Ah, it has happened, daniel has finally migrated over to the tube forum. Time to stock up on bottled water and non-perishables. 🙂
Thank you for the suggestions. The polystyrene and foil for input cap suggestion is especially suitable.
There's a problem. . .
Tube input 10k~12k
The tube's input cap probably needs to be fairly large due to the 10k potentiometer (to play with a level response from the largest variety of computer sound cards) and also because the input cap may become in series with the output caps of a wide random selection of computer sound card.
Tube output 33k~100k
The power amp has a 100k input, but it can be set as low as 33k, and in this case, that would actually be a convenience. I can try for smaller figures, but it won't go all the way down to 10k and still play well.
Put the 10 KOhm level control in front of the I/P cap. and keep the 100 KOhm grid to ground value. Then, a low WVDC 0.33 μF. part will work well. The I/P cap. connects between the wiper of the pot. and the grid circuitry.
The taller the I/P impedance of the downstream load, the smaller the value of the O/P coupling caps. can be. Since a 33 KOhm power amp I/P impedance is convenient, self bypassing 1.0 μF. PPFX MultiCaps will do nicely.
It's a good idea to mix cap. dielectric materials up. Doing so avoids having any particular plastic film dominate the circuitry's sonic signature.
Hi Dan,
Thanks for your information; I'm afraid I still don't understand half of the things you say, but that's probably due to my ignorance. Nevertheless, it seems to me that building the circuit Eli proposed might just accomplish what you need; indeed, it is not intended for accurate reproduction, but will certainly 'color' the sound.
I'm looking forward to the example schematic (the 6n3 one); perhaps clears things up for me!
Thanks for your information; I'm afraid I still don't understand half of the things you say, but that's probably due to my ignorance. Nevertheless, it seems to me that building the circuit Eli proposed might just accomplish what you need; indeed, it is not intended for accurate reproduction, but will certainly 'color' the sound.
I'm looking forward to the example schematic (the 6n3 one); perhaps clears things up for me!
Okay, well, before I lose track, its about time for a shopping list. 🙂
If I understand tube heat dissipation correctly, a ceramic with metal flange tube socket is needed to help get some heat out of this hot little tube. Is there a standard name for its size socket?
Resistors:
What resistors exceed 1/4w?
2 of 3.3k
2 of 1k (big carbon)
2 of 5k trimpot
2 of 100k metal film (input load)
10k Stereo pot (or a balance volume 2-pot config with similar load)
Diode
4 of 5mm red led with clear lens
Input Capacitor
polystyrene film and foil, perhaps 4.7uF//0.47uf
What minimum voltage rating is necessary for input capacitor?
Output Capacitor
2 of 1.0 μF. PPFX MultiCap
What minimum voltage rating is necessary for output capacitor?
Is this shopping list correct?
If I understand tube heat dissipation correctly, a ceramic with metal flange tube socket is needed to help get some heat out of this hot little tube. Is there a standard name for its size socket?
Resistors:
What resistors exceed 1/4w?
2 of 3.3k
2 of 1k (big carbon)
2 of 5k trimpot
2 of 100k metal film (input load)
10k Stereo pot (or a balance volume 2-pot config with similar load)
Diode
4 of 5mm red led with clear lens
Input Capacitor
polystyrene film and foil, perhaps 4.7uF//0.47uf
What minimum voltage rating is necessary for input capacitor?
Output Capacitor
2 of 1.0 μF. PPFX MultiCap
What minimum voltage rating is necessary for output capacitor?
Is this shopping list correct?
I recommend using 1/2W resistors or larger. 1/4W are nice for low voltage applications, but they always make me anxious about longevity in tube circuits.
In any case, assuming a 9mA anode current (assuming Eli's guesstimate is correct; I didn't check!), the 3.3k and 5k (var) resistors should be rated for at least 0.5W. So a regular trimpot might not work. I would consider splitting the 5k up into two parts: a fixed resistor of something like 3.5k or 3.8k and a 1.2-1.5k pot. That way, about two thirds of the dissipation will occur on the 3.5-3.8k resistor (which will be the 'lower' one, which is connected to ground) and you can expect less issues with dissipation in the upper half of the divider (i.e. the 1.2-1.5k pot). You could probably even get away with using a 500Ohm pot and a fixed 4.5k resistor.
You don't necessarily need a clear lens led, unless you want one for aesthetic reasons. I think a 5mm opaque red led is about the cheapest you can get, and it's perfectly adequate for this application (just helped you save $0.20 which you can now go and spend on expensive coupling caps; all of it! 😉).
As to the coupling caps voltage rating: 100V should do the trick. Assuming an input impedance of 33k (as per Eli's post), 0.47uF will be large enough (corner frequency 10Hz), but you could opt for 1uF to be on the safe side.
In any case, assuming a 9mA anode current (assuming Eli's guesstimate is correct; I didn't check!), the 3.3k and 5k (var) resistors should be rated for at least 0.5W. So a regular trimpot might not work. I would consider splitting the 5k up into two parts: a fixed resistor of something like 3.5k or 3.8k and a 1.2-1.5k pot. That way, about two thirds of the dissipation will occur on the 3.5-3.8k resistor (which will be the 'lower' one, which is connected to ground) and you can expect less issues with dissipation in the upper half of the divider (i.e. the 1.2-1.5k pot). You could probably even get away with using a 500Ohm pot and a fixed 4.5k resistor.
You don't necessarily need a clear lens led, unless you want one for aesthetic reasons. I think a 5mm opaque red led is about the cheapest you can get, and it's perfectly adequate for this application (just helped you save $0.20 which you can now go and spend on expensive coupling caps; all of it! 😉).
As to the coupling caps voltage rating: 100V should do the trick. Assuming an input impedance of 33k (as per Eli's post), 0.47uF will be large enough (corner frequency 10Hz), but you could opt for 1uF to be on the safe side.
Clears things up? Exactly! This circuit's crisp clear treble enhancement will clear up a Sony HD radio, MP4 and also AAC movie sound. It will clean up any dull, lackluster source.
It can also be used at the input to a dull amplifier from Best Buy, and together they will sparkle.
It has an error. You'll have to cut the gain by 40% or else it will clip. Fortunately, it comes with a volume pot for this.
I wonder if one couldn't simply connect a 28vac transformer to fix the clipping issue?
This circuit requires an extra cheap 6n3. It won't make treble enhancement with GE 5670, because GE 5670 will increase midband instead, and I cannot find a practical use for that.
But, if you've got a super cheap 6n3 and a terrible Sony HD radio. . . this circuit posted literally will clear things up for you: 🙂
Attachments
Nope; it would squeeze the tube (i.e. drastically reduce its headroom) and only worsen the problem.
The circuit you posted probably calls for a 6n3p; the 6n3 is a rectifier tube. Due to the low voltage and relatively small load resistance, I expect this circuit to add the same sort of 'coloration' (i.e. some harmonic distortion) just like Eli's proposal. You could consider using your 5670 as a buffer in the same way and implement it as a cathode follower, making for low output impedance and limited gain. It won't work as a phase splitter this way, of course.
Well, there's subjectivity for you: in my book, this circuit won't 'clear' anything up; it will 'soften', make for a 'warmer sound'. That's exactly my problem with subjective terminology. To you, it means something different than it does to me, while both of us may think that we have a common understanding...
You want a 9 pin miniature, AKA Noval, socket. The socket is for ease of replacement. The envelope is convection cooled. Make sure you have good air flow.
You omitted PSU parts, including the power trafo I previously mentioned. A pair of 600 PIV Schottky diodes in the full wave center tapped (FWCT) B+ supply would be very nice, but 2X UF4007s will do, especially if they are snubbed by 10 nF. capacitors. You need a filter network. At this low current, CRC is acceptable, but CLC would be better.
I'd take no prisoners and use 250 WVDC rated parts in the O/P coupling positions. For the I/P caps., 100 WVDC parts will be quite adequate.
Remember, the 10 KOhm volume control is cap. coupled to the 100 KOhm grid to ground resistor. The relevant high pass pole is formed by the I/P cap. and the 100 KOhm part. Therefore, 470 nF., without a bypass, is more than good enough. Speaking of the stereo volume control, DigiKey stock # KKA1031S28-ND is very nice and does not carry an outrageous price tag.
Mastodon made a good suggestion in combining a fixed resistance, along with a trim pot. in the cathode leg. Use a 1/2 W. rated 2.7 KOhm fixed value part. and a 2.5 KOhm variable part.
Metal film parts are appropriate in the grid to ground positions, for low noise. Grid stoppers should be Carbon composition, as only that construction is both non-metallic and non-inductive. Carbon film resistors are fine, elsewhere. Again, take no prisoners and use 1 W. rated parts for the 3.3 KOhm plate loads.
Don't want warm!
The right effect for cancellation of distortion in MP4 types would make an Increased Amplitude at 8khz and Up (high TREBLE INCREASED) due to triode type harmonic distortion. The 6n3p example does this more elegantly than usual.
The only forms of warm that are acceptable are those which are unavoidable:
1). The temperature of the diodes, tube, heater regulator, the amplifier heatsink and the elements of the capacitive multiplier.
2). The MP4 type input signal and Sony's HD radio's dull AF section, and/or a worn out tape.
Please don't think of the needs with a CD player source--that is opposite of what is needed.
Instead, computer MP4 and similar sound just needs a greater amount of richer treble.
It certainly doesn't need the intelligibility reduction of an unnecessarily warm mid-bass boost.
The right effect for cancellation of distortion in MP4 types would make an Increased Amplitude at 8khz and Up (high TREBLE INCREASED) due to triode type harmonic distortion. The 6n3p example does this more elegantly than usual.
The only forms of warm that are acceptable are those which are unavoidable:
1). The temperature of the diodes, tube, heater regulator, the amplifier heatsink and the elements of the capacitive multiplier.
2). The MP4 type input signal and Sony's HD radio's dull AF section, and/or a worn out tape.
Please don't think of the needs with a CD player source--that is opposite of what is needed.
Instead, computer MP4 and similar sound just needs a greater amount of richer treble.
It certainly doesn't need the intelligibility reduction of an unnecessarily warm mid-bass boost.
Uhm, this seems a bit odd; could you provide a reference for this? In my mind,
'Increased Amplitude at 8khz and Up' is effectively a high-pass filter. This has little to do with even order ('triode' type) harmonic distortion, which to the best of my knowledge is present throughout the audio band in any reasonably well-designed amplifier. If you want to lift anything above 8kHz, I suggest implementing a simple RC high pass filter.
Note that no amplifier can restore any high-frequency data that is absent from the signal source, which is often seen as the major drawback of digitally compressed audio, even though some 'MP4' compression methods are in effect lossless and therefore offer precisely the same audio quality as the original source (note that the DST compression used for SACD is also part of the Mpeg-4 complex).
I still don't buy that, since several high-end audio cards (still within the price range of the average consumer) offer more or less perfectly linear frequency responses up to 20kHz. But if you feel it's necessary, then go ahead and implement a high-pass filter. But personally, I would recommend investing in a decent sound card (and not use the onboard solution in your off-the-shelf PC or laptop, which can be crappy indeed) before trying to solve quality issues further downstream.
Having said that, I'll promise not to nag about this any further and just try to help you out where I can 😉
Thank you very much for all of your help!
Do you have any specific preferred model and brand of diodes that you'd like to see used?
Fast diodes can vary greatly when used on contraindicated application, such as linear power supplies (so I wouldn't put any into a linear supply unless I knew specifically which to purchase). The noise of fast switch diodes is loud enough that their voltage peaks (when noise is rectified by caps of the power supply) can bother power amps and can reduce the far fringe reception of tuners. So, I'd either have to use a specific known good, or use an ordinary diode (with an optional polyester cap helper). 🙂
On power supply, I'm trying to avoid "warm sound" in unintentional locations, such as power circuit. Therefore, I'd like to go for a regulated heater and a capacitive multiplier as well. What voltage for heater?
Thanks again. 😀
6.3V for the heaters. I find a regular 7805 with two regular silicon diodes (e.g. 4001) between ref and ground works nicely for one or two tubes (< 1A total current) and is a super-easy DC heater implementation. Maybe it's overkill, but again, it's easy. Make sure you allow for a couple of volts of headroom for the regulator, and implement cooling as necessary (small heat sink will do, or even no heat sink at all if you have just one tube to heat and not too much voltage drop on the regulator).
Last edited:
First, Thank you for your help!
Of two preferential sound cards:
M-Audio Audiophile 192
Creative Labs X-FI Music aka X-FI Gamer
The Creative Labs product contains Harmonic Treble Enhancement re-branded "Crystalizer" and their example misses the mark because its too low pitched and increases amplitude at 4k, which is a mistake according to Fletcher Munson. That error limits effectiveness because its barely a break-even prospect (their fix is about as bad as the original problem).
Had their Harmonic Treble Enhancement actually worked right, then it would be able to fake "CD Sound" from MP4, Transcodes, AAC, Itunes, 96k~160k MP3, Sony's HD Radio, and worn out tapes.
The increase of intelligibility from Harmonic Treble Enhancement is also frequently used in prosound. Some examples are much more tunable than others. Its "common knowledge" that an equalizer doesn't substitute.
By any chance are there pre-existing designs that are known to be "extremely helpful to liven up the sound of a worn tape"?
The VCR, Audio Tapes, and other sorts of worn out tapes, can suffer from the same compressor + damage as typical computer sound.
So, perhaps a pre-existing design, for tape, may also work on computer sound?
P.S.
If everyone agreed with me, then I'd never learn anything.
P.P.S.
Sorry to be creating an amp that's unsuited for use with audiophile CD player.
For these, there's a free solution. Its 2x pepsi can bottom, then cut to shape, flattened, and with the edges fangle cut (feathered) so that heat can be conducted into the air. Its a bit sharp, but its a nice aluminum deal that costs nothing. 🙂
Yes, that's not completely gap-less, but there's also GC Waldham Type 44 Thermal Compound. 😀
P.S.
It seems odd if one were to spend less on the power supply than on the AF coupling caps. What do you think of that?
Infineon Silicon carbide Schottky diodes (600 PIV) get my nod. They are "noiseless" (no snubber needed) and not especially costly. Mouser stock # 726-IDT02S60C covers this bet. ALL PN junction diodes exhibit a reverse recovery spike, as a consequence of minority carrier injection. That spike is the nasty switching noise. Being majority carrier only devices, Schottky diodes don't exhibit the unwelcome switching spike.
This is a "line" level application. AC heating handled correctly should be highly satisfactory. The suggested power trafo has a 1 A./6.3 V. filament winding. "Synthesize" a center tap for the filament winding by installing a series connected pair of 3 W. rated 5 Ω metal oxide resistors across it. Construct a resistive voltage divider across the B + supply such that the tap is in the 50 to 70 V. range. Use resistor values that limit current to less than 1 mA. Connect the tap of the voltage divider to the synthetic heater supply CT. The synthetic CT gets connected to ground via a 15 μF./100 WVDC 'lytic.
Thank you very much!
This fast switch is different than my previous experience with On-Semi's MUR series that throw out enough noise to charge up caps with their peaks, making audio amplifiers support more midrange than bass and also practically block all distant stations from the tuner.
Line level? When driven by the tube phase splitter, the power amplifiers obtain an effective gain that's adjustable in the range of 74 to 94. It seems that this is about 300% different than line level. Yes, its a lot of gain, but that will effectively exaggerate dynamics and yield some more involving imaging.
On the power supply, thank you again!
Is there perhaps a handy sketch or a shopping list for the power supply?
Well, its time to ask about a shopping list for the power supply section.
Is there a handy example? What all is preferred?
It is important to combine purchases, when possible. Otherwise, shipping charges can eat you alive. Since the power trafo comes from Allied Electronics, it makes sense to order the CLC filter's inductor from them too. $7.93 buys stock # 967-1009 which is a 10 H./50 mA. Triad C-3X.
Contact Jim McShane about 'lytics for the project. Jim may be able to help with other parts, including the 600 PIV Schottky diodes, too. Point the man at this thread. Jim's shipping charge policy is favorable and his prices are otherwise quite fair.
- Status
- Not open for further replies.