isolation between layers

Can you remove the metal from the layers in the HV region? Laser printers commonly use regular old PCB materials with large spaces between components for HV (corona wire) circuits that seem reliable enough. That circuit board doesn't have groundplanes near the HV bits, though.

FR4 should be fine. There is plenty of data available about its dielectric strength. A value of 20kV/mm seems quite commonly quoted.

If you are specifying a multi-layer board with high voltages between adjacent layers it would be a good idea to ask the manufacturers to use at least two sheets of pre-preg between each copper layer and to derate the design as if one of those layers has a hole in it. Thay way you will cope with pinhole defects in any one prepreg layer. (Being infrequent, they are unlikely to overlap.)

Dielectric strength between adjacent tracks in the same layer is better than for surface tracks but not as good as that between different layers - European safety standards for some products make this distinction.

Don't forget to consider surge voltages and to use the peak rather than the rms voltage for ac signals as dielectric strength data is usually measured with dc.

If this is for a commercial product the relevant safety standards may give additional guidance.

John

Yes and no. If you need thin isolation, Kapton (TM) is nice - about

Rob L schrieb:

Hello,

where is the high voltage, between tracks on the same layer, or between crossing tracks on different layers?

You may order a multilayer PCB with more than the standard thickness of

Bye

Rob L writes

You might be better off using wire links across the PCB with, say, PTFE insulation. No tracking issues keeping them away from other tracks, unless they're carrying fast edges. Assuming this is a commercial product, this would also be easier to get through safety certification because you won't have to lock the PCB layout down to a revision which has been inspected and approved by a 3rd party; the isolation would depend on a UL recognised wire type specified in the parts list.

Thank you - and all the other respondants - very much.

Unfortunately zero ohm linking is not a practical proposition in this case.

The spec is very poor - e.g. it specifies a 'flash' test 4KV a.c. but does not specify any limits (time, current, frequency) so presumably s long as it does blow up itself or the planet when 4KV is applied it is deemed OK. As a silly old engineer I just want to make a proper job of it for what it must do rather than test it for leakage currents of less than 1kA when HiPotted at 4kV at 0.000001Hz for 1 nanosecond.

Regards

--- news://freenews.netfront.net/ - complaints: snipped-for-privacy@netfront.net ---

case.

does=20

as it=20

As a=20

must=20

HiPotted=20

(just 3

PTFE=20

unless=20

this=20

won't=20

inspected=20

recognised=20

Rob L, please get that flash spec defined. You are flapping in the = breeze=20 until you do. Somebody at your "customer" thinks it is well defined, = make=20 them cough up that definition.