hi, as was mentioned, rf designs and designers are weird. their weirdness is only outdone by antenna designs and designers. theirs is not a 'plug-and-play' world, so tweaking may be inevitable. phil -----Original Message----- From: Marty Brooks [mailto:[log in to unmask]] Sent: Thursday, March 29, 2001 3:12 PM To: [log in to unmask] Subject: Re: [TN] RF Assembly Issues Thanks everyone for your input. To respond to some of the comments: Phil - Yes, we removed the chip and resoldered the same chip to the pwb. Problem is How can I tell the vendor to perform a labour intensive rework procedure in a mass production environment without the vendor charging me for it. I will argue their process is causing the problem, they will argue it's our design. And yes, we did replace the poorly performing part with a 'new' one and performance was much better with the new part. This all leads me down the path of solder joint quality - I just need some way to prove that to the vendor. Steve - I think the bottom only terminations of the chip does pose a problem. Have other assembly shops had problems with devices like this and how have they overcome them? The chip is a ceramic base with gold plated terminations. Dave - I've considered the trace dimensions. They are all optimized as best they can, except for the Xceiver chip which has pads 10mil longer than recommended by the manufacturer. I wonder if that extra length, as small as it is, would be enough to throw the impedance out of whack? Something that I will look into. Ingemar/James - I'm no RF engineer either. The device operates at 900MHz. Board is FR4 with a dielectric constant of 4.1 or 4.2. Board is impedance controlled to 50ohms. Surrounding components are 0603 resistors and caps - except for the RF output line which goes through a 15nH wire wrapped inductor. We have optimized the layout and moved all the components as close together as possible to minimize any transmission line anomalies that may want to appear. As far as your other questions, naturally I don't have those at my fingertips, but I'll see what I can dig up and present them to the first RF guru I can track down. You have provided some new points for consideration though. Thanks Marty -----Original Message----- From: Ingemar Hernefjord (EMW) [mailto:[log in to unmask]] Sent: Thursday, March 29, 2001 6:51 AM To: [log in to unmask] Subject: Re: [TN] RF Assembly Issues Marty, sounds like you have to use a fishbone modelling. So many parameters that could have impact. You don't say what kind of part you are making, something for under 2GHz or much above? And nothing about dielectric and surrounding components, ground planes, distriubution kind of traces, noise level requirements, spurioses, I/O-loading etc. Not easy for an RF engineer to give advice. Have you modelled the circuit and simulated variations in a computer? If you are talking very high frequencies, even length of bondwires on inside of package plays a role. And what's the package made of, ceramic or plastic? If you ask an RFspecialist to engage you need prepare to give information about: epsilon as function of temperature for your board material, as well as for all the components RF-parameters (S e.g.)as function of temperature possible hysteresis of your ceramic caps, if you use such (behave different on way up and down T) identification of the parameter that gives your freq shift (Temp sensitive part on inside of 20leg) checking if you really use valid test set up (microwave is tricky) checking possible ferroelectric phenomenons in ceramic components, if you use such what kind of tuning components do you use? Are they temperature sensitive? As James said, check all return paths, also crosstalking and feedback traces. and many more questions, depending on the complexity of your circuit diffusion phenomenons that can change chip data Tin, Silver migration that can cause bridging I recommend you to see nearest RF engineer in person. RF problems are rarely solved by mail. If it is a RF-problem at all! Try not to solder the 20pin at all. Mount some 10 boards using silverepoxy mounting the 20pin after first soldering all other components. Aftermounting with other words. If all 10 get center freq where it should be, then the 20pin could be temp sensitive. In fact your questions seems to reveal that you are not making an advanced HF circuit but something for the commercial 'low freq' market, a radio board or so. Important for you to tell TN, because the 'RF' guys are either ordinary and sound radio/TV/equipment makers..or Telecom/MIL/SPACE 'real' RF neurds with doctor's grade and at least 500 empty Coca bottles on the wall..2GHZ is called DC among the later... Ingemar -----Original Message----- From: Marsico, James [mailto:[log in to unmask]] Sent: den 29 mars 2001 14:50 To: [log in to unmask] Subject: Re: [TN] RF Assembly Issues Most of the time, RF components need the body (under the component) soldered to the PWB for proper grounding. I'm no RF engineer, (God forbid) but I understand that this grounding can be critical for performance. Perhaps this is a good place to start. Jim Marsico Senior Engineer Production Engineering AIL/Electronics Systems Group An EDO Company [log in to unmask] <mailto:[log in to unmask]> -----Original Message----- From: Marty Brooks [SMTP:[log in to unmask]] Sent: Wednesday, March 28, 2001 4:41 PM To: [log in to unmask] Subject: [TN] RF Assembly Issues We are experiencing a concerning variation from board to board in the performance of our RF product. We have investigated several causes of the problem and are currently of the opinion that soldering quality of the transceiver chip is the most significant contributing factor. As an example, on the last run of prototypes that was manufactured, the signal frequency was at the upper limit of the allowable range. We removed and replaced the transceiver chip in-house and decreased the frequency closer to what we would expect. I'm going to be making a visit to our contractor's facility next week to investigate this problem and would like some feedback on possible causes or process controls that I should look into. The board is 4 layers, HASL finish, single side smt, solder paste used is Alpha's UP78 no-clean. The transceiver chip is a 20-pin device with terminations flush along the bottom of the component. Thanks for any advice. Marty Brooks Manufacturing Engineer IDENTEC Solutions, Inc. 102, 1860 Dayton Street Kelowna, B.C., Canada V1Y 7W6 Ph:(250)860-6567, Fax:(250)860-6541 www.identec.com This email contains confidential information and is intended for the sole use of the addressee. Any other distribution, copying or disclosure is strictly prohibited. If you are not the addressee, please notify the sender by email and delete this message. ---------------------------------------------------------------------------- ----- Technet Mail List provided as a free service by IPC using LISTSERV 1.8d To unsubscribe, send a message to [log in to unmask] with following text in the BODY (NOT the subject field): SIGNOFF Technet To temporarily halt delivery of Technet send the following message: SET Technet NOMAIL Search previous postings at: www.ipc.org > On-Line Resources & Databases > E-mail Archives Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315 ---------------------------------------------------------------------------- ----- ---------------------------------------------------------------------------- ----- Technet Mail List provided as a free service by IPC using LISTSERV 1.8d To unsubscribe, send a message to [log in to unmask] with following text in the BODY (NOT the subject field): SIGNOFF Technet To temporarily halt delivery of Technet send the following message: SET Technet NOMAIL Search previous postings at: www.ipc.org > On-Line Resources & Databases > E-mail Archives Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315 ---------------------------------------------------------------------------- ----- ---------------------------------------------------------------------------- ----- Technet Mail List provided as a free service by IPC using LISTSERV 1.8d To unsubscribe, send a message to [log in to unmask] with following text in the BODY (NOT the subject field): SIGNOFF Technet To temporarily halt delivery of Technet send the following message: SET Technet NOMAIL Search previous postings at: www.ipc.org > On-Line Resources & Databases > E-mail Archives Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315 ---------------------------------------------------------------------------- ----- --------------------------------------------------------------------------------- Technet Mail List provided as a free service by IPC using LISTSERV 1.8d To unsubscribe, send a message to [log in to unmask] with following text in the BODY (NOT the subject field): SIGNOFF Technet To temporarily halt delivery of Technet send the following message: SET Technet NOMAIL Search previous postings at: www.ipc.org > On-Line Resources & Databases > E-mail Archives Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315 --------------------------------------------------------------------------------- --------------------------------------------------------------------------------- Technet Mail List provided as a free service by IPC using LISTSERV 1.8d To unsubscribe, send a message to [log in to unmask] with following text in the BODY (NOT the subject field): SIGNOFF Technet To temporarily halt delivery of Technet send the following message: SET Technet NOMAIL Search previous postings at: www.ipc.org > On-Line Resources & Databases > E-mail Archives Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315 ---------------------------------------------------------------------------------