I picked up a bunch of Jan6418's a little while ago, with the intent of doing a line stage or an amp front end (mainly for fun/curiosity; triode front end and sand power The tubes seem to be pretty happy at low voltages too, so it seemed an easier fit). There doesn't seem to be a lot of amp type circuits out there using such devices, so did you have any joy in applying them to a circuit?
Edit - yes I have seen the Oatley circuits; I meant anything else out there.
Last edited:
This is the 21st Century. If we need gain, we have chips for that, any gain you need, 19 cents.
The Application Notes show "some" voltage gain. But the output current is VERY small, and to get there they run some grid current. So there's already two halves of a 19 cent chip just to make it generally useful.
If you are a e-guitarist, "gain" means something else: distortion. The NuTube is "a tube", not the same as 12AX7 (by a long shot), but will give a "bent up sound" when pushed hard, and different from the bentness of diode-clippers or other popular solid-state distortions.
As DF96 says-- the main thing is that VFDs for car radio and microwaves have gone out of fashion, sales near zero. Noritake has cleared out a vast VFD factory for other work. But KORG asked for one machine to be left running to make NuTubes to sell for 10X what they cost to make (and IMHO, 100X what they are worth).
You know, I never really thought about the musical instrument side of things, which seems pretty dumb in hind sight.
I was of the belief that, at least for old guitar amplifiers, most of the character came more from the under sized power supply and output transformers? That transistors and valves don't clip all that differently, unless incorrectly biased if measured before the load or transformer? so this little chip doesn't make sense to me if it acts as basically a line follower? Like those cheap eBay hybrid valve amps? (nobody shoot me for saying that! [emoji23][emoji23])
I was of the belief that, at least for old guitar amplifiers, most of the character came more from the under sized power supply and output transformers? That transistors and valves don't clip all that differently, unless incorrectly biased if measured before the load or transformer? so this little chip doesn't make sense to me if it acts as basically a line follower? Like those cheap eBay hybrid valve amps? (nobody shoot me for saying that! [emoji23][emoji23])
Other way around. Most of the character comes from clipping / THD, which is generated mostly in the valves. PSU can affect the signal envelope but is a secondary consideration. OT is least important. Tubes and transistors clip very differently, particularly when comparing their often non-similar circuit topologies. Followers in guitar amps are often deliberately 'bad' followers, so they produce a lot of distortion and soft clipping, which the nu-tube is sure to do also.
"To put some numbers into it" here you have a basic triode gain stage built around a Nutube, old style graphic design using the datasheet curves , the way I learnt since 1969 and which was used earlier (no computers or simulation way back then).
trying to maximize gain, peak to peak signal out, use reasonable voltages which can be found inside an SS guitar preamp (no sense plugging this into an amp which already has *real* tubes), I chose:
* idle plate voltage: 11.3V
* bias: +3V (yes, positive grid bias, it´s hard to pull electrons from a filament at such low voltages)
FWIW both B K Butler´s "Real Tube", Mini Boogee and doubtless others also use positive grid bias, to get *some* current at such abysmal low voltages.
"You can´t beat Physics"
* idle current: 68uA
With the load line I chose, if I apply 1Vpp (swinging from +2V to +4V) signal to grid, I can swing with reasonably good results and a "somewhat" symmetrical signal (sort of) between 84uA/5V and 40uA/20V .
* so stage gain is 15Vpp/2Vpp=7.5X
As a practical comparison: 2 of these stages cascaded can provide same raw gain as a single 12AX7 stage ... which of course will need >160V supply, 6.3V@300mA filament, a socket, etc. so in some very limited situations it might be useful.
* something interesting is that gain is not symmetrical, notice that for a symmetrical +/-1V input swing, output swings +8.7v and -6.3V , an important difference.
There was an old way to graphically calculate % distortion, I forgot it long ago and I guess "modern" books don´t even mention it, but I wouldn´t be surprised at 15% distortion or so ...... which of course is *good* on Guitar Amps, even when nominally "clean".
This is NOT clipping distortion but tasty nonlinearity, mosty of it second harmonics (and even harmonics in general), good to add tube flavour to SS or even worse, digital Audio, so I am not surprised at all at Korg keeping it alive.
Almost forgot :
* 15V swing with 48uA swing means: 312k plate resistor
* so following load should not be less than 1M, so as not to load it down and lose both gain and output swing.
* since grids are positive biased, they must be pulling *some* current so input impedance can´t be high, so when cascading Nutubes, intermediate buffers are advisable.
* you need at least +32.5V supply (when plate reaches +20V peak, you still have 40uA through that 312k resistor, so you need extra 12.5V)
Ok, so now you have some numbers to decide whether Nutubes are worth it or not.
* almost forgot the last NUMBER: U$50 each
EDIT:
2) the *filament* is 600C ... and probably weighs a milligram or two ... while glass there must weigh a couple grams.
EVEN IF all filament heat were fully transmitted to glass envelope, it would only heat it up, say, a couple Degrees C .
But since there´s both excellent insulation,and to boot radiated heat (infrared rays) *mostly* pass through glass ... there´s no practical heating.
3) and in any case, filament power is just 0.7V * 17mA=11.9mW ... and as said above only a tiny fraction of it might heat the glass .... an infinitesimal temperature rise.
trying to maximize gain, peak to peak signal out, use reasonable voltages which can be found inside an SS guitar preamp (no sense plugging this into an amp which already has *real* tubes), I chose:
* idle plate voltage: 11.3V
* bias: +3V (yes, positive grid bias, it´s hard to pull electrons from a filament at such low voltages)
FWIW both B K Butler´s "Real Tube", Mini Boogee and doubtless others also use positive grid bias, to get *some* current at such abysmal low voltages.
"You can´t beat Physics"
* idle current: 68uA
With the load line I chose, if I apply 1Vpp (swinging from +2V to +4V) signal to grid, I can swing with reasonably good results and a "somewhat" symmetrical signal (sort of) between 84uA/5V and 40uA/20V .
* so stage gain is 15Vpp/2Vpp=7.5X
As a practical comparison: 2 of these stages cascaded can provide same raw gain as a single 12AX7 stage ... which of course will need >160V supply, 6.3V@300mA filament, a socket, etc. so in some very limited situations it might be useful.
* something interesting is that gain is not symmetrical, notice that for a symmetrical +/-1V input swing, output swings +8.7v and -6.3V , an important difference.
There was an old way to graphically calculate % distortion, I forgot it long ago and I guess "modern" books don´t even mention it, but I wouldn´t be surprised at 15% distortion or so ...... which of course is *good* on Guitar Amps, even when nominally "clean".
This is NOT clipping distortion but tasty nonlinearity, mosty of it second harmonics (and even harmonics in general), good to add tube flavour to SS or even worse, digital Audio, so I am not surprised at all at Korg keeping it alive.
Almost forgot
:* 15V swing with 48uA swing means: 312k plate resistor
* so following load should not be less than 1M, so as not to load it down and lose both gain and output swing.
* since grids are positive biased, they must be pulling *some* current so input impedance can´t be high, so when cascading Nutubes, intermediate buffers are advisable.
* you need at least +32.5V supply (when plate reaches +20V peak, you still have 40uA through that 312k resistor, so you need extra 12.5V)
Ok, so now you have some numbers to decide whether Nutubes are worth it or not.
* almost forgot the last NUMBER: U$50 each
EDIT:
1) as said above, vacuum is a very good insulator (the best ).
2) the *filament* is 600C ... and probably weighs a milligram or two ... while glass there must weigh a couple grams.
EVEN IF all filament heat were fully transmitted to glass envelope, it would only heat it up, say, a couple Degrees C .
But since there´s both excellent insulation,and to boot radiated heat (infrared rays) *mostly* pass through glass ... there´s no practical heating.
3) and in any case, filament power is just 0.7V * 17mA=11.9mW ... and as said above only a tiny fraction of it might heat the glass .... an infinitesimal temperature rise.
Last edited:
DM70
Luminescent sub-mini triode with decent curves, Mu 8
Good for 300 uA, 25 mW, and only $12 (1952 provisional datasheet)
Later 1954 datasheet shows 600 uA, 75 mW
http://tubedata.milbert.com/sheets/030/d/DM70.pdf
Vintage Mullard NOS 1M3 DM70 Magic Eye Tube in Original Box | eBay
And EM800 Luminescent will get you 5 mA and 0.6 Watt
http://tubedata.milbert.com/sheets/124/e/EM800.pdf
Luminescent sub-mini triode with decent curves, Mu 8
Good for 300 uA, 25 mW, and only $12 (1952 provisional datasheet)
Later 1954 datasheet shows 600 uA, 75 mW
http://tubedata.milbert.com/sheets/030/d/DM70.pdf
Vintage Mullard NOS 1M3 DM70 Magic Eye Tube in Original Box | eBay
And EM800 Luminescent will get you 5 mA and 0.6 Watt
http://tubedata.milbert.com/sheets/124/e/EM800.pdf
Attachments
Last edited:
Thanks DF
Care to post/link how´s that calculated? , I forgot and my Time Machine is out of fuel
A large number of classic guitar amps offer considerable (and adjustable) distortion in preamp stages between a Gain and a Master knob. You can moosh the end of the preamp yet play the power section so soft it is nearly clean.
This is not the same as flogging a pair of 6V6 into a cheap OT. But for many players it is musically useful.
Nelson Pass in the other section posted a plot showing up to ~~10% at clipping. This can be tweaked with bias and loading. A conventional general purpose amplifier triode shows 5%--- so this shows what a "bad" tube the NuTube is. It really is optimized for good glow at low-low current and on/off operation, not linearity.
Process (and ruler) for estimating 2nd Harmonic from loadline asymmetry is in RDH (3rd or 4th).
Wow! Thanks for all that!
Merlinb, I read that too cute costs on early guitar amps, they would typically use vastly underrated transformers. The power transformer would sag it's voltage allot adding a kind of compression to it. The power supply just couldn't cut it at full tilt.
Also, main output transformers where typically fairly underrated aswell. These would reach magnetic saturation and add a weird compression effect that would dynamically compress the bass and treble ends more so than the mid range.
They also mentioned non linear biasing, which would clip one half now than the other.
Finally, they mentioned how the output transformer would filter out allot of the really harsh clipping, especially once the transformers reach magnetic saturation.
I read this over on peavey's tech note section, which to be fair is largely marketing, but interesting none the less. Maybe they are exaggerating? What are your thoughts?
Merlinb, I read that too cute costs on early guitar amps, they would typically use vastly underrated transformers. The power transformer would sag it's voltage allot adding a kind of compression to it. The power supply just couldn't cut it at full tilt.
Also, main output transformers where typically fairly underrated aswell. These would reach magnetic saturation and add a weird compression effect that would dynamically compress the bass and treble ends more so than the mid range.
They also mentioned non linear biasing, which would clip one half now than the other.
Finally, they mentioned how the output transformer would filter out allot of the really harsh clipping, especially once the transformers reach magnetic saturation.
I read this over on peavey's tech note section, which to be fair is largely marketing, but interesting none the less. Maybe they are exaggerating? What are your thoughts?
+8.7V to -6.3V is an average swing of +- 7.5V, with 1.2V of distortion. Assuming this is second (it will be almost entirely second) then this will be 0.6V of second harmonic and 0.6V of DC. At the signal peak you get 0.6V 2nd plus 0.6V DC. At the zero crossing you get 0.6V DC minus 0.6V 2nd. So there is 7.5V signal plus 0.6V 2nd, which is 8%.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.