ground
the
A !
It's not so much about splitting the plane as keeping the digital return path separate from the analog return path, i.e., the grounds are in common at the chip, and all of the digital return current flows on one side of the layout, and the analog current return flows on the other side.
And, of course, lots and lots of bypass/decoupling capacitors. :-)
Cheers! Rich
pins !
;~)
A lot of the cell phone and WiFi chip designs I've done use multiple wirebonds down to the "power pad" to minimize the ground cross-talk. I've done it as high as 5.5GHz.
...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC\'s and Discrete Systems | manus |
| Phoenix, Arizona Voice:(480)460-2350 | |
| E-mail Address at Website Fax:(480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Audio can't ever have a real 110 dB range, unless maybe you're recording explosive blasts. Neither recording studios nor listening rooms have anything like 100 dB of acoustic isolation. My NMR drivers run a few PPM noise in a 50 KHz bandwidth and aren't all that hard; the usable s/n is real here (we're dealing with resonances with Q above 1e9) and they certainly don't need split planes.
John
Well, he got this much right...
"Note that connecting DGND to the digital ground plane applies VNOISE across the AGND and DGND pins and invites disaster!
The name "DGND" on an IC tells us that this pin connects to the digital ground of the IC. This does not imply that this pin must be connected to the digital ground of the system."
but his star grounding thing is insane.
currents
What do you reckon a drum produces ? You don't want the electronics clipping.
A good studio should be no noisier than 20dB acoustic and hopefully less. 130dB noises are quite common. Condenser microphones with noise figures around 10dB acoustic equivalent are quite common too.
In any case, if it's not that good it won't sell. As much as anything because the decent competition is that quiet too.
Also, it's normal to use up to 20dB of the dynamic range as headroom so there's no danger of signal clipping. So that 110dB dynamic range is only 90dB s/n.
Graham
strange
that's a normal setup I use, as long as both grounds are tied at or underneath the A/D, it's a good design (or one I am familar with!), effectively both grounds are tied to the same point and are not split (from the A/D point of view)
but on this design both ground planes appear to be shorted to each other right at the A/D and again on the top of the board, that allows a path for the digital ground noise to enter onto the analog side( the system ground appears to be lower right). I would remove the top connection. But I may just be misinterpreting the layout again.
"Keith" a écrit dans le message de news: snipped-for-privacy@News.Individual.NET...
low-level
your
'digital currents'
(if not
signal
plane -
you
DGND pins !
;~)
He he, reminds me something. What are you working on?
-- Thanks, Fred.
"Didi" a écrit dans le message de news: snipped-for-privacy@m73g2000cwd.googlegroups.com...
That sure stay to prove. If you use split planes, that not for connecting the input side to chassis ground. Right? Otherwise your evil ground currents... Ok, now you probably have an input to get signal into your ADC.
What's your planes connecting path inductance? 1nH at best, typically a few.
Now have , say a level 3 ESD event at your input ground side. That's a
6kV/330R pulse 1ns rise time, giving about 20A/ns di/dt.Question: what would be the voltage between your digital and analog grounds?
Certainly not from this POV.
-- Thanks, Fred.
Hello John,
pins !
Same with flip-chip mounting but in our case that were our own designs so you get to select where to place the grounds. And there we only had one ground for both analog and digital even though they had a receiver channel on board.
On these I guess they had to, considering that they burn over a watt just idling. Thye must run pretty hot.
-- Regards, Joerg http://www.analogconsultants.com
Hello Jim,
Don't they use flip-chip if it's really critical?
Question, kind of OT in this thread: Along with other functions, do you sometimes integrate switched-capacitor filters on some of your designs?
-- Regards, Joerg http://www.analogconsultants.com
pins !
;~)
Yup. The back of the board is a big heat sink, and the power pad on the pc board is full of thermal vias. These chips are $190 each, and we use 4 per box.
John
Hello John,
Can be even worse. I have seen stuff go BZZZT because of it. Those wee bond wires can only take so many milliamps and the devices on the die that straddle this area can only take so many hundred millivolts.
-- Regards, Joerg http://www.analogconsultants.com
This is cheap stuff, plus it needs the thermal path.
Yes, but I bring in another consultant... I'm not expert. Likewise for serious digital.
For the sonar chip I did for qrk/Mark I used gm-C filters with an auto-tuning procedure.
...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC\'s and Discrete Systems | manus |
| Phoenix, Arizona Voice:(480)460-2350 | |
| E-mail Address at Website Fax:(480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Hello John,
Shhhht... you are destroying part of my business turf here :-)
-- Regards, Joerg http://www.analogconsultants.com
!
That too. A smallish spike somewhere in the system can blip one plane relative to the other, setting off an SCR latchup in the chip if the grounds are wired to separate planes. A lot of these mixed-signal chips are very prone to latchup, so you could lose a couple hundred dollars worth of fancy ADCs in a few milliseconds.
John
Hello Richard,
Oh, just imagine this thread if it would have been an election year here in the US...
-- Regards, Joerg http://www.analogconsultants.com
Duh! It is. Mid-terms.
...Jim Thompson
-- | James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona Voice:(480)460-2350 | | | E-mail Address at Website Fax:(480)460-2142 | Brass Rat | |
oops, sorry.
So, what's the number-of-dollars disaster you've been called in to straighten out?
John
This rings a bell that we may have been talking about different things.
Split ground plane = one solid layer with some cuts on it, but without any DC discontinuities - or perhaps with, but bridged on other layers by 1-2mm long thick traces and several vias (typically I do the former). There are no impedances which would allow latchup capable energy buildup on either subtype.
Separate ground planes - two or more solid layers, connected somewhere somehow. If the connection between the two layers has a sufficiently high impedance, I could imagine it could cause latchup under extreme circumstances.
This is my understanding of the terms - perhaps different from what is commonly accepted, I don't know.
I talked only about the "split" type - and suggested to use it only in case the designer knows why he is doing it and how to do it.
Separate ground planes is something I would not do neither would I suggest it is useful - at least not in any application I can think of. It would only create problems without solving any.
Dimiter
------------------------------------------------------ Dimiter Popoff Transgalactic Instruments
------------------------------------------------------
John Lark> >
pins !
This rings a bell that we may have been talking about different things.
Split ground plane = one solid layer with some cuts on it, but without any DC discontinuities - or perhaps with, but bridged on other layers by 1-2mm long thick traces and several vias (typically I do the former). There are no impedances which would allow latchup capable energy buildup on either subtype.
Separate ground planes - two or more solid layers, connected somewhere somehow. If the connection between the two layers has a sufficiently high impedance, I could imagine it could cause latchup under extreme circumstances.
This is my understanding of the terms - perhaps different from what is commonly accepted, I don't know.
I talked only about the "split" type - and suggested to use it only in case the designer knows why he is doing it and how to do it.
Separate ground planes is something I would not do neither would I suggest it is useful - at least not in any application I can think of. It would only create problems without solving any.
Dimiter
------------------------------------------------------ Dimiter Popoff Transgalactic Instruments
------------------------------------------------------
John Lark> >
pins !
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.