I normally don't choose to connect ground between stages, instead returning both to my star point to avoid ground loops. I therefore have just one signal cable connected between (unbalanced) stages.
As valve stages are typically relatively high in impedance, I don't believe you need to use heavy cabling for most duties. I like to use heavy cabling for the ground scheme though.
I think it is wise to use heavy cables for your heaters. These should be tightly twisted to reduce induced currents into nearby wiring.
As valve stages are typically relatively high in impedance, I don't believe you need to use heavy cabling for most duties. I like to use heavy cabling for the ground scheme though.
I think it is wise to use heavy cables for your heaters. These should be tightly twisted to reduce induced currents into nearby wiring.
You probably should not use shielded cabling for your tube work. The shield will create capacitance to ground. This is not much capacitance mind you, but it is on a par with the capacitances that are inherent to the tubes themselves.
A good amp design takes all these capacitances internal to the tubes into account. Adding to that capacitance with shielded wire won't help things.
In a tube amp, layout is the secret to avoiding cross talk and coupling. Keep outputs away from inputs, put successive stages in a row, such that the tube with the most signal is farthest away form the tube with the least signal.
Start with a clean, well layed out schematic. Your physical layout should mirror your schematic. This is a gross generalization, with many exceptions, but a very good rule of thumb.
A good amp design takes all these capacitances internal to the tubes into account. Adding to that capacitance with shielded wire won't help things.
In a tube amp, layout is the secret to avoiding cross talk and coupling. Keep outputs away from inputs, put successive stages in a row, such that the tube with the most signal is farthest away form the tube with the least signal.
Start with a clean, well layed out schematic. Your physical layout should mirror your schematic. This is a gross generalization, with many exceptions, but a very good rule of thumb.
I've used PTFE (teflon) insulated silver-plated copper equipment wire. Really easy to work with, pretty cheap and the insulation won't melt or burn when you solder it. I've used 20AWG, which costs me about 1€/metre. It's rated 600V and 17A. If you open a bunch of commercial tube amps, you'll find that most don't use anything special for internal wiring. Just keep the PSU and amp sections as far apart as possible within reason, and make sure all AC cabling is done with firmly twisted pairs.
My wire favourite is 'as little as possible' brand. I normally use a full copper PCB as a groundplane and wire point to point using, to the greatest extent possible, device leads only. Components with one side to ground are soldered directly to the groundplane PCB oriented to minimize path lengths. Half the number of solder joints, an extremely low impedance ground connection between nodes, no funky terminals. And frugal!
poobah said:
I disagree. There is a place for shielded cables. In my amps, I use RF coax to run the gNFB from the speaker terminal back to the summing node. In this case, the external shield is grounded at the summing node only, so that it serves as a Faraday cage, and not a ground return. This helps to keep noise out of the summing node, and the cable capacitance contributes to compensating the amp. In most cases, any additional compensating capacitor isn't even necessary.
For one very high gain (120db(v) ) design I did as part of a TRF longwave receiver, I used shielded cable quite liberally, including not only signal paths (from the front end to a panel switch for switching a narrow BPF in or out) but also DC lines. Got that sucker stable despite the huge amount of gain.
Like rdf, I also tend to favor using double sided PCBs in a modified "Dead Bug" construction where connection pads are etched out. I use a Dremmel with a fine etching point to clear away the copper to form the pads. The rest of the board provides a ground plane for that particular module. These are returned to the PS star point individually. It may not be the prettiest method of construction, but it sure does work well, and allows for the use of parts on hand, and easy circuit modification.
Start with a clean, well layed out schematic. Your physical layout should mirror your schematic. This is a gross generalization, with many exceptions, but a very good rule of thumb.
For sure.
Miles Prower said:
Hi Miles. Is this in regards to circuits employing multiple boards? So far I've been fortunate to squeeze everything onto one. In this case (a long dead-and-gone, modified 50 times Mullard 3-3) it included tacking the power supply cap terminals, RCA input shell and negative speaker binding post all directly to the PCB.
http://www.diyaudio.com/forums/showthread.php?postid=542523#post542523
And ease of modification? Too easy!
Re internal cabling
If your chassis is well shielded i.e aluminum or steel then I see little use for shielded cable from stage to stage of an amp. I use Litz wire for signal grid leads and try to keep my leads as short as possible. I'd only consider shielded wire if the first input stage was close to a power transformer but I would try to lay out my chassis where this wouldn't be the case. Tristed pair is used for transformer input coupling using 600 ohm balanced lines in which you are dealing with two signals 180 degrees out of phase. But unless you are in a radio or TV broadcast station or with a radio or Tv transmitter next door to you then IMHO you really don't need balanced lines. Ray Hughes
If your chassis is well shielded i.e aluminum or steel then I see little use for shielded cable from stage to stage of an amp. I use Litz wire for signal grid leads and try to keep my leads as short as possible. I'd only consider shielded wire if the first input stage was close to a power transformer but I would try to lay out my chassis where this wouldn't be the case. Tristed pair is used for transformer input coupling using 600 ohm balanced lines in which you are dealing with two signals 180 degrees out of phase. But unless you are in a radio or TV broadcast station or with a radio or Tv transmitter next door to you then IMHO you really don't need balanced lines. Ray Hughes
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.