Hi there,
possibly stupid question that came to my mind last night:
Has anyone ever tried coaxial cable for tube heater wiring? With the outward and return current flowing coaxially, electromagnetic interference should cancel out perfectly.
Of course, the relatively small conductor thickness would require wiring each valve separately instead of daisy-chain.
Stupid idea?
Greetings,
Andreas
possibly stupid question that came to my mind last night:
Has anyone ever tried coaxial cable for tube heater wiring? With the outward and return current flowing coaxially, electromagnetic interference should cancel out perfectly.
Of course, the relatively small conductor thickness would require wiring each valve separately instead of daisy-chain.
Stupid idea?
Greetings,
Andreas
I don't think there are max specs for the current given anywhere. One probably has to estimate from the cross section of the thin wires the shield consists of. For standard RG58, the outer conductor becomes a quite thick bundle if straightened and twisted together - at least 1-1.5mm² I would guess...
Greetings,
Andreas
Greetings,
Andreas
A co-ax is only effective for un-balanced signals, so I suppose it would be good if one side of your heater referenced that way. The idea is (At RF anyway) that current doesn't flow in the screen.
My Mic amp has 70bB gain, AC heaters (referanced at 100 volts) and is silent. It just has a tight twisted pair.
My Mic amp has 70bB gain, AC heaters (referanced at 100 volts) and is silent. It just has a tight twisted pair.
Definite overkill there. You have the problem of wire size and expense. RF co-ax might work for small signal VTs, but it's definitely dubious for lighting up the big finals. It's also a good deal pricier than plain old wire that you twist. At these low frequencies, twisting the wire and paying attention to routing will do as much for you in terms of hum suppression.
The only place I've used RF co-ax is to bring the output back to the gNFB summing node. Here, just the one side of the shield is grounded (at the summing node) and the speaker end is left floating. That way, the shield acts like a Faraday cage to reduce electrostatic effects and capacitive coupling. Doing otherwise (grounding both ends) might allow a ground loop that adds power supply hash to the signal chain (what we're trying to avoid by star grounding).
If you are saying that coax means the screen must be grounded (i.e. no CT heater, or humdinger) then you are correct. Otherwise you still get magnetic cancellation, but capacitive coupling instead. You could be worse off than using twisted pair.
At RF the idea is that the current on the inside of the screen exactly matches the current on (the outside of) the inner, so you get magnetic cancellation. In addition, to avoid radiation, you want the current on the outside of the screen to be zero. A good coax guarantees the former. Correct use provides the latter. This separation of screen current into two separate flows only works when the frequency is high enough that the skin depth is much less than the screen thickness.
No, coax will cancel too. Coax will cancel the magnetic field better than twisted pair, but it introduces other issues (see above).
Hi there,
thanks for the input! So I will probably go the simple "twisted wire" way.
The question arose from different ideas in order to avoid the job of manually twisting wires for the heaters - twisted pairs readily available are all from the small signal or telecommunication area (Cat.5 Cat.6 etc), with wire sizes inappropriate for heater currents...
Greetings,
Andreas
thanks for the input! So I will probably go the simple "twisted wire" way.
The question arose from different ideas in order to avoid the job of manually twisting wires for the heaters - twisted pairs readily available are all from the small signal or telecommunication area (Cat.5 Cat.6 etc), with wire sizes inappropriate for heater currents...
Greetings,
Andreas
Manual wire twisting does not take long. It does not have to be done perfectly (unlike Cat 5 etc.) as the wavelength at 50/60Hz is rather larger! Remember that magnetic induction goes like 'loop area' x 'frequency' x 'current'. We already have 'frequency' on our side as 50/60Hz is low. Twisting reduces loop area. Current we keep low by ensuring that the output valve heater wiring keeps well away from the input stages.
If you want additional shielding, you could run normal twisted heater wires through some copper or aluminum tubing that is attached (and electrically connected) to the inside corners of the chassis. Just exit the tubing near the tubes. But there doesn't seem to be too much problem with plain old twisted pair, so why bother?
That technique would be more useful to run (unbalanced) signal wires between the preamp tubes and the back of the chassis for the RCA jacks. But shielded cable would probably provide the same outcome and be much easier.
..Todd
That technique would be more useful to run (unbalanced) signal wires between the preamp tubes and the back of the chassis for the RCA jacks. But shielded cable would probably provide the same outcome and be much easier.
..Todd
Last edited:
I don't think that statement is correct for typical audio amplifier situations. As DF96 indicated, the external magnetic induction is proportional to effective loop area of a cable pair. For comparable twisted pair and current carrying coax that audio people would use, I reckon it is much more likely for the twisted pair to be significantly better. If possible, the situation can be further improved by using thin-walled insulation cable for the twisted pair, without even getting in to the twists/turn and angle of dangle topics.
One relevant reference is 'MAGNETIC FIELD SHIELDING EFFECTIVENESS OF A COAXIAL CABLES HIGH-PERMEABILITY SHIELD ON TWISTED', by Diakin and Derewiany, 1992, who were looking at magnetic shielding effectiveness of cable for use in submarines. Although their choice of twisted pair and coax were mil spec and I can't get details on the twisted pair VE3452, the coax was RG214 which is not too far off a normal coax.
Ciao, Tim
Your two sets of three need to be twisted together in order to maximize magnetic canceling - untwisted sets of conductors don't cancel well.
Sorry, the reference title I posted didn't paste in properly - the relevant reference is 'MAGNETIC FIELD SHIELDING EFFECTIVENESS OF A HIGH-PERMEABILITY SHIELD ON TWISTED PAIR AND COAXIAL CABLES, by Diakin and Derewiany, 1992.
The paper "Predicting the Magnetic Fields from a Twisted-Pair Cable", by MOSER & SPENCER, 1968, includes a graph showing the reduction in external field with the twist pitch as a function of distance from the cable.
Ciao, Tim
The paper "Predicting the Magnetic Fields from a Twisted-Pair Cable", by MOSER & SPENCER, 1968, includes a graph showing the reduction in external field with the twist pitch as a function of distance from the cable.
Ciao, Tim
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.