Agilent 1670G - I Killed it - Opinions

[CoalCreekPlastics]

Hi all,  I hate that this is my first post (I need to be in the beginner section) but I had scored a reasonably priced Agilent 1670g logic analyzer, intending to get into FPGA's and microcontrollers.  I was running through some self-tests last night and everything was passing and "worked a treat," when I noticed the fans were dusty.  Well me being me, I decided a shot of compressed air couldn't hurt, so while the unit was powered on I shot a a short blast into the power supply fan.  I saw a spark, heard a loud pop and smelled burning and got a little bit of the magic smoke and the unit powered off. 

Did my shooting compressed air in that fan really kill the PS? 

I could return the unit, it has a "90 day warranty" but if I am responsible and it has a possibility of being repairable I don't necessarily want to do that, it was a great price.

Should I try to repair that power supply?  How difficult is that for a beginner that has an aversion to electrocution, is reasonable well equipped and can solder?  The power supply is Agilent part number 0950-3403 which is a Celestica 5011-eaac00-111 500 watt power supply.  Have not yet found any schematics for it.  The only replacement I have found was one for $299 asking price on eBay.  That's more than I paid for the analyzer.

I'm hoping it's just a fuse and a burned up fan but Murphy looms large here.  I planned to take it apart tonight to get the supply out.  I appreciate any input you folks might offer.  Thanks in advance for your time an input.

[helius]

I can see several possible consequences.
1. By forcing air into the fan, you turned the fan motor into a generator and caused a surge on the power rail.
2. The "compressed air" was actually fluoroethylene and it ignited.
3. It cooled the electronics suddenly, which either fractured a component through freezing, or else caused a step change in electrical resistance leading to rapid power consumption.

Clearly more went wrong than just a fuse, but the smoke could have come from a fusible resistor, and repair may be straightforward.

[abraxa]

My gut feeling tells me that the magic smoke wasn't merely the fluoroethylene burning some dust but some electronics dying instead. If you like it so much and bought it for cheap then I'd keep it around instead of returning it. Either you can fix it now after looking at the power supply or you can fix it later, when you gained more experience and don't feel shy around line voltage anymore.

What you did falls into the "lesson learned, will never ever do again" category of mistakes. We all have some of those on our belts

[alm]

Most people here will know the sinking feeling of having destroyed something expensive due to a brief moment of stupidity. I certainly do.

A loud pop does not really sound like just a fuse to me. A warranty claim does not seem fair to me (since this is 95% likely your fault, like making a warranty claim after dropping something and smashing it). How is the availability of service documentation? It may have basic troubleshooting steps like measuring power supply voltages to determine if the power supply is faulty, for example the recommended load resistors to apply if running the power supply unplugged from the logic analyzer. Probably no component-level information like schematics. If you power it up again, watch for more smoke or strange noises, and in that case power if off immediately and switch to a visual inspection.

As a beginner, I would not do any measurements on the off-line switch mode power supply with the mains power applied. Wear safety glasses when powering it up without its case (or close the case before powering up). Make sure it is unplugged from mains and all caps are discharged before you touch it (especially the primary ~320 VDC filter capacitor can be nasty). I would look for obviously burnt components, burnt traces or semiconductors with a hole blown into them (blow hole). If you are lucky, it may work again after replacing damaged part(s) and or repairing damaged traces. If you are not so lucky, it might blow up the parts again after you power it up .

Since you live in the US where shipping may cost less than the price of the unit, you could keep your eyes peeled for a cheap parts unit with a likely working power supply (e.g. powers up, but self test reports errors).

[CoalCreekPlastics]

So I did a quickie lookup of the MSDS on the Ultraduster that I used:

SECTION I - CHEMICAL INFORMATION
Trade Name: HFC-152a
Chemical Name: Difluoroethane, R152a
Product Supplied By: AW Distributing, Inc.
Address & Phone #: 2024 Middlefield Rd., Redwood City, CA 94063

These are sold in Wal-Mart stores so not sure if that's flammable.  It didn't feel like a detonation, just a nice crisp spark/pop and yeah some component burned up, I could smell it.  I didn't get any of the "freeze" effect you get from those kind of dusters.  Can't speak in absolutes of course.

I hope the power supply will give up it's secrets and not the ghost when I pull it later.  I'll post pics this could be a good case for other newbies like me.  I guess you could say I screwed up proper.

Yeah definitely will never do it again, I feel like an idiot.  That's what happens when you let line-of-business software guys around the electronics. 

[CoalCreekPlastics]

Most people here will know the sinking feeling of having destroyed something expensive due to a brief moment of stupidity. I certainly do.

A loud pop does not really sound like just a fuse to me. A warranty claim does not seem fair to me (since this is 95% likely your fault, like making a warranty claim after dropping something and smashing it). How is the availability of service documentation? It may have basic troubleshooting steps like measuring power supply voltages to determine if the power supply is faulty, for example the recommended load resistors to apply if running the power supply unplugged from the logic analyzer. Probably no component-level information like schematics. If you power it up again, watch for more smoke or strange noises, and in that case power if off immediately and switch to a visual inspection.

As a beginner, I would not do any measurements on the off-line switch mode power supply with the mains power applied. Wear safety glasses when powering it up without its case (or close the case before powering up). Make sure it is unplugged from mains and all caps are discharged before you touch it (especially the primary ~320 VDC filter capacitor can be nasty). I would look for obviously burnt components, burnt traces or semiconductors with a hole blown into them (blow hole). If you are lucky, it may work again after replacing damaged part(s) and or repairing damaged traces. If you are not so lucky, it might blow up the parts again after you power it up .

Since you live in the US where shipping may cost less than the price of the unit, you could keep your eyes peeled for a cheap parts unit with a likely working power supply (e.g. powers up, but self test reports errors).

I agree wholeheartedly with you alm on the ethical issue of returning.  I just put that out there in case what I did was harmless and there was another issue that this just facilitated.  It appears that is not the case and this is in fact due to my stupidity.  So let's take the return idea off the table.  I want this unit working again.

There is a service manual with instructions on how to test the power supply, and I will follow your common sense guidance.

[abraxa]

I hope the power supply will give up it's secrets and not the ghost when I pull it later.  I'll post pics this could be a good case for other newbies like me.

Please do upload pictures of the power supply after you give it enough time to discharge after pulling the plug (30 minutes should be enough but you never know, so best to treat it as if it were live). Maybe it's something that's obvious to one of us, like a cracked diode.

[WastelandTek]

Well, OP, we learned a lesson today didn't we?

I wasn't going to wade into this thread, but you won me over with your, increasingly uncommon it seems, sense of honor regarding returning the unit.

Do open it up and "have a squizz"  Some well lit macro shots of the circuit boards would help us a lot.

[CoalCreekPlastics]

Well, OP, we learned a lesson today didn't we?

I wasn't going to wade into this thread, but you won me over with your, increasingly uncommon it seems, sense of honor regarding returning the unit.

Do open it up and "have a squizz"  Some well lit macro shots of the circuit boards would help us a lot.

That's kind of you to jump in.  Thanks.  Boy howdy did I learn a lesson.  In retrospect I don't know what I was thinking.  Good thing it wasn't a much more expensive piece of equipment and it didn't belong to an employer or someone else!

I did "have a squizz" and it's a pretty dense arrangement so I'm going to need to de-solder a daughterboard that's at right angle to the main board (see on the left in the photo.)  I'll re-shoot some good photos with the Andostar, the ones I got with the iPhone were too big to post and didn't show enough.

I did smell something singed.

More to come.

ETA: Any advice on which side to de-solder that connection from, the bottom of the main board or at the daughterboard connection?

[Specmaster]

What is that brown gooey mess in the top left corner of the photo? Looks like a cap may have blown and shot its electrolyte out.

On the topic of returning it, I believe that you could have done and feel justified in doing so. I find it really hard to believe that giving it a shot of compressed air from a can would do that, especially given that the fan was running, unless you held the can upside down it is extremely unlikely that any liquid "freeze spray" shot into the fan and even then the blades would be chopping this up into a fine mist anyway.

I think it was a coincidence that it happened while you were spraying it. Its a strong possibility that either a smoothing electrolytic cap has decided to let go or a mains filter / suppression cap has blown and in so doing the arc has blown the fuse.

[cncjerry]

I think the air blowing was coincidence unless you blew some conducting material in there.  I have to be careful around my metal working machines with compressed air for that reason.  I think a cap or something in the power supply let go coincidental to the air.

You need to find the part that blew and we can tell you the odds of repair.

Jerry

[tablatronix]

Does your spray say non conductive?

[amyk]

So I did a quickie lookup of the MSDS on the Ultraduster that I used:

SECTION I - CHEMICAL INFORMATION
Trade Name: HFC-152a
Chemical Name: Difluoroethane, R152a
Product Supplied By: AW Distributing, Inc.
Address & Phone #: 2024 Middlefield Rd., Redwood City, CA 94063

These are sold in Wal-Mart stores so not sure if that's flammable.  It didn't feel like a detonation, just a nice crisp spark/pop and yeah some component burned up, I could smell it.  I didn't get any of the "freeze" effect you get from those kind of dusters.  Can't speak in absolutes of course.

I hope the power supply will give up it's secrets and not the ghost when I pull it later.  I'll post pics this could be a good case for other newbies like me.  I guess you could say I screwed up proper.

Yeah definitely will never do it again, I feel like an idiot.  That's what happens when you let line-of-business software guys around the electronics. 

That's a hydrofluorocarbon. Yes, it's flammable. (The old fluoro/chlorocarbons were not, but the environmentalists didn't like the effects they had on the environment...)

[cncjerry]

I blew 'dustbuster' into a cooktop once to clear out a clogged gas igniter hole.  Big mistake as it exploded and I inhaled the flames.  I often wondered if God knew that causing us to inhale when surprised was a good idea?  Flames?  Cold water?  Ha!

But I learned my lesson with the stuff. 

We need to see what blew.  Take more pictures.

[CoalCreekPlastics]

What is that brown gooey mess in the top left corner of the photo? Looks like a cap may have blown and shot its electrolyte out.

On the topic of returning it, I believe that you could have done and feel justified in doing so. I find it really hard to believe that giving it a shot of compressed air from a can would do that, especially given that the fan was running, unless you held the can upside down it is extremely unlikely that any liquid "freeze spray" shot into the fan and even then the blades would be chopping this up into a fine mist anyway.

I think it was a coincidence that it happened while you were spraying it. Its a strong possibility that either a smoothing electrolytic cap has decided to let go or a mains filter / suppression cap has blown and in so doing the arc has blown the fuse.

I thought that was some adhesive/buffer material used in-between components.  There are several deposits of it around the board.

I can't find an obvious fuse...

[CoalCreekPlastics]

I blew 'dustbuster' into a cooktop once to clear out a clogged gas igniter hole.  Big mistake as it exploded and I inhaled the flames.  I often wondered if God knew that causing us to inhale when surprised was a good idea?  Flames?  Cold water?  Ha!

But I learned my lesson with the stuff. 

We need to see what blew.  Take more pictures.

Wow!  Good to know.  More pictures are the plan tomorrow!

[abraxa]

I did "have a squizz" and it's a pretty dense arrangement so I'm going to need to de-solder a daughterboard that's at right angle to the main board (see on the left in the photo.)  I'll re-shoot some good photos with the Andostar, the ones I got with the iPhone were too big to post and didn't show
ETA: Any advice on which side to de-solder that connection from, the bottom of the main board or at the daughterboard connection?

Removing the daughterboard from the PSU for maintenance of the unit seems a really time-consuming (and not straightforward) thing to do, so my guess is that there's something you must be overlooking to get the entire assembly out, including the daughterboard.

[CJay]

What is that brown gooey mess in the top left corner of the photo? Looks like a cap may have blown and shot its electrolyte out.

On the topic of returning it, I believe that you could have done and feel justified in doing so. I find it really hard to believe that giving it a shot of compressed air from a can would do that, especially given that the fan was running, unless you held the can upside down it is extremely unlikely that any liquid "freeze spray" shot into the fan and even then the blades would be chopping this up into a fine mist anyway.

I think it was a coincidence that it happened while you were spraying it. Its a strong possibility that either a smoothing electrolytic cap has decided to let go or a mains filter / suppression cap has blown and in so doing the arc has blown the fuse.

I thought that was some adhesive/buffer material used in-between components.  There are several deposits of it around the board.

I can't find an obvious fuse...

Yes, that brown stuff is some sort of adhesive, not the guts of a cap.

No, spinning the fan with freeze spray would not cause 'a surge on a power rail' that's just nonsense.

it's possible you've cooled some component enough to throw the PWM out of toleracne and made an expensive bang, it's also possible you've got it cold enough to cause condensation then it could be absolutely anything in there.

There will be a fuse, it may be wire ended and covered in heatshrink.

You need to check for shorts on the power devices and input filter as a first point of action, that will give you a rough idea of how much damage you're looking at.

We need more pics of the board, both sides, several angles on the component sides.

[Towger]

Bottom right of the photo appears to show my least favourite mains suppression capacitor.  Check it for cracks in the clear epoxy case.

[oldway]

I would like to take advantage of the subject to tell you what happened to me when I was doing the maintenance of industrial no breaks.

I was blowing the dust from the radiators of the power semiconductors with a jet of compressed air coming from a Nilfisk blower motor.
I was doing this with No-break in operation.

One day I was doing this operation on a no-break of 10KVA and 120V battery voltage.
I saw something flying in the no-break.
I immediately stopped the dusting and looked at what it was : It was an aluminum chocolate packaging  ....

How has it been  there, I don't know...

I had a little retrospective fear because if this aluminum packaging had touched a power circuit, it could have caused a serious accident.

Since then, I desenergize the No-break completely before blowing the radiators. 

[IanMacdonald]

I think the most likely scenario, as mentioned, is that there was a foreign object in there already and your air blast moved it into contact with a live part.

As to repair, depends what damage it's done. You might be lucky and find it's just blown the internal fuse. Bear in mind that If it's taken out the switching transistor(s) it may also have taken out any current sensing resistors in series with these.  Or. driver ICs.

Testing a SMPS generally involves having an (earth-free) isolating transformer and some form of current limiting device. A household bulb will do at a pinch though a variac is better. One thing is certain though, if you replace blown components and test at full voltage but have missed one.. BANG and you start all over again. Thus, a current limiter of some kind is more or less a must for testing.

[CoalCreekPlastics]

I think the most likely scenario, as mentioned, is that there was a foreign object in there already and your air blast moved it into contact with a live part.

As to repair, depends what damage it's done. You might be lucky and find it's just blown the internal fuse. Bear in mind that If it's taken out the switching transistor(s) it may also have taken out any current sensing resistors in series with these.  Or. driver ICs.

Testing a SMPS generally involves having an (earth-free) isolating transformer and some form of current limiting device. A household bulb will do at a pinch though a variac is better. One thing is certain though, if you replace blown components and test at full voltage but have missed one.. BANG and you start all over again. Thus, a current limiter of some kind is more or less a must for testing.

You know I'm really thinking you and the others that suggest a foreign object may be right.  The spark was directly at the terminal connection on the fan side, and it was a short puff of air.  No "freeze juice" came out (you know with the canned air sometimes you get the propellant or whatever that freezes the thing you're dusting"  frozen components just didn't happen here.  I'm questioning if the fan was manipulated enough to cause a charge like that.

After sleeping on this another night, I really just don't think I have the experience to troubleshoot this.  I am still working on the fundamentals.  Power supply with mains voltage just isn't something I am comfortable doing yet.

[WastelandTek]

Are you, by chance, located in King County?

[Specmaster]

I think the most likely scenario, as mentioned, is that there was a foreign object in there already and your air blast moved it into contact with a live part.

As to repair, depends what damage it's done. You might be lucky and find it's just blown the internal fuse. Bear in mind that If it's taken out the switching transistor(s) it may also have taken out any current sensing resistors in series with these.  Or. driver ICs.

Testing a SMPS generally involves having an (earth-free) isolating transformer and some form of current limiting device. A household bulb will do at a pinch though a variac is better. One thing is certain though, if you replace blown components and test at full voltage but have missed one.. BANG and you start all over again. Thus, a current limiter of some kind is more or less a must for testing.

You know I'm really thinking you and the others that suggest a foreign object may be right.  The spark was directly at the terminal connection on the fan side, and it was a short puff of air.  No "freeze juice" came out (you know with the canned air sometimes you get the propellant or whatever that freezes the thing you're dusting"  frozen components just didn't happen here.  I'm questioning if the fan was manipulated enough to cause a charge like that.

After sleeping on this another night, I really just don't think I have the experience to troubleshoot this.  I am still working on the fundamentals.  Power supply with mains voltage just isn't something I am comfortable doing yet.

Then on that basis I would look at returning it for a refund as I really find it too mind boggling to think that you may have blown some foreign object into it, it would still have to get past the spinning fan blades which would be even greater odds than having the winning ticket for the lottery. Personally I'd say it was a fluke that whatever happened, happened at the same time as you spraying compressed air into the fan. I'd be inclined to tell the seller that it just went bang on powering it up and  say nothing else.

[rsjsouza]

I also imagine it was a foreign object violently blown into the equipment. In the past I have seen many fan assemblies with tiny specs of metallic material ingrained into the heavily thick dust - especially on its protection grid. The fan itself did not have enough air pressure to blow them, but a spray could easily do that.

Given there was smoke and the "Ampère smell", I would carefully inspect the circuit that could be in the path of the fan assembly airflow. Pay special attention to tiny cracked transistors (especially the large ones mounted on the heatsink) and diodes - short circuits would have a tendency to blow these first.

As others have said regarding the spray can itself, just keep in mind what kids use to make blowtorches... 

[Armadillo]

Better not classify it as foreign object, could be native resident object like poorly fastened screws or loose wires, conductive spills whatsoever.

Poor seller, but if it will to cut your heart into pieces everyday, then just bite it. Too bad you enjoyed the seconds of operation.
Poor buyer, just return it as non operational then negotiate for some relieves.
 

[TerraHertz]

Definitely do not attempt to unsolder that row of pins to separate the two boards. If you are a beginner at this, you are very likely to damage the PCB traces while doing that. Which then eliminates the option to return it to seller as faulty.

It could have been just dust that arced when you disturbed it, though from your photo it doesn't appear dusty. I had a PC power supply that blew up due to dust accumulation only.
Another possibility is tin whiskers. Those heatsinks look like they are tin plated, and maybe have some tin surface roughness developing. It's an obscure problem, google 'tin whiskers'. It has been discussed a few times here. Tin plated metal surfaces tend over time to grow ultra-fine long whiskers of pure tin. You might have blown one of them loose.

Or, you were just super unlucky and blew an old iron filing or something into the wrong spot.
If you remember where the flash seemed to be, closely inspect that area with a magnifying glass. There should be some evidence of what went phut. Cracked component, bright arc-points on metal, dark flash-metalization on nearby surfaces.

[tablatronix]

Try using a uv light to look for patterns also makes certain things easier to see, fluids, carbon, darkened pcb areas

[Specmaster]

Try using a uv light to look for patterns also makes certain things easier to see, fluids, carbon, darkened pcb areas

Thats a really useful tip, thanks for pointing it out.

[CoalCreekPlastics]

So I spoke with the seller.  I told them what happened and they agreed to split the cost of the unit.  I think that was fair.  So I now own this non-working 1670G for $135.  So I am going to get some additional pictures tonight, it's my "night off."

Thanks for all of your help and input.  I am going to follow as much of the advice as I can.

[CJay]

Well that's a result, good seller and a clear conscience for you as well as a bargain!

[CoalCreekPlastics]

Well that's a result, good seller and a clear conscience for you as well as a bargain!

Best of both worlds, turns out they actually do some repairs on stuff then resell, so they know the drill.  Great case of a seller also acting with honestly and integrity.  I have a lot of positive things to say about the eBay community by and large.

[Armadillo]

I think the Good Seller name deserves a mention here so that the community can benefits in patronizing his shop as well.  Win Win

[Specmaster]

You did well there then. The repair is more than likely to be something quick and simple to do as well so hopefully you'll end up with a 1670G at a bargain price.

[CoalCreekPlastics]

Winnah winnah chicken dinnah!

Had a "second squizz" and was getting some pics and I found a fuse.  Someone posted it was probably covered in heat shrink tubing and they were spot on.  I couldn't see it very well, so I buzzed it out and no continuity.  Then I put it under the Andostar microscope and got this shot.  Looks like it did it's job, and I have a starting point.

I am going to get more pics in the morning and get the hive's opinion on how to get this sucker out.  It appears to be soldered on one side but there is a "sleeve" where it is soldered for lack of a better term, but it doesn't want to come out easily, the other side was in a socket and it just slid right out of that side.

[CJay]

Weird thing to say but that fuse hasn't died 'violently', if it was a catastrohic failure I'd expect the glass to be spattered with metal or even shattered.

Chances are the failure is a minor one, if you're lucky you may even get away with just replacing the fuse BUT you must check for problems before you replace it and power on or you cold end up making matters far worse.

As someone else mentioned the filter capacitors are of a type that's prone to failure and would cause a fuse to blow

[Specmaster]

Once you have replaced the fuse, and looked around for any signs such as filter caps blown etc and corrected anything that you can see, power it up either via a variac transformer if you one or use the dim type approach by connecting a low wattage light bulb (not a low energy type) in series with it to limit any current in the event that there is still a problem. If it powers up fine, then plug it in the normal way.

[andy2000]

It wouldn't hurt to put an ohm meter across the power line input after you replace the fuse.  Also, you can check the resistance across the main primary side filter cap.  If either is a low resistance, you have shorted parts that need to be replaced. 

I agree that what you did probably wasn't the cause.  These things just aren't that fragile.  Probably a foreign object, loose part, or coincidence.  If a device is going to fail, it will often do so just seconds after powering it on. 

Just recently I was working on a computer.  Everything was fine until I flipped it on its side to access something.  Immediately, there was a bang and a flash and the system went dead.  It turned out that a large coil had come loose and rubbed against a heat sink (in shipping) until the insulation on the coil was damaged.  When I flipped it on its side, the coil touched the heat sink and shorted.  Fortunately, it only damaged the coil and blew the fuse. 

[TerraHertz]

I'm starting to think maybe 'coincidence'. That suspect power line filter cap (the one with the honey-colored clear plastic case) really is a highly suspect part. They are notorious for blowing up soon after powering up equipment that has been unused for a few years. By 'soon after' I mean within minutes. Even if they are still good, you should replace them on sight.
Their failure mode is, the plastic case shrinks and cracks, letting moisture into the roll of metalized foil. Then a short circuit forms between the metalization layers, but it doesn't draw a lot of current because the metalization is so thin it just vaporizes around the short. But the arc keeps chewing away at the plastic and metal foil. Usually they emit some foul smelling smoke (and you'd definitely know it if this happened), but it's possible the fuse blew before it got that bad. The 'gently blown by moderate overcurrent rather than a dead short' state of that fuse is what makes me suspect that cap more.

Of course it's always possible for old fuses to just fail for no good reason. If you can't find anything obvious wrong, it's worth trying a soft power-up (light bulb or variac to limit current) with a new fuse.

Anyway, you won't be doing dust-blowing maintenance on anything with the power on in future, will you! Because ambiguity.

[CoalCreekPlastics]

I'm starting to think maybe 'coincidence'. That suspect power line filter cap (the one with the honey-colored clear plastic case) really is a highly suspect part. They are notorious for blowing up soon after powering up equipment that has been unused for a few years. By 'soon after' I mean within minutes. Even if they are still good, you should replace them on sight.
Their failure mode is, the plastic case shrinks and cracks, letting moisture into the roll of metalized foil. Then a short circuit forms between the metalization layers, but it doesn't draw a lot of current because the metalization is so thin it just vaporizes around the short. But the arc keeps chewing away at the plastic and metal foil. Usually they emit some foul smelling smoke (and you'd definitely know it if this happened), but it's possible the fuse blew before it got that bad. The 'gently blown by moderate overcurrent rather than a dead short' state of that fuse is what makes me suspect that cap more.

Of course it's always possible for old fuses to just fail for no good reason. If you can't find anything obvious wrong, it's worth trying a soft power-up (light bulb or variac to limit current) with a new fuse.

Anyway, you won't be doing dust-blowing maintenance on anything with the power on in future, will you! Because ambiguity.

LOL, no, no more "live dusting."  Because...I can count to potato.

I am looking at buying a variac just for this project and because SMPS are a domain and skill I want to own going forward.

I am looking at the filter caps, and trying to figur out how to get this damn fuse out without tearing stuff up.

[CoalCreekPlastics]

Sorry for such a noob question, but how do I get this fuse out of that socket?  Do I need to de-solder it?

[Specmaster]

It looks to me like it is pushed into a cup on a floating wire, if thats the case simply pull it out like you did the fixed cup on the PCB. If its not pushed into a cup then, yes, you will need to unsolder it.

[WastelandTek]

those can be tight as hell

destruction of the fuse may be necessary

[gnif]

Winnah winnah chicken dinnah!

Had a "second squizz" and was getting some pics and I found a fuse.  Someone posted it was probably covered in heat shrink tubing and they were spot on.  I couldn't see it very well, so I buzzed it out and no continuity.  Then I put it under the Andostar microscope and got this shot.  Looks like it did it's job, and I have a starting point.

I am going to get more pics in the morning and get the hive's opinion on how to get this sucker out.  It appears to be soldered on one side but there is a "sleeve" where it is soldered for lack of a better term, but it doesn't want to come out easily, the other side was in a socket and it just slid right out of that side.

I wonder if the sudden freezing of the cold fuse would cause the metal to thermally shrink and break, or become too thin and just simply blow.

[Bushougoma]

TerraHertz beat me to it.

Replace those yellow Rifa X caps they are known to fail short circuit (usually with quite a bit of smoke) and blow the line fuse.

Look at the epoxy body and you will no doubt find fine hairline cracks. These are replace on sight parts it's not if they will fail it's when.

Seeing those on the power supply leaves little doubt in my mind that the air duster was just a coincidence.

Sorry for such a noob question, but how do I get this fuse out of that socket?  Do I need to de-solder it?

That's not a socket it's part of the fuse it's called an axial fuse it's not meant to be separated from the leads. You'll need to desolder it.

[CoalCreekPlastics]

I wonder if the sudden freezing of the cold fuse would cause the metal to thermally shrink and break, or become too thin and just simply blow.

I Seriously doubt that was possible.  We are talking about a micro-burst here and the fuse was behind the fan and case, covered with heat-shrink tubing.

I have the fan under the microscope now examining the are where I saw the spark.  The consensus seems to be the filter caps are suspect.  They are coming out and a variac is being sourced on eBay.  This won't be my last power supply work I suspect.

[CoalCreekPlastics]

Someone was expecting a more dramatic fuse blow.  I pulled out the second fuse I found after it didn't buzz out.
 Here it is.

[CoalCreekPlastics]

So suspect caps and blown fuses sucked out.  These are the before and after pics.  Need to clean the rosin off and I hope I didn't scorch the s**t out of this board.  No traces came up or anything so hoping it's not bad for a ham-handed amateur. 

Lost my bid for a sweet new-ish clean Superior Electric Variac on eBay tonight...sad panda.  Most of them seem pretty old and crusty.  Based on the variac thread trying to avoid the "One Hung Low" brand units.  Don't need a house fire right now.

Now to source replacement caps and fuses.  Thanks for all the help from the forum!

[CoalCreekPlastics]

And last picture set, here are our suspect turd filter caps.  Have not done a close inspection on these yet.  The axial fuses were 10A 250V, the small RIFA caps were 4n7 Y2 and the larger ones were 15n Y2.  I welcome any suggestions on replacements, I have found a few candidates that have the same box style package, but again, I am still learning.

Thanks again!

ETA: I went ahead and got some close-ups of the RIFA caps on each side.  I need to go back over the thread but I thought someone said to look for cracks.  I may have squeezed them getting them out but they all seem to show "hairline" cracks. 

This has been great I have learned so much.  Dave Jones and this forum have enhanced my understanding tremendously.  Just weeks ago I would not have been able to suck these components out!

[Specmaster]

I would say that once you replace these, there is a very good chance it booting up and working once again, well done.

[BradC]

Lost my bid for a sweet new-ish clean Superior Electric Variac on eBay tonight...sad panda.  Most of them seem pretty old and crusty.  Based on the variac thread trying to avoid the "One Hung Low" brand units.  Don't need a house fire right now.

Despite what others may have said earlier in your thread, a Variac is *not* a current limiting device. It will limit current to an extent due to its own internal resistances and in theory lower voltages result in lower current, but if you short the output you'll draw sufficient current to blow the fuse / smoke the Varaic (depending on whether it has a fuse). You can certainly blow up old reforming electrolytic capacitors, faulty mains filters and let the smoke out of power semiconductors sitting behind a Variac. The old "dim bulb tester" *is* a current limiting device and is good practice even if you are using a Variac. It's also a lot cheaper to start with.

[Specmaster]

We used to use variacs at college years ago as a soft start up device. A lower voltage does indeed limit current to an extent, basic ohms law will tell that much but they do have the failing that they can be turned up to much or even accidentally turned to a higher setting than you really wanted. But a bulb on the other hand, has a fixed maximum current let through and with a selection of bulbs that can be swapped in and out of the lampholder, you can adjust the amount of current flowing though to your device under test. As BradC rightly said, a lot cheaper too.

[alm]

I would also be hesitant to use a variac on a switch-mode power supply unless I was sure the under-voltage condition would not saturate their magnetics or defeat the inrush protection. And if the power supply has a proper under-voltage lockout, then it will just do nothing until you wind the variac up to 90 VAC or so.

[CoalCreekPlastics]

Okay I'm going to do some Googling and see if I can't find how the light bulb method is done.  I want to fire this thing up and test per the service manual as safely as possible.  Given my inexperience I think it is very possible to make a mistake and short something.  Thanks again all.

[rsjsouza]

Lost my bid for a sweet new-ish clean Superior Electric Variac on eBay tonight...sad panda.  Most of them seem pretty old and crusty.  Based on the variac thread trying to avoid the "One Hung Low" brand units.  Don't need a house fire right now.

Years ago I have purchased a 2kVA Mastech variac and I am pretty pleased with its quality. I provided some pictures at:
https://www.eevblog.com/forum/buysellwanted/mastechvolteq-variac-comments-on-quality/msg161815/#msg161815

[Strada916]

You just place a 100W light bulb in series to the active or neutral. Or even a 60W depends on the power rating of the unit.

[Specmaster]

You just place a 100W light bulb in series to the active or neutral. Or even a 60W depends on the power rating of the unit.

It needs a normal GLS lamp, not an energy saving one though. Please note that it is normal when using this type of current limiter that on switching on, for the bulb to glow fairly brightly for a split second as the smoothing caps charge up. Then, if everything is OK, the bulb should go and the device your testing should, if everything else is OK, switch on. If so then you can remove the bulb tester and power up normally.

If the bulb remains lit then you still have a problem somewhere that is causing more current to flow then normal to be resolved, the good thing is that because of the bulb in series you should not have done any harm to your equipment.

[TerraHertz]

Someone was expecting a more dramatic fuse blow.  I pulled out the second fuse I found after it didn't buzz out.
 Here it is.

Yep, that's vigorously blown! The only stage above that is where there isn't any glass left either.
So, was that fuse contained in opaque shrink-tube? If not, then there's the flash you saw.
Also those RIFA caps are the death-caps others & I have mentioned. The hairline fractures let in moisture, which kills them. ALL those type caps should be replaced when found.

However, it's quite unusual for them to show no visible sign (or smell, did you sniff them?) of the kind of failure than can blow a fuse like that. Btw, try them with a multimeter. You might find one short circuit. But even if all are 'open', it doesn't prove they are blameless. They can short out, then do much the same as the fuse - burn back to open circuit.

However we still don't know for sure it was one of those caps that shorted. After replacing all these and the fuses, it would still be a good idea to power up with a bulb as current limiter.

[CoalCreekPlastics]

Yep, that's vigorously blown! The only stage above that is where there isn't any glass left either.
So, was that fuse contained in opaque shrink-tube? If not, then there's the flash you saw.
Also those RIFA caps are the death-caps others & I have mentioned. The hairline fractures let in moisture, which kills them. ALL those type caps should be replaced when found.

However, it's quite unusual for them to show no visible sign (or smell, did you sniff them?) of the kind of failure than can blow a fuse like that. Btw, try them with a multimeter. You might find one short circuit. But even if all are 'open', it doesn't prove they are blameless. They can short out, then do much the same as the fuse - burn back to open circuit.

However we still don't know for sure it was one of those caps that shorted. After replacing all these and the fuses, it would still be a good idea to power up with a bulb as current limiter.

Fuses were indeed enclosed in black heat-shrink tubing.  I also inspected the fa and can find no scoring or burn marks whatsoever where I thought I saw the spark.  The caps did have the hairline cracking of course and when it blew there was a smell of burnt something with smoke.    I am unconvinced the fan shorted, I don't know for sure.

Definitely going to replace the death-caps, fuses of course and power up with an incandescent bulb in series to probe and do the service manual diagnostics once I replace everything.  I have already pulled them out with the fuses, I haven't checked the EM filter caps for a short with a "moldymetah," I will try that, sounds like if there is continuity between the leads it's shorted?

[CoalCreekPlastics]

Okay I think I have found replacements, on Mouser for everything, can someone check my work?

Fuses: 10A 250V (Marked F 10A l 250V)

Eaton replacement at Mouser (fast blow): http://www.mouser.com/Search/ProductDetail.aspx?R=AGC-V-10-Rvirtualkey50400000virtualkey504-AGC-V-10-R

RIFA 4n7 Y2 Filter Capacitors:

Original (Have these improved?): http://www.mouser.com/Search/ProductDetail.aspx?R=PME271Y447MR30virtualkey64620000virtualkey80-PME271Y447MR30

Panasonic ECQUL replacement (because all the best stuff is made in Japan) with "close" lead spacing and Class X2, Class Y2/X2 :  http://www.mouser.com/Search/ProductDetail.aspx?R=ECQ-U2A472KLvirtualkey66720000virtualkey667-ECQ-U2A472KL

RIFA 15n Y2 Filter Capacitors:

Original: http://www.mouser.com/Search/ProductDetail.aspx?R=PME271Y515MR30virtualkey64620000virtualkey80-PME271Y515MR30

Panasonic ECQUL replacement: http://www.mouser.com/Search/ProductDetail.aspx?R=ECQ-U2A153KLvirtualkey66720000virtualkey667-ECQ-U2A153KL

OR

This Panasonic one is in stock but I can't tell the difference from above (Description says 275V but in the stats on Mouser it says 250V and it has an "A" at the end of the part no.  ):
http://www.mouser.com/Search/ProductDetail.aspx?R=ECQ-U2A153KLAvirtualkey66720000virtualkey667-ECQ-U2A153KLA

And dammit both the RIFA and Panasonic ECQ-U2A153KL 15000pf caps are out of stock with unknown lead times.  I can get  a lot of 50 of the RIFA 15n's on eBay for $14 shipped but that's like bulk buying turds if they haven't improved.  I wanted to order from Mouser so I could get my Plato SMD well tips for my Hakko 888D at the same time so I can be cool like Dave and solder SMD components I can't even see.

Also what are the transparent plastic sleeves/sheathing that serve as insulation on leads called?  Google was worthless, it can't just be called insulation?

Any suggestions as far as brands to consider for replacements and advice is of course welcome. You guys have been incredible.  I know you'll tell me if anything is "a bit how d'ya do there." It's been fun researching and looking at datasheets.

[rsjsouza]

This Panasonic one is in stock but I can't tell the difference from above (Description says 275V but in the stats on Mouser it says 250V and it has an "A" at the end of the part no.  ):

The manufacturer says 275V.

And dammit both the RIFA and Panasonic ECQ-U2A153KL 15000pf caps are out of stock with unknown lead times. 

This one is in stock...

[CoalCreekPlastics]

This one is in stock...

Okay so I have a similar problem here, this says 275VAC in the description, and 250V in the "Voltage Rating AC:" in the mouser data below.  I can't find this part number on Panasonic's site or data sheets. What do you guys do in that situation?  I don't like to ask this stuff for fear of the "here let me google that for you" thing, I'm pretty self reliant but this is unclear.

[rsjsouza]

This one is in stock...

Okay so I have a similar problem here, this says 275VAC in the description, and 250V in the "Voltage Rating AC:" in the mouser data below.  I can't find this part number on Panasonic's site or data sheets. What do you guys do in that situation?  I don't like to ask this stuff for fear of the "here let me google that for you" thing, I'm pretty self reliant but this is unclear.

Oh, I see what you menat by the part number not found. There is an "A" letter suffix. According to page 74 of their Product catalog, the "A" suffix means it is cut lead instead of "straight".

[Specmaster]

This one is in stock...

Okay so I have a similar problem here, this says 275VAC in the description, and 250V in the "Voltage Rating AC:" in the mouser data below.  I can't find this part number on Panasonic's site or data sheets. What do you guys do in that situation?  I don't like to ask this stuff for fear of the "here let me google that for you" thing, I'm pretty self reliant but this is unclear.

I found a picture of that part number on on Ebay, it is a Panasonic and the photo clearly shows 250VAC on the side of it. If you have the schematic, check the declared voltage on the rail, if its 250V or less then that one will be fine.

[JFJ]

Okay so I have a similar problem here, this says 275VAC in the description, and 250V in the "Voltage Rating AC:" in the mouser data below.

A Panasonic ECQUL capacitor datasheet specifies the 'Voltage Rating' as:  275 V.AC (250 V.AC on CSA C22.2 No.8 spec.).

The CSA C22.2 No.8 specification is the Canadian standard for potted RFI filters, and is equivalent to UL1283 in the United States.

[CoalCreekPlastics]

This one is in stock...

Okay so I have a similar problem here, this says 275VAC in the description, and 250V in the "Voltage Rating AC:" in the mouser data below.  I can't find this part number on Panasonic's site or data sheets. What do you guys do in that situation?  I don't like to ask this stuff for fear of the "here let me google that for you" thing, I'm pretty self reliant but this is unclear.

I found a picture of that part number on on Ebay, it is a Panasonic and the photo clearly shows 250VAC on the side of it. If you have the schematic, check the declared voltage on the rail, if its 250V or less then that one will be fine.

No schematics, I am basing my replacement selections off the old parts.  I hope this one is right and that cut leads aren't more difficult to solder!

[Specmaster]

Hmm, I'm not too sure just what cut leads means but I'm assuming that it means the long leads are cut down a bit. If this is not correct I'm sure that someone will advise us just what it does it mean.

Assuming I'm correct, then as long as the lead goes through the hole on the PCB then will be able to solder just the same with no problems.

[CoalCreekPlastics]

Hmm, I'm not too sure just what cut leads means but I'm assuming that it means the long leads are cut down a bit. If this is not correct I'm sure that someone will advise us just what it does it mean.

Assuming I'm correct, then as long as the lead goes through the hole on the PCB then will be able to solder just the same with no problems.

Specmaster, you were right on, cut just means shorter leads which must be usable for pick and place etc.  They are still solderable.  I have discovered a bad habit in working on this, since I bend the leads at a 45 deg angle to get them to stay on the board, my cut off's leave them at 45 degrees possibly making things difficult for a de-soldering tool!  Fascinating journey this has been. 

I did solder in the replacements, and no joy on powering it from mains with a 100 watt incandescent bulb attached in series, the lights dimmed and then came back up so something is shorted and/or I didn't have the bulb wired in properly.  I need to suck the fuses back out and continue troubleshooting it.

In the meantime I was able to source a replacement power supply, so repair of the original supply will shift to the back burner.  Thank you all for your help, I need to get moved to a new house and I will yank this supply back out!

[Specmaster]

Hmm, I'm not too sure just what cut leads means but I'm assuming that it means the long leads are cut down a bit. If this is not correct I'm sure that someone will advise us just what it does it mean.

Assuming I'm correct, then as long as the lead goes through the hole on the PCB then will be able to solder just the same with no problems.

Specmaster, you were right on, cut just means shorter leads which must be usable for pick and place etc.  They are still solderable.  I have discovered a bad habit in working on this, since I bend the leads at a 45 deg angle to get them to stay on the board, my cut off's leave them at 45 degrees possibly making things difficult for a de-soldering tool!  Fascinating journey this has been. 

I did solder in the replacements, and no joy on powering it from mains with a 100 watt incandescent bulb attached in series, the lights dimmed and then came back up so something is shorted and/or I didn't have the bulb wired in properly.  I need to suck the fuses back out and continue troubleshooting it.

In the meantime I was able to source a replacement power supply, so repair of the original supply will shift to the back burner.  Thank you all for your help, I need to get moved to a new house and I will yank this supply back out!

Good to hear that you have sourced a new power supply, SMPS units can be difficult to work on, give me a linear one any day. I have power problems with a scope, tried just about everything and it still doesn't work. Even purchased a refurbished one from a service agent, still doesn't work. I have put it on the back burner for a while so I can get on with other things and I'll come back to it later.

You made a good point there about bending the leads at 45 degree angle, it does indeed make using a desoldering tool a bit more difficult. It can still be used if you can angle the tool to get the tool over the lead and then as you heat the solder up, you can straighten the lead in the process of wiggling the lead and sucking the solder, using the tool to do the straightening.

[CoalCreekPlastics]

Boom.  Running self-tests.  Further testament that these are simply not designed to be repaired at the component level.  I will however use that old power supply as a learning platform.  Glad the acquisition board wasn't fried!

[Specmaster]

Yes, those switch mode supplies are tricky to work on, nice to see it working again. Glad that the replacement psu worked out for you, unfortunately it didn't for me, the fault that knocked mine out in the first place is still there. Circuit resistance readings all seem fine, so for me its back to the drawing board. 

[CoalCreekPlastics]

Yes, those switch mode supplies are tricky to work on, nice to see it working again. Glad that the replacement psu worked out for you, unfortunately it didn't for me, the fault that knocked mine out in the first place is still there. Circuit resistance readings all seem fine, so for me its back to the drawing board.

Good luck to you Specmaster I wish I had more experience to lend a hand.  Thank you again for your help and participation.  And that goes for all of the board that jumped in to help!

[Nemesis1207]

Bumping this thread because it's the best source of hardware info on this system from google. I recently got myself a 1670G as well, and wanted to pre-emptively deal with the RIFA caps and generally get the unit future proofed. One thing I discovered - there are actually 5 RIFA caps in the power supply, not 4. We have these four near the back of the unit which were shown previously:


There's also a 250V 1nf Y2 cap towards the back of the main board:


I had ordered new Y2 caps to replace the four previously mentioned, but I didn't have one on-hand to replace this fifth one, so I removed it and left it out. I've ordered a replacement which I'll fit when it arrives. I didn't put RIFA caps back in, I went with TDK ones with the blue plastic outer housing. Here are the caps I used to replace the originals:
2x https://au.mouser.com/ProductDetail/871-B32021A3472M
2x https://au.mouser.com/ProductDetail/871-B32022A3153M
1x https://au.mouser.com/ProductDetail/871-B32021A3102M
These have the same important characteristics (with slightly better voltage rating), while having compatible sizes and lead spacings to be easy drop-in replacements.

If you're doing this work, it's also worth noting there's a CR1225 button cell lithium battery for the main board which is worth replacing while you're in there.