different voltages.
"One slip & the termination is toast".

The whole application note seems sloppy,compared to the HP & Tektronix ones of yesteryear.