Topics

2465A Repair Question

Matthew Wilson
 

A friend of mine gave me his old 2465A scope, which he said was working (as of a few years ago when he last turned it on).

When I turned it on, nearly all the button lights come on and I see a bright dot on the CRT, but none of the controls do much at all. Focus control will focus and unfocus the dot, and "beam find" button brings the dot the center of the screen, but the other functions do not seem to have any effect. All of the button lights stay on no matter what. No text appears on the screen.

Thought I'd check the low voltage power supply voltages so here's what I got from J1119:

Pin Spec Measured
1 -15v -17.6v
2 +5v 4.25v
4 +10v 10.01v
5 -5v -5.8v
6 +15v 14v
8 +87v 79.9v
9 +42.4v 39.2v
11 -8v -8.25v
12 +5v 4.98v

Seemed a little off - and after reading some posts on this forum and others, I decided to try to recap the two power supply boards (A2A1 and A3). I know I should have done some troubleshooting to find the problem rather than shotgunning, but I figured I'd want to do that anyway at some point. At any rate, I recapped the two boards using the recommended list of Nichicon electrolytics, film safety caps, and the few resistor upgrades.

The good news is that it appears that I did a decent job and didn't screw anything up further.
The bad news is that I didn't fix the problem: scope operates the same (all buttons light up, no functions) and the power supply voltages, though they shifted to a very minor degree (-5.81v to -6.05v, for example), have not really changed that much. They're still off.

Where do I start from here? What can I test that will help me isolate the stage and then component?

I've worked on simple tube scopes and successfully restored a Heathkit IO--102 - nothing this complex though. I have the service manual.

Any advice?

John Griessen
 

On 10/11/19 10:36 PM, Matthew Wilson wrote:
Thought I'd check the low voltage power supply voltages so here's what I got from J1119:
Pin Spec Measured
1 -15v -17.6v
2 +5v 4.25v
4 +10v 10.01v
5 -5v -5.8v
6 +15v 14v
8 +87v 79.9v
9 +42.4v 39.2v
11 -8v -8.25v
12 +5v 4.98v
I would fix volts at pin
1
2
5
6
8
9

to be closer to spec first.

Chuck Harris
 

It is just a power supply.

I would start with fixing the +/-15V supplies, as they are used
extensively throughout the rest of the supply for infrastructure
purposes (eg. they power the power supply's regulators and opamps).

Also, the supply won't work safely or properly out of the scope
without proper loading... see the manual.

There is a nice diagnostics section just for the power supply in
the manual.

-Chuck Harris

John Griessen wrote:

On 10/11/19 10:36 PM, Matthew Wilson wrote:
Thought I'd check the low voltage power supply voltages so here's what I got from
J1119:

Pin Spec Measured
1 -15v -17.6v
2 +5v 4.25v
4 +10v 10.01v
5 -5v -5.8v
6 +15v 14v
8 +87v 79.9v
9 +42.4v 39.2v
11 -8v -8.25v
12 +5v 4.98v
I would fix volts at pin
1
2
5
6
8
9

to be closer to spec first.



Matthew Wilson
 

I am working through the diagnostic section and am on the "Regulator Troubleshooting Procedure" 12 and 13. I've got the boards back inside the unit all hooked up and am able to test voltages on the A2A1 board just fine.

My +10v reference is working fine, but the +5v at pin 1 of P232 only measures 4.25v. Varying R1292 does not vary this voltage reading at all.

Procedure advises "Repair path through U1371C and U1040"

First I went to U1371C and measured Pin 8 and 10.

Pin Spec Measured
8 +4.4v +12.27v
10 +5v +4.87v

The weird thing is that when I measure the voltage at R1284, which is directly connected to pin 10 of the U1371C, I see exactly 5v. Somehow by the time it gets to pin 10, we lose 0.13v?

How common is it to have U1371 go bad? From the parts list, this is an LM324J quad op amp.

Chuck Harris
 

When dealing with opamps that have negative feedback,
there should be exactly 0V between the + and -
inputs. If the voltage difference is even 1mV in
either direction, the output of the opamp will be
hard stuck at the appropriate power supply rail.

(opamps have open loop gains that are effectively
infinite.)

If you find the voltage between the + and - inputs
is truly zero, and the output is stuck at the rails,
the opamp is bad.

And, yes, they do go bad.

Looking at your circuit, the opamp is driving the
control for the inverter output, and the direction
of that stuck at the rails output shows that the
non-inverting input is higher than the inverting
input (eg. output is at positive rail), so that
means the measured 5V is lower than the reference
5V going into the "+" input. Which is exactly
what you should expect given that the 5V supply is
at 5.87V.

Follow the diagnostic routine's directions, and
follow the signal from the output of U1371C to the
Opto-isolator U1040.

Be careful! The transistor side of U1040 is on
the mains side of the power supply. You must not
ground anything on that side!

One thing, when you replaced the electrolytics, did
you use non polar electrolytics for the few 4.7uf
and 1uf NP electrolytics? It matters.

-Chuck Harris

Matthew Wilson wrote:

I am working through the diagnostic section and am on the "Regulator Troubleshooting Procedure" 12 and 13. I've got the boards back inside the unit all hooked up and am able to test voltages on the A2A1 board just fine.

My +10v reference is working fine, but the +5v at pin 1 of P232 only measures 4.25v. Varying R1292 does not vary this voltage reading at all.

Procedure advises "Repair path through U1371C and U1040"

First I went to U1371C and measured Pin 8 and 10.

Pin Spec Measured
8 +4.4v +12.27v
10 +5v +4.87v

The weird thing is that when I measure the voltage at R1284, which is directly connected to pin 10 of the U1371C, I see exactly 5v. Somehow by the time it gets to pin 10, we lose 0.13v?

How common is it to have U1371 go bad? From the parts list, this is an LM324J quad op amp.




Chuck Harris
 

I mistyped, 5.87V should be 4.87V.

-Chuck

Chuck Harris wrote:

When dealing with opamps that have negative feedback,
there should be exactly 0V between the + and -
inputs. If the voltage difference is even 1mV in
either direction, the output of the opamp will be
hard stuck at the appropriate power supply rail.

(opamps have open loop gains that are effectively
infinite.)

If you find the voltage between the + and - inputs
is truly zero, and the output is stuck at the rails,
the opamp is bad.

And, yes, they do go bad.

Looking at your circuit, the opamp is driving the
control for the inverter output, and the direction
of that stuck at the rails output shows that the
non-inverting input is higher than the inverting
input (eg. output is at positive rail), so that
means the measured 5V is lower than the reference
5V going into the "+" input. Which is exactly
what you should expect given that the 5V supply is
at 5.87V. <<<<<< 4.87V!

Follow the diagnostic routine's directions, and
follow the signal from the output of U1371C to the
Opto-isolator U1040.

Be careful! The transistor side of U1040 is on
the mains side of the power supply. You must not
ground anything on that side!

One thing, when you replaced the electrolytics, did
you use non polar electrolytics for the few 4.7uf
and 1uf NP electrolytics? It matters.

-Chuck Harris

Matthew Wilson wrote:
I am working through the diagnostic section and am on the "Regulator Troubleshooting Procedure" 12 and 13. I've got the boards back inside the unit all hooked up and am able to test voltages on the A2A1 board just fine.

My +10v reference is working fine, but the +5v at pin 1 of P232 only measures 4.25v. Varying R1292 does not vary this voltage reading at all.

Procedure advises "Repair path through U1371C and U1040"

First I went to U1371C and measured Pin 8 and 10.

Pin Spec Measured
8 +4.4v +12.27v
10 +5v +4.87v

The weird thing is that when I measure the voltage at R1284, which is directly connected to pin 10 of the U1371C, I see exactly 5v. Somehow by the time it gets to pin 10, we lose 0.13v?

How common is it to have U1371 go bad? From the parts list, this is an LM324J quad op amp.






Matthew Wilson
 

I tested the input voltages on U1371C:
Pin 9 = 4.23v
Pin 10 = 4.98v
As you can see, there definitely is more than 0v difference.

I also measured one of the other amps on that chip and also found a difference in input voltages:
Pin 12 = 2.502v
Pin 13 = 2.464v

This is testing voltage measured on pins with pos lead and with neg lead on chassis (ground).

The 1µf caps that I replaced were WIMA film caps - so non-polarized, yes. I did not replace C1034 - the 4.7 µf tantalum cap, as I forgot to order that one.

Does this then mean that U1371 is bad? Is there any other test I can run that might confirm this before I look into replacing it?

Or should I continue troubleshooting and follow to U1040? Accessing pins on that board will be very tricky.

Chuck Harris
 

No, it doesn't mean U1371 is bad. It means that the
power supply output is low, and the feedback loop which
uses U1371 as its comparison element is trying as hard as
it can, but still can't force the supply up to 5V.

Look down the line to U104o. It could be a bad connection,
it could be a bad opto, or it could be something wrong on
the Line side of the opto. Follow the diagnostics tree.

Do not replace any tantalum capacitors in the 2465A. Every
tantalum in this scope is conservatively rated, and needs
to be a tantalum capacitor.

There are no WIMA film capacitors of that value in the
power supply. If there were in yours, someone already
replaced the NP electrolytic capacitors.... or you have a
power supply unlike any I have ever seen, and I have seen
a lot of 2465 supplies.

You have to understand the electronic parts you are measuring
for the measurements to be useful.

An opamp can be thought of as being a differential amplifier
with essentially infinite gain. That means that the difference
between the "+" amd the "-" inputs is multiplied by essentially
infinity to make the output voltage. And, the output voltage
cannot go higher than the power supply rails no matter what
the input. So, if you can measure *any* difference between
the "+" and "-" input, the output should be as close as it
can get to the appropriate power supply rail.

You have to make a dummy load for the power supply, and work
on the supply outside of the scope. It is the only safe way.
In your poking and prodding, if your probe slips, you can burn
out your scope with a high voltage surge.

-Chuck Harris



Matthew Wilson wrote:

I tested the input voltages on U1371C:
Pin 9 = 4.23v
Pin 10 = 4.98v
As you can see, there definitely is more than 0v difference.

I also measured one of the other amps on that chip and also found a difference in input voltages:
Pin 12 = 2.502v
Pin 13 = 2.464v

This is testing voltage measured on pins with pos lead and with neg lead on chassis (ground).

The 1µf caps that I replaced were WIMA film caps - so non-polarized, yes. I did not replace C1034 - the 4.7 µf tantalum cap, as I forgot to order that one.

Does this then mean that U1371 is bad? Is there any other test I can run that might confirm this before I look into replacing it?

Or should I continue troubleshooting and follow to U1040? Accessing pins on that board will be very tricky.

Matthew Wilson
 

First of all, thank you for your expertise and help! As I mentioned, I've done a lot of work on tube gear but nothing in the solid state world this complex so I very much appreciate your patience and guidance!

I replaced the non-polarized electrolytic 1µf 50v caps that were in the unit with new 1µf WIMA film caps rated to 63v. So no - there were no WIMA caps in there to begin with. Sorry for the confusion on that.

I will leave the tantalum cap in there.

I will read the section on making the dummy load carefully and proceed with checking U1040. The diagnostic tree at this point simply says "Repair path through U1371C and U1040" so I will do this.

Chuck Harris
 

Hi Matthew,

The Feedback signal from U1371C leaves the one circuit board, and
passes through the "stab" connectors to the other circuit board.

At that point, it simply goes to the LED in the Opto Isolator U1040.

Measure the voltage at the Opto Isolator's pin 1, and it should be
about 2V... maybe a little more, maybe a little less.

Measure it at The 4.3K that connects to the stab connector and it
should be the same voltage as you measured on the opto's pin 1.

If the stab connector is missing from your re-cap, all sorts of
funny stuff will happen.

Opto isolators do go bad... not very often, but it can happen.

-Chuck Harris

Matthew Wilson wrote:

First of all, thank you for your expertise and help! As I mentioned, I've done a lot of work on tube gear but nothing in the solid state world this complex so I very much appreciate your patience and guidance!

I replaced the non-polarized electrolytic 1µf 50v caps that were in the unit with new 1µf WIMA film caps rated to 63v. So no - there were no WIMA caps in there to begin with. Sorry for the confusion on that.

I will leave the tantalum cap in there.

I will read the section on making the dummy load carefully and proceed with checking U1040. The diagnostic tree at this point simply says "Repair path through U1371C and U1040" so I will do this.



 

Sounds like progress.  Lets tackle whats left one at a time.  The voltage at pin 10 of U1371 and at the junction of R1284 with R1285 MUST be the same and they MUST measure +5 Volts.  Don't waste time until they do.  Since the voltage at the junction does not match the voltage at pin 10 of U1371 you either have a measurement problem of the PCB has a fault.  Start by putting a wire jumper between the junction and Pin 10.  Now check the measurement to see if they are equal.  Change measurement technique until they are equal.
Once you have 5 volts at pin 10 of U1371 you are ready to proceed to fault 2.  U1371C serves as to integrate the error between the +5 on pin 10 and the +5Vd on pin 1 of P232.  I think your U1371 is working just fine and is giving an error voltage output "asking" to raise the voltage on pin 1 of P232.  That request is routed to pin 1 of U1040.  U1040 is an optical isolator whose output is used to drive U1030 and slew the voltage on pin 1 of P232.
So fix the voltage on pin 10 of 1371 then see why U1040 is not driving changes in U1030.  I have seen at least one case where U1040 was wide open and unable to drive anything.

On Saturday, October 12, 2019, 10:57:39 PM CDT, Matthew Wilson <@cecilkleakins> wrote:

I am working through the diagnostic section and am on the "Regulator Troubleshooting Procedure" 12 and 13. I've got the boards back inside the unit all hooked up and am able to test voltages on the A2A1 board just fine.

My +10v reference is working fine, but the +5v at pin 1 of P232 only measures 4.25v. Varying R1292 does not vary this voltage reading at all.

Procedure advises "Repair path through U1371C and U1040"

First I went to U1371C and measured Pin 8 and 10.

Pin      Spec      Measured
8        +4.4v      +12.27v
10      +5v        +4.87v

The weird thing is that when I measure the voltage at R1284, which is directly connected to pin 10 of the U1371C, I see exactly 5v. Somehow by the time it gets to pin 10, we lose 0.13v?

How common is it to have U1371 go bad? From the parts list, this is an LM324J quad op amp.

Matthew Wilson
 

On Sun, Oct 13, 2019 at 04:39 PM, machineguy59 wrote:

Change measurement technique until they are equal.
Ok - I checked my Fluke 85 and battery indicator was on - Pulled the 9v battery out and it measured 6.2v. Put a fresh one in and took another measurement.

I tuned R1292 slightly to get 5v exactly.
10v Reference = 10.02v
Junction of R1284 and R1285 = 5v
U1371 Pin 10 = 5v
U1371 Pin 9 = 4.23v
U1371 Pin 8 = 12.09v

Now on to U1040 and the other board.

 

You have to be very careful when working with "the other board" aka The Inverter Board.  The Inverter Board has mains power and at least two separate "commons" none of which are the same as your instrumentation commons.  A battery powered, none grounded, multimeter may be suitable but with limited utility.  You can test U1040 by placing a 10K resister between pins 4 and 5, then see if +5 Vd has increased in value.  If it does, it means the inverter responds to U1040 requests but U1040 is likely unable to make valid requests.  Remove the resistor and replace U1040.

On Sunday, October 13, 2019, 10:46:35 PM CDT, Matthew Wilson <@cecilkleakins> wrote:

On Sun, Oct 13, 2019 at 04:39 PM, machineguy59 wrote:

Change measurement technique until they are equal.
Ok - I checked my Fluke 85 and battery indicator was on - Pulled the 9v battery out and it measured 6.2v. Put a fresh one in and took another measurement.

I tuned R1292 slightly to get 5v exactly.
10v Reference = 10.02v
Junction of R1284 and R1285 = 5v
U1371 Pin 10 = 5v
U1371 Pin 9 = 4.23v
U1371 Pin 8 = 12.09v

Now on to U1040 and the other board.

 

On Mon, Oct 14, 2019 at 01:23 AM, Chuck Harris wrote:


The Feedback signal from U1371C leaves the one circuit board, and
passes through the "stab" connectors to the other circuit board.
A while ago, I worked on a 2465B which behaved funnily. Turned out several of these "stab" connectors were missing the tiny spring in the square bushings, resulting in intermittent contact.

Raymond

Chuck Harris
 

Good point Raymond. Installing the stab connector at
a slight angle to vertical, can wipe the spring right
out of the mating connector.

Stab connectors are made for only a couple of mating
cycles over their entire lifetime.

It would be a good idea to do a continuity check on
the pins of this stab connector before going much
farther.

-Chuck Harris

Raymond Domp Frank wrote:

On Mon, Oct 14, 2019 at 01:23 AM, Chuck Harris wrote:


The Feedback signal from U1371C leaves the one circuit board, and
passes through the "stab" connectors to the other circuit board.
A while ago, I worked on a 2465B which behaved funnily. Turned out several of these "stab" connectors were missing the tiny spring in the square bushings, resulting in intermittent contact.

Raymond

Matthew Wilson
 

I tested all the stab connectors - all tested fine.

So I pulled the boards out to attach two mini hook leads to U1040 pins 4 and 5 (via R1046 and R1044 respectively). I carefully reinstalled the boards with the leads still attached and had a 10K resistor ready to connect. I turned the unit on.

It sprang to life! A full horizontal trace appeared and text appeared at the bottom of the screen. Focus, intensity, position, channel selectors, they all work now!

Not sure what I even did - I hadn't even connected the 10k resistor to test U1040 yet. Guess it just decided to come back to life. I didn't re-assemble it any differently than before. Hmmm...

Thanks everyone for your help on this one - I'll continue the diagnostic process and then move on toward calibration. I'll let everyone know how it goes.

 

Seems you have encountered the Heisenberg Uncertainty Principle (the act of observing a particle changes the particle).  Its possible that checking the stab connectors caused one to seat and make better connection.  Lets hope it doesn't re-occur.

On Tuesday, October 22, 2019, 10:05:24 PM CDT, Matthew Wilson <@cecilkleakins> wrote:

I tested all the stab connectors - all tested fine.

So I pulled the boards out to attach two mini hook leads to U1040 pins 4 and 5 (via R1046 and R1044 respectively). I carefully reinstalled the boards with the leads still attached and had a 10K resistor ready to connect. I turned the unit on.

It sprang to life! A full horizontal trace appeared and text appeared at the bottom of the screen. Focus, intensity, position, channel selectors, they all work now!

Not sure what I even did - I hadn't even connected the 10k resistor to test U1040 yet. Guess it just decided to come back to life. I didn't re-assemble it any differently than before. Hmmm...

Thanks everyone for your help on this one - I'll continue the diagnostic process and then move on toward calibration. I'll let everyone know how it goes.

Tony Fleming
 

I'm happy for you!
One step I do, when I reassemble back to original state... I plug in the
power but I don't push the power button! Than I unplug the power cord and
push the power button for few seconds. After this I plug in power cord and
push the power button once more.
Great to hear that you are saving one more Tektronix!
Good luck!

On Tue, Oct 22, 2019 at 10:05 PM Matthew Wilson <@cecilkleakins>
wrote:

I tested all the stab connectors - all tested fine.

So I pulled the boards out to attach two mini hook leads to U1040 pins 4
and 5 (via R1046 and R1044 respectively). I carefully reinstalled the
boards with the leads still attached and had a 10K resistor ready to
connect. I turned the unit on.

It sprang to life! A full horizontal trace appeared and text appeared at
the bottom of the screen. Focus, intensity, position, channel selectors,
they all work now!

Not sure what I even did - I hadn't even connected the 10k resistor to
test U1040 yet. Guess it just decided to come back to life. I didn't
re-assemble it any differently than before. Hmmm...

Thanks everyone for your help on this one - I'll continue the diagnostic
process and then move on toward calibration. I'll let everyone know how it
goes.



Matthew Wilson
 

On Wed, Oct 23, 2019 at 08:20 AM, machineguy59 wrote:
Lets hope it doesn't re-occur.
Funny you should say that!

I was concerned about that so when it was working I turned it off for a while, then turned it on to see if it started up correctly. I did this test a few times. Each time it started up fine and a trace appeared with all controls functional.

So I removed the boards to remove the mini-hook leads (they are great for hooking onto PCB components but sometimes are tricky to remove in a tight space). I removed the leads, then reassembled everything and turned the unit on.

No trace. We're back to a single bright dot with all buttons lit up and no functionality!

I think we can rule out the stab connectors because throughout the last two disassembly/reassembly rounds, I have not taken the two boards apart at all.

Wondering if the PCB might be cracked somewhere? This is weird.

Or... the ribbon cable connectors - wonder if they can come loose internally or not connect well after being moved around during disassembly and reassembly. I think I'll start by testing ribbon cable continuity.

Any other ideas?

Tony Fleming
 

The connector will oxidise and I try to clean them all and lubricate them
with QD Electronic Cleaner and lubricate with Non-Flammable RadioShack
64-4315 called: Control/Contact Cleaner & Lubricant.
I also jigle the connectors, pull them in and out to clean the inside of
the contact plugin.
It sounds like a cold weld or it even can be a crack, but I would resolder
some spots - if you have a microscope, give it a nice detail scan with your
eyes.
Have a great day!

On Wed, Oct 23, 2019 at 8:30 PM Matthew Wilson <@cecilkleakins> wrote:

On Wed, Oct 23, 2019 at 08:20 AM, machineguy59 wrote:
Lets hope it doesn't re-occur.
Funny you should say that!

I was concerned about that so when it was working I turned it off for a
while, then turned it on to see if it started up correctly. I did this test
a few times. Each time it started up fine and a trace appeared with all
controls functional.

So I removed the boards to remove the mini-hook leads (they are great for
hooking onto PCB components but sometimes are tricky to remove in a tight
space). I removed the leads, then reassembled everything and turned the
unit on.

No trace. We're back to a single bright dot with all buttons lit up and no
functionality!

I think we can rule out the stab connectors because throughout the last
two disassembly/reassembly rounds, I have not taken the two boards apart at
all.

Wondering if the PCB might be cracked somewhere? This is weird.

Or... the ribbon cable connectors - wonder if they can come loose
internally or not connect well after being moved around during disassembly
and reassembly. I think I'll start by testing ribbon cable continuity.

Any other ideas?