These are questions that lend themselves to experiment and measurement (I think). Since I'm not in a position to do it myself, here's three grateful cheers in advance for the first volunteer.
Meanwhile, I have read the opinion elsewhere (Rane?) that resisting the temptation to make traces any wider than needed for current/thermal reasons lets you have more distance between traces which reduces capacitance between. I.e, some cats can be skinned more than one way.
@sam9: no that's a fact that doesn't need sophisticated measurements. The capacitance of parallel planes depend on their distance (not changeable in pcbs) and the area of the plane. Choosing a hatched groundplane reduces the area significantly and thus capacitance is reduced by exactly this amount.
@AndrewT: so, if I understood correctly, you think hatched ground planes only improve mechanical stability?
I know that at least Wayne from passlabs prefers hatched ground planes (though I think I read somewhere that he just likes their appearence, ahem) and that Nelson himself used the stray capacitance of the solid ground plane to add compensation capacitance (some pFs) to get nice sine waves in the XA-series (instead of soldering in a small cap) - very elegant.
Otherwise I read in this forum that somebody experimented in his job a bit with solid and hatched ground planes and found hatched ones in almost all cases to perform equally but with the added benefit of reduced stray capacitance.
But I hope somebody more experience could add his thoughts on hatched ground planes!
All the best, Hannes
@AndrewT: so, if I understood correctly, you think hatched ground planes only improve mechanical stability?
I know that at least Wayne from passlabs prefers hatched ground planes (though I think I read somewhere that he just likes their appearence, ahem) and that Nelson himself used the stray capacitance of the solid ground plane to add compensation capacitance (some pFs) to get nice sine waves in the XA-series (instead of soldering in a small cap) - very elegant.
Otherwise I read in this forum that somebody experimented in his job a bit with solid and hatched ground planes and found hatched ones in almost all cases to perform equally but with the added benefit of reduced stray capacitance.
But I hope somebody more experience could add his thoughts on hatched ground planes!
All the best, Hannes
My main problem with hatched is how hard it is to produce at home.... tried before... mine always comes out like crap. I am quite fond of ground planes... and reading with interest the groundside electron thread, which even claims the ground plane could act as a sort of electron reservoir...
@Nordic: Dunno what the discussion in the other thread is about, but a capacitive element is always an electron reservoir.
ground planes:
A more sophisticated statement about hatched ground planes here
and Wayne's opinion here
rule of thumb for magnitude of stray capacitance here
So reading above links I've the impression that for high-gain applications (I'm always refering to audio!) like phono stages better avoid ground planes. Preamps and power amps seem to be less sensitive and could benefit.
I still hope somebody more knowledgeable shares his opinion, since the main problem of us DIYers is that we just can't build always a prototype with ground plane, measure its behaviour and adapt then the ground plane.
We need it working right out of the box ;-))
All the best, Hannes
ground planes:
A more sophisticated statement about hatched ground planes here
and Wayne's opinion here
rule of thumb for magnitude of stray capacitance here
So reading above links I've the impression that for high-gain applications (I'm always refering to audio!) like phono stages better avoid ground planes. Preamps and power amps seem to be less sensitive and could benefit.
I still hope somebody more knowledgeable shares his opinion, since the main problem of us DIYers is that we just can't build always a prototype with ground plane, measure its behaviour and adapt then the ground plane.
We need it working right out of the box ;-))
All the best, Hannes
AndrewT said:
There are three reasons for the use of hatched ground planes:
- Roughly balancing the copper area on both sides of the board to ensure both sides etch at the same speed.
- Balance the copper area on both sides so the board doesn't curl when the temperature changes (due to the differing thermal expansion rates of copper and fibreglass).
- Reduce the thermal conductivity of the groundplane, to make soldering components easier.
The etching speed is much less of an issue these days, as everyone uses spray etchers, which etch at the same rate on all parts of the board regardless. Soldering issues can be overcome by simply using thermal reliefs at the pads.
The problem with hatched planes is that they have significantly higher resistance than a solid plane. The capacitance of a solid plane isn't an issue at audio frequencies, so I don't bother with hatching.
Especially your last statement is in contradiction to the previous posters!
Well I think this is the usual situation, one can argue either way and unless one is himself in a position to judge what gives benefits, one simply does not know and hopes for the best
Anyway, if I make the ground traces nicely fat and put the pcb into a metal enclosure, are there any additional benefits I loose by not using a ground plane?
Cheers, Hannes
Well I think this is the usual situation, one can argue either way and unless one is himself in a position to judge what gives benefits, one simply does not know and hopes for the best
Anyway, if I make the ground traces nicely fat and put the pcb into a metal enclosure, are there any additional benefits I loose by not using a ground plane?
Cheers, Hannes
AndrewT said:
Actually, this may not be better than a ground plane. Even though we are talking audio frequencies here, some of the things that happen in an amp are way above that band - in particular half-wave rectification in output stages of class AB amps, correction signals for NFB loops, etc. At these frequencies, you need to consider the full current loop, not just the ground line part of it. A ground plane (properly implemented) has the advantage that return paths tend to form a shadow of he outgoing path on the ground plane, hence keeping impedances at a minimum, but you do, to an extent, exchange lower inductance for higher capacitance.
I have used ground planes on audio designs extensively and I must say I have never had particular problems with capacitances. The exceptions to this rule would of course be high impedance circuits, where point to point over a ground plane (air dielectric) works better, or alternativelt, one needs to cut the ground plane suitably, or even drill holes in the PCB to lower local Er.
On the plus side for ground planes is that their use can considerably simplify routing, and if you are working in a home DIY environment, as oposed to having your PCBs made by a PCB manufacturing house (when it is pretty much the same what is drawn on the board), you actually don't etch the ground plane side, instead you remove the ground connection around pins of components that don't need it, with a large drill bit.
Of course a ground star would be best, but a more common approach which I believe works nearly equally well but is considerably easier/more handy is to do signal grounding separately from power grounding and connect them in one point.
This is also some kind of star grounding, it also avoids the noise the power ground introduces, but avoids the need of one star point which may be difficult to design.
Cheers, Hannes
This is also some kind of star grounding, it also avoids the noise the power ground introduces, but avoids the need of one star point which may be difficult to design.
Cheers, Hannes
This is very true.panson_hk said:
I've worked with an OEM project, which is a DAC board designed with the ground plane jointed at power supply only (like the Burr-Brown EVM approach). It radiates very badly around 700MHz, clearly the harmonics of I2S MCK. Later I solved the problem by adding wide junction across AGnd/DGnd plane, just underneath the I2S bus.
@panson_hk: thanks a lot for the article! It's a pity that interesting information like this can only be found scarcely!
Anyway, it mainly argues against the isolation of ground planes and recommends physically separated layouting (and a common ground plane in general).
However there's no hint of the usefullness or potential disadvantages of using ground planes per se. Oh dear.
All the best, Hannes
Anyway, it mainly argues against the isolation of ground planes and recommends physically separated layouting (and a common ground plane in general).
However there's no hint of the usefullness or potential disadvantages of using ground planes per se. Oh dear.
All the best, Hannes
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.