Circuit board insulation and high voltage

I am having a problem with leakage currents through a circuit board.

A product I am working on has a dc-dc converter that converts 12V from
a battery to 350Vdc. The two power buses are mostly isolated, where
the only connection between them is a differential amplifier with a
1MOhm input impedance. The input impedance of the amplifier centers
the high volt output relative to the battery ground. So relative to
battery ground the high volt supply is +/- 175V.

I recently added a safety circuit to detect leakage currents between
the high volt lines and battery ground. This looks at the voltages on
the high volt bus relative to battery ground, and if they are not
centered, there is a leakage current. This way the system can detect
a fault.

This worked great through development, but now in production I am
seeing a 3-5% fallout due to this fault. The boards that pass are
dead nuts on. There is no discernible leakage current to the limits
of my measurement equipment. The boards that fail show an obvious
leakage current, the impedance of the defect is within 100kOhm to a
couple MOhm. This only happens with high voltage, a regular DVM shows
an open circuit. This appears to be due to a defect within the
circuit board itself. When I look at it with a scope I can see the
voltages moving around, maybe due to partial discharges. On a couple
of the boards I have watched the fault clear itself, and have not been
able to get it to reappear.

I am looking for high voltage specifications for PCBs. The design
follows the IPC-2221A guide lines for creepage and clearance. Are
there other material specs i should be calling out?

Does anybody have an idea of what the failure mechanism is? If a
board shows a fault that clears itself, presumably burning out
whatever contamination is causing the fault, is it more likely to fail
in the future? Are there processing issues that could cause a large
leakage current? I have seen a board go through a lead free soldering
process and come out charred, but I am pretty sure that is not what is
going on here.

Thanks,

Ethan

 

Water. What process are you using to solder and clean the assemblies?

Also, you should bake them at 60C for an hour at least, if not 70C,
after they have been thoroughly cleaned. A few seconds in a vacuum
chamber will also release hidden water molecule entrapments.

After that, a good coating of conformal media might be a consideration.

The other thing is that all components on the HV <sic> side should be
kept clean from human contact during assembly and all other handling.
That is everything from finger oils to "coffee breath". Such residues
cause leakage in real HV circuits, perhaps they do also in your medium
voltage application. Any capacitor bank you place across the output also
should be elevated from the PCB such that you do not provide the path
where trapped VOCs etc. can set-up said leakage paths.

 

The solder process is a regular leaded solder, for now we are only
selling in the US, so it is not RoHS compliant. It is a surface mount
reflow process then a wave solder process for the through hole
components. The solder uses a water soluble flux, then the board goes
through a normal wash and drying process. It seems the drying process
is not drying enough.

There are some fairly big capacitors on the board. I am wondering
about their ability to capture moisture. It looks possible.

The assembly is eventually potted. I want to make sure everything is
clean and dry before potting captures whatever contaminents.

What I need right now is a way to solidly confirm that this is a
problem with moisture.

Thanks,

Ethan


I am going to try leaving the next batch of board in the oven at 65C
overnight, to see if that helps. I also need to try to find a board
that fails consistently, and see if your vacuum chamber suggestion
works, if so that would be good confirmation of water, or some other
volatile on the board.

 

The main problem I have had is where flux gets trapped under a
component and can then cause leakage when it absorbs moisture.

I was using voltages up to 150V with 10Meg 1206 resistors and the
circuit was sensitive to about 10nA of leakage (equivalent to a
leakage resistance of ~10G Ohm).

One approach that seemed to work in this scenario was to relieve the
solder mask under the resistor and put a grounded trace between the
two pads.

Removing the solder mask gave a few mil of clearance for the water
wash to get in and clean the area while the grounded trace diverted
any leakage currents.

The level of leakage you are getting seems extremely high. Are you
sure that the active circuitry is not causing a problem - what defines
the 1 meg input resistance of the amplifier?

kevin

 

Do you have the option of redesigning the PCB? For very high voltages,
you sometimes see PCB's with slots (air gaps) in them to absolutely
guarantee there's no leakage from surface contamination.

Most PCB's these days are FR4 material with some kind of UL flame
retardant spec. Your board manufacturer (or layout engineer) will have
loads of advice about materials - maybe you're using a cheap material.
Perhaps just changing to a thicker PCB will fix it.

How do you know it's not your leakage monitoring circuit? Oh yes - it is
measurable at high voltage. Maybe the leakage monitoring circuit has a
component that needs a higher voltage rating... 350V AC rather than
DC...?

 

It is not a problem with the voltage monitoring circuit. I have
tested the circuit and it is working correctly. The voltage rating of
the components have plenty of margin, The input resistance to the diff
amp is built from four resistors in series, each is rated for 200V.
The creepage distances on the board were maintained at 2.5mm, which is
the IPC-2221A recommendation.

The IPC guidelines for voltage creepage and clearance in my experience
are pretty reasonable. They are conservative enough that everything
works reliably, but not as conservative as something like UL
"inherently safe". I use the UL specs in places where someone could
get a shock, and the IPC specs if all I need to do is make the
equipment work reliably, and it is not a safety issue.

I have become convinced that my problem is due to moisture
contamination. Baking out the boards at 65C overnight fixes the
problem. This is not the best solution, adding a 12 hour step to the
production flow is not exactly lean manufacturing. Also, if there is
moisture trapped under a component, it is likely to have not been
cleaned adequitly and leaves some ionic contamination even after it
has dried out. So moisture due to high humidity or something would
allow a leakage current again. Earlier someone suggested raising the
large capacitors off the board a little bit to allow the wash to get
under them, and then to allow them to dry out seems to be the way to
take care of the problem.

Ethan

 

It IS the best solution, but 12 hours is more than a bit of overkill.
When you are talking about HV design, baking out of assemblies is a
required step and should be factored into a cost analysis. That IS as
lean as it gets.

The voltage across your multiplier, assuming there is one, would
determine whether or not the multiplier section of the device should be
potted. Also human contact circumstance, as you indicated., but that
cannot be avoided at the output lead or port of an HV supply case. Only
as it relates to the guts, and that should not be having human proximity
anyway.

A few things that will insure a long lived circuit in WallyWallWhackr's
HV circuit realm are:

Hard anodized cases. Hard anodized Aluminum yields a non-conductive
surface, reducing any coronal attraction that a case might normally
cause.

Use of an insulative partition. Using 'Nomex' insulator paper
(transformer paper) or a suitable polymer plastic sheet, or Kapton tape
layering is needed in the HV section of a design typically for any
un-potted HV device, and has benefits when properly positioned in potted
cases as well. Potting only the HV end is 'doable' and allows for a bit
more serviceability on the control/drive section.

Leave any multiplier area free of any solder mask (yes, bare board).
Solder mask adheres to potting very poorly. It adheres to bare FR4 or
G10 much better, IF the board is clean and any primer step followed.

Use slots between nodes that are separated by high potential
differences. Have the PCB maker use his 0.035 mill to cut slots, just
like those used when arraying boards out on a palette setup. Between the
nodes of the HV caps, and diodes. I can post a picture in abse
(alt.binaries.schematics.electronic) if you want to see a good multiplier
design, if you get that group. The slots allow potting to flow into them
and that makes partitions between the nodes that are nearly impossible to
breach.

Clean the boards VERY well both before and after the assembly / solder
process. Keep fingers OFF of HV components and the PCB. ALL of the PCB.
circuit boards are hygroscopic. They grab water like a sponge, and
finger oils as well. If your process is aqueous, the oils cause a
problem in the HV area. Even 'coffee breath' causes a problem without a
subsequent cleaning step. Aqueous usually gets air knifed fairly dry,
but a good two hour bake should evaporate any and all remaining water.
A few seconds in a vacuum doesn't hurt either for pulling water out of
trapped pockets (another reason to use no mask)

If your process uses solvents instead of water, you should consider a
hot alcohol bath and rinse before you bake the assembly out, and a vacuum
should be looked into as the board will want to suck up water right away.
For integration into the case, you should either be careful of the
humidity level in the room or do a quick bake again while in the case.
All bakes are at your temp of 65C.

I would not go the 12 hour bake though as that shortens your EL cap life
a bit, and the water evaporates even at less than boil temps, so the job
gets done pretty quickly.

If you are not potting, then a conformal coat will seal in a well baked
board so it no longer sucks water like Bill Gates sucks corporate cash
coffers.

WallyWallWhackr

Bakings? We don' need no stinking bakings!

 

Elevating parts is good (shoot for less than 2mm). It usually violates
IPC-610, but in the HV realm, we do all kinds of things not done
elsewhere (solder blob nodes, for example). What you get is a clean wash
and low VOCs. A good conformal coat, particularly a vacuum aided
conformal coat session will seal it up real good too.

Try Dolph's for sealing things up to industry specs. That coating
application will also end up adding stanchion to the elevated parts.