| athos56 |
Hello all
I'm building a Je labs St2a3dx and documenting my progress here, I went over my ground scheme on the last page, here. I would apreciate it if could get feedback on the ground layout.
Sorry about the links but I thought it would be a waste to repost all of the pictures here. Thanks for any assistance. |
|
|
| SY |
| I get a message saying that the images are not available. |
|
|
| radtech |
| no, those don't work either, looks like you have to register and login to see the pics. |
|
|
| athos56 |
| How about this! One... |
|
|
| athos56 |
I should have done this first! Sorry
Text...
First picture - center tap to the first 20uf cap, the cap is connected to the ground at the adjacent center ground lug, the 12 gauge copper wire bus starts there at the cap as well. The next connection is the center tap of the 6.3v, then the 200k resistor, the bus then makes a stop at each cap in order, 100uf, 50uf, 50uf.
Second Picture - Bus wire travels across the chassis and ends above the two 20uf/500v filter caps (big ones in the center)
Picture three - Each channel procceeds the same. Input jack to 100k resistor to 1k resistor to 22k cathode and 100uf bypass cap to the ground bus wire after the filter caps.
Input jack is isolated, ignore were it appears that each channels run through a ground lug in picture 3, only 2 ground points exist, one next to the 20uf cap at the power supply end and one next to the main power plug safety ground.
Please let me know if you see something which is gount to cause me problems. Thanks! |
|
|
| SY |
| I'm so gassed from the renderings that it may take me a while to suggest anything about grounding. That is very, very cool looking drawing. Is it software usable by ordinary mortals? |
|
|
| athos56 |
| Its called sketchup pro by google, trial version (free) when the trial runs out it still lets you use the basic features! And thanks! |
|
|
| SY |
| One more request: a schematic. |
|
|
| athos56 |
| This one is a mess, without the camera controls its impossible to see where everything goes. But I labled one half of the components. In all the pictures the ground is green. |
|
|
| SY |
Drawing-challenged me still couldn't make heads or tails of it. But anyway, I marked up the schematic to show how I would do the grounding and added a cap which might save you some grief. Also, I would strongly consider putting either a reverse-biased diode or a neon lamp between the grid and the cathode of the driver tube. The segment between the output stage star and the power supply star should be short and thick.
edit: I forgot to move the chassis ground. Corrected. |
|
|
| athos56 |
| Thanks for the input, it'll take me a bit to make heads or tails of it. Maybe my pictures are not as illustrative as I thought. But it does lose a bit in the 2D. Thanks for the time, and for giving me another option. |
|
|
| SY |
| Think of currents running downstream. We start at the top with the supply (being careful not to break into a high ripple current loop). We then cascade down, with successive stages dumping less and less current in until we reach the earth/safety ground. It makes logical sense since the stages that draw the most current are closest to the power supply ground node, i.e., we've minimized IR drops along the bus. |
|
|
| Roger Hill |
Question for you, Sy:
What diodes and what values for resistor and cap would you typically use for the safety/chassis ground? Thanks,
Roger Hill |
|
|
| SY |
Diodes: I use 1N4007 because I have about 10,000 of them.
Resistor: Not terribly critical. 15R, 33R, something like that.
Cap: Not critical, either, but use one with good RF properties. 0u01 ceramic disc is what I use because, again, I've got about 10,000 of them.
This is how I did the grounding in my power amps and preamp, and not a hint of hum or buzz despite three-wire line cords and no ground isolating switches.
Also, Nelson Pass has written something about this recently, but apparently he uses a bridge rectifier instead of the two diodes. Poke around www.passdiy.com a bit and I'll bet you'll find his suggestions. If they contradict mine, believe him, not me.:D |
|
|
| athos56 |
So I went to digi-key and picked up the following.
(4) 1N4007FSCT-ND DIODE GPP 1A 1000V - $0.44
(10) 1460PH-ND CAP .01UF 500V CERAMIC DISC - $1.66
(5) PPC15BCT-ND RES 15 OHM .50W 5% MF FUSIBLE - $2.57
Pretty cheap. Your way seems easier then running a bus everywhere. So I gather all of the power grounds at the first cap, then with a thick wire run a line to the other stages and then to the diodes, resister, and disc cap which is connected to the chassis near the inputs?
Thanks again for your assistance. |
|
|
| SY |
| Yes, connect your grounds in descending order as you move from the power supply to the input. It really is pretty easy. In the RLD amp, I used a 10 gauge solid copper wire to form the short (15 cm or so) bus. Dead silent. |
|
|
| salas |
| Bussed sub stars is what I use too in stereo amps even and its best. Follow SY on this one. He is correct (again). |
|
|
| salas |
| P.S. Remember that input jack must be isolated from chassis. |
|
|
| nkg |
I am also building this amplifer the extra cap you suggested from the anode load resistor to ground, is 20uf as used by the type 76 valve acceptable.
Many thanks
Nigel |
|
|
| athos56 |
| Is it 20uf? A value wasn't listed so I didn't add it. I wonder what it would change? I have a little work to do still, my hum balance pots are all wrong, i acidently got 50k ohm pots not 50ohm so I'm waiting for the mail man to bring me the rest... If a value on that cap is determined let me know. Thanks |
|
|
| lquintas |
| This extra cap only changes the value of the 50 uf cap in the power supply circuit, because they are in paralell. Is it necessary? Thanks |
|
|
| nkg |
I'm still a few weeks from finishing my amp just curious as how far off you are from finishing your amp maybe compare notes.
Nigel |
|
|
| bibster |
| quote: | Originally posted by SY
Think of currents running downstream. We start at the top with the supply (being careful not to break into a high ripple current loop). We then cascade down, with successive stages dumping less and less current in until we reach the earth/safety ground. |
This is what I needed to 'get it'... Thanks Sy!
I'm finishing my aikido now, based on JB's 5687 board.
Few related Q's though:
My RCA's grounds are 'bussed' per IN and OUT (1 bus each thus)
Should I link these two busses, and go with one wire to the Aikido PCB? If yes, then: Which of the many Gnd's on the pcb? (2 on the IN side, 2 on the OUT side)
Or should I rather put the OUT bus on the OUT side and the IN bus on the IN side?
Or should I link them one by one to the gnd bus? (I think so...)
Than: Where to go from ground bus to earth (PE, wall plug green/yellow) NOTE: My chassis is 100% wood (For the moment being) Well.. 99% wood, 0,8% glue, 0,2% paint
Should I pick a point 'further downstream' (cf. Sy's analogy) than the RCA's
So far, I thought of this bus like this:
CT---PSU C's----PSU regulator board ---Heater PSU Ground Ref.--- Aikido PCB 'ground' conn.----INPUT RCA's-----Output RCA's------diode/resistor/capacitor network to PE
(The two PSU C's are connected together, thus only one ground wire, the volume control's GND to the same spot is the input RCA's)
Does anyone GET this, and Is it 'the way to go'?
(See attached image file)
Paul
PS: Could this BUS be a piece of 2.5mm2 stripped copper, and could it make a nice(?) long (?) sightseeing tour underneath my chassis? |
|
|
| SY |
If you're using the Aikido PCB, make use of the very nice ground connections that John provides; his grounding is star rather than the bus grounding I was talking about before.
In my build using John's board, the power supply return was connected to the board star, the input commons were separate for each channel (but all input jack commons connected together within a channel) and returned to the board Input Ground, as was the bottom of the volume control pot. The twisted pair output leads were connected to the respective output Cx and GND terminals on the board. The R/C/diode network was connected between the star ground and the safety (earth) ground on the board, and a thick wire was run from the board's safety ground to where the green wire from the IEC socket was attached to the chassis. All input and output jacks were isolated from the chassis.
One other thing: when looking at spectra of the output noise, I found that it was worthwhile to trim the resistor divider that provides the nulling signal. The null is a pretty sharp one, and given the tolerance in tube mu... |
|
|
| bibster |
Thank you Sy,
So I'll make a PSU star (CT, C's, heater and HT regulator boards), run that to the GND connection on the board.
The output RCA's commons are allready bussed together (L+R and all 6 inputs together) so I'll run 1 wire to one of the input gnd connections, using the other for the vol.pot.
Same for the outputs: I'll use either of of the provided gnd's. (Or put some cat5 TP and thus use both...)
I'm not whether too use the board's safety gnd, as the board is attached to some wooden case... Well, It doesn't harm, does it? If I decide to add some shielding later...
Thanks, Paul |
|
|
|
