Debugging oscilloscope issue.
 

Looking at the last few posts this seems to be a common topic recently.

I'm working on a passive preamplifier using optocouplers that are controlled using a PWM signal from a microcontroller. As part of the project I need to analyze the PWM output before and after I run it through a low pass filter.

Recently I purchased a used JDR 3500 35mhz dual trace oscilloscope for $50. The seller claimed it was in good working order, but when I got it home and started testing things out I ran into an issue. I'm following the "setup" instructions in the owners manual and I can't get a steady trace to appear on the screen.

What is strange is that if I put it in single sweep mode and hit the trigger button it flashes the trace on the screen for a fraction of a second. But, if I put it in normal sweep mode I don't see anything. My assumption is that either the normal mode position of the switch is not working, or that something is wrong with the circuit that does continuous sweeping in normal mode.

I really need to get this project finished, in fact I we hoping to finish it before BurningAmp next weekend, so I think I'm going to spend $100 on a DSO Nano just to get moving. I would still like to get the JDR scope working at some point.

Any ideas on what else could cause this problem? The short summary is that single sweep mode seems to work properly, but in normal sweep mode the display is blank.

Thanks.

- Matt

Look for an Autotrigger mode. Most good scopes have it. Otherwise in "Norm" you need to manually adjust the Trigger level to set trigger point. Scope could have a trigger problem, of course. But if single trigger is triggering then should be able to adjust trigger control to get steady trigger. Check what your trigger source is set to. Many scopes had ch1 trigger only. But worst case "Line" trigger should display something.... just not locked to your verticle signal.
Doc

According to an ebay image I found, pull on trigger level knob to set Autotrigger. May still need to adjust trigger point.
Doc

'Normal' waits for a trigger pulse to start it up. Have you selected the proper channel as the trigger source? Have you selected the proper trigger mode AC / DC/ HF reject / LF reject ? (I don't know what your scope has as I 've worked exclusively Tektronix for the last 35 years) And after those selections are correct, the trigger level has to be set. THEN you'll see something - that may not make sense. If you're looking at a serial data stream you nay find triggering to be a problem. If there is a 'start of transmission' or something similar you could trigger on that pulse to look at the data which may then start to make sense.

'Auto' (as thaumaturge mentioned) will trigger even if there is no actual trigger pulse and show you _something_. That's usually a good place to start - and often end. I don't often use 'normal' trigger.

Learning to use a scope effectively is one of the most useful skills you'll acquire as it is the most useful tool you have.

G

Trigger circuits are funny beasts. Most I encountered in Tek and HP scopes used Tunnel diodes, which are rather fragile. The Autotrigger circuit usually fed of a small internal oscillator (not line sync'd). Line is just as name implies; a sample of line wafeform is obtained from a main transformer secondary. Selecting it makes any line related noise stand still.
I normally run in Autotrigger mode, which on the Tek scopes internally switches between "Norm" an the internal oscillator I mentioned... So trace is always visible (assuming in verticle range). I occasionally flip to Norm to select a specific trigger point. One exotic trigger option that has now become virtually worthless is TV sync modes. Those in the Tek 2400 series would let you select a specific video line. Most exotic is probably word recognizer trigger Tek offered on 246x series. That could be set to trigger on a specific logic state. Useful for finding CPU problems.
Doc

Thank you for the input. My previous testing was using autotrigger mode with both line level and input A as the trigger point, neither resulted in a trace appearing on the screen.

- Matt

Sounds like a bad trigger diode. Look for a small gold tophat diode. -Tunnel Diode- (Like regular small sig size with a metal skirt). Wonder if JDR3500 was first produced by JDR microdevices in San Jose (Cupertino)? If you could find a schizmo I could maybe point to the most likely parts.
Doc

p.s. Hey, look what I found! http://nebula.deanza.edu:16080/~noro...cilloscope.pdf

Okay, likely suspects are D1, D2 (1N4448 - Like 1N4148 only lower Vfwd) No Tunnels. If you cant find 1N4448 then go with one of new high speed shottky diodes that have ~.6 Vfwd. That HA1127 Transistor array is iffy. as is the dual FET (1a-1b)

Doc

Doc,

Wow, major thanks for the help. I will try to find time to check the diodes out tomorrow and will report back.

Thank you again.

- Matt

I got sidetracked getting ready for BurningAmp and just had time to sit down with the oscilloscope.

Using a multimeter across D1 I read .648Mohm in one direction and .602M in the other. Across D2 I read .622M and .629M.

Visually the HA1127 chip look suspect. It is a hazy gray color and has a few spots on it, whereas the rest of the chips look a normal black color. I looked up the datasheet and got the pin out for the 5 transistors. Here are resistance values between the pins.

T1 C-B = open
T1 C-E = open
T1 B-E = 5.4k

T2 C-B = 0
T2 C-E = 0
T2 B-E = 0

T3 C-B = 300
T3 C-E = 3.8k
T3 B-E = 4k

T4 C-B = 302
T4 C-E = 3.8k
T4 B-E = 4k

T5 C-B = open
T5 C-E = open
T5 B-E = 5.4k

Based on these readings T1 and T5 are the same, and T3 and T4 are the same. What is strange is that on T2 I measure a dead short in all three measurements. This, coupled with the facet that the chip looks suspicious, make me wonder if it has burned out.

Can you tell from the schematics if T2 being burned out would cause the normal sweep mode to be broken? I'm assuming there is nothing in the circuit that would be an short circuit between C/B/E of T2.

Thank you again for your help, I really appreciate it.

- Matt