Where to ground the first filter stage.

[Mike_J]

In reply #8 of my 2016 post entitled "where to place the humdinger pot?" PRR said "it is advisable to bring the PT HV CT lead directly to the first filter stage. This was for a different amp than the 5f6a I am currently rebuilding but I was wondering if you still think it would apply? As it stands right now all the ecaps are grounded before the bias grounds and if I place a humdinger pot in this amp those grounds would be between chassis ground at the PT ground lugs where the PT HV CT is grounded and the first filter stage ground. My guess would be you haven't changed your mind and decided it is no longer advisable but thought I would recheck since it has been five years.


Regarding the humdinger pot and circuitry to bring DC to the heater supply. The DC would be fed from the second stage which feeds the 470 ohm screens resistors. The reason I am mentioning this is I put some devices that would hold the DC wire that would be sent from the DC voltage divider circuitry to the wiper of the humdinger pot. It would go through the center of the power tubes to the back corner of the amp and stay in the back corner until it comes out between the last power tube socket and the rectifier tube socket where it would meet up with the wiper of the humdinger pot. Picture is attached.


The reason I am rambling like this is I have no idea whether or not I will use the DC circuitry or humdinger pot. What I do know is I will isolate the input jack and run a ground from it and the OT secondary ground to the same bolt the PT HV CT is grounded to. Would like to use the same devices that would hold the wire to the humdinger pot wiper to run the OT ground wire from the output jack ground to the lug on the PT bolt. Don't know whether this is inadvisable or not but would appreciate any advice anyone could offer. Basically, the question is can I run the OT ground wire with about 75VDC of filtered power wire?

[Mike_J]

The problem I think I have is I need to find a way for the blue and brown OT primary wires to coexist with the wire from the 470r resistor for V5 and the black ground wire from the OT ground not to mention the grid wire for V5. Pin 4 of V4 has about an inch or so run from its 470r spot on the board. V5 is not as simple. Have the grid wire for V5 running along the edge of the board and doubt that anything but a 90 degree crossing there would be inadvisable. Even then I would want to use as much of the 1/2" standoff as possible to provide even less interaction.


Not so sure about the interaction between the screen wire and the OT primary wires. Is this an advisable 90 degree situation as well for best practices? Many ways to skin this cat I am sure but would appreciate any thoughts concerning this matter.


Thanks
Mike

[Mike_J]

Have given a little thought to the grounding of the ecaps. Have read in enough places that it is a good idea to ground the first stage of filtering to the PT CT, so might as well. What can it hurt? Then I need to decide what filter cap should have its caps negative terminal nearest to the last one ohm resistor tie in? There I am leaning towards the second stage getting priority over the third stage. Although I suppose it could be argued either way. My argument is the second stage cathodes come in line before the grounded side of the bias caps related to the third stage. Believe it is a sound strategy if nobody has an argument against it.


As far as the OT ground wire. Probably should keep it far away from AC or DC. Don't know why but it just makes sense. If I move the wire to the other side of hole where the OT primary wires come in then there will be no voltages at all too close to them. Solves that potential problem. As for the wire to nowhere which may someday feed DC to a humdinger pot. It still has its dedicated spot. Last is one lonely wire going to V5-4. If the only problem I have is figuring out how to run one wire without getting excessive influence from OT primary wires, isn't it a cozy life?

[thetragichero]

if not using global negative feedback, ot secondary common need not even be connected to other grounds
if you are using negative feedback, then connect to the power/phase inverter ground. i keep those separate from the preamp ground. preamp connects to chassis ground at input jack, power amp somewhere on the other side of the chassis

[Mike_J]

Use negative feedback. The cathodes for V1 and V2 are grounded at the input jacks. The power amp including the ecap feeding the pi section are grounded to one of the PT bolts.


About five years ago I constructed an amp with a DC voltage divider feeding a humdinger pot to put DC under the AC for the heater supply and of course balance it. At the time PRR told me I should run the PT CT to the first power stage ecap ground. I have read a number of times it is good to separate the ground of the first power stage ecap.


Right now I have three stages, Pi, screens and first stage all on the same ground wire wired in order from the Pi to the first stage to one of the PT bolts which is backwards to my way of thinking. This is just one more thing I don't understand but would like to. At one time I took effort to get every ecap as close as possible to the cathode resistors that were fed by the cap and ground them all together. Don't know if it was a trend at the time or valid practice but it seemed to help keep my amps pretty quiet. So that is the basis for the question.


Thank you for replying. I agree with everything you mentioned but still not sure about how is best to ground the power and Pi ecaps.

[HotBluePlates]

... PRR said "it is advisable to bring the PT HV CT lead directly to the first filter stage." ... I was wondering if you still think it would apply? ...

Start thinking of a circuit as a circle (because it is).

Where does the first filter cap connect?  In a push-pull amp, the positive terminal usually connects to the output of the rectifier, and to an output transformer.  The negative terminal connects to ground.

     - What feeds the rectifier?  The power transformer high voltage winding.

     - What's at the other end of the high voltage winding?  When the rectifier is not a bridge, it's the grounded center-tap (only half of the entire winding conducts at any one moment).

     - The high voltage winding center-tap should connect directly to the filter cap's negative terminal to "complete the circuit."

     - What about the output transformer?  What's on the other end of that?  Power tube plates, so probably a good idea to ground the power tubes at the same spot as the 1st filter cap.

     - Any other items we missed?  How about the tube screens: their current is part of the "output tube cathode current" that we connected to the ground of the 1st filter cap & power transformer.  Better ground the filter cap feeding the output tube screen to the same spot as the 1st filter cap & high voltage center-tap.


We do all this because there are big currents (so big noise potential) at these circuit elements.  Current around other loops ("circuits") in the preamp are very much smaller.  If the preamp grounds accidentally rode on voltages cause by power transformer/rectifier/output tube currents, the preamp could very easily have hum added to the intended signal.

"Ground" is not a black hole where voltage disappears, but the return-path of currents circulating in a circuit.  We generally keep noisy grounds away from quiet grounds, and connect them in one spot to avoid hum/noise.

[Mike_J]

This is exactly what I was looking for. Then the components for the bias stage that need to be grounded could be connected to the third supply cap negative side and then sent to the same ground spot (a lug connected to a PT bolt). These components are biasing the power tubes and all of those components should be grounded far away as possible from the preamp as possible if I am understanding you.


Thanks
Mike

[tubeswell]

Here's a picture - only one chassis ground connection needed. YMMV

FYI, what I call 'AC micro wobble in the ground return' = unwanted source of noise

Red Wires = High Tension supply

Green wire = the daisy-chained ground return

Grey boxes = the PS node groups

[Mike_J]

Only thing I have for the screen supply is one lonely little 100 volt ecap for a possible voltage divider to put DC under the AC in the heater supply. Since it would be getting fed from the same supply point the screens are supplied from that is where the negative side of that cap should go and then head onward to the PT bolt. Think I am starting to get this. Just need to redirect a few wires and I will be there. Of course for now all there will be is a turret to connect to since the cap won't be connected unless the DC supply is needed.


 
Thanks
Mike

[Mike_J]

Here's a picture. YMMV


FYI, what I call 'AC micro wobble in the ground return' = unwanted source of noise

Hoffman layouts are really brilliant in that they form a star ground system of sorts that is very organized. I installed the ecap that will supply both V1 and V2 of the preamp right under the pots. Will be a very short distance from its connection at the board. Other caps will be in their dog house but I have a very good idea how they will be connected. Can't really impact distances because the dog house is where it is and the transformer bolt is where it is. Will have to deal with the realities that are there but pretty sure we can make the most of it.


Thanks
Mike

[PharmRock]

Here's a picture. YMMV


FYI, what I call 'AC micro wobble in the ground return' = unwanted source of noise

This is all great info and very timely for my Plexi/JCM800-ish build I am in the middle of doing.  I just installed all the "infrastructure" hardware and about to start wiring the filter caps, transformers and grounds this evening, heaters tomorrow.

Couple of questions...
1. If I am correctly interpreting the picture that tubeswell posted, the phase inverter should be grounded along with with the preamp/input ground (V1, and V2 for plexi)....but I've seen other references where the cap supplying the PI would be grounded at the PT ground. Does it make a difference?
2. Maybe this is splitting hairs, but is there any preference for grounding the the power supply filter caps and center taps to a dedicated chassis bolt vs. one of the power transformer mounting bolts?  I've seen some recommendations for having a dedicated bolt, and other reputable build pictures showing the ground point being a PT mounting bolt.  I'd prefer not to drill another hole in my chassis, but if a dedicated ground bolt is the way to go then I'd rather do it that way.
3. In some of the kit builds I've done, it has always been recommended to have the AC supply ground (e.g. power cord) have its own dedicated bolt.  But...I've seen this ground attached to the same ground as the rest of the power supply grounds.  Any preference of one over the other?

Thanks...and sorry Mike if this was a hijack of your thread here...figured these questions fit in with the general discussion.

[tubeswell]

Here's a picture. YMMV


FYI, what I call 'AC micro wobble in the ground return' = unwanted source of noise

1. If I am correctly interpreting the picture that tubeswell posted, the phase inverter should be grounded along with with the preamp/input ground (V1, and V2 for plexi)....but I've seen other references where the cap supplying the PI would be grounded at the PT ground. Does it make a difference?

The PI has its own PS node in that schematic. In which case it has its own tie-in to the galactic ground return buss. I put it closer to the pre-amp node than the output stage node, just because the PI current is similarly low to the pre-amp node, but its not so critical with the PI because the S:N ratio in the PI isn't as high as it is in the input stage, because the signal has been amplified through the input stage already. But if you were running the PI off the pre-amp node, then the PI ground return would also go there

2. Maybe this is splitting hairs, but is there any preference for grounding the the power supply filter caps and center taps to a dedicated chassis bolt vs. one of the power transformer mounting bolts?  I've seen some recommendations for having a dedicated bolt, and other reputable build pictures showing the ground point being a PT mounting bolt.  I'd prefer not to drill another hole in my chassis, but if a dedicated ground bolt is the way to go then I'd rather do it that way.

I prefer a single ground buss wire. This is because, while you can ground the output stage to a separate chassis bolt, you don't have as-good-a-control on the current path that the output stage ground return current 'chooses' to take as it moves through the sheet metal of the chassis. Whereas, if its all in a single wire, then you get to decide which bits of wire go where. YMMV

3. In some of the kit builds I've done, it has always been recommended to have the AC supply ground (e.g. power cord) have its own dedicated bolt.  But...I've seen this ground attached to the same ground as the rest of the power supply grounds.  Any preference of one over the other?

The mains safety ground should ALWAYS have its own dedicated secure connection to the chassis. And it should be a type of connection that is mechanically sound, so that it stays connected even in the unlikely scenario that the unit catches fire and melts the solder (e.g. struck by a lighting bolt). To my way of thinking, this should therefore be a ring terminal that can be tightly and securely clamped to bare (unsoldered) stranded safety ground wire, and then securely held tightly against the chassis with a locking nut. If you've clamped it tightly to ensure a good mechanical connection, then you can solder it afterward if you want (to enhance the electrical connection), but you shouldn't do any solder tinning etc before you clamp it.

That's my story and I'm sticking to it

[PharmRock]

thanks tubeswell this all makes sense and I appreciate the advice.

[Mike_J]

Not so sure connecting the V3 ecap negative lead to the end of the copper buss wire where the presence pot ground wire goes is a bad idea. My reading of the schematic says the presence pot is the end of the tail resistor and looks like the place where the cathodes for V3 would be grounded. Would require lifting grounds on the input jacks with those handy washers Doug sells for the job. Can't decide whether to run that wire directly to the PT ground bolt or running it by the components in the bias circuitry that need to be grounded on the way to the PT bolt. Makes it possible to ground everything at the PT bolt. Any advice would be greatly appreciated.

[Mike_J]

Aiken says, "If you have a long path back through the chassis to get from the input jack ground to the cathode resistor ground, it can pick up all sorts of stuff along the way. Keep it short, and use quality shielded cable, with the input jack isolated from the chassis, and the shield grounded at the ground side of the cathode resistor for the first stage."
Not sure what he means by the shield grounded at the ground side of the cathode resistor for the first stage. Does that mean the cathode resistor, ground wire and shield for the ground wire are all grounded on the same turret? If the answer is yes to that I doubt that I would need to do it because the wire would be very short but it doesn't require that much effort so any benefit at all and I am in for a try.
Where I was wondering if this shielded cable idea had merit is for the first stage ecap that I will be sending to the power tube cathode grounds. Any reason why it would be beneficial there? For that matter if I am going to run the ground from the presence pot side of the ground buss doesn't it make sense to use shielded cable there since all the preamp grounds are tied to it?

[Mike_J]

Just thinking, I like to use 18awg stranded for my ground wires. Those shielded cables are pretty puny in comparison. Adequate for carrying a little guitar signal to a grid but not so sure they are up to the job for much more? Any advice again would be appreciated.

[Mike_J]

Haven't been able to find single conductor shielded 18awg cable. Comes in two conductor which might be okay seeing as how I would be sending the wires to the same general area on the circuit board.


Things are getting out of control. Have so many ground wires going to the PT bolt special care must be taken to ensure a good ground. Took a picture of the vertical terminal strip that holds the heater wires and connects the 100 ohm resistors to ground. Sluckey wanted me to run three wires to V5 but I nearly went insane trying to get two wires connected to the heater pins on V5. Three would have had me certified insane.


Needed an alternative which was the three lug vertical strip. Took a picture of it but it is hard to tell anything with all those wires in the way. That is why I took an additional picture where I will use them to hold up the 68K resistors and grid wires. They are missing the bottom lugs because they weren't needed and were just in the way. However it is needed for the 100 ohm resistors so they have a ground point.


Why all the rambling you ask? Probably what I should do is use the vertical terminal strip only for the heater wires (already busy enough there) with a ground wire connected to a two hole lug tied to the PT bolt. The other hole could be used for a horizontal strip where the grounds are ordered from input to output consistent with the schematic. Should make it quiet. Any comments would of course be appreciated.


 

[thetragichero]

use smaller wires

[Mike_J]

use smaller wires

How small? Looked for 18 and 20awg single conductor shielded cable. Doesn't exist as far as I can tell. Want multiple conductors then take your pick but single conductor not that I can find.


Have a theory. Make your ground wire a nice 18awg stranded and the electrons says hey there is a super highway let's go that way. No idea if it makes even a little sense but it seemed like it did at the time and I have been doing since.

[thetragichero]

i don't like single conductor anyway. i use anywhere from 18 to 24awg stranded for heaters, depending on if they're big power tubes or little bottles. probably a bit overkill on the lower gauge side. 18 or 20awg generally for the ground wires, ~24awg for signal wires (besides input grids which are much smaller shielded wires)

[Mike_J]

Story that is a little related. Built a house ages ago. Had them put the meter in next to a main breaker. Well I want to install the breaker panel in a closet that is probably 25 or 30 feet away. Well I go to the store and buy this 000 wire and some grey 2" PVC conduit. Call the electric company and they send out these linemen to turn of the power at the pole. Took two large men to get that 000 wire in that conduit. More than a little tight. Father-in -law wired the box. It was a thing of beauty. Finally got it buttoned up and called the power company to have the power turned back on.


These linemen come back to inspect things and give me that look, do you have way more money than you need or are you just crazy. Probably had never seen wire as thick as 000. Positive the line coming in from the pole wasn't that large.


Had two inspectors come in when I sold the house and both said they wish they knew it was for sale because it was built so well and the panel box was so beautifully done. Of course I had to take credit for it.

[Mike_J]

Giving thought to how I am going to order the grounds on the horizontal terminal strip. Start off with the OT secondary ground probably in the same lug with the first filter cap ground which will connect with the power tube cathode grounds on its way to the lug. Next lug is for the ecap ground for the screen supply that went to the turret that may never be used and on to the lug we are discussing. Next lug is from the bias circuit grounds. The other end of the bias circuit grounds will go to the buss bar next to where the presence pot is grounded. At the same point the third stage ecap ground for the pi section will be connected. Will remove the connection at the V1-3 resistor ground and take my ground to the buss bar from V2-8 and connect it to the same area the bright channel volume pot is grounded The ecap for the V! and V2 tubes will also connect at the V2-8 cathode ground. If my count is correct that is three lugs. This should ensure the grounds go from one end of the amp and never cross over themselves and only meet the chassis at one point which is the PT ground bolt. This system will require that all input jacks and the output jack are isolated from the chassis. Any comments regarding this system would be appreciated?


Probably will go back to using two connectors each with two holes in them and screw the vertical terminal strip to the chassis somewhere in the area. Will send the ground for it to the last hole in the ground lugs. This assumes I can get the OT secondary ground wire and first filter cap wire in the same hole. Will need to investigate that likelihood first. If I use the horizontal terminal strip I will tie the lugs together with some 10awg bare copper wire and connect it to the hole in the lug on the PT bolt with some 14awg wire. Should be adequate.



Thanks
Mike

[scstill]

Here's a picture - only one chassis ground connection needed. YMMV
FYI, what I call 'AC micro wobble in the ground return' = unwanted source of noise

This is a great diagram in #7. Thanks.
Very timely for my build which is point to point wiring.
The build is housed in an old metal car radio about 12" square (pic below Pwr supply left side, Preamp right side)

Two questions:

1) I have already built the power supply filters on a connector strip with a single chassis ground point (pic below).
The power supply is located on the opposite side of the chassis from the preamp and PI (both powered from same node). Is it better to leave the power supply ground alone or divide the filters into separate grounds (like pic in #7) and run a line to the preamp ground which would be about 8" on the other side of the chassis.

2) the input jacks are grounded to chassis with coax running to the Preamp grids (2 channel). Is it better to leave alone or run a ground line from jack to preamp ground (about 6")?

I suppose I could finish the build and if its unusually noisey I could try some ground tricks. I do like the single ground buss idea. but again to get the ground in the same node I would have to separate and run a long ground wire.

[tubeswell]

1) I have already built the power supply filters on a connector strip with a singe chassis ground point (pic below).
The power supply is located on the opposite side of the chassis from the preamp and PI (both powered from same node). Is it better to leave the power supply ground alone or divide the filters into separate grounds and run a line to the preamp ground which would be about 8" on the other side of the chassis.

Try it as you built it, then it you think its noisy, try it the other way. Otherwise leave it. Its one of those 'if it works' things

2) the input jacks are grounded to chassis with coax running to the Preamp grids (2 channel). Is it better to leave alone or run a ground line from jack to preamp ground?

The schematic ground point is only representational. If the input jack sleeves are the grounding type (i.e. Switchcraft)  then run the buss bar to the Hi-gain jack sleeve terminal lug, and use a decent toothed-spring washer to make tight chassis contact.