really messed up HP3310B repair probability? *revived*

[coppercone2]

Ok so I have a HP3310B  that had a long ride tumbling down a steep road (it rolled about about 450 feet on a STEEP hill), having fell off a car roof about 5 years ago. It has been sitting for a while pissing me off (in a fucking rush in the morning) and I Have decided to take a look inside of it. I meant to take it to work to test something out because engineering was not really funded but it did not end in my favor.

It looks like the mechanical switches are intact. there is a few transistors that show over heating on a thermal imaging camera, I think I traced it down to 2 blown transistors. I think that happened is I turned it on after getting the dents out but a card edge connector broke through the edge and misaligned the card. It looks like the power supplies are in current limit mode. When I follow the manual the 25V rail test says its in Ilimit (0.7V) but the -25V rail has 1.1V on the junction (so its like -16V) so I dunno whats going on there.

Do you think its possible there is no damage other then the short that happened during power on after a 450 feet bounce roll? I thought to replace the over heating transistors. The only other parts that are suspicious are some hot resistors but I think they are being driven by the shorted transistors. The hottest part is 105C but its a big ass power resistor that still had the correct 39 ohm value, its connected to a transistor that is hotter then all the rest.  The voltage rails look to all be pulled down from 25/10/-10/-25v to lower voltages . I want to say the blown heat sink power transistor had shorted leads because the card it was on slid and i did not notice the leads were shorted before I turned it on, so I think the base got HV on it.

It is tempting to throw away like some other HP devices that I do not like but this unit is actually pretty cool and it might make me a little sad.

Do you think based on a card being off by 1 position in this unit there could just be a few parts that died that can be replaced and make it functional again? or is there other likely damage from the grievous beating it took?

[Gyro]

That's a shame, the HP3310A/B are nice little function generators.

The one thing you have in your favour is that the service manual is still readily available (thank you Keysight!)... https://www.keysight.com/en/pd-3310A%3Aepsg%3Apro-pn-3310A/function-generator

The big possible killer is that it contains a pair of Tunnel Diodes which are almost impossible to source. If those have been damaged - or you bugger them trying to test them or with ESD, then the unit is probably scrap.

You need to go through the manual methodically rather than powering up and looking for the hot bits. It sounds as if you're just making it worse at the moment.

[SilverSolder]

This sounds like a challenging repair!  Pretty much anything could be wrong after that unfortunate mishap.

I would start with a magnifier and inspect all the boards thoroughly under a bright lamp - looking for cracks, larger components lifting pads or track, damaged connectors, that kind of thing.

It is normally super helpful to have another unit like it available to compare with...

[coppercone2]

yea a nice auction ended like 30 min after I posted this thread I thought I made it worse but on the other hand my new DDS systems pretty much outclass it, it would be useful because its small on a small power electronics bench or something

Can you destroy those tunnel diodes with 34401A diode tester? I am working on a ESD mat and just probing,

I decided to fix a few more pieces of equipment, so far the thermal camera has been bang on on finding faults on the sorensen dual psu, just replaced the hot IC's and I got both channels working. (these work just fine with a different op amps).

105C might not be bad, functional sorensen units have a resistor running normally at 130C with no faults. When I looked at the PCBs from what I can tell their not damaged,


Pertaining to the unit:
Any idea what the replacement for RCA 7634 is? and rca CA 7629? They are on the PCB but heatsinked to the chassis and its a PCB that slid and the leads look bent, one of these guys is getting mad hot, so I will defiantly need to replace this.

is this table accurate?
https://www.schematicsforfree.com/archive/file/Components/Semiconductors/Transistors/Transistor%20Equvalants.pdf

For sure according to the manual I need a replacement, there are visibly damaged
sp8438
SP8442

unlike the power supplies this unit is extremely unpleasant to work on because of its construction. I know people don't like thermal but I just fixed 2 sorensen units with just thermal with various failures :-/ but I did work enough on PSU design and stuff that I know what it was doing and could see what could be causing what there (like failure tree possibilities for what can be causing heat), this thing is a mystery and much more complicated and its mad old. But given that the rails are being pulled down and stuff is hot, its likely shorted and the heat is not escaping to some hard to find area. I would put a little bit of money on the SP84 parts being a problem. If anyone can suggest a modern replacement I would be happy to try. I need it to stop being a bench a warmer before I go through systematic analysis.

Unless its like a random ASIC amoung normal parts I don't like buying weird replacements from online stockers because chances are with this thing there is more problems and its a wild goose chase.. I would need to take it apart and separately power the boards I think, damn I don't wanna do that. Just randomly nosing around looking at where to get some of these parts makes me think its beyond economical repair for the value it gives me. But again if it stopped making so much freaking heat I would maybe feel different...

[coppercone2]

you come across the 5 point extra credit question about music instruments written with scrambled hieroglyphics on a physics test that is mostly about horses and some stuff related to space and antennas. the test duration is.. the rest of your life, in the twilight zone

great substitute parts recommendations for these old transistors  :
https://www.ebay.com/itm/Fever-Sultry-Scarecrow-Costume-Woman-Sexy-Fancy-Dress/142281888551?epid=137150581&hash=item2120a95327:g:0XAAAOSwMYJfMf~J

[Gyro]

No, don't risk the tunnel diodes on a normal diode test setting, they are very unusual devices that undergo a negative resistance step at a specific forward current. I wouldn't advise even unplugging them unless / until there is some reason to suspect that they are actually a problem in circuit operation.

I just looked through my HP3310 folder and found a reference that 2n2904A and 2n2218 are suitable for the output transistor pair. One thing to check - the heatsinks for those transistors are electrically isolated from the chassis (ceramic washers). It is possible that if a transistor gets pushed in too far, it might contact the end of the chassis fixing screw.

EDIT: I knew I'd discussed this before... https://www.eevblog.com/forum/repair/hp3310-repair-output-power-amp-and-feedback-path/

also... https://www.eevblog.com/forum/repair/hp3310b/

(Link fixed)

[coppercone2]

forget it, I am looking at how bent the circuit boards are and all the cinch connectors are messed up and to make this thing feel like it would not randomly break I would need to spend at least half and replace too many parts that I think might have been mechanically damaged. if it was the last oscillator on earth maybe but I am just gonna scrap it, I have a feeling other components can be ready to go, a 450 foot roll on asphalt is just too much for a side project.. it needs significant panel beating and straitening also.

junked for parts. this is moby dick territory to continue this repair.

dont put anything on your car roof, the pizza boxes are bad too because they scratch it up. its just a damn bad habit.

[Gyro]

It might make somebody a useful parts mule - particularly things like the previously mentioned Tunnel diodes. It would be nice to offer somebody the chance rather than binning it.

[coppercone2]

oh yea I will keep the stuff but I am getting rid of the messed up chassis etc, not going to throw it in the trash outright might get another one some day but I have enough FG currently to not really need that one.

the control knob for frequency is quite nice it needs to be put on something else lol

if you are curious I cut the transistors off that were getting hot and its not a nice unit to fix, they seem fine with diode test. it looks like something else was shorting or there is a distributed short circuit /overload making it not show up on pin point thermals.. I would have totally let it live if it was like 2 pcbs not a box made of four pcb.. or at least if it did not have card edge. there is just too much work. IMO this gets a low repairability score. 4 boards, diode and transistors connected to the chassis at weird angles, mixture of solder and pin wires, etc.  If I took it all apart I would still need to work on a cube or etch my own PCB cable adapters for the cinch connectors. Its not robust enough to handle that kinda damage. + it needs 4 new knobs because they were all destroyed. if it had a cable harness it would be a game changer even with all the soldering.


I remember watching it fall down the road hoping it does not fly into someones parked car wind shield (thankfully low speed limit).

[SilverSolder]

LOL yes you are right, it is probably too far gone to fix, but would be an excellent parts unit if you get hold of a non-working unit from elsewhere.

[coppercone2]

i dunno if i wanna start a 3310 obsession , with how its built ima end up with 2 non working units.

I suppose I can make a antistatic bag labeled 3310 parts

[vk6zgo]

oh yea I will keep the stuff but I am getting rid of the messed up chassis etc, not going to throw it in the trash outright might get another one some day but I have enough FG currently to not really need that one.

the control knob for frequency is quite nice it needs to be put on something else lol

I have a dead one in the shed, & I was thinking about using that setup for the tuning of a regenerative radio receiver

if you are curious I cut the transistors off that were getting hot and its not a nice unit to fix, they seem fine with diode test. it looks like something else was shorting or there is a distributed short circuit /overload making it not show up on pin point thermals.. I would have totally let it live if it was like 2 pcbs not a box made of four pcb.. or at least if it did not have card edge. there is just too much work. IMO this gets a low repairability score. 4 boards, diode and transistors connected to the chassis at weird angles, mixture of solder and pin wires, etc.  If I took it all apart I would still need to work on a cube or etch my own PCB cable adapters for the cinch connectors. Its not robust enough to handle that kinda damage. + it needs 4 new knobs because they were all destroyed. if it had a cable harness it would be a game changer even with all the soldering.

I remember watching it fall down the road hoping it does not fly into someones parked car wind shield (thankfully low speed limit).

My one has some sort of PSU problem.
I haven't looked really closely, but it appears to have a multi-layer PCB-----for a Power Supply!
I toyed with the idea of packing it up & sending it to you, but the postage/freight would probably too high from Oz.

[coppercone2]

thats not it, my blood pressure is the problem. yesterday I spent all working on 2x sorensen power supplies and today I was working on my 2nd rack mount. I had to find all sorts of parts and do all sorts of testing and I was ok but as soon as I start working on this 3310 it makes me feel crazy because of how its built. I almost bought another one yesterday but deep down I think I knew what would happen from betting on those hot transistors (a red herring).

It is premium and the distortion spec is impressive but just even probing this thing is a pain in the ass because of the construction. You need alot of patience for this unit I have not run across anything quite so bad.. even the older HP vacuum tube units are easier to work on.. this thing is in another league

I trashed a Saturn? brand dynamic signal analyzer for this reason.

Its getting hard to imagine to fix this with a return on financial investment. Perhaps I should for posterity? I fixed more modern AWG before but that to me was easy because at least i could identify the parts and figure out what it was doing (50MHz wavetek unit). But that thing was in line with how I would design a circuit myself.. it seemed logical because I went on a long endeavor to design some AC/RF control gear before with modern techniques. Maybe I will print out the schematics and have one more look to see how much the cards interplay

i will revisit this later.

Maybe 3d print some parts to fix the edge connectors and glue that together with some dp8810 so the cards just sit right first and check the chassis for bends with a mini starett square when its not occupied and do a full disassembly. i cut the transistors nice so i can put it back together, check the chassis elements, regrease the thermal junctions then have a good look at the PCB.  I was really hoping the dmm would show a problem between them when its disconnected

lost my cool there, also found a torn wire that might have been shorting something out on one of the coaxial connectors, i thought it was a bodge wire to the board but it was sitting on the leg of another part :-x

i don't envy accident investigators.. I am starting to think i can pay off the heat sink clips by desoldering every active component and testing it on a bread board  .. need to stop thinking about return on investment lol, if i could fix this and the a big fried 30A supply it would be some serious cash that i would never sell.


I think I had a change of heart because I setup another rack mount in the last weeks and when I am finally clearing the wire shelves off the heavy RF & MW crap the lab is starting to feel workable again so i can do some crazy shit like this. i think clutter was making me frustrated , maybe i can even setup a fluke hydra to work on this... I think I will take a crack at the 30A sorensen because someone on the forum fixed one before so this is the last thing to work on.

i know something bad is gonna happen if i start this without being tidy

[vk6zgo]

The horrible layout is why I gave up on it & chucked it in the shed!

[coppercone2]

actually I want to formally apologize to HP, because I was down in the basement and my old fluke 9something scopemeter caught my eye. Fluke can out do anything in terms of being non repairable. The thing is like the battlestar galactica after a nuclear missile barrage hit, its literary so fucked up that the parts are held together inside of the rubber casing. I fixed it once or twice and the  cracks keep propagating, I even got premium plastic friendly glues the last time I fixed it. I think it has like 4 major cracks. I think 60% of the screw fasteners disintegrated and most major plastics on the bottom part are cracking. I keep it around because I thought I could give a go at plastic welding or advanced glue techniques repair one day. Realistically I think I will put it inside of a metal project box one day to make a scope laptop because I am pretty sure that housing is going to completely come apart eventually.

The odd part is that it works, but the design is lethally bad because once the negative banana jack came off the PCB, so you could plug in a multimeter and if you don't test the probes for continuity it would plug in fine but there would be no connection made, give you a low voltage on mains. The banana jack is press fit into plastic that is no longer tight enough so the insertion of a probe broke the solder joint. I fixed it by resoldering and gluing it into place with plastic safe threadlocker but I don't trust it for any safety work. It is press fit into a smooth bore.. so even if you do continuity it could come loose afterwards depending on how you are holding it, if the solder joint cracks, so I recommend for safety using both a no touch pen and a multimeter. I don't handle that one when its plugged in anymore, i set everything up and do a test then disconnect power then play with the meter settings since I dunno what can happen inside of it.

[coppercone2]

took it fully apart,

the cinch connector had severe internal damage that was invisible without disassembly. Maybe there was a no-contact I missed. I want to fix it. I need to get new machine pin sockets to put on the wires because the wiring is really shoddy, I don't trust some of the wires.

The chassis is bent up too.

[coppercone2]

I think I see that each of the edge connectors has a via on it, I think that I can carefully solder small diameter wires to all of these so I can lay out all the PCB on a esd flat surface connected without the box for easy probing, I am hesitant to buy replacement card edge connectors because they are kind of expensive, my investment in this unit was $50... and there is too many weird transistors and diodes that are difficult to replace. I just want it because its small. If it works fully I will fix the mechanical damage.

I desoldered one of the hot barrier diodes and it seems fine because I was curious if it was a damaged part or if it just looked weird (its big and kind of weird looking).

What the hell are the parameters of these weird diodes in this thing anyway? (never heard of a hot barrier diode)

Are they all just low capacitance diodes?

[Gyro]

It's hot carrier, not barrier. They're basically fast, low capacitance small signal Schottkys (remember how long ago it was designed). People have successfully replaced them with 1N5711s without needing adjustment. It's the Tunnel diodes that are the unobtanium ones.

[coppercone2]

Ok, so what is the safe way to test them? They happen to be the last part on my ESD mat by chance, I see its 80$ each for a replacement. I almost don't wanna do the work until I know they are OK.

Red and blue wires going to the diodes into an aluminum bar grounded by the chassis.

https://w140.com/tekwiki/wiki/Tunnel_diodes

It just says heat, so I should be careful not to break the wires. So if I try a diode test from a 34401A they can break?

edavid wrote here
https://www.eevblog.com/forum/repair/tunnel-diode-polarity/


I don't see any reason why a DMM diode test function would hurt a tunnel diode, unless it were a very tiny low current device (in which case it would be useless in a pulse generator).  It's more likely that 30-40 year old surplus Russian electronic parts are poorly tested/flaky/unreliable.

Given that they are power diodes on a giant heat sink, I think ESD is the only threat ? So they should remain on the mat.

So I think this is how it goes, I measured it
They are roughly reading 0.6V on the diode test. They are reversed polarity to the bus bar. I think the diodes are identical but they are reversed inside of the junction so that you can install both of them in the same way or so they sink heat properly.

One is 0.64 and the other is 0.63

I could not resist. 

So it seems this is fixable. And I would not be suprized if you can put some kind of other part network to mimic that diode there because there is room in the chassis and power supply rails to spare, maybe thats wishful thinking to make a replacement diode from something.

So I think I am good to work on this since those would be the biggest traps?

 I can't handle that much stress not knowing if they were shorted lol, at least I think now it has a good chance of working. Looking at the pictures it seems that there is some kind of other application for these diodes at high frequencies where they are mounted to the PCB in some kinda on board blob thing, where most of this caution comes from, this looks like pretty robust.

If the wire breaks during assembly I will crimp a ferrule with some strands of wire as padding around the tip of the diode so I don't heat it again, then again I would like to see if a faberge egg is bouncy, this thing aint that priceless !!

And at least if that test broke them, you will know what it does so you are not tempted with your own diodes to see what they read.

[Gyro]

Are you actually bothering to refer to the manual schematics at all when identifying these diodes? 

The two big diodes on the Aluminium bar are power Zener clamping diodes, not the Tunnel diodes. These Zeners were introduced in an ECO (which you can find it in the manual) to improve protection of the output stage against inductive transients and back-driving. They are on the output stage schematic.

The Tunnel diodes are tiny, with tiny junctions! They are little gold things plugged in on the top board (irrc), their component idents are clear so you can identify them on the schematic. There's no reason to remove them unless there is actually a problem with the generator actually failing to toggle (change ramp direction).

Diode testing on a 34401A won't tell you anything useful, there is just no point. As I explained before, Tunnel diodes undergo a sharp negative resistance change at a certain forward threshold current. A standard DMM diode test simply will not replicate these conditions or see the transition. The best way (if necessary) to verify them is to very carefully scope them in actual circuit operation. Their negative resistance transition time is very fast (their purpose in triggering), an accidental short while probing will kill them. As I said before, they are fragile. Given their cost and scarcity there is absolutely no point in manhandling them, potentially stressing their leads and damaging their hermetic seals, unless they are actually failing to operate as expected.

Anyway, luckily you seem to have been carefully measuring a pair of bog-standard Zeners rather than the Tunnel diodes, so no harm done! 


P.S. There is absolutely no problem soldering to the zeners, you are more likely to crack their lead seals with the mechanical force of crimping.

[coppercone2]

I think I looked at pictures someone posted and the diodes were featured in a big picture so I thought they were a highlight

the schematic is also super horrible because the PDF is all over the place, I never looked at it too much because I need to do a giant quilt job with it

I see it now, next to TP9 there is something like you describe

I must of missed it because the capacitor was bent over hiding it

Lets just say with all the oxidation, rubbed off text and shitty schematics, I am not having fun

I bought replacement card edge connectors for 15$, maybe if I assemble it on the work bench without the chassis in something I can actually probe with a few supports I can find out if the tunnel diodes actually work with the test point.

[coppercone2]

https://groups.io/g/TekScopes/topic/7637086

so it seems if you do this, it should read close to a short circuit
https://w140.com/tekwiki/wiki/Tunnel_diodes

They don't seem too scared but this component looks much more like something that I would expect to break rather then those giant zener diodes.

so its the first semiconductor based randsomware.

This reminds me of the episode of Seinfeld with Elaine and the sponges

who is lining their sock drawer with boxes of these diodes?

[Gyro]

The GE Tunnel Diode Manual referenced in the TekScopes thread is useful, I hadn't seen it before... http://www.archive.org/details/GeTunnelDiodeManual

Yes, if you had measured them I think you would probably have tossed them as faulty, they are weird devices. Quantum tunnelling is a curious effect (see the Tunnel Effect section in the GE manual). Also note the low forward current limits for most of the parts listed.

[coppercone2]

Hmm, that is interesting, since its a quantum device. Perhaps this is the most unique instrument I have then, next to the YIG spheres.

For some reason it went over my head that it is quantum tunneling, I thought it was something else like avalanche.

Don't freak out now but when I think hard back to 2016? I MAY have already measured this diode when I first took it apart after it rolled down the hill and failed to turn on, I do recall something reading ~0.04 V when tested, but at the time I assumed there is a parallel circuit or that it was damage from a short circuit. It is slowly coming back to me what I did before giving up, but this golden component seems some how familiar. perhaps I bent the capacitor over it to see if there was like some kind of short circuit related to fracture. I had much more interesting things going on so I threw it on a shelf for 5 years. I think I expected there was a component that would cleave in half when gently pushed, but it seems much more robust then anticipated.

Not that I did not learn about it before but it was just a serious annoyance up to this fact, it is now interesting. No one ever gave me the bright happy world of tunnel diodes, more like IF FREQUENCY IS LOW ENOUGH COMPUTER!!! seemed to be the trend in education.

Now that I think about it I do have a forgotten Tektronix 284 tunnel diode oscillator. I ever paid for the GR to BNC adapter, but I think free electron drowned it with his pulse circuit.

[coppercone2]

Ok, I used my new solder extractor to put in new cinch connectors, now I got it put back together and powered up and

It is working better then before,the wire ripped on my normal level, but the low-level shows the appropriate waveform, however the top of the waveform is clipped

The voltage rail is also of (6V instead of 10V).

Can anyone verify that if its showing all functions besides the top of the waveform (appears to be clipping) that its not the fancy diodes?

I think I found some bad transistors, there are 2 little metal cans running at 80C and they stand out like crazy from anything else on the PCB. They are on the power board near the relay, I thought it might be dragging a rail down with a fault of some kind.

Also, does anyone know what the crimps that are used in this? They appear to be round pins. I can invest in another crimp tool because I do like this style of crimp. Is it aviation or something? Q5 & Q6 on A5 PCB SS2077 and SM3197, figure 7 of schematic, on the phase circuit

They call this a 'limiter'. Can that be it?

Ok so I see the supplies are 14V and 6V, the difference would make +-10V instead of -6 +14. hmm so balance is upset. It looks like the adjustment won't cover it. SO the regulation is bad. It is low noise though, so I wonder what can be causing too high a voltage.

And +28 and -20..

damn these rails are freaking messed up.

Can those giant pass transistors be the problem? Q21 etc? There is 4 of them on a bar. That took severe damage but when I clipped them off and measured them, i got something akin to transistors.. but I don't trust them one bit. The chassis moved but the board did not do they were bent to all hell.

I am not sure what to do because.. 4 bad rails is a bit much.

Is the problem with the high voltage rails contained to the transistors? I don't know what would cause them to disregulate high other then damage. They are still insulated.