|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
69.12.241.189
Due to real estate and weight issues, I've started making separate power supplies and amps. It also helps for the endless tweaker, as I try to build "killer" supplies that can be used for a variety of different amps, with little or no mods. I make my own cables and use Jones plugs.
I mix the HT (or several) with AC filaments, all in the same bundle.
I use hum pots, and have quiet amps. BUT, I read here somewhere that this is not "advised", because the B+ will pick up hum from the AC filament wires. Really? So I made two separate cables, one for filaments, and the other for well filtered B+, and I couldn't hear any difference. I'd rather not use two cables, and am wondering what negative impact there really is in using one, ala heathkit W2, W3 etc...
Would it be advised to somehow shield the AC or DC in the power supply cable? Or is all of this concern "silly"?
Follow Ups:
I would trust your ears. I use the same practice as you do, grouping the HT voltage with twisted AC filament and it's quiet. I've also run my twisted AC outside the HT group but I couldn't hear any difference with filament VAC at or below 6.3. I can imagine things changing at higher VAC but that doesn't mean I'l actually hear anything.
Matt
The idea is that of voltage level, electric field strength, and circuit impedances. Typically, an AC heater string will, by way of induction, inject ac hum into the sensitive (high impedance, low voltage) signal path. Shielding the signal path and/or heater string will normally neutralize this affect.However, here we are dealing with a high voltage, low impedance power supply line, being tickled by a low strength E/M field from your heaters. There is simply insufficient field to produce electron flow in the B+, as the low source impedance of B+ is the dominant player. Most likely, you would not experience any 60Hz inductive pickup from a heater line.
What would be more of an issue would be high AC current, I suspect. Higher currents produce correspondingly stronger magnetic fields. The magnetic fields are more problematic than electric fields, as they can actually permeate an electrostatic shield. But again, looking at source impedance, I find it hard to believe you will pick up any hum.
Past tense.
After reading your link (thanks) I fail to see what part of your Low everything PS could not also include a separate chassis. Sure, the extra wire would add a little dcr, but less than one of the chokes, if made properly short as possible with overkill gauge wire. I agree that output stage ripple, can be "worse" than we used to think, and be fine.
So please help me understand what part of your experimenting lead you to
the state of "past tense"? It appears that the very positive results
WERE amidst the usage of separate supplies and cables. Yes? no?
Hi bobbyj,Greg's amp, built about three years ago, actually uses umbillicals now, and we heard all this great stuff, despite the umbillicals being in-circuit. But now, I want them " outta there". I learn over time, we all do.
A BIG factor is if the amp is P-P or SE, as Greg's single 245 stage is. You see, in SE, the ground IS the negative AC signal of ther AUDIO circuit, and the whole supply, ( warts and all ), is "seen" by the SE amp. So, the can of worms presented by SE design IMHO PRECLUDES use of umbillicals, and indeed, I have very recently heard a HIGHLY developed SE 2A3 amp in Montana, where a single 12 gague ground or B+ wire was reduced by just an inch, and the resultant configuration played a lot better.
In P-P , the rules are much more lax, and umbillicals may be used, but FOR ME, I'll no longer do so in P-P because I feel it can't help or improve performance.
When using umbillicals, its beneficial to split up the capacitance some, so the audio chassis has some local storage, maybe a LOW DCR L and a under-50 uF cap after it, feeding the audio circuit at "close" to its point of use. This is my personal take, YMMV. Regards.
Thanks Jeff. I see your point, I prefer single chassis mono designs as well, yet sometimes it seems impossible, due either to size/weight or lack of two matching sets of components. I've been curious about the "controversy" regarding Low everything power supplies, and now am prepared to make one and see for myself. I have 4 chokes:.33H/2.2 ohms
.30H/3 ohms (x2)
.5H/15 ohmsMy plate transformer has 11 ohms DCR each side of CT, so 22 ohms total across the winding. It's the lowest I've got. It's 570VCT. I'm guessing that even though I'll use choke input, as you suggest, I will still end up with a "cap" input B+ value, or close? I was going to do:
.33H/4uf/.3H/8uf/.3H/16uf/.5H/30uf. I wanted to operate an SE 2A3, but I think I'll end up with a little too much B+, so I may have to use a 300B, which then forces me to use DC filaments. BTW, I'm guessing that
after the effort to create a low dcr supply, that self-bias is kind of contradictory, so does that mean you used fixed bias on your friends 245 amp? just curious. And what about the separate supply for the driver, which I assume needs a much lower ripple factor? Does it "pay" to make a similar supply here as well? Or? and why? thanks
Hello,You have asked some good questions. I prefer the sound of self bias over fixed bias on filamentary triodes. So I personally would never ever use fixed bias on such tubes.
The self bias supply must be done well. I'd ONLY use Mills MRA-12 resistors, and in many cases, it is beneficial to parallel two EXACTLY matching Mills Rs to get your desired value, as this sounds good. For a bypass cap, consider an oil in a round can if you are on a budget, or maybe a Dynami Cap film with similar caps in parallel for bypassing if finances allow.
IMO, I hate, absolutely hate to see you opt for a 300B when a 2A3 is in your grasp to build. IMHO, 2A3s sound better - and 300Bs are only for people who don't " get it " and HAVE to use a 300B - due to their use of inefficient speakers.
The PT you have, 570 VCT at 22 ohms, will be WAY too high for the intended 2A3 stage. My ESTIMATE is that a low DCR input L such as 35 to 320 mHY, with an optimized value of C1 after it, will give you WELL OVER .9 times ( VAC to VDC, for a choke input meeting critical inductance at 60 HZ ), and SLIGHTLY UNDER 1.41 times ( VAC to VDC, pure cap input with attendant splke generation )....Probably 1.35 is a number I like to use in my estimates, minus the drop for ( tube or SS ) rectifiers and the DCR loss of the L1 and L2, etc filter chain and theoutput trannie's DCR.
Crunching numbers, half of your 570 VCT PT is 285 VAC, times 1.35, is 384.75 VDC before rectifier and iron losses...too high. Ideal sonic operating parameters for most 2A3s, in an idealized SE amp, will be 240 VDC P-K and about 42 to 45 mA MAX !! This necessitates a sub 300 VDC supply for that stage on the 2A3 plate, with self bias.
Hey, what about an inexpensive industial control power trannie offa eBay? Lets look at that. If its 120 to 240 ( or any similar 1:2 step ratio, like 240 to 480 ), you get the following VDC ( assumes 120 VAC house line ) : 240 times 1.35 equals 324 VDC. BUT, your line might be different than 120, many are, and, if the trannie is lightly loaded, it may produce 5 to 10 % higher VAC and VDC voltages.
Enter the Signal Transformer Company's DU-1/2. It has four windings, two on the primary and likewise for the secondary. Each winding is tapped at 104 VAC, 110 VAC and 120 VAC. THINK of all the hook-up possibilities !! Its really low in DCR, even if seriesed ( under 5 ohms as I recall ) and it is beautifully made. Alas, taps on a PT are NOT as desireable as no taps, but IF you are a DIY experimenter, this trannie is useful. NO, I wouldn't use it on a production item. Does SkyCraft Surplus have any DU1/2s left at $40 a piece??
Your suggested filter configuration is too complex, sorry, and your C values are too low. I would suggest, as I have before, to keep the L HY values at 1/2 HY MAX and under 10 ohms, Cs at 50 uF MAX, with 40 uF being nicer. For 2A3 or 245 output stages, I have not yet fully ascertained if a double Pi filter is going to be better overall than the triple Pi filter as outlined in Greg's amp write up. If it were double Pi, I would just eliminate the first L1/C1 and keep all other values the same. This would have a quicker step response I would GUESS, than the triple supply.
The front end supply needs far more filtering than the output stage supply. However, low DCR Inductors, which are DRIVEN, is certainly desireable, from years of my listening experiments. But, you can relax the DCR requirements SOME. Use your 570 VCT / 22 ohm PT for the front end. Use 5U4GBs or 5V3's as rectifiers. If you are on a budget, for wiring, use a single strand of Kimber TCSS for audio signals, a lightly twisted double pair for B+ wiring, and a single TCSS run as ground, in PARALLEL with a single 10 or 12 gague run of ROMEX with the insulation left on. That is a good wire compromise IMO. Anything less in SE is a waste of time and money.
On the input stage B+ supply, I would not be totally adverse to using a dropping R, but the R might be 1K or so, with 40 uF of 50 uF max after the R. If the input stage is low current draw, a brute force shunt regulator ( Mills MRA-12s B+ to ground ), would be possibly worthwhile.
Have fun and feel free to email me if you need off-Forum help.
Thanks again Jeff. It all makes sense, except I would have guessed a fixed bias. I have always used self bias for SE DHT's as well, but I thought a fixed one "might" "compliment" the low everything situation,
have you tried it together with an "optimized" supply?
I prefer a 45 or a 2A3 mainly because I can use AC filaments, and with my (guessing) 98db speakers, that works out fine. Poor 300B's have such a mediocre reputation, yet I wonder how much of this is truly deserved,
and how much is anti-fashion nose picking. I honestly don't know. I have a couple NAVY marked 300B's, that test good, they seem like a big 2A3 to me. Which brings up the starved filament concept. Perhaps you read Steve Bench's article/research. 385VDC minus a bias, would be a relatively low OP point for a 300B, maybe 60% of max dissipation, which then "might" allow a reduced filament voltage of even 3.8VAC ? which in turn might be quiet enough (lower HD too).
Also, have you applied the low everything concept to a filament supply? Why do we always use "big" cap values in this situation?
I have found that a "shitty" filtered supply (with a hum pot) can be made quiet, Maybe a 300B just needs "slighlty" rectified and filtered
(and lowered) filament supply? maybe just getting 5VDC to have maximum 2.5VAC ripple, is "enough" for quiet operation? Maybe a 300B, if surrounded by an "ideal" PS would sound fine? With basically the same Rp and mu as a 2A3, what could explain the perceived "inferiority"?
I may throw together a test amp, and use the PS parts I have (with bigger cap values), which does spell 300B for the moment, until I can come up with the right power transformer for a 2A3. Bench suggests that you can lower the filament voltage by an amount that is a "similar" percentage value less than whatever % your op point is less than maximum. I have tried this to a 20% degree with other DHT's, and yes it worked fine, and had measurably lower harmonic content. Long term tube life? don't know, yet Bench did "some" tests here, and found no premature death.
And then there are those individuals who believe that "stated" maximum ratings are actually at that 62% "golden ratio" point already, and that you can run a 2A3 at 400V on the plate, 50ma all day....?...?
I know that this has been tried successfully, well, "without failure".....
Hi,Why postulate?
All ya need to do is LISTEN to a 2A3 at Manufacturer's rated op point, vs; about 62% of that, in a GOOD amp circuit, and you will easily KNOW the correct answer. : <)
I used to use about -5% on my DIY Efs. Value selected BY EAR.
AC on a 45 or 2A3 is VERY DESIREABLE, and yes, 'ya can do it at 2.5 V Ef. Its another reason for that tubes' selection in audio amps.
My original triode amp audio mentor, Mr Robert W. Fulton ( 1925-1988) would NEVER EVER allow me ( in the 70's and 80's) to use 300Bs. He was incensed and said " everybody KNEW, back in his time, the 2A3 was superior. It was common knowledge !! "
None of these older tubes BTW, I'm told ( by a smart/experienced Montana audiophile ), can compare to the current Emission Labs 45s and 2A3s.
This post is made possible by the generous support of people like you and our sponsors: