Noticed any recent bare PCB leakage problems?

[Geoff C]

Hi,

Has anyone noticed a recent increase in board leakage in FR4 Double sided
boards? We have some high gain preamps that have worked for 8 years no
probs, then found leakage on boards from an Aussie and a Malaysian
supplier. This first happened around November.

 

[Greg the Grog]

Hi,

Has anyone noticed a recent increase in board leakage in FR4 Double sided
boards? We have some high gain preamps that have worked for 8 years no
probs, then found leakage on boards from an Aussie and a Malaysian
supplier. This first happened around November.

Are you saying that your circuit boards have incontinence issues (like Phil
;-)
or are you suggesting that the dielectric constant of FR4 material is no
longer
as constant as you would like?

Should a high gain preamp design be so reliant on substrate material
properties
that manufacturing variations can adversly effect the circuit performance?

More info pls.
Is it a matching issue, stripline calcs are no longer valid or dieletric
losses are
making your preamp not such high gain?

You say leakage, so I assume you are not flashing an LED with a 555, but
perhaps a little faster. FR4... not above 1GHz if you can get away with it.

Testing the boards is the only option and make sure there were no
component / supplier changes around about the same time.

Matching / Q gone out the window.... I hate it when that happens.

cheers,
GtG

 

[David L. Jones]

Hi,

Has anyone noticed a recent increase in board leakage in FR4 Double sided
boards? We have some high gain preamps that have worked for 8 years no
probs, then found leakage on boards from an Aussie and a Malaysian
supplier. This first happened around November.

What values are we talking about here?

I find it hard to believe that a bare PCB could have leakage that low
to be a problem, even with very high value impedances (100Mohms+).
Are you sure it's the bare board and not a board loading/washing
contaminant?

Dave

 

[The Real Andy]

Are you saying that your circuit boards have incontinence issues (like Phil
;-)
or are you suggesting that the dielectric constant of FR4 material is no
longer
as constant as you would like?

Should a high gain preamp design be so reliant on substrate material
properties
that manufacturing variations can adversly effect the circuit performance?

More info pls.
Is it a matching issue, stripline calcs are no longer valid or dieletric
losses are
making your preamp not such high gain?

You say leakage, so I assume you are not flashing an LED with a 555, but
perhaps a little faster. FR4... not above 1GHz if you can get away with it.

He said high gain, not high frequency.

Testing the boards is the only option and make sure there were no
component / supplier changes around about the same time.

Matching / Q gone out the window.... I hate it when that happens.

I would be interested to know exactly what 'high gain' means.

With 8 years of success, I would be looking at
manufacturing/components before blaming the substrate.

 

[Mark Harriss]

I would be interested to know exactly what 'high gain' means.

With 8 years of success, I would be looking at
manufacturing/components before blaming the substrate.

This can be a problem with some opamp circuits: Nat Semi
suggests lifting one IC leg into the air and wiring the
components in air.

 

[Geoff C]

Are you saying that your circuit boards have incontinence issues (like
Phil ;-)
or are you suggesting that the dielectric constant of FR4 material is
no longer
as constant as you would like?

Should a high gain preamp design be so reliant on substrate material
properties
that manufacturing variations can adversly effect the circuit
performance?

More info pls.
Is it a matching issue, stripline calcs are no longer valid or
dieletric losses are
making your preamp not such high gain?

You say leakage, so I assume you are not flashing an LED with a 555,
but perhaps a little faster. FR4... not above 1GHz if you can get away
with it.

Testing the boards is the only option and make sure there were no
component / supplier changes around about the same time.

Matching / Q gone out the window.... I hate it when that happens.

cheers,
GtG

Frequencies are low, 100 Hz samplin and sqare wave, sensitive to a few
kHz and DC. I measured the baord with a Kiethley electrometer and it is >
100Gohm but I will test more rigorously next week. Yesterday I drilled
the PCB and air connected the opamp node, and all was OK. The f/b R is 10
meg, input is photodiode ground connected. Opamp is OPA404.
Sure, FR4 is not always suitable for high gain, but I would rather stick
with low cost FR4 and air connect the 4 connections to the node. The
board has worked fine for 8 years. Different suppliers no prob. No known
component changes. All contending components were swapped b/n a good and
bad board but prob followed PCB.
Bottom line is I have a fix. I will crunch numbers next week. Good to
have a solid explanation, but at least production is running. BTW, tried
baking and solvent cleaning, and latest baord had guarding artwork on the
sensitive node. Of course, this only works for surface currents.