Tektronix 7000 series power supply, sluggish start

[Joerg]

Ok, guys, without taking it all apart again, does anyone know a typical
"it's nearly always this part" type of cause?

When turning a Tex 7704A on the power supply lets off a faint bzzzzt ..
ttzzzt ... brrp ... bzzzzzt .... until after several minutes it come to
life. Is the magic neon bulb past its prime? Or something else?

I can still use the scope but I see the day coming when it no longer
cranketh over.

 

[Joerg]

It's cold ;-)

67F in the office. But thanks for the reminder, got to reload the wood
stove. My wife ain't gonna be happy when she comes back and the house
has dropped to 65F

Anyhow, one minutes ago ye olde 7704A has decided to call it quits :-(

 

[tm]

Ok, guys, without taking it all apart again, does anyone know a typical
"it's nearly always this part" type of cause?

When turning a Tex 7704A on the power supply lets off a faint bzzzzt ..
ttzzzt ... brrp ... bzzzzzt .... until after several minutes it come to
life. Is the magic neon bulb past its prime? Or something else?

I can still use the scope but I see the day coming when it no longer
cranketh over.

Do you have the service manual? There is a very good theory of operation
on the power supply. It is a combination switcher and linear system.

Most of the time, it's a shorted capacitor. There is current limiting on all
the
low voltage supplies in the form of current shunt emitter resistors on the
series pass transistors.

The good news is there are many parts units available for the 7704(A)
mainframes.

You might want to check the "Tekscopes" group on yahoogroups. A lot of help
is
available there.


tm

 

[Joerg]

Do you have the service manual? There is a very good theory of operation
on the power supply. It is a combination switcher and linear system.

I do have that. The switcher is one of those "hairball" circuits with
self-start, SCRs, and whatnot. If I have to I'll dig in. Was trying to
avoid that if someone knows a part that is the culprit in most cases.
Like on those AOR scanner receivers where once they are 5+ years old
usually the backup battery in there croaks and the thing quits. The
battery is soldered onto a board, sort of hidden.

Most of the time, it's a shorted capacitor. There is current limiting on all
the
low voltage supplies in the form of current shunt emitter resistors on the
series pass transistors.

Yes, I saw that. The post-regulators are on the acquisition board.
They've cut it quite close, 2V dropout on the 15V rails.

The good news is there are many parts units available for the 7704(A)
mainframes.

Yes, I could just buy a "new" one and be done with it. But somehow that
goes against the grain, from an environmetal POV. And I'd be throwing
away a piece of history. It's not quite a 2465 but those 7704A are darn
good scopes.

You might want to check the "Tekscopes" group on yahoogroups. A lot of help
is
available there.

Thanks, I'll have to check that one out.

 

[Joerg]

Even 62F isn't bad with a few layers of clothing... below 60F the
problem gets to be that fingers just don't want to work very well
(nimbly), even if overall you're still warm and comfortable!

In the mornings it can be 58F in the office and I actually like that.
Problem is, I don't feel it being too cold but my wife does. So when she
is gone for a few hours I set a kitchen timer. It reminds me to look at
the thermometer every 1-1/2h and reload the wood stove if too low.

On the other hand, as I recall, in summer you're all for having the
office at something like 94F before you turn on the AC?

Not anymore. Now we have a small evap cooler which keeps the living area
under 85F and the office under 90F. If I could figure out a way to mount
one near the office it would both go to less than 80F. But the way this
house is built that's nearly impossible. It is one of those Frank Lloyd
Wright style houses.

Second life as yard art? Or do you plan to fix it?

If it won't turn into a major science project I want to fix it.
Admitted, I rarely use it these days but when I do it always performs.
Or has :-(

Seems like old lab gear isn't immortal. The Wavetek 23 that I really
like went thoroughly kaputt half a year ago. But Tektronix did a much
better job omn this scope. Nothing gets too hot in there.

 

[tm]

I do have that. The switcher is one of those "hairball" circuits with
self-start, SCRs, and whatnot. If I have to I'll dig in. Was trying to
avoid that if someone knows a part that is the culprit in most cases.
Like on those AOR scanner receivers where once they are 5+ years old
usually the backup battery in there croaks and the thing quits. The
battery is soldered onto a board, sort of hidden.



Yes, I saw that. The post-regulators are on the acquisition board.
They've cut it quite close, 2V dropout on the 15V rails.



Yes, I could just buy a "new" one and be done with it. But somehow that
goes against the grain, from an environmetal POV. And I'd be throwing
away a piece of history. It's not quite a 2465 but those 7704A are darn
good scopes.



Thanks, I'll have to check that one out.

--

You might also try it with most of the plugins pulled just in case it is a
problem
there. You can take one known good module (vert) and try it in different
slots.

On the backplane there are about five electrolytic, most are tantalums and
they
are known to short out. You will need to unscrew the boards to see them.
That
is done much easier if the panels are all removed and the bottom guide rails
are pulled.

The power supply can be pulled out the rear and fully exposed while running.
Use care not to stress the cables too much.

The primary inverter has a circuit that requires all the voltages to balance
out
or it will go into a "tick" mode while it tries to restart. The frequency of
the tick
can give clues to the problem.

There are some voltages in there that can teach a lesson though.



tm

 

[Joerg]

tm wrote:

[...]

You might also try it with most of the plugins pulled just in case it is a
problem
there. You can take one known good module (vert) and try it in different
slots.

I had tried that. It won't even start with all modules pulled.

On the backplane there are about five electrolytic, most are tantalums and
they
are known to short out. You will need to unscrew the boards to see them.
That
is done much easier if the panels are all removed and the bottom guide rails
are pulled.

Aha, tantalums. That's often bad news. Looks like a job for a rainy
weekend then. Thanks for the hint.

The power supply can be pulled out the rear and fully exposed while running.
Use care not to stress the cables too much.

Yes, that was fairly easy, had done that one before.

The primary inverter has a circuit that requires all the voltages to balance
out
or it will go into a "tick" mode while it tries to restart. The frequency of
the tick
can give clues to the problem.

It's not a tick, more like a rattlesnake in the distance. Irregular
bzzzzt intervals. It used to start up after doing that for a couple
minutes but not anymore. Maybe it didn't like the long periods where it
wasn't turned on and wants to punish me for that now.

There are some voltages in there that can teach a lesson though.

Yep, it says so in the manual

 

[Joerg]

Umm, it's the 21st century! Has been for a while now.

Doesn't mean we have to abandon all the good stuff. I won't throw out a
nice bottle of wine just because is was from 1999

Rigol has a 1 GHz digital scope now. They aren't toys any more.

I've got a 200MHz and a 1GHz scope but they are digital. DSOs are what I
now use maybe 99% of the time. However, in my line of work there is that
1% where DSOs just can't cut it, no matter how fancy they may be. I have
been asked by clients "How on earth did you find this tiny random pulse
noise?"

 

[tm]

Umm, it's the 21st century! Has been for a while now.

Rigol has a 1 GHz digital scope now. They aren't toys any more.

John

Yeah, but can you plug a curve tracer, spectrum analyzer, logic analyzer,
sampling
plug in, etc in the Rigol?

In 20 years, can you replace IC-1 in the Rigol?



tm

 

[legg]

Ok, guys, without taking it all apart again, does anyone know a typical
"it's nearly always this part" type of cause?

When turning a Tex 7704A on the power supply lets off a faint bzzzzt ..
ttzzzt ... brrp ... bzzzzzt .... until after several minutes it come to
life. Is the magic neon bulb past its prime? Or something else?

I can still use the scope but I see the day coming when it no longer
cranketh over.

That's the self-start circuit functioning.

The main converter is resonant (L3037 C3037). Capacitive rectification
on prereg outputs.

The only thing that's going to get better as it warms up is
electrolytic esr, but these things used wet tantalum. If they're
leaking that could cause high ripple and reduced output voltages.
The 50V lines are most suspect.

RL

 

[Joerg]

That's the self-start circuit functioning.

The main converter is resonant (L3037 C3037). Capacitive rectification
on prereg outputs.

Yes, looks like 30kHz series-resonant. In those days they couldn't go
much higher.

The only thing that's going to get better as it warms up is
electrolytic esr, but these things used wet tantalum. If they're
leaking that could cause high ripple and reduced output voltages.
The 50V lines are most suspect.

Thanks. Then I guess the drill is to get the power supply out,
disconnect the various connectors for the output rails. If it still
won't start the short must be on the module itself. Which would be good
because prying out the backplane doesn't look like fun.

Strange thing is, sometimes when I let it try for a while the scope will
power up and work fine. Except that a very faint ticking of a couple
Hertz or so can be heard when I press my ear against the side of the
scope. But that could be the neon bulbs which are very tired.

 

[tm]

Yes, looks like 30kHz series-resonant. In those days they couldn't go
much higher.



Thanks. Then I guess the drill is to get the power supply out,
disconnect the various connectors for the output rails. If it still
won't start the short must be on the module itself. Which would be good
because prying out the backplane doesn't look like fun.

Strange thing is, sometimes when I let it try for a while the scope will
power up and work fine. Except that a very faint ticking of a couple
Hertz or so can be heard when I press my ear against the side of the
scope. But that could be the neon bulbs which are very tired.

--

It will not run without a load. Also, there is a supply cable and a sense
cable that come together on the backplane. The ticking is the voltage
balance circuit shutting the supply down. Come in from the front and pull
the modules from the left three slots and you will see the caps. They are
just dipped solid tantalums. It sounds like one of them is getting ready to
flame out. There is a summing point on the supply that you can connect
a scope to and based on what direction the voltage goes too during one
of the ticks, it will tell you what supply is failing. For example, if the
+15 shorts
to ground, the voltage will go negative.

Also, the HF signal from the inverter feeds the high voltage power supply in
the display unit and that load must also be there.


tm

 

[Nico Coesel]

I do have that. The switcher is one of those "hairball" circuits with
self-start, SCRs, and whatnot. If I have to I'll dig in. Was trying to
avoid that if someone knows a part that is the culprit in most cases.
Like on those AOR scanner receivers where once they are 5+ years old
usually the backup battery in there croaks and the thing quits. The
battery is soldered onto a board, sort of hidden.

Tektronix has a forum on their website where there is plenty of room
for talking about component level repairs.

This website may also help to get diagrams:

http://bama.edebris.com/manuals/tek/

 

[Joerg]

It will not run without a load. Also, there is a supply cable and a sense
cable that come together on the backplane. The ticking is the voltage
balance circuit shutting the supply down. Come in from the front and pull
the modules from the left three slots and you will see the caps. They are
just dipped solid tantalums. It sounds like one of them is getting ready to
flame out. There is a summing point on the supply that you can connect
a scope to and based on what direction the voltage goes too during one
of the ticks, it will tell you what supply is failing. For example, if the
+15 shorts
to ground, the voltage will go negative.

Also, the HF signal from the inverter feeds the high voltage power supply in
the display unit and that load must also be there.

Thanks. Looks like there are some daughterboards that are on the
backplane that I'll have to unscrew and remove. Can't see any tantalums
to the side of those. This looks like a tear-down job because there's
tight-fitting gray coaxes coming off of those "mezzanine boards".

Surprisingly, this morning the scope turned on immediately, like in the
old days, without delay and without any rattlesnake hisses. Weird.

 

[tm]

that is the sign of the PS load becoming marginally too high;the supply
current limits and shuts down,then restarts again and again.
(this series-resonant supply has several control loops,and current limit
circuits on some of the regulated supplies,too.)
It's probably some electrolytics(ESR),both on the PS itself,and on the
mainframe.
it could also be HV related.

One more thing;I've seen the Vb-e rise on the heat-sunk flat-pack(TO-126?)
power xstrs on the cap-regulator board and cause a current limit condition
on it's supply. Those xstrs are the ones that mount to the rear heat
sink,several of them in a row. Replacing the xstrs fixed the problem.
Of course,I had a curve tracer to check those xstrs......how to
troubleshoot that without one,I dunno. the current limit circuit on those
supplies is a foldback type,IIRC.

--

I have also seen the current sense resistors drift higher causing a lower
trip
point. These are easy to check in circuit. If you measure a higher
resistance
than the marked value, you have found it.


tm

 

[Joerg]

that is the sign of the PS load becoming marginally too high;the supply
current limits and shuts down,then restarts again and again.
(this series-resonant supply has several control loops,and current limit
circuits on some of the regulated supplies,too.)
It's probably some electrolytics(ESR),both on the PS itself,and on the
mainframe.
it could also be HV related.

One more thing;I've seen the Vb-e rise on the heat-sunk flat-pack(TO-126?)
power xstrs on the cap-regulator board and cause a current limit condition
on it's supply. Those xstrs are the ones that mount to the rear heat
sink,several of them in a row. Replacing the xstrs fixed the problem.
Of course,I had a curve tracer to check those xstrs......how to
troubleshoot that without one,I dunno. the current limit circuit on those
supplies is a foldback type,IIRC.

Thanks, Jim. I could test those by hand. Takes longer but done it
before. From what RL wrote the supply won't start without a sufficient
load. I do not have many modules because I only use the 7704A as a plain
vanilla analog scope. This morning it decided to work normal again,
after thinking about it for a night. Beats me why.

Looking into the 7A26 module I see that there's tons of tantalums in
there as well. Maybe one of the modules is acting up.

 

[Joerg]

For $400 or less, why bother to fix them? ...

For me, environmental reasons are number one. If we all follow the
unhealthy habit of "chuck it and buy a new one" that is not a good
thing. When we moved across the pond we needed a new clock radio that
didn't need 50Hz. $9.95 plus tax. One digit didn't work. Drat! Thought
about it for 5secs, opened it up, fixed it, done. Took 20mins or so off
our Saturday. Driving back to Walmart where they'd just given us a new
one would have cost at least 45mins plus a gallon of gasoline. And
they'd have chucked the "broken" radio which just had one bad solder joint.

The best thing is, whenever I do that I always learn something new about
cost efficient design. Most of all packaging, and with some projects
packaging can be >50% of an EE's real work.

[...]

I used to use a bunch of 7000 series scopes, but most have died by
now. Long-term, older tube units, like the 547, are more reliable and
more repairable. We do have one 7104 that we use now and then.

But you can really see the meter spin up when firing up a 547

Also, I found that while tube stuff is great it becomes harder and
sometimes painfully expensive to find NOS replacement tubes. That has
brought a nice generator to grief over here because it was built with
WW-II era steel tubes.

[...]

 

[tm]

Thanks. Looks like there are some daughterboards that are on the
backplane that I'll have to unscrew and remove. Can't see any tantalums
to the side of those. This looks like a tear-down job because there's
tight-fitting gray coaxes coming off of those "mezzanine boards".

Surprisingly, this morning the scope turned on immediately, like in the
old days, without delay and without any rattlesnake hisses. Weird.

--

Take a picture or two so you know how the cables go back together. The
tantalum caps are behind the left two boards. Much easier if you pull a few
of the black rails out of the bottom and stand the scope on the rear. Go in
through the bottom and you can remove the bad cap(s). Some suggest just
clipping the leads next to the old cap and just soldering the new cap to the
old leads. The board is multilayer and very old so pulling the plating out
of
the holes is easy and ruins the board.

Be gentle around those edge connectors. The plastic covers pop off easily
and don't always go back correctly. They are very necessary.



Good luck,

tm

 

[tm]

For $400 or less, why bother to fix them? None of our color digital
scopes, two Rigols and about a dozen Teks, have broken so far.

I used to use a bunch of 7000 series scopes, but most have died by
now. Long-term, older tube units, like the 547, are more reliable and
more repairable. We do have one 7104 that we use now and then.

But if I want a spectrum analyzer, I can buy a nice new digital one.
In inflation-adjusted dollars, they are a lot cheaper than the old Tek
plugins were, and work a lot better. And don't need their own wheeled
vehicle.

The 7A22 diffamp was cool, but I use an AM502 to front a digital scope
and get the best of both worlds.

I do still use 11801 sampling scopes, wonderful boxes. They seem
pretty reliable, and used ones cost a few percent of what a new
sampler would cost. The sampling plugins for the 7000 series were
pretty barbaric.

John

Hey, at least they are less likely to walk away from your bench.

The Rigols are cheap right now because of the Chinese dumping. When the
environazis get their way and make them follow all the laws we need to,
those
prices will get to more like ours. They sell a digital scope for $400
because
they don't worry about osha, epa, unions, human rights, etc. And pay what, a
dollar a day to the labor force?

Then you will want to repair the broken scope.



tm

 

[Joerg]

Take a picture or two so you know how the cables go back together. The
tantalum caps are behind the left two boards. Much easier if you pull a few
of the black rails out of the bottom and stand the scope on the rear. Go in
through the bottom and you can remove the bad cap(s). Some suggest just
clipping the leads next to the old cap and just soldering the new cap to the
old leads. The board is multilayer and very old so pulling the plating out
of
the holes is easy and ruins the board.

Be gentle around those edge connectors. The plastic covers pop off easily
and don't always go back correctly. They are very necessary.

That is a good idea. I was thinking along the same lines, just clipping
and soldering some non-tantalums in there. Hoping that the black rails
don't break, they feel as if they are totally hardened.

To my surprise the scope decided to just work again. Not sure if the
constant pulling and re-inserting of modules did that or what.