HP 33120A 15MHz Arbitrary/Function Generator - Burned Output Amplifier Resistors

[SMB784]

Howdy everyone, so a few days ago I found a nice arb/function gen for the lab I work in at the surplus property office.  It was an HP 33120A 15 MHz arb/function generator, and it is in quite good condition, at least outwardly so.  A cursory inspection of the internals looked pretty good - the unit was clean and looked like it had been taken care of pretty well (see below):



I hooked it up and turned it on, and connected the output to my SDS1104X-E ("upgraded" to 200MHz bandwidth) to test the output from the main output BNC terminal.  What I got was a flat waveform with a slight DC offset and no oscillitory behavior whatsoever.  I verified this over the entire frequency band of the instrument (1 Hz to 15 MHz), and on all channels of my scope:



I first tried to remove the offset, but found that the offset was set to 0 VDC (upper left panel in image below).  I figured maybe the amplitude was set too low and was unable to be viewed against this unremovable offset without the waveform being off screen, so I set the frequency to 15 MHz and adjusted the amplitude from its lowest setting to its highest setting, yielding no change in the waveform except for a higher DC offset that corresponded to the 6 attenuation ranges:



Now that I was certain this behavior was not user error, I took a look at the main board more closely in order to find a possible culprit for this behavior.  When I glanced over the main output amplifier section, I saw something that made me sad: someone had let the magic smoke out, and in a big way!  Four of the resistors in the main output amplifier section had exploded! (Here's a link to the service manual where I found the schematic)



This is clearly the culprit behind the missing output, although it is heartening to see that the attenuation relays still work.  Hopefully this has not damaged any of the downstream power amplifiers, but I am certainly not optimistic about this, and I am at a loss as to what could have caused this level of destruction.

The sync output functions flawlessly, and I tested it over the whole range of frequencies with a 50 Ohm terminator.  The rise time appears to be beyond the ability of my scope to measure:



Hopefully this can be fixed, I am thinking I will just solder in replacement resistors and try the output again.  Any advice would be greatly appreciated.

[bd139]

That's a nice little mess. Got to find out why that happened rather than just replace or you'll just smoke the replacements as well. I'd check all the surrounding components and bias voltages as well.

Then scrape out the burned bits and see what is left.

Then decide what to do.

Not sure how many layers those boards are but it looks like a relatively doable repair. The vias close to the resistors may be damaged as well.

[SMB784]

That's a nice little mess. Got to find out why that happened rather than just replace or you'll just smoke the replacements as well. I'd check all the surrounding components and bias voltages as well.

Then scrape out the burned bits and see what is left.

Then decide what to do.

Not sure how many layers those boards are but it looks like a relatively doable repair. The vias close to the resistors may be damaged as well.

The board appears to be a simple two layer board, thankfully.

[duak]

I would check all of the components from Q709 & Q710 to the right before powering up again.  Four resistors charring like that indicates a fair bit of current in a circuit that shouldn't have that much.  I suspect one of the other resistors to be open or have a change in value.  I also expect at least one of the transistors to be shorted or open.  This should be easy to test on the diode range on a DMM.

With any luck, Q713 & Q718 will be OK.  I haven't cross referenced them, but my experience with function generators tells me these parts can be hard to get.  All the other parts are pretty much jelly beans.

Cheers,

[SMB784]

I would check all of the components from Q709 & Q710 to the right before powering up again.  Four resistors charring like that indicates a fair bit of current in a circuit that shouldn't have that much.  I suspect one of the other resistors to be open or have a change in value.  I also expect at least one of the transistors to be shorted or open.  This should be easy to test on the diode range on a DMM.

With any luck, Q713 & Q718 will be OK.  I haven't cross referenced them, but my experience with function generators tells me these parts can be hard to get.  All the other parts are pretty much jelly beans.

Cheers,

I was worried about Q713 and Q718 as well, but it turns out that they too are jelly bean parts (see relevant links on part numbers earlier in this sentence).  They are a little bit expensive, perhaps $25 for the pair, but at least I don't have to worry about finding some unobtainium part if it turns out the power transistors are toast.

[SMB784]

I would check all of the components from Q709 & Q710 to the right before powering up again.  Four resistors charring like that indicates a fair bit of current in a circuit that shouldn't have that much.  I suspect one of the other resistors to be open or have a change in value.  I also expect at least one of the transistors to be shorted or open.  This should be easy to test on the diode range on a DMM.

With any luck, Q713 & Q718 will be OK.  I haven't cross referenced them, but my experience with function generators tells me these parts can be hard to get.  All the other parts are pretty much jelly beans.

Cheers,

I tested the voltage supplies with the power on, and they all read within their tolerance.  If there is a short, its low enough amperage that it isn't loading down the +/- 15 supply.  This makes me think that perhaps the problem here was someone shorted the output to ground and just left it on for long enough to toast the resistors in the output section of the board.  I am currently investigating other components around these resistors, and I will report back my findings soon.

[andy2000]

Someone probably connected the output to a voltage source.  I repaired one with a damaged output a few years ago.  All of the damage was confined to the output circuit with no difficult to find parts needed.  I would check all of the transistors in that circuit for shorts and then replace the resistors. 

[SMB784]

So I used my diode test setting on my multimeter to test all the front end transistors, and found that the two main output transistors appear to be blown.  On the Q718 transistor (a 2n5583), the CE junction measures 19 both directions and has a 0.018V drop in both directions.  From my own meager understanding of transistors, this voltage drop should be much higher (if not read as totally open in both directions) for this junction.  For the Q713 transistor (a 2n5943), the BC junction measures 5 and has a 0.005V drop in both directions.  The resistors with a red slash through them are the fried resistors from the images in my first post.



I havent tested their associated zener diodes yet (not sure how to test those), but I will be ordering some replacement transistors to swap out the blown ones.

Any advice is greatly appreciated!

[duak]

Yep, the output transistors are bad.  Any low resistance, especially with both polarities indicates a shorted junction.

The zener diodes can be tested in situ with the output transistors removed; just measure the voltages of the +15V and -15V nets to circuit common with the unit's power on.  If the particular voltage is low, the zener is open.  If the voltage is high, say 18 V, the zener is shorted or is leaking.  Your component tester or DMM should be able to confirm this with the unit's power off.

Have you checked the condition of resistors R731 to R735?  One or more is probably bad too.

Cheers,