What's a good PCB surface creepage clearance for 1200 volts?
--
John Larkin Highland Technology, Inc picosecond timing precision measurement
jlarkin att highlandtechnology dott com
What kind of enviroment?
Rough number would be between 12mm and 19mm.
This is neat
Cheers
Den onsdag den 3. februar 2016 kl. 22.14.10 UTC+1 skrev Martin Riddle:
this looks like similar numbers
~6mm x 1.5-3.2 for functional isolation and creepage
-Lasse
A cut in the PCB material, with at least 1/2-inch travel to make it around the cut. :-)
--
Thanks,
- Win
Inside an optics box, so it should be cool and clean. Available energy is low, under a joule.
Yikes, I was figuring 160 mils, about 4 mm.
The 2SK4177 mosfets that I'm using, D2PAKs rated for 1500 volts, have about 145 mils clearance from the source and gate pads to the drain pad. I could cheat the pad sizes a bit and pick up a little more. But there's no point in making other clearances much bigger than that.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
I really like that 40 volts/mil number.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
This board is packed with parts, 231 to be precise, and there's HV stuff everywhere. I posted the power supply schematic, and there's 6 more sheets of HV stuff.
I was figuring 175 mils maybe. I don't plan to UL or CE test this board, but my customer might do the system. At 12 mm clearance, the board would be the size of an XL pizza.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Conformal coating can increase voltage withstand considerably.
That is, except for connector pin spacing which of course can't be coated.
boB
A clean enviroment, 5 - 7mm And Milling between the pin pads helps alot.
Cheers
In the end it all depends on the standard it has to fulfill. And of course the environmental exposure such as altitude.
For coated traces that can suffice, depending on application.
One important issue is to give everything a radius. No sharp corners in traces, not even 45 degree, but rounded. Then make the corners of footprints slightly round. IOW, you would have a special "HV footprint" for 1206 and similar.
Watch for working voltage ratings of resistors. Not breakdown voltage but working voltage. HV resistors with high tolerances that do not have laser trim cuts are usually rated much higher so it's an advantage to use 5% where possible.
Milled out air gaps can also help.
Production: Watch out for flux build-up underneath parts with high voltages across them.
-- Regards, Joerg http://www.analogconsultants.com/
Some parts I'm using:
2SK4177 D2PAK mosfet rated 1500 volts 145 mils between pads (calculates to 10.3 volts/mil)ACPL227 opto rated 3KV RMS 4 mm between pads (26 v/mil) claims IEC60747
TLP191 opto rated 2.5 KV RMS 5.6 mm between pads (16 v/mil) claims UL1577
LND150 sot-23 depl mosfet rated 500 volts 33 mils between pads (15 volts/mil)
My tentative 1200 volts and 160 mils is 7.5 v/mil!
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
The amount of energy would seem to be the big factor here. Under a joule is small, so the smaller creepage distances to match the terminals of your device seem sound. If this was 1200 volt power substation switch gear that spacing would not be acceptable. The sealed-ish box would prevent dust from building up as well. Even traces under 1200 volts can accumulate that soot-like dust over time.
Referring to this board,
I forget where the trace all goes, but it's at something like 1-1.5kV, for pulsed focus or something like that, in a Trinitron monitor. It eventually goes to the CRT socket.
For scale, the triangular footprints are TO-220 transistors, so the legs are 200 mils apart. Clearance looks to be on a similar scale. Anywhere it's closer, they've routed the creepage away.
The monitor has been in fine operating condition since mid 1999, under low pollution conditions (just a thin smear of dust on the PCB surface).
I don't know that silkscreen over the trace is anything the UL would pay attention to, but obviously, Sony figured "why not".
On a related note: I recently did an IEC 60950-1 review, and can't seem to find any indication whether soldermask meets their definition of "suitable coatings". The PCB manufacturer didn't know either. Such coatings are subject to regular testing and inspection anyway, so wouldn't be a good idea for a cheap product.
Tim
-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website: http://seventransistorlabs.com "John Larkin" wrote in message news:m4q4bblgsq0k1icibju8vhpuktj1846t52@4ax.com... What's a good PCB surface creepage clearance for 1200 volts?
Be careful not to mix up working and rated voltage. For example, the ACPL227 is rated for 3kV at a maximum of one minute. That appears not to be the working voltage.
Datasheets are often vague and one has to cajole the full ratings out of the manufacturer, preferably in writing. I always prefer parts that have complete datasheets, in this case with a stated working voltage and the standard per which it is declared.
This is a properly spec'd optocoupler with clearly stated working voltages:
-- Regards, Joerg http://www.analogconsultants.com/
Air pressure ? Sea level, 4 km, 10 km, 30 km ? Relative humidity ? 0 % or 100 % ? Coating material used ?
That 1200 V doesn't sound very much, the Octal base could handle that.
However, any mains connected system would require clearance for 1500 or 2500 V for a minute or so.
For IEC60950, you really need to know the "pollution degree" of the location, and "Comparative Tracking Index" of the PCB material as that will determine the creepage. It could give distances anywhere from 4.2mm to 20mm depending on the CTI and pollution degree. The PCB manufacturer should be able to find out the CTI of their material, and you may need to provide evidence of this if you have to go to a test house.
Probably if you pot it, you would be allowed anything down to 0.4mm (though with anything less than about 1mm, I would be worried about how to expel air bubbles during potting, and prove this has been done).
Depending on the category of product, some other standard might apply instead, which may be much more vague and/or open to interpretation (by other people too, which is usually not an advantage).
The HCNW version has recommended pad spacing of 350 mils. That's rated for 5KV RMS for 1 minute, 20 volts/mil. The IEC rating is 1414 volts peak, I assume sustained, which is 4 volts/mil.
The truth must be in there somewhere.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Looks like my transistors are the pacing items. The 2SK4177 is rated for 1500 volts and has about 150 mils between the source and drain pads.
I'm only running the LND250 sot-23 mosfets at 500 volts, but that clearance is 33 mils.
According to some rules, these parts can't be used uncoated at more than a fraction their rated voltages. Even coating seems dicey.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
The standard referenced by this website has a different table (2K) for clearance between non-safety-critical conductors.
Below 1400V, 5mm creepage between functionaly isolated traces is considered sufficient to avoid bridging tests.
Creapage between traces in similarly grounded circuits can be any distance, providing that bridging them demonstrates not to produce a hazard.
The old mil standard of 8V/mil would suggest 3.8mm for a 1200V clearance. In more recent standards this reduces, where material quality justifies it.
You should keep in mind that adjacent traces have edges that represent inreducable 'sharp' edges...
RL
RL
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.