Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Vic Lau

#10930

problem when testing the RF product | 19 June, 1999

Do anyone have any experience on assembling high frequency product, 1.8GHz or 2.4GHz.

In our production line, we manufacture 1.8GHz product with using no clean paste (RMA type) for reflow and no clean wire for rework. In our functinal test,either at 25degC or 50degC, no malfunction is observed.

However when switch to 2.4GHz product,we encounter malfunction of the product at 50deg C. The paste we are using is also a no clean one, but of ca.100ppm halide. I just wonder the dielectric constant of the flux residue will vary with the test temp!?

I look up some test book, it state for RF product, cleaning is required, is it true?

Dear all, do you have any comment? Thanks in advance

reply »

Scott Cook

#10931

Re: problem when testing the RF product | 20 June, 1999

| Do anyone have any experience on assembling high frequency product, 1.8GHz or 2.4GHz. | | In our production line, we manufacture 1.8GHz product with using no clean paste (RMA type) for reflow and no clean wire for rework. In our functinal test,either at 25degC or 50degC, no malfunction is observed. | | However when switch to 2.4GHz product,we encounter malfunction of the product at 50deg C. The paste we are using is also a no clean one, but of ca.100ppm halide. I just wonder the dielectric constant of the flux residue will vary with the test temp!? | | I look up some test book, it state for RF product, cleaning is required, is it true? | | Dear all, do you have any comment? Thanks in advance | Hmmm....sounds like someone is working with the new SiGe wireless-LAN chipsets.

I know others out there will not necessarily agree with me, but here goes....

In the frequency ranges you are working in, I personally would never ship uncleaned product.

RF is a funny world. You can nail down your processes tighter than a drum skin.....only to run up against what can only be referred to as the vodoo and black magic which is embedded with the RF monster. Yes, we've come a long way. But as much as technology has advanced, there is still art involved in designing and building RF products. A really good RF design and test engineer names their own price these days......

Using RMA no-cleans, I've done Joe Keller's recommended test using distilled water and voltage application across the traces on RF product before--hello dendritic growth.

If you'd like the test parameters, I can dig them up out of a box in my barn.......

Scott

reply »

Glenn Robertson

#10932

Re: problem when testing the RF product | 21 June, 1999

| Do anyone have any experience on assembling high frequency product, 1.8GHz or 2.4GHz. | | In our production line, we manufacture 1.8GHz product with using no clean paste (RMA type) for reflow and no clean wire for rework. In our functinal test,either at 25degC or 50degC, no malfunction is observed. | | However when switch to 2.4GHz product,we encounter malfunction of the product at 50deg C. The paste we are using is also a no clean one, but of ca.100ppm halide. I just wonder the dielectric constant of the flux residue will vary with the test temp!? | | I look up some test book, it state for RF product, cleaning is required, is it true? | | Dear all, do you have any comment? Thanks in advance | | Vic - There has been some work done in this area, but the publications are sparse. You might want to look on page 242 of Jennie Hwang's book "Modern Solder Technology", and there was a short paper on "Effect of Solder Paste Residues on RF Signal Degradation" in the SMI98 Conference that described a test setup. If you (or anyone else) find other information please post it here.

Glenn Robertson glenn.robertson@usa.alcatel.com

reply »

Scott Cook

#10933

Re: problem when testing the RF product | 21 June, 1999

Since Vic posted me privately and asked for the SIR test parameters I spoke of in my other post, I thought I'd respond on the forum.

Ok, here they are, per Joe Keller--Solder / process consultant AKA Manko.....

This is an accelerated SIR / dendrite test......Joe claims that to be valid, traditional SIR testing must cycle the PCA through it's dewpoint.

Test: Take a board or coupon with SMOBC mask. Joe recommends using pads / trace spacings of both .005" and .010".

Place a drop of 1 megOhm water so that the pads / traces both contact the water.

Apply 0.3 V /.001" spacing at < 300 micro Amps for one minute (i.e. 1.5 volts at .005", 3 volts at .010").

That's it. Watch for dendritic growth. It is dramatic to see it occur. It happens very quickly.

Scott Cook

reply »

Scott Cook

#10934

Addendum / correction | 21 June, 1999

| Since Vic posted me privately and asked for the SIR test parameters I spoke of in my other post, I thought I'd respond on the forum. | | Ok, here they are, per Joe Keller--Solder / process consultant AKA Manko.....

This SHOULD have read "ALA Manko"; not AKA Manko. Fumble fingers......

Scott

reply »

Vic Lau

#10935

Re: Addendum / correction | 22 June, 1999

| | Since Vic posted me privately and asked for the SIR test parameters I spoke of in my other post, I thought I'd respond on the forum. | | | | Ok, here they are, per Joe Keller--Solder / process consultant AKA Manko..... | | | This SHOULD have read "ALA Manko"; not AKA Manko. Fumble fingers......

Dear Scott,

Thanks for your information. It is very useful to me. As there is some problem in my email, thus I can only send you email wiht my personal account, hope you don't mind.

Cheers, Vic Lau | | Scott |

reply »

Vic Lau

#10936

Re: problem when testing the RF product | 22 June, 1999

| | Do anyone have any experience on assembling high frequency product, 1.8GHz or 2.4GHz. | | | | In our production line, we manufacture 1.8GHz product with using no clean paste (RMA type) for reflow and no clean wire for rework. In our functinal test,either at 25degC or 50degC, no malfunction is observed. | | | | However when switch to 2.4GHz product,we encounter malfunction of the product at 50deg C. The paste we are using is also a no clean one, but of ca.100ppm halide. I just wonder the dielectric constant of the flux residue will vary with the test temp!? | | | | I look up some test book, it state for RF product, cleaning is required, is it true? | | | | Dear all, do you have any comment? Thanks in advance | | | | Vic - | There has been some work done in this area, but the publications are sparse. You might want to look on page 242 of Jennie Hwang's book "Modern Solder Technology", and there was a short paper on "Effect of Solder Paste Residues on RF Signal Degradation" in the SMI98 Conference that described a test setup. | If you (or anyone else) find other information please post it here. | | Glenn Robertson | glenn.robertson@usa.alcatel.com | Dear Glenn,

Thanks for your information, I will check the Jennie Hwang's book "Modern Solder Technology".

Thanks again.

Victor Lau

reply »

Frank Frimpong

#10937

Re: problem when testing the RF product | 23 June, 1999

| Do anyone have any experience on assembling high frequency product, 1.8GHz or 2.4GHz. | | In our production line, we manufacture 1.8GHz product with using no clean paste (RMA type) for reflow and no clean wire for rework. In our functinal test,either at 25degC or 50degC, no malfunction is observed. | | However when switch to 2.4GHz product,we encounter malfunction of the product at 50deg C. The paste we are using is also a no clean one, but of ca.100ppm halide. I just wonder the dielectric constant of the flux residue will vary with the test temp!? | | I look up some test book, it state for RF product, cleaning is required, is it true? | | Dear all, do you have any comment? Thanks in advance | | Hi Vic, I posted your problem on the RF global net and here are the responses. There is a paper on this subject in the Technologies for Wireless Applications Digest, 1997 I believe. E-mail me at gcd@ghz-design.com for reference details if interested. Geoff Dawes

We had some major flux problems down in the VHF and UHF areas. It was more simplistic... a DC leakage path. It caused problems in our PLL loops by loading the voltages down in any high impedance areas. It would change as the voltage was run through it and it heated. We found people on the line using flux to solder parts that were placed onto the board after the reflow process. It never 'burned' off. Bud Simciak

reply »

Essemtec Spider SMT Fluid Dispenser