richwalters said:
Offhand, I can't remember the exact reason for this, but I do know there is an advantage an NPN pass transistor has over a PNP (but you are correct, the lower dissipation of the PNP is alluring).
I think it had something to do with output impedance of the transistors, that NPN was superior. In fact, there are those that claim the PNP low dropout regulators have inferior performance to the standard. I would have to dig a little to come up with this article, but it exists out there somewhere. Could it be Morgan Jones, anyone? Possibly National Semi?
Not that I don't like your regulator - I am a big fan of discrete regulators such as yours. Even better are they with a current source "preregulator", for additional ripple rejection. Little more tricky with a low dropout design, tho.
A little extra output impedance is no big deal here- in fact, for series heaters, a CCS can be profitably used. See Morgan Jones, for example. The important thing (IME) is getting out the common-mode noise, which a single-ended regulator won't do. Note that I used bipolar regulators in my preamp just for that reason. The new preamp of mine (AC heaters) uses balanced bypassing for that reason.
rdf, I've certainly had more than one preamp which hummed without that cap. It can be very effective at low frequencies. For example, my phono preamp uses a 220k->47k as the heater lifting string. The bypass cap is 10uF (which is a pretty small electrolytic). HF nasties tend to be common mode (most frequently, diode noise from the B+ supply coupled into heater windings), so are best bypassed in a common-mode manner.
rdf, I've certainly had more than one preamp which hummed without that cap. It can be very effective at low frequencies. For example, my phono preamp uses a 220k->47k as the heater lifting string. The bypass cap is 10uF (which is a pretty small electrolytic). HF nasties tend to be common mode (most frequently, diode noise from the B+ supply coupled into heater windings), so are best bypassed in a common-mode manner.
zigzagflux said:
M J's book does revel on using the 317 types with the inherent 1.25V in/out drop. I'm a bit suprised on this tack of simplisticity, as I feel he hadn't tackled the issue head on. This 1.25V is a thorn in the side when only a few volts are required at high current. Do the math and the dissipation rockets when current is req. In the low drop out discrete circuit shown, the rms current in the supply smoothing cap isn't small either and a tag type is best here.
r:-
SY said:
Any thoughts on the mechanism? At low frequencies an ancient Tung-Sol paper described the coupling between heater and cathode as due to 'emission' and explicitly ruled out capacitive, supported by typical KH capacitance numbers. The floating centre winding of a filament transformer shouldn't have anything but leakage reference to the other windings and therefore at most a tiny voltage to the centre taps of other secondary windings. It’s seems that the act of DC biasing AC heaters potentially substitutes one problem for another, depending on the quality of the PS transformer and topology.
I have just converted the front end of my phono preamp to AC from DC. Its based on Thorstens El Cheapo circuit and I am running a Goldring 1012 into it.
There is a very slight hum from one channel if you put your ear right up to the speaker (at least 100db sensitivity). Can't remember exactly if there was that hum there originally. Anyway just shows it can be done and it in no way diminishes the sound quality.
Shoog
There is a very slight hum from one channel if you put your ear right up to the speaker (at least 100db sensitivity). Can't remember exactly if there was that hum there originally. Anyway just shows it can be done and it in no way diminishes the sound quality.
Shoog
rdf, I don't know for sure, but if I were to guess, it was because of some LF or DC leakage between heater and cathode at high temperature. I note that Morgan Jones goes one better and recommends emitter followers for the biasing.
Now you have me wondering. The cap should do absolutely nothing regarding hum injection caused by the filament voltage swinging either side of the centre tap since by definition that voltage doesn't appear at the centre tap. If a cap is reducing hum, it would seem it can only be due a common mode voltage on the centre tap itself modulating the heater/cathode-grid voltage. Could it be a case of transformer imperfections injecting hum directly into the cathode circuit? Emitter followers would certainly take care of that. It also suggests fixed bias are less vulnerable and since they're almost all I ever use it would explain why I don't notice the issue. It further suggests a big whacking resistor between centre tap and voltage divider will also be effective. Something new to try. 🙂
Uneven wear
If this were significant, we would see it prominently in Diodes.
Matters not if a directly heated diode's filament is AC or DC.
Assuming heater and plate are always phased the same.
The conduction will always happen from the same "end".
Warning, the following may be complete nonsense:
1) Emission is based more on temperature than voltage.
2) Thermionic effect cools whatever part emits the most.
3) Emissive surfaces should tend to equalize temperature.
If this were significant, we would see it prominently in Diodes.
Matters not if a directly heated diode's filament is AC or DC.
Assuming heater and plate are always phased the same.
The conduction will always happen from the same "end".
Warning, the following may be complete nonsense:
1) Emission is based more on temperature than voltage.
2) Thermionic effect cools whatever part emits the most.
3) Emissive surfaces should tend to equalize temperature.
rdf said:
I dunnow wassupwiddat either. Nevertheless, he did say that the bypass capacitor did reduce the hum. Of course, without actual measurements, I suppose it could always be more psychoacoustic than electronic. HK capacitance is usually so low that you don't need to begin worrying about it until your signal frequency approaches VHF, or, perhaps, the 10M band.
"The floating centre winding of a filament transformer shouldn't have anything but leakage reference to the other windings and therefore at most a tiny voltage to the centre taps of other secondary windings."
Key word there: "shouldn't". These secondaries are always wound with 1/2 the turns, pull out the wire to make a center tap, and wind the other 1/2 turns. However, the coupling coefficient is never equal to 100%, and so that "center" tap is not the same point as the electrical neutral, close, but not identical. You can see this most easily if you set up for full wave rectification, but don't add the filter capacitor, O'scoping the raw FW output will show that the wave peaks aren't hitting the same Vp point. It's especially obvious with some cheap, wall wart, type xfmrs. so there wil be some residual AC at the center tap. Now, whether or not there's enough to significantly raise the hum level in the rest of the circuit is a whole 'nother story.
As for my own experience with this, for one project that included a hybrid DC supply, I was planning to add overvoltage protect Zeners to the cathodes of a follower stage that was connected to the negative rail. Since that rail comes up faster than a VT can warm up, this puts a severe overvoltage across the heater and cathode. Not having that Zener, I simply floated the whole 6.3V heater winding in order to carry out some preliminary tests. I was expecting a lot of hum, but got very little. So little that you can't tell the thing is running except by the glow of the tubes and the plate currrent meter. So I figured t'hellwiddit and never added that Zener, and never bothered to ground the 6.3V center tap. All heaters still float, and there still isn't any significant hum at all.
So who knows? Maybe it's one of those "voodoo" aspects that make audio design such a "black art"? 😕
Rectifying the imbalance in a filament secondary would appear to be an easy hurdle, snip the centre tap and create a new virtual centre reference with a couple resistors matched to any arbitrary precison. When floated the filament balance should (that word again) be perfect to the limits of the transformer's parasitics. Still, according to any references I've seen none of this should matter with DC-biased filaments. Biasing supposedly repels the emission which generates hum.
SY
Add another 50 years to the experience of yourself and ? I started with directly heated tubes like 1R5, 1T4, 3S4 etc - all 1,4V filament. How come they never failed? You could not spell AC then - batteries, gentlemen, all the time!
But seriously, no, in all the valve manuals I read (and there was quite a few), no mention was ever made of disadvantages of dc on heaters. In fact, I would venture that as far as physics are concerned, AC reversing, even at 100 (120) times/sec, puts some strain on filaments. A very long time ago, but that has been shown in physics labs where I worked. Largely academic, but still.
But as for necessity: My early pre-amps (down to 2mV input RIAAs) used AC ... but there was the merest hint of hum, and one shielded. 9000mVp next to an input sensitivity of 0,3mV (RIAA amplification at 50 Hz) - it did need some scheming. At the time with only selenium solid state rectifiers available ... need I go on?
But nowadays as said several times, compared to the cost of a complete tube amplifier, there cannot be too much problems with a dc supply, surely! I am talking about pre-amps (if this has an effect in your power amplifier - to be kind, go back to construction school, friend!) And I see mainly 6V supplies mentioned; what happened to connecting heaters in series?
The rest has been said; only to stress that "raw" dc is no dc at all, simply unreversed polarity. The supply has to be properly filtered or the switching transients will make matters worse rather than better. If only a 6V winding, it could be used in a doubling fashion which will give enough headroom for a decent enough filter. Here in RSA it is cheaper/more convenient these days to have power transformers wound, so no problem there to get any heater voltage required and use a common LM.... regulator as filter.
I am not addressing DHFs; that is a concern for those who find pleasure in that technology.
Oh yes, and BAS H.! Thanks for that link of yours; the first time I have seen that! (At my age, still learning ...)
Add another 50 years to the experience of yourself and ? I started with directly heated tubes like 1R5, 1T4, 3S4 etc - all 1,4V filament. How come they never failed? You could not spell AC then - batteries, gentlemen, all the time!
But seriously, no, in all the valve manuals I read (and there was quite a few), no mention was ever made of disadvantages of dc on heaters. In fact, I would venture that as far as physics are concerned, AC reversing, even at 100 (120) times/sec, puts some strain on filaments. A very long time ago, but that has been shown in physics labs where I worked. Largely academic, but still.
But as for necessity: My early pre-amps (down to 2mV input RIAAs) used AC ... but there was the merest hint of hum, and one shielded. 9000mVp next to an input sensitivity of 0,3mV (RIAA amplification at 50 Hz) - it did need some scheming. At the time with only selenium solid state rectifiers available ... need I go on?
But nowadays as said several times, compared to the cost of a complete tube amplifier, there cannot be too much problems with a dc supply, surely! I am talking about pre-amps (if this has an effect in your power amplifier - to be kind, go back to construction school, friend!) And I see mainly 6V supplies mentioned; what happened to connecting heaters in series?
The rest has been said; only to stress that "raw" dc is no dc at all, simply unreversed polarity. The supply has to be properly filtered or the switching transients will make matters worse rather than better. If only a 6V winding, it could be used in a doubling fashion which will give enough headroom for a decent enough filter. Here in RSA it is cheaper/more convenient these days to have power transformers wound, so no problem there to get any heater voltage required and use a common LM.... regulator as filter.
I am not addressing DHFs; that is a concern for those who find pleasure in that technology.
Oh yes, and BAS H.! Thanks for that link of yours; the first time I have seen that! (At my age, still learning ...)
Yvesm said:
I have no experience with using small chokes there nor shunted transformers for supplies. A CMC may help as mentioned.
Avoiding a large input capacitor and adding some series impedance should help. Id just use a small wirewound resistor here.
Example:
18V CT 2A transformer-full wave Schottky rectifier-1R-12mF-1R-12mF-LM317. 6.3V @ 1.2A.
Is it possible that high frequency crud is bleeding from the AC heater winding and making the valve oscillate. Oscillation often manifests as hum pickup. The small value cap takes out the high frequency hum and so the tube isn't oscillating and so doesn't pick up the hum.
Just a thought.
Shoog
Just a thought.
Shoog
DC bias on heater wires ?
How about :
http://www.netads.com/~meo/Guitar/Amps/Kalamazoo/Mods/hum.html
?
cheers,
Empee
How about :
http://www.netads.com/~meo/Guitar/Amps/Kalamazoo/Mods/hum.html
?
cheers,
Empee
Re: DC bias on heater wires ?
Yes, though I don't take the B+ from the first cap in the power supply, I just take it from the final B+. I don't think there is much danger of feeding hum back into the B+, as you are using pretty high value resistors for the divider. The difference between the two is likely to be minimal, so just do what is convenient.
Sheldon
Empee said:
Yes, though I don't take the B+ from the first cap in the power supply, I just take it from the final B+. I don't think there is much danger of feeding hum back into the B+, as you are using pretty high value resistors for the divider. The difference between the two is likely to be minimal, so just do what is convenient.
Sheldon
I made the following comment earlier in this thread:
I have been thinking about it and don't see how DC can damage the filaments on a DHT even. Oxide coated filaments(or cathodes) are designed to operate with a space charge. They emit electrons at an average rate and the space charge supplies electrons during current peaks. Depleting the space charge damages the oxide coating when the filament is called upon to supply current peaks directly. This causes damage regardless of whether you use AC or DC. If there is a space charge(the tube is being operated within its maximum current limits), the electrons should come from it, not the filament. Using AC or DC will not change how many electrons get emitted, because emission is a function of temperature, which in turn is a function of how much power is dissipated in the cathode.
That leaves the hot spot theory which doesn't make any sense, since hot spots would be caused by inconsistencies in the filament and independent of the type of power applied.
That leaves only the electro-migration possibility, which is a remote one in my opinion.
I welcome any real evidence to the contrary.
I have been thinking about it and don't see how DC can damage the filaments on a DHT even. Oxide coated filaments(or cathodes) are designed to operate with a space charge. They emit electrons at an average rate and the space charge supplies electrons during current peaks. Depleting the space charge damages the oxide coating when the filament is called upon to supply current peaks directly. This causes damage regardless of whether you use AC or DC. If there is a space charge(the tube is being operated within its maximum current limits), the electrons should come from it, not the filament. Using AC or DC will not change how many electrons get emitted, because emission is a function of temperature, which in turn is a function of how much power is dissipated in the cathode.
That leaves the hot spot theory which doesn't make any sense, since hot spots would be caused by inconsistencies in the filament and independent of the type of power applied.
That leaves only the electro-migration possibility, which is a remote one in my opinion.
I welcome any real evidence to the contrary.
Which DC supply?
I can get one of these DC power supply unit for around $30 (USD).
http://www.meanwellusa.com/seriessearch.html#ac/dc
Can someone please give me some feedback on them? I am using DH filament 2A3 SET amp with AC heater (hum pot) for my 105 db horn driver and reducing hum is somewhat of a challenge for me (novice tube amp builder). It's not loud but I'd like to reduce it further if I can. Thanks for your forthcoming help.
I can get one of these DC power supply unit for around $30 (USD).
http://www.meanwellusa.com/seriessearch.html#ac/dc
Can someone please give me some feedback on them? I am using DH filament 2A3 SET amp with AC heater (hum pot) for my 105 db horn driver and reducing hum is somewhat of a challenge for me (novice tube amp builder). It's not loud but I'd like to reduce it further if I can. Thanks for your forthcoming help.
The first thing I'd do is make sure that the filament supply is in fact the issue. Just rig up some D cells to give approximately the right voltage (2 D cells and an inline resistor), disconnect the filament supply and clip in the battery set up. If your hum is gone then a good DC supply will eliminate it too.
Pete Millet published this one, so a switcher can work: http://www.pmillett.com/hf_fil.htm
Tentlabs has a good analog supply (I've use it them): http://www.tentlabs.com/Products/Tubeamp/Tubefilament/index.html
There is another kit available (don't remember the details) which requires a higher input voltage than the tentlabs piece, but may be less expensive.
Sheldon
Pete Millet published this one, so a switcher can work: http://www.pmillett.com/hf_fil.htm
Tentlabs has a good analog supply (I've use it them): http://www.tentlabs.com/Products/Tubeamp/Tubefilament/index.html
There is another kit available (don't remember the details) which requires a higher input voltage than the tentlabs piece, but may be less expensive.
Sheldon
I've put all the DC and signal tracking on my PCB, and deliberately kept the heaters offboard, so I can twist, braid and even ferrite (is this overkill?!) the heater leads if needs be.
Keeping it apart should help!!
L
Keeping it apart should help!!
L
Does DC power cable need to be shielded when connecting from external power source?Psychobiker said:
- Status
- Not open for further replies.