Agilent 33120A, 15 MHz Generator, (teardown and repair)

[HighVoltage]

This Agilent 33120A 15 MHz Generator was broken, when I got it for Euro 100 and it must have been badly abused.

State when I got it:
- The knob was missing
- The rotary encoder shaft was broken off
- It was full of dirt and dust inside
- Nothing happens at power ON

Firstly I cleaned everything with air pressure, may be I will clean the main PCB with alcohol.
The fuse is OK and the transformer output seems OK.
This one has a full mains switch at the back.

I installed a new rotary encoder after cleaning the front panel PCB, display and buttons.
Now I need to start the trouble shooting.
Here are the first few pictures from the repair so far.

The date stamp on the inside housing is June 01, 2000

[HighVoltage]

+18 V and -18V are bad, far too low.
Either LM317 or large filter capacitors are may be bad or the PCB pulls too much current?

Here is the power supply schematics, for anyone interested.
 

[HighVoltage]

The two filter caps before the LM317 are bad
These are C1001 and C1002

They are branded KME and I don't think I have seen them before.
Is this a good brand?
Does anyone know?
They are 2000 uF 50V

I only have 2200uF, 35V Nichicon in my bin of new parts
But since the rail carries only 18V, I think I will be ok with 35V nichicon, right?
Any comments?

[edpalmer42]

HV, I think you had a blonde moment or maybe a senior moment there!  Pick whichever one applies!   

If you had googled <KME capacitor>, you would have found that KME is the model family.  The shield logo tells me that the brand is United Chemi-con aka Nippon Chemi-con which is one of the top brands.  The first Google hit is the link to the datasheet: 

https://www.westfloridacomponents.com/mm5/graphics/datasheets/3/KME-Series-Nippon-Chemi-Con.pdf

I'm suprised to see them fail under the relatively benign operating conditions of a linear power supply.  They tend to be almost bulletproof.  But I guess that after 20 years, anything is possible.  I wonder how hot those heatsinks get in normal operation.

You said they're bad, but you didn't mention how you determined that.

As for replacing them, check the AC voltage coming off the red leads.  How does the peak voltage compare to the 35V rating on your capacitors?  I suspect that there might not be a lot of safety margin there.

I'm also suspicious of that schematic.  When I look at the lower bridge rectifier and the circuitry around it, I don't see any big filter capacitors on the input of the LM337 and the lm2940.  But they claim to have polarized capacitors of 0.01uf and 0.015 uf.  Be cautious about trusting that schematic.

Ed

[Samogon]

I would not go almost half rated voltage. Take some break and order proper caps.

[HighVoltage]

HV, I think you had a blonde moment or maybe a senior moment there!  Pick whichever one applies!   
Ed

 
I probably typed "KEM" instead in to google, and got no hits,  LOL

The heat sink was "pretty" hot as to a point that you don't want to touch them.
I will measure the temperature tomorrow.

I would not go almost half rated voltage. Take some break and order proper caps.

I will try the nichicon 35V caps tomorrow and if they work, I will order the correct 50V version.

[glarsson]

When I look at the lower bridge rectifier and the circuitry around it, I don't see any big filter capacitors on the input of the LM337 and the lm2940.  But they claim to have polarized capacitors of 0.01uf and 0.015 uf.  Be cautious about trusting that schematic.

But the schematic doesn't say 0.01u and 0.015u, it says 0.01 and 0.015 when all other polarized capacitors have u. Probably 0.01F and 0.015F, i.e. 10000uF and 15000uF. Reasonable value considering their physical sizes and voltage requirements.

[edpalmer42]

When I look at the lower bridge rectifier and the circuitry around it, I don't see any big filter capacitors on the input of the LM337 and the lm2940.  But they claim to have polarized capacitors of 0.01uf and 0.015 uf.  Be cautious about trusting that schematic.

But the schematic doesn't say 0.01u and 0.015u, it says 0.01 and 0.015 when all other polarized capacitors have u. Probably 0.01F and 0.015F, i.e. 10000uF and 15000uF. Reasonable value considering their physical sizes and voltage requirements.

That sounds better!  I've never seen that labelling method.  The only time I've seen Farads is with supercaps and then they usually show 'F' instead of leaving it blank.

Ed

[edpalmer42]

HV, I see that there's a little fan in the back.  Is it still good?  If it seized up, that could explain the cooked capacitors.

You might also consider some kind of thermal shield between the capacitors and the heatsinks.

Ed

[PA0PBZ]

When I look at the lower bridge rectifier and the circuitry around it, I don't see any big filter capacitors on the input of the LM337 and the lm2940.  But they claim to have polarized capacitors of 0.01uf and 0.015 uf.  Be cautious about trusting that schematic.

But the schematic doesn't say 0.01u and 0.015u, it says 0.01 and 0.015 when all other polarized capacitors have u. Probably 0.01F and 0.015F, i.e. 10000uF and 15000uF. Reasonable value considering their physical sizes and voltage requirements.

That sounds better!  I've never seen that labelling method.  The only time I've seen Farads is with supercaps and then they usually show 'F' instead of leaving it blank.

Ed

[glarsson]

I've never seen that labelling method.  The only time I've seen Farads is with supercaps and then they usually show 'F' instead of leaving it blank.

Over the years many methods have been used to designate capacitor values in schematics. In the olden days you could find 10MF for 10uF and 10MMF for 10pF (ten micro micro Farads) and other monstrosities like using .1 for 100nF or 1uF depending on how many specks was added or removed by the xerox machine of fax machine. I started with electronics in 1973 and have seen a fair share of the madness. 

[HighVoltage]

HV, I see that there's a little fan in the back.  Is it still good?  If it seized up, that could explain the cooked capacitors.

You might also consider some kind of thermal shield between the capacitors and the heatsinks.

Ed

Yes, I forgot to mention, the fan was fully plugged up with dirt and dead!
This was definitely part-reason for overheating.

The shield idea is a good one.
I will probably also use new high temp capacitors.

[HighVoltage]

Today I installed (temporarily) the two Nichicon 2200uF, 35V caps.
And the instrument turns ON perfectly

There are the following 18V rail test points marked on the PCB:

TP                    / Reading old Caps    /Reading new caps
+ 18V                 + 12.2V                  + 18.08 V
-  18V                 -  13.6 V                 -  18.24 V

I will now select and order the correct caps and replace all electrolytic caps.
Who knows, they probably all got far too hot.
The new rotary encoder is working perfectly as well.

So, to be continued, when I have the new capacitors.

[Dr. Frank]

The two filter caps before the LM317 are bad
..

I only have 2200uF, 35V Nichicon in my bin of new parts
...But since the rail carries only 18V, I think I will be ok with 35V nichicon, right?

HighVoltage,
You're a bit confused 😖, aren't you?

The 18V are behind the regulators, but these caps see the transformers / rectifiers higher voltage.

50V seems very high though, so if the voltage on the two caps w/o  any load , circuit off/removed  , stays well below 30V, I would prefer the 35V types, as they will have lower ESR.

Frank

[amspire]

I only have 2200uF, 35V Nichicon in my bin of new parts
But since the rail carries only 18V, I think I will be ok with 35V nichicon, right?
Any comments?

The mains transformer output measures as 50V center tapped on my 33120A - that is +/- 35.3V peak from ground before the bridge rectifier with 240V input. You have to be able to cope with the peak voltages over the full specified range of mains voltages.

35V is definitely marginal - you want 50V. I am not sure that lower ESR will help at all. It could cause higher peak currents from the transformer and in the rectifiers, but you will still end up with the same amount of ripple.

[HighVoltage]

HighVoltage,
You're a bit confused 😖, aren't you?

The 18V are behind the regulators, but these caps see the transformers / rectifiers higher voltage.

50V seems very high though, so if the voltage on the two caps w/o  any load , circuit off/removed  , stays well below 30V, I would prefer the 35V types, as they will have lower ESR.

Frank

I expected something just above 20V on the input side to the regulators.
But that was a very wrong assumption!

After I installed the 35V Nichicon caps and measured the voltage at the cap directly, I saw about 32 Volts, surprisingly high.

So, I will use new, good brand, 50V, 2200 uF again with a temperature rating of 105°C