LeCroy 9450 Digital Oscilloscope
The scope has a fault that's
eluded other repairers so I turned it on to see if I could figure
out what it might be. I saw what appears to be normal working
of the various controls, as these are echoed back by the various
microprocessors to tiny LEDs, but absolutely no display. My guess
is the EHT PSU has failed, whether within its circuit board or
because it's missing a power supply feed remains to be seen.
Here's the circuit diagram
and below with less definition.. I immediately see that it's
nothing too difficult to understand. In TV receivers a stock
fault was failure of the tripler which is shown here as a BG1895
which with a fair wind I might find in my stock of spares..The
PSU is driven by an LM3524 and I notice there's an on/off switch
for the EHT built into the circuitry and of course if there's
a low voltage rail missing that would be a simple fix. Dismantling
the scope looks messy but I might be able to make some measurements
with the top cover removed? The 688 volt output is also a possible
reason for no picture as this will supply the CRT electrodes.
That feedback potentiometer comprising a 33M and 33K resistors
is interesting and must be for regulation of the EHT. I understand
the EHT is 11KV so the tripler input must be about 3,700 volts
giving 3.7 volts feedback so maybe I should check the voltage
at Pin 1 of the LM3524 if it's accessible rather than chance
getting a shock from the CRT anode?
A view of the channel
processing boards which plug into the lower motherboard. The
slot to left centre is for a heavy metal bar which is screwed
to the top of the power supply.
All the processing hardware
dates of course to the period when the oscilloscope was designed
which will be the mid-1980s. Most of the components will still
be available, but I did notice a few custom chips which will
be difficult to obtain plus lots of EPROMs carrying code that
might be awkward to reproduce although I have seen evidence of
Le Croy enthusiasts on the Internet from whom this might be available?
I was a bit surprised
once I'd detached the top cover and removed the power supply
module which is fitted over the motherboard. At first sight it
looked complicated to work on, but in fact it wasn't too bad.
Remove about a dozen screws and the power supply can be detached
. It carries three cables: a short multi-way cable to the motherboard,
a three-way cable (3 black wires terminated in a brown plug)
to the display board and a third to the fan (at the rear of the
case). To prevent a serious accident the PSU can be rested on
a metal bracket positioned to protect the neck of the CRT. As
the display board was of interest I removed this next. It's screwed
to the side of the chassis frame and cables connect to the CRT
base and the scan coils (green + yellow and red + blue) as well
as the CRT final anode (thicker red).
As I was removing the
PSU I'd noticed a couple of fuses. Both were rated 2A and when
I checked found that both were badly blown. The fuses protect
the plus/minus 15 volts supply for the display board so it would
appear I was correct in my assumption that this particular board
was faulty. The display board (below) carries a fair amount of
microprocessor circuitry at one end where a standard mult-way
connector mates with the motherboard. At the front end is the
EHT circuitry with the high voltage circuit contained in a black
metal box (below..with yellow label and red lettering). I tested
various points with my multimeter and soon found a dead short
between the minus 15 volt socket to ground. I then checked for
other shorts and was surprised to find a whole pile of transistors
carrying shorts between their middle and the pin at one end.
I need to re-examine these shorts later because I also found
a dead short at the main smoothing capacitor from the minus 15
volt supply to chassis. I wasn't inclined to start removing components
one by one because this can risk track damage so I decided on
another approach. As the minus 15 volt rail can carry up to 15
volts it can certainly carry say one volt. I connected my bench
power supply set to 0.5 volt and 1 amp and connected it across
the smoothing capacitor without anything of note. I then increased
the current trip to the maximum of 3 Amps and scanned the board
for any warmth. As I again found nothing of note I increased
the voltage to one volt so that (somewhere on the board) the
dissipation was 3 watts, but still no heat was apparent from
any area of the board which I found strange so I decided to detach
the black metal case from the EHT circuitry and see if anything
inside the box was getting hot.
The box is held in place by
4 screws that mate with a metal plate at the underside of the
board. This done, I again turned on the power but noticed zero
current drain. I cranked up the voltage a little and a few milliamps
flowed. Clearly the short-circuit had disappeared so I sat down
to ponder what I'd discovered. I then noticed a label on the
black metal cover... "Repair 5129" with the figure
1 in continental style like a 7. I thought that as the short
was now absent I perhaps could reassemble the scope and power
it up. Perhaps the short had eventually fused open with the 3
amp current? As I refitted the black metal box I noticed something
odd. The metal baseplate wasn't symetrical. One edge was 50%
wider than the other. The copper land on the circuit board was
also non-symetrical but when I fitted the base correctly I noticed
that three transistors whose resistances I'd tested previously
were now out of sight underneath the cover. It looked highly
likely that the base had been shorting to a section of track
even though most of the tracks are coated to prevent solder adherence.
However, a short might have taken place over a long period from
metal pressing on the tracks especially as the baseplate edges
are bare machined aluminium. Maybe the guy that carried out Repair
5129 was responsible?
I carried on reassembly and
then discovered the display board had been easy to detach because
a fixing screw at the lower edge hadn't been fitted. Was this
Repairer 5129 again? I then noticed a tiny grey-coloured 2Mohm
pot located on the upper edge of the display board was wobbly.
I tested the connections and found only one pin of the pot was
connected... two pins had broken off. I temporarily fitted a
2.7Mohm pot (dark yellow.. below) and completed reassembly. Alas...
still no picture.
I now wonder if someone has
attempted a repair and given up and, in reassembling the scope
made those errors I discovered? In which case, I'm now looking
at the original fault.... I'm now wondering if the first repair
ie Repair 5129 was to fit a new tripler, because the solder joints
look different to the rest, and the careless reassembly was due
to more recent activity? Maybe the tripler is a stock fault so
my best bet is to remove the tripler and test it, or at least
just disconnect it and check the board on the bench?
I found the display board
can be removed by just detaching cables and a few screws then
pulling it out... no need to take out the PSU. As long as the
EHT circuit isn't operational... no problem (by which I mean
no shock from the CRT connector!). What next? Well the 15 volt
rails are now present on the board, and not merely at the brown
plug, but clearly the EHT power supply isn't working so I studied
the circuit diagram and quickly discovered extra transistors
to those shown in the circuit diagram. The diagrams must be for
a later version as, for example, I can see Q4 but the diagram
does not have such a transistor ***. I powered the board from
a couple of bench supplies and noted the control chip, an LA3520
was in fact an SG3520, but nevertheless an identical product.
It was certainly oscillating (waveforms at pins 3 & 7) but
nothing could be seen at its output (pins 11 & 14). As I
had a new spare I fitted this but to no avail... exactly the
same problem.. The EHT circuit consists of an oscillator running
at around 40KHz which feeds an amplifier with a step up transformer
at its output. This is used to generate around 650 volts for
the CRT electrodes and 60 volts for the odd scan circuit. The
transformer also supplies about 3.5KV which feeds a tripler for
the CRT final anode. A resistor potentiometer across the 3.5KV
output feeds a differential amplifier in the SG3520 "chopper"
chip with around 5 volts which is compared with the voltage at
a preset pot to drive the output of the SG3520. A pin at this
chip carries the output of the differential amplifier to a smoothing
circuit used to damp feedback. The method of operation is feedback
from an output voltage changes the waveform of the 40KHz oscillator
(pulse width modulation). By this technique the monitored voltage
can be kept to exactly the right value. In this case the feedback
will ensure the voltages to the CRT will be maintained to precisely
the correct values and brightness settings will not vary inadvertently.
In summary...I could see oscillation
taking place at the chopper chip and the differential amplifier
inputs at pins 1 & 2 seemed to be fed correctly but there
was nothing at the SG3520 output pins 11 and 14. These outputs
are via open collector/open emitter transistors, used in this
scope as emitter followers driving a grounded base transistor
Q8. This in turn feeds an amplifier connected to the power transistor
Q35 feeding the transformer. I'd disconnected the tripler, then
removed and tested several key transistors, all of which tested
OK. This leaves three possibilities... first an open circuit
track and second a short circuit perhaps at the pin used for
smoothing the output from the differential amplifier, or let's
not forget... something I'd missed..
*** I found two transistors
inc Q4 used as reference diodes with pins shorted that may no
longer be fitted because transistors were closer in spec to earlier
Something I'd missed...
everything had tested OK except I found the inputs to the differential
amplifier were set to switch off the oscillator output
gates within the SG3524 chip. When I temporarily grounded the
EHT potentiometer output (sitting at about 5 volts) the minus
15 volt currents tripped the external PSU which was set to only
20mA. Marked on the circuit is HVREGOFF which I'd assumed
merely stopped high voltage stabilising but in fact acted as
a shutdown signal. The HVREGOFF (also called HVOFF elsewhere
which is a lot more meaningful) is generated by a set of operational
amplifiers wired in an OR network. The network is arranged to
turn off the EHT if any one of a number of faults arises. First,
if the plus/minus 15 volt supplies wander too far from being
approximately equal, second, if the thermal switch operates,
and third if the on-board Vcc isn't present or falls too low.
Essentially the OR network is arranged to check that power supply
voltages are correct. So, for example, if the plus 15 volt supply
drops to say 12 volts the CRT will turn off. Clearly, with the
board removed from the scope and checked on the bench I didn't
bother arranging a 5 volt supply as it's not required for the
basic display so, no Vcc present.. no EHT.
As with all complicated equipments
it takes ages before you can find your way around the circuitry.
Now that I've investigated the display board and carried out
various tests and seen the EHT shutdown circuit, I re-measured
the PSU outputs which I must admit was one of the first things
I'd done. I'd noticed the voltages were not particularly accurate
but didn't pay much attention to the discrepancies (at that time
both display board fuses had been open!). Now, however, after
re-reading the repair manual the minus 15 volt supply
at 13.25 volts does seem too low, particularly as it was
checked off-load (ie. fuses open). The plus 14.54 looked better.
The spec gives the outputs of the PSU as +/-1% of 15 volts which
is 15.15 volts to 14.85 volts and for the neg 15 volts it says
+/-1% of -15.4 volts which works out as -15.19 volts to -14.89
volts. The measured values are actually -3% and -12% off load.
Maybe the low value of the neg 15 supply is the reason for the
lack of EHT? Time to remove the PSU and check its capacitors
(capacitors have a lifetime measured in thousands of hours versus
ambient temperature and the scope is 30 years old!).
The PSU has a circuit
board with a pair of large heatsinks, one of which mounts four
pairs of power transistors and the second four high power bridge
rectifiers. I tested all the electrolytic capacitors and all
were in perfect shape with very low ESR values and accurate values,
so I'll look further. The chopper circuits, of which there are
four , are controlled by a small subsidiary mains supply and
a small plug-in circuit board (seen in the picture opposite)
carrying four chopper chips type UC3526AN and including for each,
two potentiometers... one multi-turn pot appears to provide an
overvoltage setting and the other accessed from the side allows
control of the output voltage. This pot is very sensitive and
allows a very high range of adjustment so take extreme care if
one is adjusted whilst the PSU is connected to the scope. With
the PSU on the bench with no load I adjusted both the 15 volt
outputs to 15.15 volts. The 5 volt outputs were fine. Presumably
it has been merely a case of component aging that had resulted
in the low voltages?
Now to ressemble everything
and see if there's a display....
After reassembling everything
I switched on and after a worrying warm-up period of around 10
seconds a nice orange display appeared and all the functions
appeared to work. That dark yellow circular component top right
is the temporary potentiometer that I fitted to replace the broken
one. It sets the CRT general brightness level. There are some
slight problems however.... the traces are a bit wavy as if there's
ripple on one of the power supplies or pickup within the circuitry.
I haven't yet fitted the aluminium
bar which holds the PSU clear of the bracket protecting the CRT
and that might be inducing magnetic interference affecting the
Below you can see the connections
to the CRT. Scan coils are the red/blue and green/yellow leads,
Centre is the power lead carrying the plus/minus 15 volt supplies
with centre wire ground, the thick red cable carries EHT and
a horizontal plug carrying the connections to the CRT base. Top-centre
is a Klixon 1822F02-2 thermostatic switch which monitors the
temperature of the scan transistors and if this is too high the
switch turns on and switches off various sections of circuitry
via the OR network of op amps mentioned previously that monitor
When I hooked one probe to the
front panel test point and randomly fiddled with lots of front
panel controls both channels displayed a square wave so I guess
I need to read the operators manual, if I can find one, as this
looks like there's a function "Channel 1 plus Channel 2"
The date shown on the screen is 31st
December 2001 which might be when the scope left the factory.
This error will be due to a flat battery as 30 years is too much
to expect from a small lithium battery so I'll need to discover
on which board this is located before I can fit a new one. In
fact, studying the repair manual revealed the lithium cell is
located on the processor board. If this is a 3-volt type I can
fit a new wire-ended AA size spare marked 3.6 volts that I have
in stock. Although nothing like the original (the square orange
thing below), there's enough room on the processor board to fit
the new "Omnicell". I measured the terminal voltage
of the orange lithium cell and it read 11mV.
I then worked out how to enter the menu
and fix the date. Not easy as the buttons aren't marked and although
I fixed the new date as 27th Feb 2019 fixing the time will have
to wait. Interestingly, during the date setting I noticed the
year could be changed to 2500. I guess the new lithium battery
will be past its best by then...
Here's a couple of pictures
of traces... focus is a bit dodgy in the second because of trying
to avoid reflections.. and controls are not adjusted properly.
Le Croy still make oscilloscopes
under the Teledyne-Le Croy name with their most expensive
scope fetching $1,000,000. A trifle more than mine is worth..
and theirs doesn't need any connections to make it work.