day shot to hell
in [Design]
|
Prev: Audio VCO design
Next: Fantastic new audio amp !
From: Joerg on 20 Nov 2009 21:30 a7yvm109gf5d1(a)netzero.com wrote: > On Nov 20, 7:22 pm, John Larkin > <jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: >> OK, I just got the first board from production this morning, for this >> spectroscopy controller thing. >> >> ftp://jjlarkin.lmi.net/First.JPG >> >> It gets 12 volts in, which runs an LTM8023 switcher brick to make 3.3 >> volts. The 3.3 runs most of the logic on the board (including a >> Spartan 6 and a PLX PCIe bridge, both BGAs) and also drives four >> secondary switchers and some LDOs to make 1.2, 1.5, 1.8, 2.5, and -5 >> for various uses. >> >> So when I powered it up everything went nuts. The PLX chip was >> obviously fried. After that was pulled, the Xilinx was running hot, >> and the 3.3 volt supply was bogged down to about 2.6. The LTM >> regulator was hot. >> >> Pulled the Spartan BGA next. >> >> Now the 3.3 volt rail wants to run at 5 or so. >> >> After much head scratching, I discovered this: >> >> ftp://jjlarkin.lmi.net/Swapped.jpg >> >> The resistor that's screened "R127" is actually R129. And vice versa. >> So the switcher was programmed wrong, told to run at an absurdly low >> frequency and an absurdly high voltage. The ref designators somehow >> got misplaced during layout. We usually check for this. >> >> Apparently our production people, when semi-auto placing dense parts, >> double-check the ref designator and plop the part into the "correct" >> place, even if the machine coordinates are a little off. I'll have to >> warn them to be suspicious about cases like this, especially on first >> articles. >> >> TGIF >> >> John > > I usually produce an "assembly drawing" where another layer of text > contains the refdes inside the part outline. > This refdes text is defined at the library level and is never moved by > hand except when the component moves. > This drawing is produced separately from the silkscreen and is shipped > to the assembly house. > This is the reference, not the silkscreen. Typically this is done via the XYRS file which feeds into the placement machine. So theoretically all this could not possibly have happened, but .... then there's reality. On my last one they had one chip 180 degrees rotated. On all boards :-( -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
From: John Larkin on 20 Nov 2009 21:57 On Fri, 20 Nov 2009 17:06:55 -0800, "Joel Koltner" <zapwireDASHgroups(a)yahoo.com> wrote: >"John Larkin" <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote in message >news:u0beg512h0fcels3arn2i6s9f8vcc6cqa9(a)4ax.com... >> The resistor that's screened "R127" is actually R129. And vice versa. > >I (cough!) think most of us have had that happen... > >What's the big yellow silkscreened square around the oscillator-in-a-can for? >Metal shield maybe? > >---Joel > There will be an aluminum cover over the OCXO, more to keep air currents off of it than as an EMI shield. That yellow thing was supposed to be bare copper, not screened, but it is kind of cute this way. The cover will get grounded by the mounting hardware anyhow. It's impressive how nice of a phase noise improvement you can get by keeping air flow off XOs, even cheap ones. Ditto reducing 1/f noise in opamps. We've found a few mistakes on this board so far... luckily none are fatal and none require serious kluges. Mostly minor mechanical things that annoy manufacturing. I didn't ground HSWP_EN on the FPGA (we used it as a signal pin) so all the lines that drive the many TTL interface chips float until the FPGA is configured. Some of them randomly go linear and get quite hot (fun to watch on the FLIR) and almost bog down the 3.3 volt switcher. Luckily I can pull down HSWP with a resistor, which adds weak pullups at powerup time, fixing that problem. I will breathe serious sighs of relief when the PCIe link works and the FPGA configures. John
From: John Larkin on 20 Nov 2009 23:07 On Fri, 20 Nov 2009 17:26:06 -0800, Joerg <invalid(a)invalid.invalid> wrote: >don wrote: >> John Larkin wrote: >> >>> ftp://jjlarkin.lmi.net/First.JPG >>> ftp://jjlarkin.lmi.net/Swapped.jpg >>> > >Hey John, you wasted lots of real estate there. Wish I could have some >of that. Doing an EMC fix on a client design right now and I can't even >shove one more 0402 part in there :-( > We have lots of space for this one. We're replacing an older-generation board that is about 6x or so of our board area. You will be not-pleased to know that the switcher section has its own rectangular ground plane section that is connected to the rest of the plane through a number of thinnish slivers. The rows of inductors, incoming and outgoing, straddle the plane gaps. The idea is to keep the various circulating currents in the switchers from leaking into the main ground plane where the analog stuff is. I did the spread-spectrum thing on all the switchers, too. John
From: Joerg on 21 Nov 2009 13:04 John Larkin wrote: > On Fri, 20 Nov 2009 17:26:06 -0800, Joerg <invalid(a)invalid.invalid> > wrote: > >> don wrote: >>> John Larkin wrote: >>> >>>> ftp://jjlarkin.lmi.net/First.JPG >>>> ftp://jjlarkin.lmi.net/Swapped.jpg >>>> >> Hey John, you wasted lots of real estate there. Wish I could have some >> of that. Doing an EMC fix on a client design right now and I can't even >> shove one more 0402 part in there :-( >> > > We have lots of space for this one. We're replacing an > older-generation board that is about 6x or so of our board area. > > You will be not-pleased to know that the switcher section has its own > rectangular ground plane section that is connected to the rest of the > plane through a number of thinnish slivers. ... Interestingly, one of the line items in my recommendations for this one is to pepper a similar isolated plane with vias to the ground plane. For EMC purposes, since this one must pass much stricter rules than the usual class B. > ... The rows of inductors, > incoming and outgoing, straddle the plane gaps. The idea is to keep > the various circulating currents in the switchers from leaking into > the main ground plane where the analog stuff is. I did the > spread-spectrum thing on all the switchers, too. > Oh, how I wish I could do spread spectrum. But with the EMC measures so far I am already at a full 100% of available real estate. Unless someone knows a self-contained oscillator in an SC75 package :-) -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
From: krw on 21 Nov 2009 13:08
On Fri, 20 Nov 2009 19:27:10 -0500, Phil Hobbs <pcdhSpamMeSenseless(a)electrooptical.net> wrote: >John Larkin wrote: >> OK, I just got the first board from production this morning, for this >> spectroscopy controller thing. >> >> ftp://jjlarkin.lmi.net/First.JPG >> >> It gets 12 volts in, which runs an LTM8023 switcher brick to make 3.3 >> volts. The 3.3 runs most of the logic on the board (including a >> Spartan 6 and a PLX PCIe bridge, both BGAs) and also drives four >> secondary switchers and some LDOs to make 1.2, 1.5, 1.8, 2.5, and -5 >> for various uses. >> >> So when I powered it up everything went nuts. The PLX chip was >> obviously fried. After that was pulled, the Xilinx was running hot, >> and the 3.3 volt supply was bogged down to about 2.6. The LTM >> regulator was hot. >> >> Pulled the Spartan BGA next. >> >> Now the 3.3 volt rail wants to run at 5 or so. >> >> After much head scratching, I discovered this: >> >> ftp://jjlarkin.lmi.net/Swapped.jpg >> >> The resistor that's screened "R127" is actually R129. And vice versa. >> So the switcher was programmed wrong, told to run at an absurdly low >> frequency and an absurdly high voltage. The ref designators somehow >> got misplaced during layout. We usually check for this. >> >> Apparently our production people, when semi-auto placing dense parts, >> double-check the ref designator and plop the part into the "correct" >> place, even if the machine coordinates are a little off. I'll have to >> warn them to be suspicious about cases like this, especially on first >> articles. Our automatic insertion tools would never make this mistake. They make a *lot* of others, but not this one. :-( >Better complain to the head of Engineering. ;) Fire the manufacturing manager. ;-) |