Merlin's next book

[tubeswell]

New Valve Wiz book on Power Supplies out soon

http://www.freewebs.com/valvewizard/index.html

[David Root]

That will be interesting as that is one area where solid state technology has brought real benefits to builders, with soft start hexfred diodes, double diodes and bridge rectifiers with power capabilities so big that heatsinks are hardly necessary any more, and with practical voltage handling to +700V to boot.

[Structo]

I read the bit he has about power supply grounding last night.
I found it a bit confusing and that some of it goes against what I have learned.

I don't understand how attaching the grounds along a buss closer to a certain point would change anything since electrically, it is the same.
He claims that you should have things closer to the nodes filter capacitor for lowest noise.

Or that the filter caps should be located physically close to the section they are filtering, which I know some amps did it that way.

But on this diagram, why would it change anything if the actual points of connection are closer to the component?

[img:400:319]http://lh3.ggpht.com/_MHnI-B0pJ8k/TLH67 ... tering.JPG[/img]
Fig. 13.12: a. A standard bus ground system is possible in low-gain
designs, but is far from optimum. b. An improved bus scheme. Connections are made closer to the smoothing capacitors to minimise the interaction of signal and power currents; grid-leak connection are made to their respective cathode circuits; the speaker ground is moved downstream of the power valve ground currents; the chassis connection is made right at the input jack socket.

[JamesHealey]

Merlin lives about 25 miles away from me I'd love to pick his brain some time. Met him once in a guitar shop.

[Structo]

Does he really wear one of those pointy hats?

[Cliff Schecht]

I can explain this grounding scheme. We use this style of grounding extensively at Paia and I use it in my personal designs to much success. I can explain it tonight when I'm not typing a phone.

[Colossal]

Does he really wear one of those pointy hats?

Merlin was educated at Hogwarts; graduated Cum Loudly Rumor has that there is an odd blue glow that sometimes emanates from the sole window of his tower.

Can't wait for his new book. Tom, as for the grounding scheme presented, as counterintuitive as it might seem to have the sole ground point at the input (you might think it should be at the "other" end), many of the guys on AX84 use this and I have heard firsthand how quiet they can be.

[PlinytheWelder]

I did my grounding scheme on my express close to that way and it's very quiet. Aiken's site has another take on it... http://www.aikenamps.com/StarGround.html

[Cliff Schecht]

I can explain this grounding scheme. We use this style of grounding extensively at Paia and I use it in my personal designs to much success. I can explain it tonight when I'm not typing a phone.

[paulster]

It's all about current loops. There was some good discussion about it here a few years ago with some great contributions from Paul Ruby.

He gets it!

[tubeswell]

I used it on a BFPR build (with separated filter caps) - works like a charm - and very low noise floor

[Alexo]

It's much more about the order in which the ground points meet up than it is the physical distance. At first, those two diagrams look insignificantly different, but if you notice the little things, like the way the grid reference resistors tie directly to Rk, before meeting up with the grounds from the next stage, etc., things starts to become clear.

Once you get the ground loop thing, and what's really going on in the "negative universe," I swear, it's like having your blinders removed.

[JamesHealey]

His business card has a wizards hat on it! Does that count? Very polite and helpful chap.

[Cliff Schecht]

Don't know why my message posted twice earlier.. Sorry about that.

I agree with Merlin's method and really it makes a lot of sense. Essentially what you are doing is directly controlling the ground currents and sending them to the lowest impedance point (the "-" side of the cap). These currents have to eventually flow here anyways and so sending them directly to where they need to go keeps any current from flowing where it shouldn't and causing unwanted noise (hum). Really it doesn't matter where you tie the ground to the chassis because you aren't relying on it to carry any current. You don't want to leave the circuit floating relative to the chassis ever though, so make sure that ground tie-off point is a secure one!

If you do decide to go with a chassis grounding scheme, isolating the input from the chassis is absolutely essential. Every single amp that I have converted to an isolated input grounding scheme has gone from overtly noisy to dead quiet. This is the one spot where it's very easy to pick up hum and keeping any ground noise out of here is undoubtedly important. I try to ground this to the cathode of the first stage which should ideally have a low impedance path to its decoupling cap. Latter stages have a much better SNR and usually aren't as picky about grounding, but it's still good practice to control the ground currents wherever possible.

I think the only reason KF got away with his grounding scheme is because of how he separated the power grounds from the signal grounds. While both of these signals do use the chassis for grounding, each section is clearly separated by the decoupling cap ground tie off point. The power currents flow towards this point, not towards the input section. I wouldn't go with this scheme in my amps but Ken was obviously smart enough to make it work.

[M Fowler]

Here is an example of a quiet Stangray amp. Notice both caps on the board, can cap, CTs and IEC are all connected to one chassis point.

Notice that the input jack and pots are not isolated. Notice that the buss wire is used for ground termination for the board.

Many ways to get a quiet amp.