How to trace a current fault.
 

Hi,

I been beating my brains out for the last few days
trying to run down a problem on a 20-year old Icom-740
transceiver with not a lot of luck. Perhaps someone here in
the group has a trick or two up their sleeve to get me out
of the hole I've been digging myself?

The situation is this -

The rig won't receive but transmits OK and there are no
other fault indications.

The apparent cause is the regulator for the 8V receive
line shutting down due to current overload (i.e. more than
100mA which is the limit of the MB3756). At present it is
running at about 3.3V with around an 80mA load (lower than
100mA but the foldback circuit on the chip is operating).

All other supplies are normal; in fact the other two
supplies out of the MB3756 are on the button and I have
changed anyway it just in case. If I isolate the line and
apply an external 8V the current goes to over 400mA in a
manner that suggests there is a semiconductor junction or
two in the fault path.

As these rigs are a mass of interconnections (the main
board alone has 25 plugs/sockets on it), normal fault-
finding procedure would be to monitor the current and yank
the plugs one by one until the excess load disappears and
then to have at it with a VOM. Unfortunately the system of
on-board links that Icom use makes this a more difficult
task than you would think. Over the last few days I've been
sure I'd found the leak several times but was badly
mistaken on each occasion. Of course, the fault may be an
above ground supply line to supply line leak which wouldn't
show up as an earth leakage path.

The tool I really need is a miniature Hall effect DC
amp-clamp but there doesn't seem to be one available. In
the past I've used a hp current tracer on PCB tracks but
even if I had one now it wouldn't be easy to use here.
However, I do have a bulky hp-1110A (2mV/mA) scope current
probe and a 10526T pulser, so tomorrow I'll try that route
with or without the power on.

Sorry about the windy posting friends but if anyone has
any idea on ways to trace wayward current leaks, please let
me know.


73 de Joe, G3LLV

Joe McElvenney wrote:
Hi,

I been beating my brains out for the last few days
trying to run down a problem on a 20-year old Icom-740
transceiver with not a lot of luck. Perhaps someone here in
the group has a trick or two up their sleeve to get me out
of the hole I've been digging myself?

The situation is this -

The rig won't receive but transmits OK and there are no
other fault indications.

The apparent cause is the regulator for the 8V receive
line shutting down due to current overload (i.e. more than
100mA which is the limit of the MB3756). At present it is
running at about 3.3V with around an 80mA load (lower than
100mA but the foldback circuit on the chip is operating).

All other supplies are normal; in fact the other two
supplies out of the MB3756 are on the button and I have
changed anyway it just in case. If I isolate the line and
apply an external 8V the current goes to over 400mA in a
manner that suggests there is a semiconductor junction or
two in the fault path.

As these rigs are a mass of interconnections (the main
board alone has 25 plugs/sockets on it), normal fault-
finding procedure would be to monitor the current and yank
the plugs one by one until the excess load disappears and
then to have at it with a VOM. Unfortunately the system of
on-board links that Icom use makes this a more difficult
task than you would think. Over the last few days I've been
sure I'd found the leak several times but was badly
mistaken on each occasion. Of course, the fault may be an
above ground supply line to supply line leak which wouldn't
show up as an earth leakage path.

The tool I really need is a miniature Hall effect DC
amp-clamp but there doesn't seem to be one available. In
the past I've used a hp current tracer on PCB tracks but
even if I had one now it wouldn't be easy to use here.
However, I do have a bulky hp-1110A (2mV/mA) scope current
probe and a 10526T pulser, so tomorrow I'll try that route
with or without the power on.

Sorry about the windy posting friends but if anyone has
any idea on ways to trace wayward current leaks, please let
me know.

I had a similar problem with a shorted supply rail in an IC-746.

Since it was more accurately a *near* short, and the voltage regulator
was pushing out quite a large current, there were actually several
millivolts all along that rail. By using the DVM on its very lowest
range, and taking the least significant digits seriously, it proved
possible to trace the voltage gradient downward to the short.

The fault is worth mentioning too. It was one of those Icom ribbon
jumpers, that I had inserted slightly off-centre in its slot. The
surprising thing was that the transceiver had worked OK for about a
month... but all that time, the ribbon had been creeping slowly
sideways... until finally it had moved far enough to cause a short.


--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)

http://www.ifwtech.co.uk/g3sek

Joe McElvenney wrote:
Hi,

I been beating my brains out for the last few days
trying to run down a problem on a 20-year old Icom-740
transceiver with not a lot of luck. Perhaps someone here in
the group has a trick or two up their sleeve to get me out
of the hole I've been digging myself?

The situation is this -

The rig won't receive but transmits OK and there are no
other fault indications.

The apparent cause is the regulator for the 8V receive
line shutting down due to current overload (i.e. more than
100mA which is the limit of the MB3756). At present it is
running at about 3.3V with around an 80mA load (lower than
100mA but the foldback circuit on the chip is operating).

All other supplies are normal; in fact the other two
supplies out of the MB3756 are on the button and I have
changed anyway it just in case. If I isolate the line and
apply an external 8V the current goes to over 400mA in a
manner that suggests there is a semiconductor junction or
two in the fault path.

As these rigs are a mass of interconnections (the main
board alone has 25 plugs/sockets on it), normal fault-
finding procedure would be to monitor the current and yank
the plugs one by one until the excess load disappears and
then to have at it with a VOM. Unfortunately the system of
on-board links that Icom use makes this a more difficult
task than you would think. Over the last few days I've been
sure I'd found the leak several times but was badly
mistaken on each occasion. Of course, the fault may be an
above ground supply line to supply line leak which wouldn't
show up as an earth leakage path.

The tool I really need is a miniature Hall effect DC
amp-clamp but there doesn't seem to be one available. In
the past I've used a hp current tracer on PCB tracks but
even if I had one now it wouldn't be easy to use here.
However, I do have a bulky hp-1110A (2mV/mA) scope current
probe and a 10526T pulser, so tomorrow I'll try that route
with or without the power on.

Sorry about the windy posting friends but if anyone has
any idea on ways to trace wayward current leaks, please let
me know.

I had a similar problem with a shorted supply rail in an IC-746.

Since it was more accurately a *near* short, and the voltage regulator
was pushing out quite a large current, there were actually several
millivolts all along that rail. By using the DVM on its very lowest
range, and taking the least significant digits seriously, it proved
possible to trace the voltage gradient downward to the short.

The fault is worth mentioning too. It was one of those Icom ribbon
jumpers, that I had inserted slightly off-centre in its slot. The
surprising thing was that the transceiver had worked OK for about a
month... but all that time, the ribbon had been creeping slowly
sideways... until finally it had moved far enough to cause a short.


--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)

http://www.ifwtech.co.uk/g3sek

You might be able to spot it by frosting each board with freeze spray
then applying power to see what's getting hot. That won't work if the
short's a really good one, and of course parts that normally dissipate a
lot will show up too. But if it's a good short, a hot PCB trace might
show up to reveal the location of the problem.

Roy Lewallen, W7EL

You might be able to spot it by frosting each board with freeze spray
then applying power to see what's getting hot. That won't work if the
short's a really good one, and of course parts that normally dissipate a
lot will show up too. But if it's a good short, a hot PCB trace might
show up to reveal the location of the problem.

Roy Lewallen, W7EL

A common way to trace problems like this is to measure voltage drop
along the power supply traces and wires. They do have resistance, and
they do show an I*R voltage drop. You need a fairly sensitive meter
range to see the effect, but it's a very useful technique. Clip one
side of the meter to the regulator output, and start probing places
that it's supposed to go. Start with things connected close to the
regulator, and move out...when the voltage drop stops increasing,
you've passed the culprit.

Another way that's a bit more extreme is to connect another supply in
parallel, with current limit set to something that won't burn up PC
traces but will make the problem easier to find. Then you can trace
the current even by looking at (or feeling) temperature rise in the
traces! ;-) Or--look for the magic smoke escaping (but I suppose
something has already lost its magic smoke anyway).

Cheers,
Tom

Joe McElvenney wrote in message ...
Hi,

I been beating my brains out for the last few days
trying to run down a problem on a 20-year old Icom-740
transceiver with not a lot of luck. Perhaps someone here in
the group has a trick or two up their sleeve to get me out
of the hole I've been digging myself?

The situation is this -

The rig won't receive but transmits OK and there are no
other fault indications.

The apparent cause is the regulator for the 8V receive
line shutting down due to current overload (i.e. more than
100mA which is the limit of the MB3756). At present it is
running at about 3.3V with around an 80mA load (lower than
100mA but the foldback circuit on the chip is operating).

All other supplies are normal; in fact the other two
supplies out of the MB3756 are on the button and I have
changed anyway it just in case. If I isolate the line and
apply an external 8V the current goes to over 400mA in a
manner that suggests there is a semiconductor junction or
two in the fault path.

As these rigs are a mass of interconnections (the main
board alone has 25 plugs/sockets on it), normal fault-
finding procedure would be to monitor the current and yank
the plugs one by one until the excess load disappears and
then to have at it with a VOM. Unfortunately the system of
on-board links that Icom use makes this a more difficult
task than you would think. Over the last few days I've been
sure I'd found the leak several times but was badly
mistaken on each occasion. Of course, the fault may be an
above ground supply line to supply line leak which wouldn't
show up as an earth leakage path.

The tool I really need is a miniature Hall effect DC
amp-clamp but there doesn't seem to be one available. In
the past I've used a hp current tracer on PCB tracks but
even if I had one now it wouldn't be easy to use here.
However, I do have a bulky hp-1110A (2mV/mA) scope current
probe and a 10526T pulser, so tomorrow I'll try that route
with or without the power on.

Sorry about the windy posting friends but if anyone has
any idea on ways to trace wayward current leaks, please let
me know.


73 de Joe, G3LLV

A common way to trace problems like this is to measure voltage drop
along the power supply traces and wires. They do have resistance, and
they do show an I*R voltage drop. You need a fairly sensitive meter
range to see the effect, but it's a very useful technique. Clip one
side of the meter to the regulator output, and start probing places
that it's supposed to go. Start with things connected close to the
regulator, and move out...when the voltage drop stops increasing,
you've passed the culprit.

Another way that's a bit more extreme is to connect another supply in
parallel, with current limit set to something that won't burn up PC
traces but will make the problem easier to find. Then you can trace
the current even by looking at (or feeling) temperature rise in the
traces! ;-) Or--look for the magic smoke escaping (but I suppose
something has already lost its magic smoke anyway).

Cheers,
Tom

Joe McElvenney wrote in message ...
Hi,

I been beating my brains out for the last few days
trying to run down a problem on a 20-year old Icom-740
transceiver with not a lot of luck. Perhaps someone here in
the group has a trick or two up their sleeve to get me out
of the hole I've been digging myself?

The situation is this -

The rig won't receive but transmits OK and there are no
other fault indications.

The apparent cause is the regulator for the 8V receive
line shutting down due to current overload (i.e. more than
100mA which is the limit of the MB3756). At present it is
running at about 3.3V with around an 80mA load (lower than
100mA but the foldback circuit on the chip is operating).

All other supplies are normal; in fact the other two
supplies out of the MB3756 are on the button and I have
changed anyway it just in case. If I isolate the line and
apply an external 8V the current goes to over 400mA in a
manner that suggests there is a semiconductor junction or
two in the fault path.

As these rigs are a mass of interconnections (the main
board alone has 25 plugs/sockets on it), normal fault-
finding procedure would be to monitor the current and yank
the plugs one by one until the excess load disappears and
then to have at it with a VOM. Unfortunately the system of
on-board links that Icom use makes this a more difficult
task than you would think. Over the last few days I've been
sure I'd found the leak several times but was badly
mistaken on each occasion. Of course, the fault may be an
above ground supply line to supply line leak which wouldn't
show up as an earth leakage path.

The tool I really need is a miniature Hall effect DC
amp-clamp but there doesn't seem to be one available. In
the past I've used a hp current tracer on PCB tracks but
even if I had one now it wouldn't be easy to use here.
However, I do have a bulky hp-1110A (2mV/mA) scope current
probe and a 10526T pulser, so tomorrow I'll try that route
with or without the power on.

Sorry about the windy posting friends but if anyone has
any idea on ways to trace wayward current leaks, please let
me know.


73 de Joe, G3LLV

In article ,
Joe McElvenney wrote:

Hi,

I been beating my brains out for the last few days
trying to run down a problem on a 20-year old Icom-740
transceiver with not a lot of luck. Perhaps someone here in
the group has a trick or two up their sleeve to get me out
of the hole I've been digging myself?

The situation is this -


snipped for brevity


Sorry about the windy posting friends but if anyone has
any idea on ways to trace wayward current leaks, please let
me know.


73 de Joe, G3LLV



Look for a shorted bypass Cap along the supply rail. These
small tantilum caps are notorious for shorting out, and giving
your symptoms. Just follow the rail on the schematic and isolate
each board one at a time untill you isolate which board is drawing
down the supply. Then look for the bypass caps on that board, for
any discolor or sign of excess heating.


Bruce in alaska
--
add a 2 before @

In article ,
Joe McElvenney wrote:

Hi,

I been beating my brains out for the last few days
trying to run down a problem on a 20-year old Icom-740
transceiver with not a lot of luck. Perhaps someone here in
the group has a trick or two up their sleeve to get me out
of the hole I've been digging myself?

The situation is this -


snipped for brevity


Sorry about the windy posting friends but if anyone has
any idea on ways to trace wayward current leaks, please let
me know.


73 de Joe, G3LLV



Look for a shorted bypass Cap along the supply rail. These
small tantilum caps are notorious for shorting out, and giving
your symptoms. Just follow the rail on the schematic and isolate
each board one at a time untill you isolate which board is drawing
down the supply. Then look for the bypass caps on that board, for
any discolor or sign of excess heating.


Bruce in alaska
--
add a 2 before @

Hi again,

Thank you for your useful comments which were all duly
noted. However I have been able to clear the problem after
a deal of head-scratching. It became likely that the leak
was between supply lines and not simply to ground when the
PA cooling fan started-up on receive.

The fault turned out to be a s/c between the tops of two
zero-ohm resistors (links) on the IF board just a little
north-east of the speaker plug. As I was the last person to
plug and unplug that particular lead, I guess that I must
be the guilty party.


Thanks again - Joe