Thanks so much for all that. I poked around at my amp and I believe I have all those in place: the grounds from the Amp boards go to the ground of the PSU. A line from this ground (PSU) goes through a CL-60, which then goes to the chassis, and then to the IEC's ground (dirty).
The heatsinks are also grounded to the dirty ground- should they stay "clean?" If so this would mean that I should put the CL-60 between the chassis and the IEC ground.
And I believe the RCA ground is kept clean. I've never used a coax cable- does it have 2 or 3 leads (shielding)? Where is the shielding connected- clean ground?
Thanks
The heatsinks are also grounded to the dirty ground- should they stay "clean?" If so this would mean that I should put the CL-60 between the chassis and the IEC ground.
And I believe the RCA ground is kept clean. I've never used a coax cable- does it have 2 or 3 leads (shielding)? Where is the shielding connected- clean ground?
Thanks
Getting a hum free amp in DIY is tough if one is not careful to the strict adherence of some key points.
Chassis ground/earth ground/mains ground are “dirty”. You want to never let your analog ground (0 volts ref from source and speaker and PSU “GND”) ever directly touch the “dirty” ground. They can be connected via a ground loop breaker (GLB). Usually a CL60 or 8D-20 NTC and optionally a 22nF MKP 400v film cap. Make sure you never inadvertently connect clean ground to dirty ground where the RCA jacks penetrate the panel. Some lower cost ones don’t have insulator bushings. Use star topology for tying your dirty grounds to a single point on the chassis. Use star topology to tie your 0v analog clean ground.
Tie the heatsinks to chassis ground. Keep your linear trafo as far from the input stage of the amp as possible. Use shielded coax from the RCA to the amp input. Twist all AC lines to reduce emission. Twist all signal lines to reduce pickup. Don’t run the signal wires near or next to the trafo or in parallel to wires from the mains switch.
One of the worst culprits is a mains switch on front panel that requires mains wire to traverse from IEC in back to front and exposing its EMI to all the sensitive low signal wires.
Hope that helps. There are many threads on this in the Solid State forum. Member Bonsai has really studied this and put together some nice guides. Here is one:
How to wire up an Amplifier
Ground Loops
Curious.. I built My F6 and DIY store PS..Exactly... as per Guide instructions and it has Zero noises, even at full power with ears next to the compression drivers .
Power on / power off, No differences.
Take a serious look at what you did differently.. and fix it
PS ; imo mounting the Transformer on it's side is less than ideal... noise wise.
Power on / power off, No differences.
Take a serious look at what you did differently.. and fix it
PS ; imo mounting the Transformer on it's side is less than ideal... noise wise.
FWIW, my F6 build is also dead quiet. Don't know if that's just luck, but I did follow the build thread description pretty much to a T. And my power transformer is also mounted vertically as far from the input transformers as possible.
I’m curious whether anyone has gone the route of putting their power supply in a separate box. Looking down the road I think it’s likely I might build more than one of the first watt builds and having the power supply the external would seem to make it easier to change amplifiers. Also you could put the amp in a smaller box more like the ACA. And just run a power cord to it from your power supply. Does this idea have any merit or is there something I’m missing that would make it a really bad idea? Have any of you tried this?
I don't know the percentage of builds incorporating separate PSUs, but it's certainly been done quite a bit. Your general concepts are sound. You can find a number of examples for implementation throughout the threads. You'll find quite a bit of chat re: suitability of connectors and wire, safety, and where to place caps when using a CLC or CRC etc.
Advantages may include... modularity / ease of DIY swapping for mix-and-match, layout for your rack, performance vs. space constraints, and lower noise / improved sonics.
Disadvantages may include... complexity, total volume / space of the dual chassis, and cost.
I use two PSUs across all of my FW-based amplifiers (currently ~8 of them). So, cost is hit and miss. Using two PSUs is cheaper than 8, but the initial cost may be higher just due to an additional chassis for one or two amps.
Have fun!
Advantages may include... modularity / ease of DIY swapping for mix-and-match, layout for your rack, performance vs. space constraints, and lower noise / improved sonics.
Disadvantages may include... complexity, total volume / space of the dual chassis, and cost.
I use two PSUs across all of my FW-based amplifiers (currently ~8 of them). So, cost is hit and miss. Using two PSUs is cheaper than 8, but the initial cost may be higher just due to an additional chassis for one or two amps.
Have fun!
It’s a fine idea.
But you can’t put the amplifiers into a smaller box as you need the heatsink area. (That said, you could put them into a narrower box…)
Anyway, as the chassis and PSU is actually the most expensive part of these builds, most people just swap the amplifier boards in/out of the same chassis.
The separate PSU route adds a lot of complexity and expense, which is no reason not to do it… this is DIY after all and it’s a lot of fun to build things the way you want them.
I do, however, suggest that you build the first one in a more conventional manner, and work towards something more fancy as when you have that experience.
But you can’t put the amplifiers into a smaller box as you need the heatsink area. (That said, you could put them into a narrower box…)
Anyway, as the chassis and PSU is actually the most expensive part of these builds, most people just swap the amplifier boards in/out of the same chassis.
The separate PSU route adds a lot of complexity and expense, which is no reason not to do it… this is DIY after all and it’s a lot of fun to build things the way you want them.
I do, however, suggest that you build the first one in a more conventional manner, and work towards something more fancy as when you have that experience.
Gianluca and his Modushop team are preparing Monoblock case and adequate (so heatsink-less) PSU case, both narrower than existing ones
dissipation info - I believe it'll be 4U/400deep, so good for easy 80W of heat, 100W with care
dissipation info - I believe it'll be 4U/400deep, so good for easy 80W of heat, 100W with care
External, separate power supplies work quite well. Especially if there may be signal transformers in the front end of the amplifier. And if you like to use really beefy power transformers and ginormous capacitors, then in can be nice to split up the weight as well. 😉
While I haven't been able to take advantage of a reusable PSU, the dedicated one I built for my P-channel VFET amp has made a wonderful difference in the sound. Definitely consider building two channels of PSU into that separate chassis.
While I haven't been able to take advantage of a reusable PSU, the dedicated one I built for my P-channel VFET amp has made a wonderful difference in the sound. Definitely consider building two channels of PSU into that separate chassis.
I did read somewhere about not having all your capacitors in the PSU box if building it separate can anyone verify and the reason ?
You want local decoupling capacitors close to the amplifier circuit boards in your amplifier case to store a charge locally to maintain voltage and also to provide filtering of any noise that may be picked up in the lines between the power supply case and the amplifier circuit case.
So,
If I have the complete dual mono power supply in a separate box including cap bank.
Feeding DC via umbilical to the amp boards. (Umbilicals are three wire, +, - and ground braided together and less than two feet long.)
Everything silent now.
What would be good decoupling caps to fit inside the amp?
If I have the complete dual mono power supply in a separate box including cap bank.
Feeding DC via umbilical to the amp boards. (Umbilicals are three wire, +, - and ground braided together and less than two feet long.)
Everything silent now.
What would be good decoupling caps to fit inside the amp?
The same type caps you have on the PSU board.
Ah, but… you said, “Everyting silent now.”
Which is a fantastic time to completely stop, call it good, realize the project is finished, and sit back and enjoy the music.
I never, ever touch a project if it gets to genuinely quiet. DIY is an exercise in fighting hum, if you have none, it’s time to smile, declare victory, and start the next thing. 🙂
Ah, but… you said, “Everyting silent now.”
Which is a fantastic time to completely stop, call it good, realize the project is finished, and sit back and enjoy the music.
I never, ever touch a project if it gets to genuinely quiet. DIY is an exercise in fighting hum, if you have none, it’s time to smile, declare victory, and start the next thing. 🙂
Maybe just some 70 uF motor run capacitors. These are actually big oil filled polypropylene, and can fine tune the sound of the amp.
But yeah, 6L6 is not wrong.
But yeah, 6L6 is not wrong.
Hoping following requests are in line with OP's request:
- would it be ok to use a +20-0-20VAC transformer (of proper size 300/400VA, which I already own) for either a stereo F5 or F6 amp?
- would it be ok to enclose an (Hypnotize) B1 and volume control with its separated transformer in the same case as the F5/F6?
- would it be ok to use a +20-0-20VAC transformer (of proper size 300/400VA, which I already own) for either a stereo F5 or F6 amp?
- would it be ok to enclose an (Hypnotize) B1 and volume control with its separated transformer in the same case as the F5/F6?
"Coax" cable can have one, two, three leads or five leads (for a single signal channel). Strictly speaking, only one and two lead coax is actually coaxial. The others are cable. I am also not discussing true coaxial that has two wires coaxially located within the shield.And I believe the RCA ground is kept clean. I've never used a coax cable- does it have 2 or 3 leads (shielding)? Where is the shielding connected- clean ground?
One lead coax has one lead in the center for signal, and surrounding shield. Shields, as a general matter, can be either foil, woven copper, or both (the best). Foil shields always have a wire in intimate contact with the shield, so as far as I know there are no foil shielded one-lead coax. To use the shield, one may tease out and twist a portion of the woven shield to create a lead that is then soldered or crimped to the connection tab/terminal/PCB pad. Or, one may carefully solder one end of a short length of copper wire to the shield, and then solder/crimp the other end of the wire to the connection tab/terminal/PCB pad. Alternatively one can uses a connector (like some Neutrik XLR products) that make contact directly with the shield (in the Neutrik products, one expands and fans open the woven shield, which is then clamped in place and makes electric contact with the shell by mechanical assembly of the connector). The advantage of the Neutrik approach is that there is no break in the shield between the end the cable's shield, and the connector housing; the fan-out provides full shielding with no gap to allow FR/EMI to leak in. If I understand it correctly, the Neutrik XLR also provide a circular continuous capacitor to isolate DC voltage that may be on the shield from entering signal return/ground and causing a group loop. The capacitor, however, allows RF/EMI energy to conduct to chassis ground and thence out through the safety ground in the power cord. That is why you may often see small ceramic capacitors (must use ceramic capacitors, which are effective at RF frequencies) connecting from an RCA chassis jack to the chassis via a small screw and toothed lock washer. I do not like one lead coax because the shield must perform double duty as the signal ground/return as well as the RFI/EMI shield. This is how ground loops can occur.
Two lead coax is just like one lead coax but there is a wire in intimate contact with the shield so that one solders the shield wire to the connection tab/terminal/PCB pad. You can ignore the shield wire if you want to use the Neutrik-style connector that makes a direct connection to the shield. It has the same issue as one lead coax: the shield performs double duty as signal return/gound, and RFI/EMI shield.
Three lead cable has two wires typically twisted together) located within the woven or foil/woven shield. An example of such signal cabler with twisted pair is Canare L-2E5. You could use this type of cable in two different ways. First, you could use the two wires to connect both the signal hot and signal return/ground. The shield itself is NOT tied to the signal return/ground but terminated to the chassis. This way, the shield has only a single function: shield against RFI/EMI. The second way that it can be used is as a "twisted pair" to reduce common noise pickup. Both wires are used to conduct the signal hot, and the shield performs double duty as two and one lead coax. Again, not my preference.
Finally, there is five lead cable which contains two twisted pairs within the shield. An example is Canare L-4E6S. In this approach, you can use one twisted pair for signal hot, and the other twisted pair for signal return/ground. The shield performs only a single function: shield against RFI/EMI and is tied to chassis ground.
FWIW, I only use five wire cable in microphone and external interconnect cables.
- Home
- Amplifiers
- Pass Labs
- F5, F6 Power supply guidance