R & S HMP 2030 with 2 blown channels

[SparkyBruce]

Hello , I have just bought a R&S HMP2030 power supply that has two of it's three output channels blown.

I see that people have had issues with shorts in the PCB near the main bridge rectifier but what I have on the two channels appears to be both power N FETS blown / shorted in the switching pre-regulator circuit.

The fets are Infineon IPD135N08's (presumably the n3G version) and they are controlled by a LTC 3703 switching PSU controller.

I have already found one manufacturing fault on one of the front panel PCB's with a tomb stoned resistor in one of the remote sense circuits.

So I am wondering what would have caused the fets to go S/C - do I have two faulty controllers?
Thankfully the two 7A SMD input fuses popped before any other magic smoke could escape.

Thanks

Bruce

[SparkyBruce]

I have made some progress with both of the blown channels having a shorted 3A Schottky diode on them (SS39's).

I managed to damage one of the NFET's on one channel so I swapped in some 60V 45A ones I had kicking around and when fed off an external DC supply (rather then the 57V built in rectifier system) it is running properly - but it does show a high standby  current (when compared with the working channel) so I will dig out the thermal camera tomorrow to see what is getting hot.

Whilst I wait for new fets and diodes to arrive I am trying to use the EEZ software (flow) to automate some control tasks for me.

Does anyone have a beginners guide to EEZ flow by any chance?

Thanks
Bruce

[Jens01]

I have a similar unit (HMP4040) with exactly the same symptoms. I also found more users having the same problems:
https://community.element14.com/members-area/personalblogs/b/gough-lui-s-blog/posts/roadtest-extra-rohde-schwarz-hmp4040-04-teardown-attempted-repair
https://community.element14.com/challenges-projects/project14/restoration-repair/b/blog/posts/a-shocking-variac-made-a-little-less-shocking

I've had shorts on the output rail (caused by blown FETs), and after changing all the fuses/schottky's/FETs the channel worked until i attached a serious load to it. I suspect the LTC3703 driver IC is also shot, but haven't had any time to change it. I would suggest that part as a good starting point for pinpointing the high standby current.

[SparkyBruce]

Thanks,

I have new FET's and some more diodes with some of the LTC3703's and more diodes coming soon.

One of the gate drive series 10R resistors had smoked so I will check out the controllers against my working channel to see if they are causing the issues.

Other people had reported actual PCB shorts which I have not found to be the case - just multiple diode and fets failed short.

There is a suggestion in the LTC datasheet to improve shoot through protection by running the 3703 off a slightly negative rail (down to -5V rather than 0V) - but I'm not sure if this PSU uses that or not.

The converters are mostly working as they do control the pre-regulated output to a couple of volts above the output voltage - I just don't have the smoking gun of why the fets/diodes failed in two channels.

All the final linear control (voltage and current) is fully working so I don't think this was something externally causing the damage as the final output is fine.
I will be back on the case tomorrow.
I have taken some high res pictures of the board which I can upload.

[SparkyBruce]

I have swapped both fets on both of the failed channels and replaced the Schottky diodes (the ones across the lower Fet).

The original working channel has a quiescent current of 20mA off the 57V supply voltage and both repaired channels have 110mA quiescent current.

The current is stable and the repaired channels are fully working with the only difference being the extra thermal dissipation.

I have now received the replacement LTC controllers - but I am going to check things like switching frequency between channels and gate drive signals before I swap them.

I have taken the following picture comparing the original channel against a repaired channel (The repaired channel on the right hand side) , and you can see that the LTC controller IC is working harder together with the 10R gate drive resistors (hence why I am wondering if they might be running faster for some reason).

The PSU has been working for a couple of days without any issues - but I still don't really know what happened to cause the fault (possibly the lower diode shorted - which caused the upper FET to short).

Only time and some scoping will tell and for mow I only run it with non critical loads when I am present in the room.

Thanks Jens01 for the pointers

[eurgenca]

Hello.
Just quriosity, which model of thermal camera are you using?

[SparkyBruce]

Hi ,

it is a Fluke Ti9 which is a 160 x 120 res.

Bruce

[SparkyBruce]

Today I have scoped the three channels and looking at the inductor waveform all three channels are running at about 174Khz with the on time varying with the output demand voltage.

I was wondering if some of the gate drives were "slow" causing the increased quiescent current - but that does not appear to be the case.

I will next check the other support components and even the diodes types used (as I do have some identical replacements coming).

[Jens01]

Interesting! I expected the same "higher frequency = higher losses" scenario. Have you checked the gate drive voltages, including the VCC rail? I will see if i can find some time to reproduce your measurements, as i have never been able to repair the broken channel.

[SparkyBruce]

I will probe the gate drives and VCC's tomorrow being careful not to break my working channel.

I do have the replacement LTC parts now - but swapping them will be my last resort as I'm not sure that they are actually faulty.

I am seeing the low side diodes getting warm so it could conceivable be that the low fets aren't actually being turned on at all and it isn't working in synchronous mode at all (that would warm up the diode and reduce the efficiency).

[SparkyBruce]

Well I have checked the gate drives and VCC's of all three channels and they are identical.

So apart from the different low side diodes the three outputs are the same.

Hopefully the replacement "proper" diodes will arrive soon (exactly the same as the original parts).

Now I am wondering if there is something different elsewhere in the PSU channels that might be causing the quiescent current difference.

[electr_peter]

Has repair affected the accuracy? Most likely not based on changed parts, but it would be a bummer to repair it and loose accuracy. HMP has digital calibration, only R&S has special SW for that.

[SparkyBruce]

Hi,
As the repair is in the pre-regulated switcher supply I have not touched the linear output regulator so both voltage and current settings are still good (which is very lucky).

The pre-regulator runs a couple of volts above the output voltage demand to reduce the wastage in the linear output circuit.

Bruce