30-10-2018, 08:27 PM
Hi Trevor
When you know them they are easy enough to fault find. But like with anything a intermittent fault would be a different matter.
Hi Jeffrey
Diagnosing at a distance is fraught with difficulty. What just takes a few minutes to diagnose when right in front of you can be almost impossible at a distance.
Point taken about X1 probe. Old habits die hard!
If faultfinding around a seemingly dead microcontroller etc. The first thing I would usually do is scoot around with a probe to see if there are any sings of life. I realize that it could possibly load a crystal oscillator enough to stop but to be honest that hasn't happened yet. In the case of the 27MHz clock the probe will greatly effect the wave shape(the scope bandwidth will also) and cause timing errors on the circuit under test particularly up stream. But at this stage I just want to know if the clock is present and looks reasonable. It is by no means a conclusive test, just quick. But I suppose it would be best not to proliferate bad practice.
When I was looking at the oscillograph of the bad 27 MHz clock I thought if I was less experienced and didn't know better I might be wondering is that clock normal? maybe it's frequency should change with picture content or something. So I thought I would post what a good one would look like using that same probe.
If I was looking for more accurate measurement I would use a probe that I seen described on one of the forums by you. That is a length of coax terminated with 50R and with a 1K resistor at its tip. The end of the resistor soldered to the point being tested and the braid earthed and kept as short as possible. That of course can only be used if the point being measured can tolerate a resistance of 1K between it and ground. I used that when needed but the coax that I have can be unwieldy.
Frank
When you know them they are easy enough to fault find. But like with anything a intermittent fault would be a different matter.
Hi Jeffrey
Diagnosing at a distance is fraught with difficulty. What just takes a few minutes to diagnose when right in front of you can be almost impossible at a distance.
Point taken about X1 probe. Old habits die hard!
If faultfinding around a seemingly dead microcontroller etc. The first thing I would usually do is scoot around with a probe to see if there are any sings of life. I realize that it could possibly load a crystal oscillator enough to stop but to be honest that hasn't happened yet. In the case of the 27MHz clock the probe will greatly effect the wave shape(the scope bandwidth will also) and cause timing errors on the circuit under test particularly up stream. But at this stage I just want to know if the clock is present and looks reasonable. It is by no means a conclusive test, just quick. But I suppose it would be best not to proliferate bad practice.
When I was looking at the oscillograph of the bad 27 MHz clock I thought if I was less experienced and didn't know better I might be wondering is that clock normal? maybe it's frequency should change with picture content or something. So I thought I would post what a good one would look like using that same probe.
If I was looking for more accurate measurement I would use a probe that I seen described on one of the forums by you. That is a length of coax terminated with 50R and with a 1K resistor at its tip. The end of the resistor soldered to the point being tested and the braid earthed and kept as short as possible. That of course can only be used if the point being measured can tolerate a resistance of 1K between it and ground. I used that when needed but the coax that I have can be unwieldy.
Frank