If this smps has phase loss detection, you may want to look for a cheap 3 phase motor and a few caps, maybe put together a small rotophase for this sort of thing. Ya, the third phase isn't perfect...but good enough to do the job.... And quite well too.
Problem with running on single phase is ripple will be much higher - if it has PFC this might cause problems even at fairly low powers. On a PSU like this it may also have some sensing to make sure all phases are present.
+mikeselectricstuff Yes it would suck if it has a sensor. In that case I can't do it without a 3 phase supply. Worth feeding in the single phase and trying though.
+EEVblog Actually, come to think of it, the small three phase wires (near the MOV's) going over to the smaller PSU board looks suspicious. No reason to do that really unless that board is doing some sort of phase detection.
+mikeselectricstuff I work with a similar sorensen unit at work, 'only' 3kw though. Its happy with whatever you feed it, 240V or 208V single phase? Just stick line and neutral on any 2 of the line inputs. 208 3 phase? Use all 3 obviously. 190V@30A still provides 5700 W, and those heatsinks dont look upto snuff for 700w dissipation. Guess its time to RTFM. Like the first user should have 8) Mind the ground levels of the DB-25 for the programming input, no metal shell on it, so it might be a common ground with the output - Thats how it is on the above mentioned Sorensen
A little thing to note, fans don't blow air in a linear path, rather fan it out in a cone, pun unintended. So the position of that heat sink is ok, since the air would be pushed right over it. You can see this if you mount your fans back wards in your PC case and have them blow from the back to the front. The dust will be in a truncated cone pattern where visible over the MoBo.
+rhkips I think there's somewhat of an analogy between 3-phase and a straight 6 internal combustion engine. From what I've read, in a straight 6 all your secondary forces cancel out and that makes the engine really smooth. Also, from what I read about three phase: if you balance it properly, in a star configuration (ground in the middle), the current on the ground wire is zero.
+plonkster The ground connection is only ever used for protection. This supply has 3 phase delta input as opposed to 3 phase wye. The wye configuration includes a neutral which carries no current if the 3 phases are balanced. This supply has no neutral.
+Waxing My Knee: ᶃᶉḁᵱḥᵻḉ ᵭᵻᴤᵻᶃᶇ You can see it from the size of those heatsinks. If 5KW smps has 90% efficiency, thats 500W of heat now lets divide that by 2 (2 power boards in parallel) and we get 250W of heat for those small heatsinks which wouldnt last wery long so i think its probably going to be much more efficient, like 95% or more otherwise there would have been much bigger heatsinks.
My 2,5kW smps with 50V output have about 93% efficiency. As we know, higher output voltage = smaller current = smaller losses and higher efficiency. So I also think this smps have at least 95% efficiency.
This unit has been qualified to the MIL-STD-810 vibration requirements, which explains the silicone to mount those capacitors you pointed out early in the video. This is a tough power supply!!
+Deer Viehch Well if you watch enough, you'll learn a lot of them as well especially if you're just talking about "this is a resistor, this is a diode, this is a capacitor " You can do most of that without any real knowledge. But with individual parts, you'll start to recognize patterns. 4 digit part numbers beginning with 78 followed by two numbers are almost always a linear voltage regulator with the output voltage being the last two numbers. Ones beginning with 79 have the output voltage negative in respect to the ground. 74xxx are mostly logic chips, 4xxx are also mostly logic chips. Products by Altera, Xilinx, and Lattice are usually programmable logic. Atmel will usually be microcontrollers and flash memory. Microchip is frequently microcontrollers but they also have a lot of other products like chips for interconnecting to different things (CANbus transcievers, IO Expanders stuff like that)
stonent **head* *explodes** But seriously, the most advanced electric thing i've ever done is build a game where you have to contact one of 16 solder spots in a 4x4 matrix in the right order to complete it, but the order was integrated and not random ;_;. also it was professionally(not at all) build on a wood rack :D
Back in 2005, Lambda generously donated 4 big power supplies to our arts group:24v 600w bridgable units which very reliably ran our 50 SmartMotor servo actuators all over our sculpture! They are indeed remarkably well constructed 3-phase units. I bypassed all the explody drama by triple-checking the available power before fitting the units up!
Great video! You certainly know your stuff. I work for TDK Lambda and repair the 10 & 15kW Gen's from 10V 1000A to 600V 25A. Those 1000A 3U units really are beasts 👍
That dust near the input fuses sure looks to me like the dust from when an HRC fuse is broken. Either from catestrophic overload beyond the capacity of the HRC limits, or from physical damage.
... 400V in three-phase, yes, between the phases! If you connect the same phase to all three inputs then you have zero volts between the phases! And there is no place connect the Neutral, since the fourth terminal is PE. This beast runs in delta-configuration!
+John McFerren just that there are only four connectors on the back for L1, L2, L3 and PE. You cannot connect N to PE in an equipment like this and thus 3-legged star is not possible.
+Alexander470815 yes, you can probably connect L to one of the L1, L2, L3 and N to any other of the three and this power supply should power up. Just that Dave commented in the film that he wanted to connect his single phase to all 3 input phases, which works in a 3L+N device, but this unit lacks the N input.
+Uwe Zimmermann You can run a wye device without a neutral because if the wye device is ballanced no current will flow in the neutral. Since this is one device it can easily be balanced. Many large industrial motors do not have a neutral. If you want to single phase this device you simply attach a Neutral to one of the Line inputs because it just uses a 3phase rectifier so you are just using one of the three rectifiers. It will certainly increase your ripple and reduce your max output power.
I have seen 220V rated soviet equipment operate on 245VAC without problems. That is surely a nice PSU. Edit: The negative track under the bridge rectifier is quite close to the other pin at 25:41.
There are cheap VFD inverters designed to operate three-phase gear from a single-phase supply. You should be able to find a 2KVA unit from China for ~$150. Get one designed for 120V/208V and operate it from your variac. Job done. That in itself would make an interesting video.
Few other interesting things on the output board. Check 24:20 - there are vias on the positive rail (bottom of the board, around capacitor C83), that go to the other side of the board, most likely to double amount of copper for a current handling, but you do not see the vias on the negative side (around C84). Also the common mode choke area, has a footprint for a bigger common mode choke L81, probably for higher current models. Another interesting thing is the footprint for one more capacitor, around configuration resistors, it is even labeled C92 with silkscreen. Very smart modular design. Also the suppression cap C81 look to have two different footprints. The parallelisation of the capacitors C83 and C84, is done using JW81/JW82 jumpers most likely. Also the bus bars are probably an overkill in this 8.5A model. You could easily pass the current with few cables. But they most likely left it there, as it is shared design of other components, and also the busbars act as some form of radiator and current collector for transistors, so they left it there. Also, as you can see the output board is only for 150, 300, 600 V models. The lower voltage and higher current models, must do have different output modules. The connector itself would need to be very different, and the amount of conductor would need to be increased too to handle few hundredths amps or so.
What about using a single to three phase VFD inverter? I've only ever used them on equipment with 3 phase motors in, not 3 phase electronics, so not sure if it would work. I'm also going to assume it would need to be pure sine not modified, but again don't know enough about this sort of thing. I look forward to seeing the repair.
Ouch. Just went to the keysight web site to look up the specs on that supply and the first thing noted is that the voltage input is set at the factory and CAN NOT be changed. Basically, you order the supply and at the time of ordering, specify what input voltage the thing will have available. If you make a mistake, then you need to send the unit back and get a new unit with the correct voltages.
Nearly 247 volts out of your socket? I live in the UK in a well built up city, meant to be 240V. Much closer to 255-260 and can go up at peak times aswell.
Dave, you need the 3 phase input. 3x1 phase will not work because of missing neutral. rectifing a 3phase source to DC is also easyer. I know that certification labs use a 6 phase transformer to create "rimpelless" DC :)
Dave you will need a true 3-phase supply, i.e three separate 240v phases (415v phase-to-phase), each separated by 120 degrees. You won't be able to just parallel the single phase 240V supply up. Lee
+Eviltech Not necessarily because they're probably 600-1000V rated. I'd check the input filtering caps (2.2nF 275V). Even if they are the self-healing ones, their capacity may have gone down significantly. The "middle" MOV probably done it's job, but I'd check the other two rails going to the PFC, since their fuses have blown and there was a significant surge of current going down there to the PFC. There could be some nasty damage done in the PFC on those two input rails... I'd check for shorts on those rails after taking the PFC board out, just to be sure.
Jasenko Cicvara A 1000V rated ones for sure! However the blown MOV is after them, so the surge current has passed trough the rectifiers. We are talking hundreds of amps, even though for just a few microseconds.
You have a point. Sure, It took a considerable current to blow the hell of of that MOV. But, it makes me wonder why the fuse is still intact. with that amount of energy going through it, it should've been blown to kingdom come, like the others. It probably happened that way because the voltage at that phase was the highest at the moment of connection to the power outlet, so the other two phases took out only the fuses. Either way, a catastrophic mistake...
25:07 Nope, it got 2 turns on each side :) Each pass through the centre of the toroid counts as 1 turn. Looking forward to repair vids, will be nice to see it working again
In Europe you could safely use this power supply, but you would need to connect it to a 32A CEE blue plug...since 3 phase power is also 400V. You simply hook up a neutral and phase to any two of the three phase inputs on the power supply...or over in Australia you would hook it up to a single phase 32A outlet....Over in the USA 3 phase domestic power is usually 240V (not 208V since the most transformers are delta wired with one winding used as a split phase to achieve 120V for normal sockets) but you can also hook it up to a single phase NEMA 240V 30A or 50A outlet....using the two ''hot'' wires as they call it. But the normal home power supply is split phase (google it)...
ΩM, Dave. I notice the two outside fuses and the center mov blew. If they correspond in their order I imagine that the unit blew before it even went through a third of a complete phase. In other words, instantly.
That MOV that exploded was probably on the line that had the HRC fuse that didn't explode. The other two MOVs were probably fine because of the shattered fuses.
+muh1h1 i remember having to consider vibration on electronic components on a job i did, looks like that might be to stop basic waggling, without it might not be long before the legs work harden and fail. in the end i think that some stuff was resin in-cased, not good for repairs then though.
Kind of makes sense for there to be glue on the side of that component... it was near the fuses. People are likely to be slipping tools/fingers in there and could accidentally give it a shove. Not that paste/glue is going to do a lot, but some engineer probably thought it was a solution.
If you dont put it in low z, putting it on ohms will on most multimeters discharge pretty quick, you will know when the cap is discharged if the ohmage stops jumping around.
+Roger Lamb I work on military COTS hardware and we use a fair amount of similar power supplies on the platforms I support. Don't recall exactly when, but the exact same power supplies (at least externally) started arriving with the Agilent brand instead of the TDK-Lambda brand a couple years ago. Never took a look inside to see if there were any other changes. They're not cheap and breaking one wouldn't exactly go well for me.
+tychosis Yes our Lambda supply is an older one I think that it is like some of the other electronics test equipment there has been a condensing of manufacturers that make the same or nearly the same supply with identical appearance except maybe paint but different branding. especially in the high power supplies. I don't think there is a difference in quality between Aglient/Keysight and Lambda as they are both powerhouse names the industry. Not like the difference you might find between Agilent and Rigol.
6:57 the overall construction looks very similar to a lot of Hitachi power supplies for computer servers. Something makes me think a lot of these high-powered devices share architecture
Nice BIG PSU. ~$8k nice price. Still waiting to see the repair video :) Maybe it's just like: replace 2 missing HRC and one MOV? Somebody suggested burned trace near MOV, so maybe it's easy-to-fix?
I've been checking one damaged of the same series, main PSU working fine (turns on and gives a heat/fan error) I replaced the old dirty fans for ones hoping these will send the right rpm feedback, but no, same error. Finally I removed the output board closer to the main PSU and i found a couple of burnt ICs :/, i haven't found replacement for those ICs i might just replace the whole output board if the price makes sense.
3 times single fase wont work, you need at least 1 line and 1 return, or 2 of either and 1 of the other, as you need the voltage difference not between fase and ground, but between the fases.
What generally happens to a lot of these power supply units when they are recycled, is the front panel gets smashed to buggery, so seeing this unit with it's front in tact is a shock. Also, especially with the APC units, the front panel is often used as a handhold, and ends up slipping off and smashing onto the ground. You assume that recycling would take better care, but these units are next to worthless, they very rarely get reused, and are only worth their weight in batteries. (if they have batteries).
Kean Maizels Yeah, I know, but the similarity is there, and often I've seen gear similar to this. I had a job which dealt with recycling government equipment, and the stuff I saw was amazing, equipment which was off the shelf for $250k. Spare no expensive is definitely how government works. I often thought that a lot of this gear would make for good material for this show. Unfortunately, now parted ways with this job, it was getting worse and worse, they had a contract to destroy eftpos units, and were shredding them inside the warehouse, with spot fires, smoke and lithium fumes everywhere.
Just FYI on the TDK labeling: I would bet these are indeed rebadged TDK Lambda GEN 600-8.5 PSUs. I work with a few of the 1.5kW model. We used to buy the Agilent N5772A PSU until we found the OEM ones straight from TDK-Lambda (Genesys GEN 600-2.6) save us nearly $1k per unit (minus the GPIB interface and Keysight brand-wank of course).
I've noticed an increase in line voltage here in the states as well, at least in my location. I've been in the same place for 20 plus years. 20 years ago, my line voltage was between 118V to 120V. Over the years, it's slowly crept upwards and now sits between 125 to 126.8. Anyone else noticed this as well?
What You cannot see over there but I see from here: my brain is blowed up :))))) I know, is only one phase and 2W system running on 1 Mhz :))))) But this white paperweight... such a massive power! You can drive a cloth iron with that :)
Dave, your building probably has 3 phase power. You may have 3 outlets in your shop which are on different phases. A carefully made power combiner could get you a 3 phase supply without having to pay for it. Of course, if there are GFI's on them it won't work. .
Anothrr great one DJ. 3-ph voltage is always specified phase-to-phase and the phase-to-ground will be whatever it is. Don't understand why its labeled 240 if in fact it's 400-whatever phase-to-phase...
Talking about outlet voltages, my voltage in the UK sits at around 246-248v and goes as high as 250v sometimes! I wonder how those 220v rated appliances like it, seem ok so far!
No point proceeding further if you do not have a 3phase supply and step down transformer. You cannot effectively test the supply. BTW the repair fee for 3 phase equipment is generally much higher because setup with 3 phase normally has a higher cost of operation, beside they are also much higher current equipment that uses less common parts.
Not an EE, but here's my naive analysis of the failure. The fuses for 1 and 3 blew because those two were high, 1 was higher and the fuse blew so hard that it dislodged the holder. The previous fellow cleaned up the fuses. The fuse for 2 didn't blow because it was low, and the MOV instead blew because it overloaded before the current went high enough on the phase to blow the fuse. Again, not an expert, that is just what 'makes sense' based on naive observation. Does this diagnosis have any merit?
PLCC sockets always gave me headaches on gear that was constantly powered and ran at fairly hi temps... Every month we'd have a crash and the fix was to have to shut down, pull he cards and PUSH THE DANG ics back home! The little buggers were famously walking out of their sockets.... (probably due to changing ambient heat)...
I feel compelled to share with everyone that this power supply was neither designed nor manufactured by Agilent/Keysight. Believe it or not, this was actually designed by power supply company called Lambda, now known as TDK Lambda. If you pause at 9:56, you will see the PWB says Lambda. This particular unit is a rebadged GEN 600-8.5, which is very unlike Agilent to do. I suppose the rationale was to enter the high power density system DC power supply market quickly by not having to design their own. One thing is for sure, TDK Lambda builds phenomenal power supplies, I personally know an engineer who works for that company who can attest to that. We have several of this power supply at my shop - and they perform flawlessly.
The two good movs will match up with the blown fuses, they did their job, the blown mov will be on the undamaged fuse, probably advisable to replace/check that fuse too... could you run it on 3 separate 110v power supplies? Off to jaycar... ;)
+fleuroman No, the MOV will cause the fuse to blow if it operates (clamp at rated volts and lots of current). It is possible that all the fuses blew, Dave didn't check the one that was still in the board. You can't use 3 110v power supplies, as they will all be in phase, the phase to phase voltage will be zero. Needs 120 degrees phase difference A-B-C.
Yea these are the same psu's that were being used at my last place of work for rectifier burn in ovens , they have hundreds of them there running 24/7 and to be honest the amount of them I seen fail is a real joke , not to mention after a while in service the displays on them would start going mental and eventually fail altogether .
I dont know if its just me but the trace inbetween connector CN302 and cap C312 (unpopulated) beside the third MOV looks weird. Perhaps its just the angle of the shot (@ 20:05)? Thanks for the video tho!
Dave, not sure how you can feed a single phase supply into "all three phases" because you have no common. What you would need to do is feed your single phase into L1 and L2 (or any other combination) I would have thought.
+Gerry Sweeney i was thinking that also! but, would it not be 180degree's out on each phase also ? =- judging on the fact what little E-Engineering knowloedge i still have left -=
+Gerry Sweeney Ah yes, correct. I should have looked and thought harder before waffling. There is indeed no earth common, it's just a safety earth, so have to apply 240V on any combination of the phases.
+killer1479 yeah but in a SMP all that is happening is its going into a bridge rectifier so its only important that there is a phase differential between L1 and L2
+killer1479 Look at this image, and think of the alternator windings as the power station delivering the three-phase power, all thats in the PSU is 6 diodes and a cap to get the DC that the SMP is using. www.ibiblio.org/kuphaldt/electricCircuits/Exper/05019.png
+Gerry Sweeney Maybe the PFC doesn't like it when one phase is missing. The Caps for smoothing out the DC may be to small, with three phase rectified the voltage is never zero. With 1 phase it obviously is. All this may only be a problem if you would run it up to maximum power.
190-240 VAC 3W[wire]+(earth symbol) Should have been a dead giveaway to the setup guy that its voltage phase to phase limit was 190-240VAC. Someone didn't pay attention to the EE symbols portion of their engineering course. :-p Since it's a 3phase device you should be able to apply your shop mains across any two of the line terminals and get it to fire up just fine, but it probably will end up running floating ground since tying in mains ground to the chassis has a good chance of tripping any earth leakage breaker feeding it.
you will get nothing if you feed same phase to all inputs. just power it up with 240V through L1 and L2. you couldn't get full power obviously but you can fix it :)
If my memory serves me right, I believe this series of Power Supplies is designed and Manufactured by Lambda and sold to Agilent. - Oops now I see you found it, I just jumped the gun.
With 5kW knocking around, what's the distance between those bus bars? They look pretty tight to me. I know they'd have done the maths and made sure it won't arc across, but my brain is kinda screaming for some kind of insulation between them.
If this smps has phase loss detection, you may want to look for a cheap 3 phase motor and a few caps, maybe put together a small rotophase for this sort of thing. Ya, the third phase isn't perfect...but good enough to do the job.... And quite well too.
Problem with running on single phase is ripple will be much higher - if it has PFC this might cause problems even at fairly low powers. On a PSU like this it may also have some sensing to make sure all phases are present.
+mikeselectricstuff Yes it would suck if it has a sensor. In that case I can't do it without a 3 phase supply. Worth feeding in the single phase and trying though.
+EEVblog Actually, come to think of it, the small three phase wires (near the MOV's) going over to the smaller PSU board looks suspicious. No reason to do that really unless that board is doing some sort of phase detection.
+EEVblog Try using a 1 phase motor to drive a 3 phase generator.
+EEVblog With a bit of luck, it might power up enough to display an error code or light.
+mikeselectricstuff I work with a similar sorensen unit at work, 'only' 3kw though. Its happy with whatever you feed it, 240V or 208V single phase? Just stick line and neutral on any 2 of the line inputs. 208 3 phase? Use all 3 obviously.
190V@30A still provides 5700 W, and those heatsinks dont look upto snuff for 700w dissipation.
Guess its time to RTFM. Like the first user should have 8)
Mind the ground levels of the DB-25 for the programming input, no metal shell on it, so it might be a common ground with the output - Thats how it is on the above mentioned Sorensen
A little thing to note, fans don't blow air in a linear path, rather fan it out in a cone, pun unintended. So the position of that heat sink is ok, since the air would be pushed right over it.
You can see this if you mount your fans back wards in your PC case and have them blow from the back to the front. The dust will be in a truncated cone pattern where visible over the MoBo.
+aserta exactly my thoughts
"Two phases of a 120V AC line is 208V" ...And this is why I'm a DC guy.
+rhkips It's easy. 120*sqrt(3) is 208V. Square root of 3 because there are 3 phases.
+rhkips Thank you. DC-when you need one plus one to equal 2.
+rhkips I think there's somewhat of an analogy between 3-phase and a straight 6 internal combustion engine. From what I've read, in a straight 6 all your secondary forces cancel out and that makes the engine really smooth. Also, from what I read about three phase: if you balance it properly, in a star configuration (ground in the middle), the current on the ground wire is zero.
+plonkster The ground connection is only ever used for protection. This supply has 3 phase delta input as opposed to 3 phase wye. The wye configuration includes a neutral which carries no current if the 3 phases are balanced. This supply has no neutral.
+Chris W I see. As I dug through more of the comments, that became clear to me too :-) TH-cam is anamazing resource for learning :-)
But is it 80 plus titanium rated?
+Waxing My Knee: ᶃᶉḁᵱḥᵻḉ ᵭᵻᴤᵻᶃᶇ What's this referencing?
Nope, its 600V smps, it has 90% plus efficiency...
+Waxing My Knee: ᶃᶉḁᵱḥᵻḉ ᵭᵻᴤᵻᶃᶇ hahahahahahaha... it's just "bronze"... booo
+Waxing My Knee: ᶃᶉḁᵱḥᵻḉ ᵭᵻᴤᵻᶃᶇ You can see it from the size of those heatsinks. If 5KW smps has 90% efficiency, thats 500W of heat now lets divide that by 2 (2 power boards in parallel) and we get 250W of heat for those small heatsinks which wouldnt last wery long so i think its probably going to be much more efficient, like 95% or more otherwise there would have been much bigger heatsinks.
My 2,5kW smps with 50V output have about 93% efficiency.
As we know, higher output voltage = smaller current = smaller losses and higher efficiency.
So I also think this smps have at least 95% efficiency.
This unit has been qualified to the MIL-STD-810 vibration requirements, which explains the silicone to mount those capacitors you pointed out early in the video. This is a tough power supply!!
Its cool how you can name any component, and sometimes tell a little story about them. You are my official electronics god.
+Deer Viehch Well if you watch enough, you'll learn a lot of them as well especially if you're just talking about "this is a resistor, this is a diode, this is a capacitor " You can do most of that without any real knowledge. But with individual parts, you'll start to recognize patterns. 4 digit part numbers beginning with 78 followed by two numbers are almost always a linear voltage regulator with the output voltage being the last two numbers. Ones beginning with 79 have the output voltage negative in respect to the ground. 74xxx are mostly logic chips, 4xxx are also mostly logic chips. Products by Altera, Xilinx, and Lattice are usually programmable logic. Atmel will usually be microcontrollers and flash memory. Microchip is frequently microcontrollers but they also have a lot of other products like chips for interconnecting to different things (CANbus transcievers, IO Expanders stuff like that)
stonent **head* *explodes**
But seriously, the most advanced electric thing i've ever done is build a game where you have to contact one of 16 solder spots in a 4x4 matrix in the right order to complete it, but the order was integrated and not random ;_;. also it was professionally(not at all) build on a wood rack :D
Back in 2005, Lambda generously donated 4 big power supplies to our arts group:24v 600w bridgable units which very reliably ran our 50 SmartMotor servo actuators all over our sculpture! They are indeed remarkably well constructed 3-phase units. I bypassed all the explody drama by triple-checking the available power before fitting the units up!
Great video!
You certainly know your stuff. I work for TDK Lambda and repair the 10 & 15kW Gen's from 10V 1000A to 600V 25A. Those 1000A 3U units really are beasts 👍
That dust near the input fuses sure looks to me like the dust from when an HRC fuse is broken. Either from catestrophic overload beyond the capacity of the HRC limits, or from physical damage.
+mausball Fast blow HRC fuses often have a powder inside. I suspect to support the thinner than normal fuse wire and dissipate the heat.
+Rob B The powder is to help extinguish the arc when whey blow.
... 400V in three-phase, yes, between the phases!
If you connect the same phase to all three inputs then you have zero volts between the phases! And there is no place connect the Neutral, since the fourth terminal is PE.
This beast runs in delta-configuration!
+John McFerren Why this star/delta rubbish.
190V - 240V between two phases.
No matter what system is used to supply this.
+John McFerren just that there are only four connectors on the back for L1, L2, L3 and PE. You cannot connect N to PE in an equipment like this and thus 3-legged star is not possible.
+Alexander470815 yes, you can probably connect L to one of the L1, L2, L3 and N to any other of the three and this power supply should power up. Just that Dave commented in the film that he wanted to connect his single phase to all 3 input phases, which works in a 3L+N device, but this unit lacks the N input.
+John McFerren The supply might be star, but the load connection would likely be delta(the hint being no neutral)
+Uwe Zimmermann You can run a wye device without a neutral because if the wye device is ballanced no current will flow in the neutral. Since this is one device it can easily be balanced. Many large industrial motors do not have a neutral. If you want to single phase this device you simply attach a Neutral to one of the Line inputs because it just uses a 3phase rectifier so you are just using one of the three rectifiers. It will certainly increase your ripple and reduce your max output power.
I just love looking at electronics. Cheers for the video, Dave!
I have seen 220V rated soviet equipment operate on 245VAC without problems. That is surely a nice PSU.
Edit: The negative track under the bridge rectifier is quite close to the other pin at 25:41.
There are cheap VFD inverters designed to operate three-phase gear from a single-phase supply. You should be able to find a 2KVA unit from China for ~$150. Get one designed for 120V/208V and operate it from your variac. Job done. That in itself would make an interesting video.
Holy crap! I'd be very careful around that beast if you did get it fixed! Thanks for sharing Dave
Few other interesting things on the output board.
Check 24:20 - there are vias on the positive rail (bottom of the board, around capacitor C83), that go to the other side of the board, most likely to double amount of copper for a current handling, but you do not see the vias on the negative side (around C84). Also the common mode choke area, has a footprint for a bigger common mode choke L81, probably for higher current models.
Another interesting thing is the footprint for one more capacitor, around configuration resistors, it is even labeled C92 with silkscreen. Very smart modular design. Also the suppression cap C81 look to have two different footprints. The parallelisation of the capacitors C83 and C84, is done using JW81/JW82 jumpers most likely.
Also the bus bars are probably an overkill in this 8.5A model. You could easily pass the current with few cables. But they most likely left it there, as it is shared design of other components, and also the busbars act as some form of radiator and current collector for transistors, so they left it there.
Also, as you can see the output board is only for 150, 300, 600 V models. The lower voltage and higher current models, must do have different output modules. The connector itself would need to be very different, and the amount of conductor would need to be increased too to handle few hundredths amps or so.
1k viewers an hour. Nice that your reaching more people now than when i subbed a year or so ago. Learning your catchphrases too +eevblog
+Tangobaldy I get my most views in the first 24-48hrs
Good ol' TDK-Lambda 2U system. I've built these from scratch before...
Maybe the main power supply isn't turned on until the input is verified. This would explain the damage you've described.
+Galfonz Makes sense, when the MOV blew up, it lost one phase and shut down.
if they have catched the magic smoke and put it back inside it would have worked again
+ljmike1204 lol
+ljmike1204 If they inhaled the magic smoke they might be able to shoot lightning.
+Kasper Guldmann and spell caught ;)
+ljmike1204 IBM do refills:
c1.staticflickr.com/7/6013/5990429717_a5a53c7da9_b.jpg
Even if this thing doesn't get fixed, you could salvage some rather decent bits out of it.
I once had a camera flash cap discharge into my finger which numbed it for about a minute. One can imagine what this rig would do if charged.
What about using a single to three phase VFD inverter? I've only ever used them on equipment with 3 phase motors in, not 3 phase electronics, so not sure if it would work. I'm also going to assume it would need to be pure sine not modified, but again don't know enough about this sort of thing. I look forward to seeing the repair.
27:38 Pilot G2. I'd recognize that pen anywhere. definitely the best pen you can buy for under $5. Good man.
Wow, bet they said "Ooops!" when they powered it on.
Amazin dumpster catch!! Hope it works fine after repair.
The US actually has 3 common 3-phase voltages - 208v, 240v (Delta) and 460/480v.
Ouch. Just went to the keysight web site to look up the specs on that supply and the first thing noted is that the voltage input is set at the factory and CAN NOT be changed. Basically, you order the supply and at the time of ordering, specify what input voltage the thing will have available. If you make a mistake, then you need to send the unit back and get a new unit with the correct voltages.
Now that you've taken it apart, time to TURN IT ON!
Nearly 247 volts out of your socket?
I live in the UK in a well built up city, meant to be 240V. Much closer to 255-260 and can go up at peak times aswell.
Dave, you need the 3 phase input. 3x1 phase will not work because of missing neutral. rectifing a 3phase source to DC is also easyer. I know that certification labs use a 6 phase transformer to create "rimpelless" DC :)
Dave you will need a true 3-phase supply, i.e three separate 240v phases (415v phase-to-phase), each separated by 120 degrees. You won't be able to just parallel the single phase 240V supply up. Lee
Ecellent execellent video Dave, more excellent than usual! I've been pining for a good teardown for aaagggegeeeessssssss
hi dave I'd love to see you fix this. I'm just a hobbyist but I love these repair videos
Good to grab this one , a nice bit of kit
So, is the blown MOV on the middle phase whose fuse didn't blow?
+Jan Hlavatý I wondered that. I need to get the board out to check.
+EEVblog if there is a blown MOV there will be damage on the rectifier bridges too.
+Eviltech Not necessarily because they're probably 600-1000V rated. I'd check the input filtering caps (2.2nF 275V). Even if they are the self-healing ones, their capacity may have gone down significantly. The "middle" MOV probably done it's job, but I'd check the other two rails going to the PFC, since their fuses have blown and there was a significant surge of current going down there to the PFC. There could be some nasty damage done in the PFC on those two input rails... I'd check for shorts on those rails after taking the PFC board out, just to be sure.
Jasenko Cicvara A 1000V rated ones for sure! However the blown MOV is after them, so the surge current has passed trough the rectifiers. We are talking hundreds of amps, even though for just a few microseconds.
You have a point. Sure, It took a considerable current to blow the hell of of that MOV. But, it makes me wonder why the fuse is still intact. with that amount of energy going through it, it should've been blown to kingdom come, like the others. It probably happened that way because the voltage at that phase was the highest at the moment of connection to the power outlet, so the other two phases took out only the fuses. Either way, a catastrophic mistake...
25:07 Nope, it got 2 turns on each side :) Each pass through the centre of the toroid counts as 1 turn. Looking forward to repair vids, will be nice to see it working again
at 20:03 it looks like the trace on the right of the MOV lifted but im not sure, thanks for the great vid again
Can't wait to see the follow up
It looks like a track is peeling up from the board @ 20:05 by the mov!
+Itguy92 It does look like it got a little warm.
In Europe you could safely use this power supply, but you would need to connect it to a 32A CEE blue plug...since 3 phase power is also 400V. You simply hook up a neutral and phase to any two of the three phase inputs on the power supply...or over in Australia you would hook it up to a single phase 32A outlet....Over in the USA 3 phase domestic power is usually 240V (not 208V since the most transformers are delta wired with one winding used as a split phase to achieve 120V for normal sockets) but you can also hook it up to a single phase NEMA 240V 30A or 50A outlet....using the two ''hot'' wires as they call it. But the normal home power supply is split phase (google it)...
ΩM, Dave.
I notice the two outside fuses and the center mov blew. If they correspond in their order I imagine that the unit blew before it even went through a third of a complete phase. In other words, instantly.
I think the bridge rectifier on the output is flyback protection.
One powerful unit.
That MOV that exploded was probably on the line that had the HRC fuse that didn't explode.
The other two MOVs were probably fine because of the shattered fuses.
i'm guessing they have stuck down the one component, so you can't accidently break it off chaning the fuses...
+muh1h1 i remember having to consider vibration on electronic components on a job i did, looks like that might be to stop basic waggling, without it might not be long before the legs work harden and fail. in the end i think that some stuff was resin in-cased, not good for repairs then though.
Vibration is a factor in such a large power supply. The gobs of supporting filler-goo, is just good construction.
Kind of makes sense for there to be glue on the side of that component... it was near the fuses. People are likely to be slipping tools/fingers in there and could accidentally give it a shove. Not that paste/glue is going to do a lot, but some engineer probably thought it was a solution.
If you dont put it in low z, putting it on ohms will on most multimeters discharge pretty quick, you will know when the cap is discharged if the ohmage stops jumping around.
I thought that supply looked the same as a Lambda supply I had in my shop. The Lambda silk screens on the pcb confirmed it for me.
+Roger Lamb I work on military COTS hardware and we use a fair amount of similar power supplies on the platforms I support. Don't recall exactly when, but the exact same power supplies (at least externally) started arriving with the Agilent brand instead of the TDK-Lambda brand a couple years ago. Never took a look inside to see if there were any other changes. They're not cheap and breaking one wouldn't exactly go well for me.
+tychosis Yes our Lambda supply is an older one I think that it is like some of the other electronics test equipment there has been a condensing of manufacturers that make the same or nearly the same supply with identical appearance except maybe paint but different branding. especially in the high power supplies. I don't think there is a difference in quality between Aglient/Keysight and Lambda as they are both powerhouse names the industry. Not like the difference you might find between Agilent and Rigol.
6:57 the overall construction looks very similar to a lot of Hitachi power supplies for computer servers. Something makes me think a lot of these high-powered devices share architecture
Nice BIG PSU. ~$8k nice price. Still waiting to see the repair video :) Maybe it's just like: replace 2 missing HRC and one MOV? Somebody suggested burned trace near MOV, so maybe it's easy-to-fix?
I've been checking one damaged of the same series, main PSU working fine (turns on and gives a heat/fan error) I replaced the old dirty fans for ones hoping these will send the right rpm feedback, but no, same error. Finally I removed the output board closer to the main PSU and i found a couple of burnt ICs :/, i haven't found replacement for those ICs i might just replace the whole output board if the price makes sense.
3 times single fase wont work, you need at least 1 line and 1 return, or 2 of either and 1 of the other, as you need the voltage difference not between fase and ground, but between the fases.
What generally happens to a lot of these power supply units when they are recycled, is the front panel gets smashed to buggery, so seeing this unit with it's front in tact is a shock. Also, especially with the APC units, the front panel is often used as a handhold, and ends up slipping off and smashing onto the ground. You assume that recycling would take better care, but these units are next to worthless, they very rarely get reused, and are only worth their weight in batteries. (if they have batteries).
+Beren Scott This isn't a UPS, but a high-voltage high-precision lab-grade power supply
Kean Maizels Yeah, I know, but the similarity is there, and often I've seen gear similar to this. I had a job which dealt with recycling government equipment, and the stuff I saw was amazing, equipment which was off the shelf for $250k. Spare no expensive is definitely how government works. I often thought that a lot of this gear would make for good material for this show. Unfortunately, now parted ways with this job, it was getting worse and worse, they had a contract to destroy eftpos units, and were shredding them inside the warehouse, with spot fires, smoke and lithium fumes everywhere.
Just FYI on the TDK labeling: I would bet these are indeed rebadged TDK Lambda GEN 600-8.5 PSUs.
I work with a few of the 1.5kW model. We used to buy the Agilent N5772A PSU until we found the OEM ones straight from TDK-Lambda (Genesys GEN 600-2.6) save us nearly $1k per unit (minus the GPIB interface and Keysight brand-wank of course).
I've noticed an increase in line voltage here in the states as well, at least in my location. I've been in the same place for 20 plus years. 20 years ago, my line voltage was between 118V to 120V. Over the years, it's slowly crept upwards and now sits between 125 to 126.8. Anyone else noticed this as well?
That copper sheet on the GPIB board is actually a heat spreader, not screening.
I don't get it... the label on the unit says no serviceable parts inside and there are three removable fuses! Go figure!
+simontay1984 Unless it's an apple product.
Ok that label would definitely had fooled me to hook it up ;) Would have released magic smoke in my country as well..
Every time you say "Trio test and measurement", i still hear Trio smart cal. :D
What You cannot see over there but I see from here: my brain is blowed up :))))) I know, is only one phase and 2W system running on 1 Mhz :))))) But this white paperweight... such a massive power! You can drive a cloth iron with that :)
15:24 I'm amazed that they didn't release the magic smoke.
Dave, your building probably has 3 phase power. You may have 3 outlets in your shop which are on different phases. A carefully made power combiner could get you a 3 phase supply without having to pay for it. Of course, if there are GFI's on them it won't work. .
Anothrr great one DJ.
3-ph voltage is always specified phase-to-phase and the phase-to-ground will be whatever it is. Don't understand why its labeled 240 if in fact it's 400-whatever phase-to-phase...
It could be that the silicone goop on the side of the capacitor could be there just to protect the capacitor from violent fuse explosions.
+krbruner Smart engineers :->
If it's designed properly the only blown parts are one or more mov's.
I've seen it dozens of times in my job as a service engineer.
All those Back To The Future references :D
Talking about outlet voltages, my voltage in the UK sits at around 246-248v and goes as high as 250v sometimes! I wonder how those 220v rated appliances like it, seem ok so far!
No point proceeding further if you do not have a 3phase supply and step down transformer. You cannot effectively test the supply. BTW the repair fee for 3 phase equipment is generally much higher because setup with 3 phase normally has a higher cost of operation, beside they are also much higher current equipment that uses less common parts.
Not an EE, but here's my naive analysis of the failure. The fuses for 1 and 3 blew because those two were high, 1 was higher and the fuse blew so hard that it dislodged the holder. The previous fellow cleaned up the fuses. The fuse for 2 didn't blow because it was low, and the MOV instead blew because it overloaded before the current went high enough on the phase to blow the fuse.
Again, not an expert, that is just what 'makes sense' based on naive observation.
Does this diagnosis have any merit?
PLCC sockets always gave me headaches on gear that was constantly powered and ran at fairly hi temps... Every month we'd have a crash and the fix was to have to shut down, pull he cards and PUSH THE DANG ics back home! The little buggers were famously walking out of their sockets.... (probably due to changing ambient heat)...
I feel compelled to share with everyone that this power supply was neither designed nor manufactured by Agilent/Keysight. Believe it or not, this was actually designed by power supply company called Lambda, now known as TDK Lambda. If you pause at 9:56, you will see the PWB says Lambda. This particular unit is a rebadged GEN 600-8.5, which is very unlike Agilent to do. I suppose the rationale was to enter the high power density system DC power supply market quickly by not having to design their own. One thing is for sure, TDK Lambda builds phenomenal power supplies, I personally know an engineer who works for that company who can attest to that. We have several of this power supply at my shop - and they perform flawlessly.
I would like to see that power supply fixed.
i still have so many videos to watch
The two good movs will match up with the blown fuses, they did their job, the blown mov will be on the undamaged fuse, probably advisable to replace/check that fuse too... could you run it on 3 separate 110v power supplies? Off to jaycar... ;)
+fleuroman No, the MOV will cause the fuse to blow if it operates (clamp at rated volts and lots of current). It is possible that all the fuses blew, Dave didn't check the one that was still in the board. You can't use 3 110v power supplies, as they will all be in phase, the phase to phase voltage will be zero. Needs 120 degrees phase difference A-B-C.
Yea these are the same psu's that were being used at my last place of work for rectifier burn in ovens , they have hundreds of them there running 24/7 and to be honest the amount of them I seen fail is a real joke , not to mention after a while in service the displays on them would start going mental and eventually fail altogether .
Bet they had brown trousers when that went bang, especially at that price... :P
If you cant repair it than you have a whole new parts bin of stuff and one hell of an amazing project case!
+Alyx BioHaz One heck of as scrap project indeed.
We have the same thing at work. It just says "TDK-Lambda" on the front...
Lambda, Lambda, Lambda and the Omega Mu's - epic line in an epic movie!
(Revenge of the Nerds for those of you who didn't get it ;) )
I dont know if its just me but the trace inbetween connector CN302 and cap C312 (unpopulated) beside the third MOV looks weird. Perhaps its just the angle of the shot (@ 20:05)? Thanks for the video tho!
Dave, not sure how you can feed a single phase supply into "all three phases" because you have no common. What you would need to do is feed your single phase into L1 and L2 (or any other combination) I would have thought.
+Gerry Sweeney i was thinking that also! but, would it not be 180degree's out on each phase also ? =- judging on the fact what little E-Engineering knowloedge i still have left -=
+Gerry Sweeney Ah yes, correct. I should have looked and thought harder before waffling. There is indeed no earth common, it's just a safety earth, so have to apply 240V on any combination of the phases.
+killer1479 yeah but in a SMP all that is happening is its going into a bridge rectifier so its only important that there is a phase differential between L1 and L2
+killer1479 Look at this image, and think of the alternator windings as the power station delivering the three-phase power, all thats in the PSU is 6 diodes and a cap to get the DC that the SMP is using. www.ibiblio.org/kuphaldt/electricCircuits/Exper/05019.png
+Gerry Sweeney Maybe the PFC doesn't like it when one phase is missing.
The Caps for smoothing out the DC may be to small, with three phase rectified the voltage is never zero.
With 1 phase it obviously is.
All this may only be a problem if you would run it up to maximum power.
Did this ever get repaired ?
I cant find a follow up video :(
190-240 VAC 3W[wire]+(earth symbol) Should have been a dead giveaway to the setup guy that its voltage phase to phase limit was 190-240VAC. Someone didn't pay attention to the EE symbols portion of their engineering course. :-p
Since it's a 3phase device you should be able to apply your shop mains across any two of the line terminals and get it to fire up just fine, but it probably will end up running floating ground since tying in mains ground to the chassis has a good chance of tripping any earth leakage breaker feeding it.
At the end you da never done a repair video. This would had been interested to watch. You stilla have the unit?
The shielding over the digital board is most likely to protect the MCU against the interference from the nearby high power switching action.
+yoksel99 It's actually a heat spreader -search for Chomerics T-Wing.
And if you knew what we were working on in a 2U package, it would blow your mind...
you will get nothing if you feed same phase to all inputs. just power it up with 240V through L1 and L2. you couldn't get full power obviously but you can fix it :)
If my memory serves me right, I believe this series of Power Supplies is designed and Manufactured by Lambda and sold to Agilent. - Oops now I see you found it, I just jumped the gun.
its actually from TDK-LAMBA, Genesis series
This thing is the real deal. Holy shit.
Is there a coincidence that the middle fuse isn't blown, but the middle MOV is?
Omega Mu's for the win!
+Gabbos Ironfist Not the Mu's!
could be powered by a three phase auto transformer with multiple voltage tappings the type that are used for powering machine tools.
Did this ever end up working? Just curious!
Can you test it with a three phase variable speed motor controller? Using a scope or freq counter to set it to 50hz?
No wonder it broke.
There's a kennel load of big ass 'puppies' flapping in the breeze in that power supply.
With 5kW knocking around, what's the distance between those bus bars? They look pretty tight to me. I know they'd have done the maths and made sure it won't arc across, but my brain is kinda screaming for some kind of insulation between them.
Hey Dave, my mom told me I couldn't build a car out of spagetti. You should have seen the look on her face as I drove pasta.