Not a call that can be made without specifics. Usually my plate loads are in the tens to hundreds of megs and wiring raised high above the plane to avoid coupling to ground.
Some of Henry W. Ott's thoughts on ground planes. Jim Brown (AES Committee Chair on EMC) writes about Henry W. Ott: "If you disagree with a single word,you're wrong!"
Grounding of Mixed Signal PCBs:
Grounding of Mixes Signal Systems
Slots in Ground Planes:
Slots in Ground Planes
Other Henry W. Ott "Tech Tips" papers:
Tech Tips
Grounding of Mixed Signal PCBs:
Grounding of Mixes Signal Systems
Slots in Ground Planes:
Slots in Ground Planes
Other Henry W. Ott "Tech Tips" papers:
Tech Tips
Henry Ott's is probably the world expert on the subject. His talk on mixed signal PCBs at the 2006 AES in SF was outstanding. You san tell when someone knows their stuff whan they can present it in a way that's easy to grasp.
Notice in his talk that when the absolute best isolation is needed e.g. 20 bit ADCs, the recommendation is for partitioned planes with well defined, i.e. narrow, circuit return paths under the A/D packages.
To my understanding, the problem Henry is solving is not how to make a better analog circuit, but rather how to integrate analog circuits with the necessary evil of a high speed digital ground plane when a mixed signal assembly is required.
Highly recommended, along with Bill Whitlock's papers on shielding and grounding.
Michael
Dear All,
Like to put this old topic back under the attention. I am not a tube guy at all, but what a nice discovery is Mr. Millet. Never seen any one using full groundplanes in tube amps. There goes most of the theories. In Millets designs, Power supply, main cap's low voltage, high voltage. Al on one groundplane layer without hum and superb performance. It is possible
Like Guido Tent and Bruno putzeys, It is all about routing discipline and knowing where the currents flow.
http://www.pmillett.com/hybrid_amp.htm
With kind regards,
Bas
Like to put this old topic back under the attention. I am not a tube guy at all, but what a nice discovery is Mr. Millet. Never seen any one using full groundplanes in tube amps. There goes most of the theories. In Millets designs, Power supply, main cap's low voltage, high voltage. Al on one groundplane layer without hum and superb performance. It is possible
Like Guido Tent and Bruno putzeys, It is all about routing discipline and knowing where the currents flow.http://www.pmillett.com/hybrid_amp.htm
With kind regards,
Bas
Last edited:
Thanks I know. While there is often said here we can ignore the EMI and HF guidelines for audio, because audio is only low frequency and those "HF things" are not important for audio, I do feel that listening to the HF guys can actually teach us a lot and following the HF guidelines can improve a circuit. Just my feeling
And I think Mr Millett is spot on.With kind regards,
Bas
Dear All,
Like to put this old topic back under the attention. I am not a tube guy at all, but what a nice discovery is Mr. Millet. Never seen any one using full groundplanes in tube amps. There goes most of the theories. In Millets designs, Power supply, main cap's low voltage, high voltage. Al on one groundplane layer without hum and superb performance. It is possible
Hybrid amp with adjustable damping
With kind regards,
Bas
That's all very cool that a ground plane can be made to work with a tube amp, but I didn't find the place where Pete explains the specific advantages of a ground plane vs. star ground.
"There goes most of the theories"? Where did they go?
Cheers,
Michael
Dear Michael,
Sorry let me define. I am a bit stubborn from nature, but I try always to keep my eyes and ears open and catch as much information I can. Some share the theory "that one can NEVER apply groundplanes for amplifiers where high and low currents are involved"
That is why it was good to see that on this particularly design even the power supply cap's with dirty currents are on the same board.
With kind regards,
Bas
Hi Bas,
Yes, this I can agree with.
My whole point is that I don't see a compelling reason to go off and figure out how to apply ground planes as a general improvement to audio amplifiers.
The main reason to figure out ground planes for audio is IMO so one can build mixed signal PCBs or single-PCB amps like Pete's hybrid, in spite of the extra work involved in layout and analysis.
I still wonder though if the same audio circuit would perform even better if a star-on-PCB wiring scheme were to be used instead of a ground plane...
Cheers,
Michael
For myself, I know that soon
This subject triggers me for a very long time already. And after months of reading every single thing I could possible find I bite the bullit, and made from a previous 3 channel solid state star ground design a design with all groundplane. But with very carefull component placement. I figured out, this is the only way to find out. Measure it, listen to it. I get the boards soon so for now I don't know yet.
With kind regards,
Bas
I just completed an RH84 build using a common ground-plane for the driver and output sections. This amp is DEAD quiet through a pair of Klipsch RF3's (~98 dB). You can bet your bunz I'm going to use this method on my next build!
Ground plane construction has been championed by the amateur radio community for decades, and for circuits that handle microamp-level rf signal currents. I've read more than one QST article where the author built a prototype using a ground plane before building the final version on a PC board, only to admit that the "dead bug" prototype performed marginally better.
Just my $0.02...
Ground plane construction has been championed by the amateur radio community for decades, and for circuits that handle microamp-level rf signal currents. I've read more than one QST article where the author built a prototype using a ground plane before building the final version on a PC board, only to admit that the "dead bug" prototype performed marginally better.
Just my $0.02...
Great to hear! I wish there where more articles directly related to amplifiers and Ground Planes.
It means in your design, you also put the main PSU capacacitors with a wire direct on the same ground plane?
With kind regards,
Bas
I once helped a colleague to troubleshoot his DIY audio amp. He was a professional RF designer, and good one. He made lots of compartments from copper clad boards dividing all stages from each other, and added lots of capacitors under each tube soldered to that copper grid. But the amp oscillated wildly!
What I did, I discarded all that shields and capacitors, and made all earth connections by a wire that goes from power transformer to first filter cap, from this filter cap to the second after a choke filter cap, from which separate wires were going to all his stages (well, almost: input sockets were connected through shielded cables already, and shields were soldered near volume control pots, then were going to first tubes).
Chassis and ground were connected directly near input sockets: we did not have then requirements to connect safety ground. We did not have it in outlets.
As the result, all wild oscillations stopped, and an amp become dead quiet.
He was surprised a lot! Before that he tried to teach me that ground wires I use are too thin.
What I did, I discarded all that shields and capacitors, and made all earth connections by a wire that goes from power transformer to first filter cap, from this filter cap to the second after a choke filter cap, from which separate wires were going to all his stages (well, almost: input sockets were connected through shielded cables already, and shields were soldered near volume control pots, then were going to first tubes).
Chassis and ground were connected directly near input sockets: we did not have then requirements to connect safety ground. We did not have it in outlets.
As the result, all wild oscillations stopped, and an amp become dead quiet.
He was surprised a lot! Before that he tried to teach me that ground wires I use are too thin.
Hey Wavebourn, I most certainly agree that you can take the ground plane deal too far!
Actually all I did was to use a piece of copper-clad board to solder all the ground connections to. Nothing fancy at all on this build. In fact, my '84 was a "quick & dirty" build, inspired by my finding a couple of 125ESE's that I bought years ago.
The important thing to remember here (I think) is that the increased flexibility afforded by the ground plane is not an excuse for poor component placement.
This amp sounds so good I'm considering rebuilding it with Lundahl iron. Hmm... K&K Audio is just down the road!
Actually all I did was to use a piece of copper-clad board to solder all the ground connections to. Nothing fancy at all on this build. In fact, my '84 was a "quick & dirty" build, inspired by my finding a couple of 125ESE's that I bought years ago.
The important thing to remember here (I think) is that the increased flexibility afforded by the ground plane is not an excuse for poor component placement.
This amp sounds so good I'm considering rebuilding it with Lundahl iron. Hmm... K&K Audio is just down the road!
In the past it was very common to use the chassis as a signal ground. Would that count as a "ground plane"
In fact it was so common that if you look at almost any solder lug terminal at least on of the lugs is always an extension of the support leg that holds the terminal strip to the chassis.
I've seen many Hammond organ amps that used the chassis for signal ground although they would backup that connection with bus wire.
I'd go so far as to say that successfully implementing a ground plane requires even more thought given to layout than PTP+star ground due to the additional coupling mechanism of the ground plane itself. With PTP ground wiring I only worry about the magnetic loop area of the high current conductors e.g. the filter caps.
Bas, I'm curious as to what advantage you are hoping to gain by using a ground plane. Can you elaborate a little?
Cheers,
Michael.
You are right!
Ground plane is not an excuse for messy placements of components with sporadic grounding. If it is used as a wire, all current paths have to be carefully considered, the same like in case of a straight wire.
Such a way it can work well. But it is not wise to close rectification loop through the same wire that carries input signals: peaks of high charging currents cause small voltage drops on that wire that cause different types of rectifiers to sound differently: in pauses it can be nearly inaudible, but on loud passages it can intermodulate with the signal resulting in fuzzy sound and uncertain soundstage.
Also, if an input stage is between output stage and power supply, ground plane current goes from output stage to power supply through input stage ground causing very small, but sensitive voltage drops that add undesired feedback.
It is easier to isolate such currents from each other using a peace of thin wire than a huge massive chassis.
Ground plane is not an excuse for messy placements of components with sporadic grounding. If it is used as a wire, all current paths have to be carefully considered, the same like in case of a straight wire.
Such a way it can work well. But it is not wise to close rectification loop through the same wire that carries input signals: peaks of high charging currents cause small voltage drops on that wire that cause different types of rectifiers to sound differently: in pauses it can be nearly inaudible, but on loud passages it can intermodulate with the signal resulting in fuzzy sound and uncertain soundstage.
Also, if an input stage is between output stage and power supply, ground plane current goes from output stage to power supply through input stage ground causing very small, but sensitive voltage drops that add undesired feedback.
It is easier to isolate such currents from each other using a peace of thin wire than a huge massive chassis.
Please forgive me Bas; I forgot to answer your question!
The ground plane I used is a small strip of PC board material (approx. 2.5 cm x 10 cm) mounted with double-sided tape just behind the tube sockets. All grounded components that connect directly to the tube socket pins are soldered to this board. I placed a blob o' solder to ONE point on the ground plane. From this point is a wire to the input grounds (the RCA jacks are isolated from the chassis), one wire to the ground of the second PS filter cap, and one wire to the IEC connector.
I suppose you could say this is actually a hybrid grounding scheme, as the ground plane is acting as the center of the "star". It's also breaking a few rules, but the proof is in the pudding: This amp is so quiet that before I spin a CD I've found myself looking at the pilot light to make sure the thing is even on.
The ground plane I used is a small strip of PC board material (approx. 2.5 cm x 10 cm) mounted with double-sided tape just behind the tube sockets. All grounded components that connect directly to the tube socket pins are soldered to this board. I placed a blob o' solder to ONE point on the ground plane. From this point is a wire to the input grounds (the RCA jacks are isolated from the chassis), one wire to the ground of the second PS filter cap, and one wire to the IEC connector.
I suppose you could say this is actually a hybrid grounding scheme, as the ground plane is acting as the center of the "star". It's also breaking a few rules, but the proof is in the pudding: This amp is so quiet that before I spin a CD I've found myself looking at the pilot light to make sure the thing is even on.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.