# Obtaining O-Scope signals in a AC/DC supply - grounding issues

I need to check waveforms in the power supply from this design reference by Texas Instruments.

I essentially built the supply, but it is not operating correctly - ie, no output current, but it has around 5,17V at the output.

I've been reading the document back and forth and somewhere it says, that on startup, should everything be operational, one should see a quick 3 pulse signal at the DRV pin - soft-start.

I want to check that out so that I try to figure out what is wrong with my supply.

Now comes the scary part: I can't figure out how to connect the O-scope to my circuit, as I can't figure out were to ground the O-scope's alligator lead(s). Plus, should I need to look at both the primary and secondary in the O-scope at the same time, how should I connect it to my circuit? Would I need to ground alligators from different probes in different places?

The schematic for the circuit uses the notation for ACGND (triangle) in the primary and GND (horizontal lines) in the secondary.

I haven't tried anything yet, because I've been reading around on how to safely use the O-scope (because it's not mine) and I'm feeling scared about blowing it up (or even myself! :)

Any thoughts?

Btw, this is the O-Scope: http://www.hameg.com/710.0.html?L=0

EDIT: I mistook the link for the design reference and posted the wrong link. The supply I'm working with is actually a 10W (5V@2A max) supply. With @Cornelius' help I validated that the supply is working to some extent. When connected to a LED+resistor load, the supply lights up the LED, so at least 30mA are going through it. I also tried a 4.8 Vdc motor as the load, but I didn't know the rated power for it and it didn't work - after this I realized it was a 35W rated DC motor, which explains its failure to operate.

Right now I'm trying to obtain some DC readings with a multimeter in order to validate the readings I'm obtaining rather than checking out the waveform of the PWM signal at the MOSFET's gate. I'll try to contact TI's FAE to get some more info regarding this subject.

As for the method regarding the O-scope operation, it seems that I need one of two solutions:

• either a 1:1 isolated transformer in order to isolate the ACGND and the GND
• a differential probe for the O-scope.

Still, I'm having trouble understanding the reasons behind these options, so any insight might come in handy.

• When you say no output current what is the load? – Cornelius May 15 '14 at 9:48
• @Cornelius a small 5V DC motor... – Joum May 15 '14 at 10:09
• What about a LED in series with a 220 ohm resistor? Will it light? – Cornelius May 15 '14 at 10:11
• Yes. I have one immediately before attaching other loads and it lights up properly. – Joum May 15 '14 at 10:12
• So you have at least 20 mA current. This proves that the SMPS is working. Maybe the motor is drawing too much current. – Cornelius May 15 '14 at 10:13

Here is a cheap solution for using oscilloscope when you need to measure voltages that you are unsure about grounding or magnitude of voltages. I have built several of these. The circuit has poor characteristics in all regards, but gives you a view of voltages and waveform for low frequency circuits (less than 10kHz). I don't use it for voltages expected higher than 350v.

The differential amplifier here acts as an attenuator with a divide by 100 gain (10K/1M). The ground symbol in the schematic would go to your oscilloscopes ground.

I use carbon composition resistors for the 1M resistors R1 and R2.

Also us a bypass capacitor across the batteries.

• thank you for your help! This is roughly equivalent to a differential probe, right? – Joum May 15 '14 at 13:20
• @Joum . Yes this is a poor man's differential probe. You could even build 2 of these and use as both your scope probes. – Marla May 15 '14 at 13:24
• Can you please walk me through it? As I figure it, the V_i- and V_i+ are the AC signal terminals, and V1/V2 are 9V batteries, correct? Are other values feasible? It must depend on the OpAmp being used, right? – Joum May 15 '14 at 13:30
• @Joum . Correct on all. You can use +/- 15 VDC to power the op-amp if you wish. The 1 Meg resistors give the isolation. Carbon composition was suggested because film resistors don't like large voltages across them. Any reasonable op-amp will work. – Marla May 15 '14 at 13:37
• Also, if you are going to do this, you MUST, and I cannot emphasize this too strongly, put the whole circuit in a non-conductive housing. And you MUST pay close attention to insulating your input probes or clips. Laying a circuit board with 300 volts on it on your workbench is just asking for an accidental discharge. In response to your question, it is possible to use a single battery, but then you must use a single-supply, rail-to-rail opamp. The 301 is an oldie but a goodie, and does not work at rail-to-rail. – WhatRoughBeast May 15 '14 at 20:04

To see the PWM signal generated by the IC:

• ground probe to ACGND (hot ground)
• signal probe at pin 3 of IC

That's the only thing to do to know if your circuit is working (besides checking MOSFET).

However, to see the transformer's output waveform connect probes to the output of the transformer, before the rectifier diode.

• But do you mean with the alligator probes for each channel in different places, right? Like, in the primary, between pin 3 (the MOSFET's gate) and R11's ground pin (R11 is on of the FET's source resistors. And at the secondary, other channel coupled around R12 (output resistor before the rectifier). Do you agree? – Joum May 15 '14 at 9:53
• @Joum measure one at a time (use only one channel). – Cornelius May 15 '14 at 9:57
• Well, this didn't work. The electrical switchboard in the room just disconnected... Any thoughts on what might have happened? – Joum May 15 '14 at 10:11
• @Joum have you read the link in the answer? – Cornelius May 15 '14 at 10:12
• yes, I did. But I couldn't make any sense of it. I mean, the DRV pin is driving 14V pulses at the MOSFET's gate. It is supposedly a DC signal, but yet it's connected to the ACGND via the FET's source. The question you linked mentions isolation, but ATM I don't have a 1:1 transformer or one of those costly probes. Furthermore, won't using a voltage divider at this point, change the pulse needed to drive the FET? – Joum May 15 '14 at 10:20