DSO-X 3024A Power Supply defect

[enz]

Hi All,

bad luck today.

When I plugged in my DSO-X 3024A the circuit breaker tripped and I saw a light flash cominf out of the backside of the oscilloscope.
After opening the device and inspecting the power supply I noticed a burned part, see picture below:

I think the broken part was a varistor, but I am not 100% sure. Could also be a capacitor.
Can anyone identify the part?

Otherwise I would leave the part unpopulated.

Best regards,
Martin

[JFJ]

The part might be a NTC thermister, like this:

https://www.yoycart.com/Product/571067033190/

[soesante]

Believe its NTC Thermistor...

Sent from my SM-J730G using Tapatalk

[ArthurDent]

The blown part is obviously a NTC thermistor as others have pointed out but the question is why did it blow? It is possible that it was just that part, it happens, but there may have been something more serious that caused it to blow. In the lower left there is F1 a 6.3A fuse, is that blown?

You say the problem caused a breaker to trip. That sounds more like a direct short from one of the A.C. input leads to ground (earth) or a line-to-line short. If it was just a problem within the power supply the 6.3A fuse would blow but not the breaker. What is the breaker capacity? It is probably several amps higher. There are 2 toroidal inductors on the left side of the power supply, probably one on each leg of the A.C.. If there was an arc-over in that area, that may have caused the problem.

It would be a good idea to carefully check the power supply out and not just replace that part and plug it back in to power.

Edit: I said fuse was 0.3A and it is actually 6.3A.  Thanks to PA0PBZ for pointing out my error.

[PA0PBZ]

In the lower left there is F1 a 0.3A fuse, is that blown?

I think it says 6.3A so it could well be that the breaker was faster. When something like this happens the diodes are the first suspects. If you find one shorted replace them all!

[JFJ]

The usual purpose of a thermistor, in this type of power supply, is to limit the inrush current to the main input filter capacitor. In normal operation, you wouldn't expect the thermistor to get hot enough to break apart. Presumably, something downstream of the thermistor went short circuit (e.g. the main switching MOSFET transistor). So, as ArthurDent suggests, it would be unwise to just replace that part and plug it back in to power.

A better way, would be to connect the oscilloscope to mains power via an incandescent lamp. A 60W light bulb, for example, would limit the current to around a quarter of an amp. If there isn't a short circuit, the lamp should, initially, glow brightly and then dim, but in the presence of a short circuit, it would continuously glow brightly.

[nctnico]

The usual purpose of a thermistor, in this type of power supply, is to limit the inrush current to the main input filter capacitor. In normal operation, you wouldn't expect the thermistor to get hot enough to break apart. Presumably, something downstream of the thermistor went short circuit (e.g. the main switching MOSFET transistor). So, as ArthurDent suggests, it would be unwise to just replace that part and plug it back in to power.

A better way, would be to connect the oscilloscope to mains power via an incandescent lamp. A 60W light bulb, for example, would limit the current to around a quarter of an amp. If there isn't a short circuit, the lamp should, initially, glow brightly and then dim, but in the presence of a short circuit, it would continuously glow brightly.

I'd pull out a DMM first and check the rectifier and primary MOSFET for a short before attempting a repair or even power it.

[enz]

Hi All,

sorry for the late response, I was very busy.

Thanks for all the hints, very appriciated.

The defect part was indeed a 7Ohm Thermitor, thanks for pointing that out.

The 6.3A Fuse was not blown. This Fuse is only in Series to a varistor, so it only blows if the varistor goes defect.

The power supply is working fine now. Here is what happened:
The main capacitor (100uF/420V, Nichicon 105°C) was defect.
This caused the voltage to rise and the primary MosFET popped to a dead short.
This caused the Thermistor to overheat and blow (but maintain a low impedance connection). Due to the dead short of the MosFET the circuit breaker (16A) tripped.

I replaced the three parts and checked the rest of the power supply, but all other parts were fine.
Scope is back in business now.

The main problem seeems to be, that the supply is powered all the time. The scope is about 7 years old. Don't know why Agilent decided to put the switch on the secondary side.

Thanks again,
Martin

[HighVoltage]

I had it on a few Keysight and also Keithley instruments power supplies, that the NTC only broke.
Only exchanging the NTC solved the problem completely.
Funny: I bought a few extra NTC's and they are almost all gone by now.

I think they fail, when the instrument is turned ON/OFF multiple times, shortly after each other and the NTC is overloaded. I also noted that newer instruments have a much larger NTC installed ind the PSU.

[DEHiCKA]

Got mine 3034A blown today exact the same way as OP.
Loud electric bang with smell of an electronic death and 16A circuit breaker tripped.
Last time it was powered more than 12 hours ago from the same outlet.

Clearly, PSU is not well designed: main cap heated so much for so long so the board on the other side is black and cracked.
One of the mosfets was blown to pieces (P12NM50). And there was an arc from the PSU backside to the scope's main shield.
The 6.3A Fuse was not blown. NTC has a small crack but still good. Probably something else is dead.
Here is another DSO-X blown the same way with some dead parts on the PSU ctrl board:
http://powersuppyrepair.blogspot.com/2018/02/lineage-power-cch0125f1-z03a.html

I do not really want to repair the PSU, but can not find where to buy a direct replacement yet.
It is 5"x 3" 12V 125W. DSO-X uses remote sense lines and remote on-off.

There are bunch of 5"x 3" 120W+ PSUs with sense and on/off (pull down), but I can not find one with the same pinout.

[DEHiCKA]

Follow up on the repair.
The main 100uF/420V cap has degraded down to 9uF.
The PSU control board was also fried, so I did not bother to repair it.
Exchanged PSU for Meanwell RPS-400. Works fine now.

I'm glad the cirquit breaker was only 16A / 220V, not 32A or 40A!

[HighVoltage]

Congratulations on the successful repair.

Nice that we can buy fitting power supplies for these scopes.

[frabre]

Hi DEHiCKA,

I have the same problem of power supply.
I tried to repair but...
Which model of RPS-400: RPS-400-12 or RPS-400-12-C ?

If 400-12: no problem of shielding ?

If 400-12-C: no problem of height ?

Thanks in advance...

[DEHiCKA]

Hi, frabre
I have the cheapest RPS-400-12 without cage.
No problem with the shielding, it has a slight audible noise, but otherwise working just fine.
In/Out terminals are differ from original though, so you have to make some kinda adapter.

[Madmanguruman]

Just chiming in about shielded vs. unshielded 3x5 power supplies. I recently replaced the same PSU from a DSOX3014A scope with a Murata PQC250-12 and a bit of Mylar to create some creepage between the cable bundles and power supply itself. So far so good, no discernable pick-up or noise issue. I would argue that if you are replacing with a non-caged PSU that you should carefully dress the cables to keep them away from hotspots or from touching any primary-side parts - this is why I added some Mylar under the cable assemblies as they cross the power supply.

Also, to sanity-check the scope before you get a new power supply, you can remove the PSU and connect a 12VDC 10A lab supply directly to the 12V rail (no need for enable or remote sense) - the scope will power-up and operate without these signals. My scope was drawing between 5.5 and 6.5A with a single scope probe connected and the scope capturing data.

[salvagedcircuitry]

Hey Guys! I found a similar problem with my 3034a oscilloscope power supply. I made a repair video about fixing it. 



Hats off to @DEHiCKA for posting the power supply blog repair link. My current mode controller was completely toast as well.

List of replacement parts:
Datasheet for look-a-like lineage branded power supply. CCH125 series:
http://apps.geindustrial.com/publibrary/checkout/CCH125?TNR=Data%20Sheets%7CCCH125%7CPDF&filename=CCH125.pdf

Genuine Keysight PSU repair part: 0950-5277
https://www.keysight.com/my/partDetail/0950-5277

Original mosfet:
https://www.digikey.com/product-detail/en/stmicroelectronics/STP12NM50/497-2666-5-ND/603691

Replacement mosfet:
https://www.digikey.com/product-detail/en/ixys/IXTP450P2/IXTP450P2-ND/2328313

Two pin Diode package near mosfet:
https://www.digikey.com/product-detail/en/on-semiconductor/MUR860G/MUR860GOS-ND/919926

Current mode controller:
https://www.digikey.com/product-detail/en/on-semiconductor/UC3842BVD1R2G/UC3842BVD1R2GOSCT-ND/5802572

Replacement MOV:
https://www.arrow.com/en/products/v300la20ap/littelfuse

Replacement NTC Varistor:
https://www.arrow.com/en/products/b57237s709m/epcos-tdk

Replacement Fuse:
https://www.digikey.com/product-detail/en/bel-fuse-inc/MRT-6-3AMMO/507-1119-1-ND/4794478

Replacement 1w 22k resistor under main filter capacitor:
https://www.arrow.com/en/products/erj-1tyj223u/panasonic

Tall cap I selected:
https://www.digikey.com/products/en?keywords=ucy2w101mhd

Dimensionally compatible cap:
https://www.digikey.com/product-detail/en/rubycon/450VXH100MEFCSN22X25/1189-2028-ND/3564971

Power Supply Replacement Candidates:
Meanwell RPS-300-12
Meanwell RPS-400-12
Meanwell EPP-300-12
Will need ps_on inverted:
Meanwell RPSG-160-12
Murata PQC250-12

[mlloyd1]

so, is anybody seeing this issue on the 3000T models (yet)?

mlloyd1

[TheSteve]

so, is anybody seeing this issue on the 3000T models (yet)?

mlloyd1

The 3000t series uses a different power supply that so far doesn't appear to have any issues.

[salvagedcircuitry]

so, is anybody seeing this issue on the 3000T models (yet)?

mlloyd1

The 3000t series uses a different power supply that so far doesn't appear to have any issues.

Dave's early review shows the same lineage power supply @2:23


Granted, they could have changed it out later on (I hope they did).

[TheSteve]

Keysight has confirmed in the past that they have changed it. I know my 3024T has a different model supply. Perhaps it was after production had started.

[AndyC_772]

Just to add another data point - this morning my MSO-X3104A died 

It was sitting on the bench, powered up but not actually connected to anything, when the display suddenly went blank without warning. I removed and reinserted the power cable, there was a loud bang, and the GFCI for the house tripped.

The main reservoir cap in the PSU has vented, and was very hot indeed. The current mode controller has blown the top off its package, and the gate resistor for the MOSFET Q1 is toast too.

Ordinarly I wouldn't even consider repairing an off-the-shelf PSU module, but this one appears to be a custom part that's not readily available, even though it looks like a standard item.

Has anyone replaced one of these with something else recently? Is there anything particularly special about the Lineage module that makes it hard to substitute for another unit?

[HighVoltage]

I have always repaired the Keysight power supplies, because usually it is very simple.

What is different on the MSO-X3104A?
You have pictures?

[AndyC_772]

I don't think the MSO-X3104A has a different power supply. I've asked Keysight to quote for a replacement part, and there's always the option of fitting an alternative with a suitable wiring adapter, but since I have some spare time this weekend I figured it was probably worth seeing if the damage looks repairable.

I had a few different dead parts than others have seen; MOSFET Q1 has a series resistor which is 5.11 Ohms 0805, and R7 is 110mOhm 5% (probably 2W).

In my PSU both were open circuit; the gate resistor was blown to bits, and the larger one I suspect was damaged during extraction of C11. They're stuck together with white goop which made removing them to access the current mode controller (also blown to bits!) tricky.

The MOSFET Q1 had a big diagonal crack all the way across the package body, the main cap had vented, and the board was dark and discoloured around the 22k resistors on the underside. Part of me is curious to know what happened in what order. Another part of me is cursing a wasted day and just wants it back working again.

I was able to determine the values of the dead resistors by measuring them in a working PSU, which I borrowed out of my MSO-X3054A. Since I rely on my scope to make a living, having a spare on hand was never optional, and today my decision to keep two that are very similar has paid off. I've put the PSU from the spare scope into my 3104A which is now working fine.

Even the working PSU in the 3054A was very hot when I removed it, having been running for several hours since the 3104A died. I'm really not surprised they fail.

My MOV and NTC seemed OK, there's high impedance from L to N and about 7 Ohms across the NTC and chokes.

I should warn anyone else looking to work on one of these supplies, you'll need a powerful soldering iron and a good range of tips. The board is constructed with no thermal relief on any of the through-hole pads, and the copper draws heat out like no other board I ever remember working on. Desoldering the heat sinks and large components without damaging the board was exceptionally difficult even with my JBC, and you'll need to be both gentle and patient. The main cap C7 was especially reluctant to come out, and it's fortunate that the little copper pad on the top side of the board, negative terminal, wasn't actually connected to anything.

I wasn't able to find an exact replacement for the 110mOhm resistor, so I've ordered 100m + 10m to fit in series. All the other parts were available from stock at either RS or Farnell, including some parts which looked OK but which I had to remove to gain access to others. Given the cost of new components vs my time, I rarely re-fit used ones to anything.

Fingers crossed nothing else has failed, and when it's all reassembled I get a working PSU again. I'll be running it for some time out of the box on a decent load before I trust it with my scope.

[salvagedcircuitry]

I don't think the MSO-X3104A has a different power supply. I've asked Keysight to quote for a replacement part, and there's always the option of fitting an alternative with a suitable wiring adapter, but since I have some spare time this weekend I figured it was probably worth seeing if the damage looks repairable.

I should warn anyone else looking to work on one of these supplies, you'll need a powerful soldering iron and a good range of tips. The board is constructed with no thermal relief on any of the through-hole pads, and the copper draws heat out like no other board I ever remember working on. Desoldering the heat sinks and large components without damaging the board was exceptionally difficult even with my JBC, and you'll need to be both gentle and patient. The main cap C7 was especially reluctant to come out, and it's fortunate that the little copper pad on the top side of the board, negative terminal, wasn't actually connected to anything.

I wasn't able to find an exact replacement for the 110mOhm resistor, so I've ordered 100m + 10m to fit in series. All the other parts were available from stock at either RS or Farnell, including some parts which looked OK but which I had to remove to gain access to others. Given the cost of new components vs my time, I rarely re-fit used ones to anything.

Fingers crossed nothing else has failed, and when it's all reassembled I get a working PSU again. I'll be running it for some time out of the box on a decent load before I trust it with my scope.

My PSU repair was rather painful, but successful. Overall, I'm not too fond of the repairability of the Lineage PSU design. It definitely took a lot of patience, flux and heat to get the aluminum heatsink off and not destroy the staking pads. I think I ended up having to super-glue a heatsink pad back on at the end of the repair process. I opted to leave the daughter PCB soldered in and just desoldered the failed soic-8 current mode controller on my unit. Since the lineage design does not instill confidence, I went with a meanwell RPSG-160-12 as a replacement power supply to have on hand in case the failure occurs again. Unfortunately, I goofed and ended up with the RPS-160-12 which does not have the remote-on feature which this oscilloscope needs as the power button is designed to switch the secondary side.
https://www.arrow.com/en/products/rpsg-160-12/mean-well-enterprises

If I could do the repair all over again, I would desolder the (4) series 22k 1w SMT resistors below the main filter cap and replace them with one ~88k 4w or 5w through hole resistor and just leave the resistor spaced off the main pcb. This way the resistors can't directly heat up the back side of the filter cap. There is enough space between the bottom face of PSU PCB and the aluminum shell of the agilent oscilloscope to accommodate the resistor. Additionally I would add electrical tape on the back side of the oscilloscope casing where the resistor may come close. On closer look, It may be a set of 2 power resistors, I see a trace between the 3rd and 4th resistor on the back side.

Something like this:

[AndyC_772]

I'm a little hesitant to make a resistor modification like that because I'm not sure I can see a sufficiently robust way to mount a TH part off the PCB.

The SMT resistors may get very hot, but they are at least light weight and unlikely to detach and become dangerous if the scope takes a knock. I'm concerned that the added weight of a TH part could cause it to detach the already weakened pads from the PCB, and in the worst case this means a loose end at 400V flapping about.

As an aside, I was faced with a very similar problem here. After I posted that thread, the equipment in question did actually fail again, and it was the resistor chain exactly as I'd predicted.

I've removed the resistors from my scope PSU just to check the condition of the board and solder joints. The board is blackened and the solder mask is cracked, but the pads are still firmly attached, and following a good clean with IPA, I'd put it in the "I've seen worse" category. I'll order some replacement resistors anyway, no sense refitting old ones.

Given that the failure is entirely predictable, I've also ordered a replacement main capacitor (this one) for my other scope, by way of preventative maintenance. If I can spend half an hour swapping a single component, rather than waiting for a potentially catastrophic failure that will inevitably happen at the most inconvenient moment, that has to be time well spent, IMHO.

One thought I did have was to fill the space between the resistors and the internal aluminium chassis of the scope with a thermal gap filler pad material like this stuff. This would conduct some of the heat out of the board and into the chassis, which in turn would keep the temperature of the capacitor down and help it last a bit longer.