Cautions injecting signal into circuit?

A 10X scope probe has a 9Meg resistor in series with the inner conductor, and it makes no sense to use such a probe for signal injection. However, there are 1X probes and one could also slip and connect the ground to something. Use a resistor in series - say 10K or larger for load isolation, and a

1000V capacitor for DC isolation. That will give reasonable isolation or protection in most cases. Am assuming you do not have 10KV or 10KW devices to test....

On Wed, 19 Nov 2003 11:10:20 -0800, Robert Baer wrote (in message ):

I'd be using a common Tek 1x probe.

Thanks.

Correct, just an audio amp with +/- 65 vdc rails. I want to inject signal to different stages to see result.

I'll rig some cap/res isolation for it.

Thanks again,

I tried using a small capacitor in series with the sig gen output. The signal went to zero.

At first I used a scope probe to connect the output to the circuit, and then thinking that it might be interfering with the signal in some way, used just a BCN connector with a length of RG 58 (I think it was). (This 1 KHz, so I don't think there is any significant cable impedance issues here.)

I 'scoped the signal before (sig gen side) the cap, and get full strength (ie, 500 mV). 'Scoped on the circuit side, zero. Disconnect the cap from the circuit, full amplitude on both sides of the cap.

Obviously the cap is blocking/dissipating the signal. I've tried small (33 pf tant(?) ) and (relatively) large (0.1 uF ceramic disc) caps, with the same results.

When I cabled the sig gen to this audio amp directly earlier at the RCA jack input, it drove the circuit with no problems.

My goal is to isolate the sig gen from the potential danger of the circuit (+/- 65 vdc) while poking around the amplifier with a probe so as to not damage the generator's output stage.

What am I doing wrong?

Thanks,

Tantalum caps are usually electrolytic, e.g., in the 0.1 to 100uF range. A

33pF capacitor is probably ceramic or mica.

Reactance is Z = 1/2*pi*C*f. At 1kHz, 33pF behaves like 4.8 Meg. So, no surprise that with a 4.8 Meg resistor in series with the oscillator, you got no signal. After all, you're probably connecting it to a circuit node with an impedance to ground of, say, 100 ohms? Of course it depends where you're connecting it - some nodes might have 10k impedance to ground. Since you're working on a power amp, there probably aren't a lot that are more than that.

A 0.1uF cap at 1kHz looks like 1.6k ohms. So, still adding a fair amount to your source impedance. What's the output impedance of the signal generator itself? Might be 50 ohms, for an RF-capable unit, or it might be more like

150 or 600 ohms for some audio-purposed signal generators. Either way, if you put 500mV at one end, and then try to drive a 100 ohm node through 1.6k, you're only going to get about 28mV at the node.

Dunno if that's what's actually going on for you; just trying to set some context.

You probably already realize this, but in case someone else is reading: remember that the power rails themselves are very low impedance to ground. You're not going to be able to inject much meaningful signal into them. It's just like trying to inject a signal into ground.

By the way, speaking of ground, are you sure that your signal generator is sharing the same ground as your circuit? It won't do to connect only the hot lead, you also need to make sure that the grounds are connected, either through the power mains or with a ground lead. It's possible that the signal generator, the amp, or both have their ground isolated from the mains ground.

Dave,

You need to inject your signal into a high impedance point in the amp. You didn't say where you are injecting it. For a "sanity check", go back to what worked originally - sig_gen===>amp at the RCA input jack. When you've verified that that setup still works, add your capacitor: sig_gen==>cap===>amp at the RCA input jack and verify that it works.

That verifies your initial setup. Then you can probe away. Anytime the signal disappears from the end of the probe farthest away from the generator, you're connected to a low impedance point. (That low impedance point could be legitimately low, or there could be a shorted component.)

With a stereo amp, you can compare results between the right channel and the left channel (that also serves as a kind of "ongoing sanity check") to find the point of failure.