DN2540
Mouser
689-DN2540N5-G
Click: dn2540
£1.12
(And don't do that pathetic girly whine about postage costs).
🙂
Danny, this and your other questions are eminently reasonable. So let me try to give you a reasonable answer, shorn of as much mysticism and speculation as I can.
First, unless you're willing to trim out a good chunk of the unquestionably audible spectrum, it will be a challenge to prevent transients with significant high frequency content from reaching the output stage. Practical filters in the analog domain are gradual, not brick wall.
Second, although the audibility/detectability of 30kHz is... questionable, the power sucking and distortion of supersonic ringing is not. That's why circuits that oscillate at VHF sound so lousy and why we use grid stoppers and similar measures to prevent these oscillations in audio circuits.
Third, there's the question of feedback, which is commonly used in competent amplifier designs. Without getting into a tutorial on stability, the inclusion of a second order rolloff at a frequency that low means that the amp's open loop dominant pole will need to be set at a few kilohertz. This inevitably results in poor HF response, increased treble distortion, and rising output impedance.
As a parenthetical note, it's a good thing, IMO, that questions be asked and things that seemingly don't make sense be brought forward, irrespective of who said them. One of the reasons that some particularly honest experts use a "nom de clavier" is specifically so that their ideas are considered and argued on their merits, without consideration of authority.
Once SS CCS sounds harsher, I don't care why it sounds that way, I care that it sounds that way.
You do not need too much luck getting wide band chokes.I tried,for the fun of it,a 10H filter choke,20 buck chinese iron,loading an 300ohm Rp tube.Flat up to 65khz.
That's the key right there. Try the same choke with something like a 6SN7 and the bandwidth, on both ends, will be unfortunate.
A simple ccs can be point-to-point wired in less than 30mins. No need for a PCB or anything like that. So why not just get a six-pack, put on some decent background music, sit down and try this for yourself. Give us a report on Monday on what YOU think, and what your friend has to say...
SY you are right again.High Rp tubes are mostly avoided,although some quite linear tubes belong to this group.
I would trade 20db more THD anyday for 10X les Rp
I would trade 20db more THD anyday for 10X les Rp
For me, that would depend on what I'm trying to get the tube to do. Tradeoffs are application and situation-specific.
I'll do that, though not this week, since I just moved and I have a lot of unpacking to do before I can set up my work bench.
Ok, fine so far.
Doesn't the feedback that you're about to get to, opposite in phase, gently damp that supersonic bell and cut off the power suckage of oscillation?
Right again.
I used the wrong lobe of my brain cell by implying that we use 2nd or 3rd order filters to roll off frequency above 20khz. The ringing at thirty in the test is caused by deliberately massaging the windings of the OT with energy at that frequency. I would still say that those frequencies included in program material are rare to say the least. I no longer use vinyl but I'm convinced there are NO supersonic transients coming off bumps on plastic disks. In the beginning (back to mysticism) when Deutsche Gramophone gave us DDD CDs there was plenty of "out of band" energy, but it came from poor recording process' not from fundamental energy. Those were awful recordings. Unlistenable.
This one immobilizes me. Shouldn't be necessary?? To set it that low I mean? I think what I'm not getting is where the supersonic energy is coming from. Is it a product of the circuitry in the chain from performance to playback, or does it exist in the performance? I remember when my son was a tot with the colic. He had a scream that did something quite extraordinary in my cochlea that never happens any more for one thing, but was I hearing harmonics caused by room acoustics or a psychological reaction designed in to the me by God to come to the aid of my distressed offspring?
Why do some of us prefer valve reproduction over silicon? I think this mix of both technologies is wild and am really anxious to get started on one but I haven't seen in my reading so far any practical way around the impedance transformation process, and yet it's the source of all the problems. What ya'll are talking about is better controlling the various deltas of circuit design. Forget OTL for a moment, it's too distracting.
Looking at Atwood's list of transformers (if we accept that only 60khz and above is good) it's apparent that well over half of the old time, accepted winners need to be thrown out (or used for filament duty) because of bandwidth and ringing problems. Atwood also claims that high frequency ringing is not affected by power input. Really? If I take a bell and strike it softly with a rubber thingy it rings a little, when I hit it with a 16 pound hammer it rings alot. It's counter intuitive. I guess a lot of stuff is, but perhaps we're not looking at the right parameter? When the 30K artifact is turned back, out of phase, and re injected it doesn't kill the ringing because the ringing is not phase related. It's a mechanical characteristic of the magnetic interactions of the transformer. Ahhhhh. So that's the rub. We are correcting early in the circuit for something that isn't found earlier in the circuit and can't be corrected by feedback because it's mechanical not LCR related. Then we amplify it. Feedback exacerbates the ringing. Hmmmmm
I got up for a midnight snack with the dog and I'm going in circles. I think I'll go back to bed and re-think this tomorrow.
Yes.
Exactly the opposite. My advice, since you're clearly a bright guy who doesn't need to be spoon fed (I'm not, for once trying to be sarcastic), is to read up on feedback, stability, and pole placement. Norman Crowhurst had some EXCELLENT charts in "Understanding Hi Fi Circuits" showing where the dominant pole had to be with respect to the next rolloffs for stability depending on the number and order of poles. Try to dig that book up if you can- it's old, but it's really good on the basics.
And you should have Morgan Jones's "Valve Amplifiers," 3rd edition. Morgan is an advocate of using solid state to extract the best performance out of tube circuits and does detailed and rigorous justification, backed up with measurement. His treatment of feedback and stability is different than Crowhurst's, but also easily understandable and useful.
The 30kHz ringing on top of 10kHz square wave is intrinsic to the output transformer and I'm not sure if it is feasible to "filter" it out. It's neither audible nor power-robbing. It's just there and is an indication of a design that has poor transient response and will probably have issues elsewhere.
John
The other way to put this is, "You have a second order filter at ~30kHz and a high Q." Stabilizing that inside a feedback loop is going to take some heroic measures and something will have to give- HF distortion or frequency response.
One has to keep in mind that John's tests, though important, useful, and correct, are very limited.
I recently ordered from Mouser and found them very reasonable if you buy 50 quids worth of stuff. They seem to have a warehouse in Germany and the stuff came next day🙂
Cheers Matt.
Ok, just bought both books. That should keep me busy. I'm also in a late in life masters program right now (Theology) so this is going to cut into my build time I guess. Thing is, Theology is pretty dry you know? The electronics my derail things. I'm easily distracted.
I'll start accumulating parts for the Baby Huey EL-84 version with Triode Electronics' SCA-35 tranny's. I'll use my own OEM power transformer.
One other thing. I've been spending like a madman lately on retirement income but I need a function generator. There is one at AES for $160 or I can play the lotto and Ebay something like a Fluke waveform generator.
Your thoughts?
Thanks
I'll start accumulating parts for the Baby Huey EL-84 version with Triode Electronics' SCA-35 tranny's. I'll use my own OEM power transformer.
One other thing. I've been spending like a madman lately on retirement income but I need a function generator. There is one at AES for $160 or I can play the lotto and Ebay something like a Fluke waveform generator.
Your thoughts?
Thanks
I'm not considering response from dc to light, seriously do an RSS on 4 cascaded stages with responses of 20kHz @ -1dB and guess what - the response of those 4 stages will be down a dB at only 7kHz.. That is why you design for higher bandwidth. I typically am happier with 200kHz or better per stage, but that is usually not attainable with choke or transformer loading. Usually the output stage in a tube amplifier will end up having the worst HF bandwidth due to the transformer. An argument against interstages incidentally unless they are very good.
Note in my earlier post (#34) I swapped the RSS, use 1/F case for HF roll off and just the RSS of F for LF when calculating the cascaded response. Sorry, the dyslexia in action! 😀
Last edited:
"When the 30K artifact is turned back, out of phase, and re injected it doesn't kill the ringing because the ringing is not phase related. It's a mechanical characteristic of the magnetic interactions of the transformer. Ahhhhh. So that's the rub. We are correcting early in the circuit for something that isn't found earlier in the circuit and can't be corrected by feedback because it's mechanical not LCR related. Then we amplify it. Feedback exacerbates the ringing. Hmmmmm"
Not mechanical, it is a function of an electrical resonance (frequently several in fact) in the windings for the transformer itself. Audio transformers are not electrically, nor mechanically perfect devices, and have stray capacitances and leakage inductances that conspire to create unintended resonant circuits..
Mechanical issues with OPTS relate to magnetostriction amongst other things which is what causes a good transformer to sing at higher power levels. Poor construction technique call result in singing at lower levels from loose windings, insulation, etc. as well.
Not mechanical, it is a function of an electrical resonance (frequently several in fact) in the windings for the transformer itself. Audio transformers are not electrically, nor mechanically perfect devices, and have stray capacitances and leakage inductances that conspire to create unintended resonant circuits..
Mechanical issues with OPTS relate to magnetostriction amongst other things which is what causes a good transformer to sing at higher power levels. Poor construction technique call result in singing at lower levels from loose windings, insulation, etc. as well.
The more I thought about it in the wee hours last night the more that becomes clear. At VHF or UHF for instance, a wave guide is the only way to travel because of all that stray stuff. In the early data networks they used some really expensive waveguide all over the attic of buildings to spread the data. Same deal.
In Atwoods paper it seems that some of the ringing is self damping anyway, so that's a good thing. But with music, it's just as likely to be started again shortly anyway.
- Status
- Not open for further replies.