Advice on HP Function Generator Repair

[Jake]

Hello all,

I have a question about where I should start in an overhaul of my
newly purchased HP 3114A function generator. It is the original model
with the hard plastic buttons, circa 1986 I believe.

A little background on me, I am an electrical engineering student,
first time post here. I have worked as an electrical engineering
technician for several years, so I am very comfortable with soldering
and useage of lab equipment, and I have a very nice scope and logic
analyzer and LCR meter and such, so I just need some basic advice on
how to improve the waveform quality on this thing.

I notice in the service manual that there are a lot of things I can do
calibration wise, but the only problem I am really having with the
unit is the waveform quality.

The higher the frequency, the worse the distortion, at ~5mhz I really
start to notice overshoot in the square wave and a little bit of
distortion in the triangle wave, and at ~20mhz, the sine and triangle
look the same (mediocre sine wave, triangle is a little dirtier) and
the square wave is completely unstable, basically really bad overshoot
and oscillation, not anywhere near flat, and it does not rise/fall as
fast as it should (has a slight slope to it)... The frequency is
pretty accurate, though I notice a little jitter on my scope.

My first thought is to just go through it and replace all the
electrolytics. I am under the general impression that the inductors
and ceramic caps should be OK.

Basically, I just want to give it a "tune-up" before I go through the
procedures in the service manual only to find that I can't get it into
spec because of old components, since the unit is almost 25 years old.

Any advice that can be given on what to watch for on these units would
be appreciated. I would like to get this thing working right so I can
start experimenting

 

[jdiaz5513]

Hello all,

I have a question about where I should start in an overhaul of my
newly purchased HP 3114A function generator. It is the original model
with the hard plastic buttons, circa 1986 I believe.

A little background on me, I am an electrical engineering student,
first time post here. I have worked as an electrical engineering
technician for several years, so I am very comfortable with soldering
and useage of lab equipment, and I have a very nice scope and logic
analyzer and LCR meter and such, so I just need some basic advice on
how to improve the waveform quality on this thing.

I notice in the service manual that there are a lot of things I can do
calibration wise, but the only problem I am really having with the
unit is the waveform quality.

The higher the frequency, the worse the distortion, at ~5mhz I really
start to notice overshoot in the square wave and a little bit of
distortion in the triangle wave, and at ~20mhz, the sine and triangle
look the same (mediocre sine wave, triangle is a little dirtier) and
the square wave is completely unstable, basically really bad overshoot
and oscillation, not anywhere near flat, and it does not rise/fall as
fast as it should (has a slight slope to it)... The frequency is
pretty accurate, though I notice a little jitter on my scope.

My first thought is to just go through it and replace all the
electrolytics. I am under the general impression that the inductors
and ceramic caps should be OK.

Basically, I just want to give it a "tune-up" before I go through the
procedures in the service manual only to find that I can't get it into
spec because of old components, since the unit is almost 25 years old.

Any advice that can be given on what to watch for on these units would
be appreciated. I would like to get this thing working right so I can
start experimenting

Electrolytics are a good start, be sure to give the unit a *good* look-
over for signs of heat stress. The PCB will be browned/darker in those
areas; I'd recommend replacing components in those areas if you really
want the thing to shine. Obviously if any components are just plain
charred black, replace those too .

Ceramics and inductors don't really have a "soft" failure mode - when
they're bad it should be fairly obvious. Same goes for resistors most
of the time, but I would recommend giving larger low-ohm power
resistors a quick measurement, it wouldn't hurt. Those are more prone
to fail due to high current.

Also check the reference voltages and the main supply voltage - too
little power will definitely give you rise time issues. If they're
adjustable via varistors, adjust them to be as close to their intended
value as possible (5v, 12v, 3.3v, etc.).

-JD

 

[Jake]

Electrolytics are a good start, be sure to give the unit a *good* look-
over for signs of heat stress. The PCB will be browned/darker in those
areas; I'd recommend replacing components in those areas if you really
want the thing to shine. Obviously if any components are just plain
charred black, replace those too .

Ceramics and inductors don't really have a "soft" failure mode - when
they're bad it should be fairly obvious. Same goes for resistors most
of the time, but I would recommend giving larger low-ohm power
resistors a quick measurement, it wouldn't hurt. Those are more prone
to fail due to high current.

Also check the reference voltages and the main supply voltage - too
little power will definitely give you rise time issues. If they're
adjustable via varistors, adjust them to be as close to their intended
value as possible (5v, 12v, 3.3v, etc.).

-JD- Hide quoted text -

- Show quoted text -

Hello, thanks for the reply. Just an update, my initial inspection was
rather hasty and I have noticed that in fact my rise time is way to
slow (leading edge of my square wave at lower freqs is very round) and
same goes for the fall time. When I get above about 5 MHz, all the
waves look the same - kind of like a sine wave, though the square wave
looks more like a triangle wave with rather rounded ends. Also, the
higher frequencies do not seem to be stable at all, no matter what I
do with my trigger, the waveform jumps all over on my Tek TDS2024B.
Could this still be related to caps or should I start inspecting other
things?

I have thoroughly inspected all PC boards and have found no evidence
of overheating. I don't know that this unit was used much, it is
spotless inside and there was no dust whatsoever, anywhere. Also the
fan looks original and is still very quiet.

I am going to order the electrolytics. I have noticed that my supplies
are off a little, and will adjust accordingly.

Hopefully this will clean the thing up, as right now it can't make a
stable waveform above a couple MHz to save its life

Thanks!

 

[Sjouke Burry]

Hello, thanks for the reply. Just an update, my initial inspection was
rather hasty and I have noticed that in fact my rise time is way to
slow (leading edge of my square wave at lower freqs is very round) and
same goes for the fall time. When I get above about 5 MHz, all the
waves look the same - kind of like a sine wave, though the square wave
looks more like a triangle wave with rather rounded ends. Also, the
higher frequencies do not seem to be stable at all, no matter what I
do with my trigger, the waveform jumps all over on my Tek TDS2024B.
Could this still be related to caps or should I start inspecting other
things?

I have thoroughly inspected all PC boards and have found no evidence
of overheating. I don't know that this unit was used much, it is
spotless inside and there was no dust whatsoever, anywhere. Also the
fan looks original and is still very quiet.

I am going to order the electrolytics. I have noticed that my supplies
are off a little, and will adjust accordingly.

Hopefully this will clean the thing up, as right now it can't make a
stable waveform above a couple MHz to save its life

Thanks!

A small advice, leave the unit powered for a few days.
Old equipment with a long storage time, often needs to
re-format electrolytes.
You might see improvement after that.

 

[jdiaz5513]

Hello, thanks for the reply. Just an update, my initial inspection was
rather hasty and I have noticed that in fact my rise time is way to
slow (leading edge of my square wave at lower freqs is very round) and
same goes for the fall time. When I get above about 5 MHz, all the
waves look the same - kind of like a sine wave, though the square wave
looks more like a triangle wave with rather rounded ends. Also, the
higher frequencies do not seem to be stable at all, no matter what I
do with my trigger, the waveform jumps all over on my Tek TDS2024B.
Could this still be related to caps or should I start inspecting other
things?

I have thoroughly inspected all PC boards and have found no evidence
of overheating. I don't know that this unit was used much, it is
spotless inside and there was no dust whatsoever, anywhere. Also the
fan looks original and is still very quiet.

I am going to order the electrolytics. I have noticed that my supplies
are off a little, and will adjust accordingly.

Hopefully this will clean the thing up, as right now it can't make a
stable waveform above a couple MHz to save its life

Thanks!

Smells like dry electrolytics to me. An ESR meter is a godsend if you
plan on debugging many circuits, it lets you test for bad caps even
while they're in-circuit!

-Julian

 

[Jake]

Not having seen a picture of the function generator, I'm presuming you're
using an external oscilloscope to look at the FG's output.

Just a question: Have you been able to verify that the strange waveforms are
actually coming from the function generator itself, rather than an artifact
of... say, the oscilloscope probe and/or the oscilloscope itself not being
able to handle incoming frequencies that high?  Just don't want you to waste
time trying to fix something that may not actually be broken.  

If it is definitely coming from the function generator, then I'd agree with
the previous posters: replace the electrolytic capacitors.  They're over 20
years old now.

I am using a brand new Tektronix TDS2024B 200MHz 2GSa/s oscilloscope
with a fresh calibration.

Thanks for all the advice!

 

[Jake]

Another thing I am noticing, has anyone else refurbished one of these
and discovered that they use capacitors with values that are difficult
or impossible to attain? I will probably just use some parallel/series
combinations...

 

[Bob Larter]

Another thing I am noticing, has anyone else refurbished one of these
and discovered that they use capacitors with values that are difficult
or impossible to attain? I will probably just use some parallel/series
combinations...

Electros are so wide tolerance (usually -10%/+20%) that you should be
able to use the next standard value up without any problems.

 

[jdiaz5513]

Electros are so wide tolerance (usually -10%/+20%) that you should be
able to use the next standard value up without any problems.

--
    W
  . | ,. w ,   "Some people are alive only because
   \|/  \|/     it is illegal to kill them."    Perna condita delenda est
---^----^---------------------------------------------------------------

You can also safely use caps that are rated at a higher voltage -
they'll just be bigger.

-JD

 

[Arfa Daily]

Not having seen a picture of the function generator, I'm presuming you're
using an external oscilloscope to look at the FG's output.

Just a question: Have you been able to verify that the strange waveforms
are
actually coming from the function generator itself, rather than an
artifact
of... say, the oscilloscope probe and/or the oscilloscope itself not being
able to handle incoming frequencies that high? Just don't want you to
waste
time trying to fix something that may not actually be broken.

If it is definitely coming from the function generator, then I'd agree
with
the previous posters: replace the electrolytic capacitors. They're over 20
years old now.

I am using a brand new Tektronix TDS2024B 200MHz 2GSa/s oscilloscope
with a fresh calibration.

Thanks for all the advice!

Not wishing to labour the point, but what you describe as happening to the
waveshapes, in that they are all heading towards a dirty sine wave the
higher up in frequency that you go, sure sounds like what you will get if
you glue a low-pass filter on the end of such a generator. Are you
absolutely sure that there is not any bandwidth filter switched in on the
'scope, and that the probe you are using is good for the frequencies that
you are looking at ? Is it a low capacitance probe for instance ? Looking at
anything above a few hundred kHz generally requires the use of a low
capacitance probe (typically a x10 attenuator probe, or a 'universal' probe
switched to its "x10" setting)

Arfa

 

[hr(bob) [email protected]]

In addition to the advice already given, I'd like to make a suggestion or
two.  Make sure that the generator's output is terminated correctly (50
ohms?).  A feedthrough termination is the generally accepted way to load the
output; place the termination at the scope input rather than at the
generator output. You don't normally use a probe in this setup; use a cable
of the proper impedance (RG-58 for 50 ohms; RG59 for 75 ohms). This will
give you a much better picture of the output signal.  If you're lookingat
the unterminated output through a low frequency probe, it's no surprise that
your square waves look like triangles and the triangles look like sines.
Next, if you don't already have one, try to get a service manual for it.  I
looked in all the usual manual vendors' sites and didn't find any listings
for your 3114A, so it might take a bit of searching.

Cheers!!
--
"In theory, there isn't any difference between theory and practice.  In
practice, there is."  - Yogi Berra

Dave M
masondg44 at comcast dot net- Hide quoted text -

- Show quoted text -

Jake:

Have you looked at waveforms with your scope from a different
generator over the frequency ranges you are looking at from your own
generator? That way you can confirm that your scope is really working
correctly and that your generator is truly the problem. Even with a
fancy new scope, unless you prove it is working correctly, you are
making a risky assumption. Better to take the time to prove the
scope is good than to spend hours fixing a non-existent problem with
your generator. After 50 years in electronics, trust me, I have
chased many a ghost looking at the wrong problem.

 

[[email protected]]

Jake:

Have you looked at waveforms with your scope from a different
generator over the frequency ranges you are looking at from your own
generator?  That way you can confirm that your scope is really working
correctly and that your generator is truly the problem.  Even with a
fancy new scope, unless you prove it is working correctly, you are
making a risky assumption.  Better to take the time to  prove the
scope is good than to spend hours fixing a non-existent problem with
your generator.  After 50 years in electronics, trust me, I have
chased many a ghost looking at the wrong problem.- Hide quoted text -

- Show quoted text -

I agree with a previous poster who advised to ESR the electrolytic
capacitors. I would definitly do that first rather than shotgun all of
them. I use the Dick Smith ESR meter. If you don't own one and you're
interested in electronics you really should. I built it from a kit
several years ago and it is one of the most used pieces of test
equipment on my bench. I hope that they're still available online.
Just another thought, if this unit is as you say 23 years old it may
need calibration. I worked in a cal lab around the time your unit was
manufactured and we did a fair amount of HP. The thing with
calibration is that your standards must be verifiable as well. You
won't be able to make any adjustments, (you shouldn't) unless your
standards either meet or exceed HP's or better yet NBS specs. Lenny.

 

[[email protected]]

Another thing I am noticing, has anyone else refurbished one of these
and discovered that they use capacitors with values that are difficult
or impossible to attain? I will probably just use some parallel/series
combinations...

Are you trying to use cheaper caps than what was put in?

I would take it to a place that calibrates HP/Agilent test equipment
before I started screwing around with it. Twenty-five years is nothing
for a piece of HP test equipment.

 

[Spehro Pefhany]

I am using a brand new Tektronix TDS2024B 200MHz 2GSa/s oscilloscope
with a fresh calibration.

Just bought one of these for use in non-critical applications.
Make sure you have not set the probe to x1. Why? RTFM, or see for
example:

http://www2.tek.com/cmsreplive/marep/13210/061434300web_2008.10.02.16.19.01_13210_EN.pdf
note 1.

There's also a bandwidth limit (20 MHz, IIRC) that can be set from the
front panel.

Thanks for all the advice!

Best regards,
Spehro Pefhany

 

[Bob Larter]

You can also safely use caps that are rated at a higher voltage -
they'll just be bigger.

Yep.