Hi,
Having had a closer look at the Gary Pimm CCS and hearing die hard tube-o-philes on the possibilities of creating all tubed CCSs I thought it worthwhile to start a discussion.
It must have been done in the past and if I look at Amperites' catalogue I see a lot of interesting stuff that can be put to good use.
Do any of you feel the same way?
I welcome your input.
Cheers,
Having had a closer look at the Gary Pimm CCS and hearing die hard tube-o-philes on the possibilities of creating all tubed CCSs I thought it worthwhile to start a discussion.
It must have been done in the past and if I look at Amperites' catalogue I see a lot of interesting stuff that can be put to good use.
Do any of you feel the same way?
I welcome your input.
Cheers,
Simple solution for a Constant Current Sink is a pentode of adequate ratings (for a LTP, a 6AU6 or similar RF amp pentode would be suitable, whereas for...Idunno, what uses a lot of constant current?). Sinks are easier in tubes due to that damn screen voltage...I suppose if you wanted to get real technical, you could build a voltage follower setup such that screen voltage tracks exactly +x volts above the cathode, *without loading the cathode*.
That last bit is where it hits you...for (e.g.) a mu stage, it's made to have some sort of load. But then, it obviously isn't a CCS. (Not like a mu-follower is a CCS anyway.)
But then again, perhaps you don't need a pentode. Use a triode and amplify the current difference with a cathode resistor, as in the mu stage. DC coupling would be difficult however; it'd need a zener for bias...unless you can find a mercury-vapor (instead of neon) lightbulb.
Tim "I hope this makes sense because I'm too lazy to proof-read right now"
That last bit is where it hits you...for (e.g.) a mu stage, it's made to have some sort of load. But then, it obviously isn't a CCS. (Not like a mu-follower is a CCS anyway.)
But then again, perhaps you don't need a pentode. Use a triode and amplify the current difference with a cathode resistor, as in the mu stage. DC coupling would be difficult however; it'd need a zener for bias...unless you can find a mercury-vapor (instead of neon) lightbulb.
Tim "I hope this makes sense because I'm too lazy to proof-read right now"
Artists at work.
Hi,
Thanks Tim.
What I had in mind is eliminating the semis from Gary Pimm's reg,
namely the LM 431 and the diode and replace it with valves/tubes.
After all,I reckoned,what can be done with sand can be done with glass...economics aside.
Then we could compare performance (theory) and sound (practice) on both topologies.
Cheers,
Hi,
Thanks Tim.
What I had in mind is eliminating the semis from Gary Pimm's reg,
namely the LM 431 and the diode and replace it with valves/tubes.
After all,I reckoned,what can be done with sand can be done with glass...economics aside.
Then we could compare performance (theory) and sound (practice) on both topologies.
Cheers,
NP. It's something I've done a little thinkin' about (needs much more ). Like I said, the problem with a pentode is the screen loading the system. Okay if there's a good bit of load to swamp it, but for pure unloaded triode mu-ness, it ain' gonna happen.
If you don't grid-bias the pentode, the screen-supply resistor can go up to 100k on a 6CL6...so like, near 1M on a 6AU6. Since Ik = Ia + Ig2, the screen itself shouldn't load down the cathode too much right?
Well anyway, this in parallel with the plate resistance might make for 500k-1M depending on stuff. Might be a 2-5x improvement over a resistor, but not worth it IMO.
/me gets to work on triode CCS...
Tim
). Like I said, the problem with a pentode is the screen loading the system. Okay if there's a good bit of load to swamp it, but for pure unloaded triode mu-ness, it ain' gonna happen.If you don't grid-bias the pentode, the screen-supply resistor can go up to 100k on a 6CL6...so like, near 1M on a 6AU6. Since Ik = Ia + Ig2, the screen itself shouldn't load down the cathode too much right?
Well anyway, this in parallel with the plate resistance might make for 500k-1M depending on stuff. Might be a 2-5x improvement over a resistor, but not worth it IMO.
/me gets to work on triode CCS...
Tim
CCS...COCA COLA SYSTEM?
Hi,
O.K. I copy.
No need to stick to the penthode with its bias prob here...I don't mind a bit of lateral thinking.
What if we define a practical sample,say a 5 to 10 mA driver stage,or you take a pick for that matter.
I think we both understand the pros and cons of penthode versus triode,don't we?
Cheers and thx,
Hi,
O.K. I copy.
No need to stick to the penthode with its bias prob here...I don't mind a bit of lateral thinking.
What if we define a practical sample,say a 5 to 10 mA driver stage,or you take a pick for that matter.
I think we both understand the pros and cons of penthode versus triode,don't we?
Cheers and thx,
I have built a few pentode sinks.
My normal procedure is to apply about +30V to g1 which enables me to raise the value of the cathode resistor accordingly. Since the sink works by multiplying its gain by the value of the unbypassed cathode resistor, this can provide a very satisfactory sink,
Capacitance is the enemy as in all sinks, so choose valves with a lowish Cout. You are also looking for nice flat curves for linearity.
6AU6 is a good suggestion, but my favourites for this service are in order of Pa: 6096/EF95; 6EJ7/EF184, Z759 and finally EL822. The latter three all have gm of 15 and very high anode resistance - just what is required.
7N7
My normal procedure is to apply about +30V to g1 which enables me to raise the value of the cathode resistor accordingly. Since the sink works by multiplying its gain by the value of the unbypassed cathode resistor, this can provide a very satisfactory sink,
Capacitance is the enemy as in all sinks, so choose valves with a lowish Cout. You are also looking for nice flat curves for linearity.
6AU6 is a good suggestion, but my favourites for this service are in order of Pa: 6096/EF95; 6EJ7/EF184, Z759 and finally EL822. The latter three all have gm of 15 and very high anode resistance - just what is required.
7N7
Forgot about this'un.
For sink - tetrode in negative-resistance region.
I have a few 24As around, between 40 and 80V (or so) the plate resistance is about -40k, depending on bias. Put a 40k resistor in parallel with this and viola, infinite ohms.
Of course, starting it may be a problem, since there are two currents where the tube is stable. A simple solution would be to shunt the plate with a neon lamp; it would clamp it to below 80V, right at the knee, at the end of the -R region. In operation, the setup would be biased for the middle of the range (60V or so), so the neon would effectively be nothing more than a few pF.
Tim
For sink - tetrode in negative-resistance region.
I have a few 24As around, between 40 and 80V (or so) the plate resistance is about -40k, depending on bias. Put a 40k resistor in parallel with this and viola, infinite ohms.
Of course, starting it may be a problem, since there are two currents where the tube is stable. A simple solution would be to shunt the plate with a neon lamp; it would clamp it to below 80V, right at the knee, at the end of the -R region. In operation, the setup would be biased for the middle of the range (60V or so), so the neon would effectively be nothing more than a few pF.
Tim
Quite nice,
this thread, as I am also looking for a tube based CCS right now... Interesting ideas above, but perhaps Sch3mat1c and 7N7 could enrich their postings with some schematics ...?
Thanks in advance
Andreas
PS. Seems i didn't notice this thread was from 2003 ... Hope the two mentioned above are still active in this forum ??
this thread, as I am also looking for a tube based CCS right now... Interesting ideas above, but perhaps Sch3mat1c and 7N7 could enrich their postings with some schematics
...?Thanks in advance
Andreas
PS. Seems i didn't notice this thread was from 2003 ... Hope the two mentioned above are still active in this forum
??
Hello.
Well times have changed rather drastically - I am now in France and no longer have any of my kit.
Nevertheless I would like to remain an occasional visitor hereabouts.
I will look to see if I have any circuit diagrams, but in fact all you need to know about pentode sinks may be found in the indispensible "Valve Amplifiers" by Mr Morgan Jones.
Best of luck
7N7
Well times have changed rather drastically - I am now in France and no longer have any of my kit.
Nevertheless I would like to remain an occasional visitor hereabouts.
I will look to see if I have any circuit diagrams, but in fact all you need to know about pentode sinks may be found in the indispensible "Valve Amplifiers" by Mr Morgan Jones.
Best of luck
7N7
Let's see here...from the simplest, a triode with a big cathode resistor. Plate Z becomes mu*cathode_R + original, grounded cathode Plate Z. Not so bad when you can get a 5k cathode R and a mu of 50.
The trouble starts when you want to pass more than a fraction of a mA. How zealous do we need to be with the SS-free voltage references. Make a 60 volt reference, and this beastie will run ~12 mA. 250kOhm impedance with 12 mA.
If a 60volt SS V_ref is OK, why not a 120V, and we'll make a cascode. Upper anode curves are hardly curved anymore....
Now the upper device does not have to be a triode. Pentode will be fine. R-C to energize the upper g2. Proper V_ref for its grid above the lower device's cathode. Shunt capacitance goes down with the use of a pentode on the upper section...that's why it is called *SCREEN* grid ( g2 that is ).
cheers,
Douglas
The trouble starts when you want to pass more than a fraction of a mA. How zealous do we need to be with the SS-free voltage references. Make a 60 volt reference, and this beastie will run ~12 mA. 250kOhm impedance with 12 mA.
If a 60volt SS V_ref is OK, why not a 120V, and we'll make a cascode. Upper anode curves are hardly curved anymore....
Now the upper device does not have to be a triode. Pentode will be fine. R-C to energize the upper g2. Proper V_ref for its grid above the lower device's cathode. Shunt capacitance goes down with the use of a pentode on the upper section...that's why it is called *SCREEN* grid ( g2 that is ).
cheers,
Douglas
uh oh,
It's generally easier to have a separate negative supply but you don't actually need it if the cathodes of the LTP are at a sufficiently high positive voltage for the pentode to operate. For a 6AU6 I wouldn't go below 100v, especially if you neded a higher current than about 2 or 3 mA.
It would probably work without a negative supply, for instance, if you had a Mullard-type circuit with a single-ended input stage directly coupled to the LTP. Then, you would need to drive the screen from the B+ rail, through a suitable resistor. Ensure the screen to cathode voltrage is no higher than the plate to cathode voltage.
A negative supply is easy enough to fix up if your power transformer has a centre-tapped secondary for the HT and you're using full-wave rectification with the CT as ground, instead of a bridge (which is also full-wave, of course, but that's the terminology). Then you can connect a pair of silicon diodes off the same HT secondary, in reversed polarity, to get the negative line. If the voltage would be too high, an old trick is to use capacitors in series with the diodes as wattless droppers.
Although you're using tube rectifiers fior your B+, it's pointless using tube rectifiers for the negative rail as well. SS diodes will OK for the negative supply, especially if you want to use it also as a source of negative bias for the OP tube grids. That will ensure that the negative rail is up before the positive rail, so the OP tubes are safely biased .
It would probably work without a negative supply, for instance, if you had a Mullard-type circuit with a single-ended input stage directly coupled to the LTP. Then, you would need to drive the screen from the B+ rail, through a suitable resistor. Ensure the screen to cathode voltrage is no higher than the plate to cathode voltage.
A negative supply is easy enough to fix up if your power transformer has a centre-tapped secondary for the HT and you're using full-wave rectification with the CT as ground, instead of a bridge (which is also full-wave, of course, but that's the terminology). Then you can connect a pair of silicon diodes off the same HT secondary, in reversed polarity, to get the negative line. If the voltage would be too high, an old trick is to use capacitors in series with the diodes as wattless droppers.
Although you're using tube rectifiers fior your B+, it's pointless using tube rectifiers for the negative rail as well. SS diodes will OK for the negative supply, especially if you want to use it also as a source of negative bias for the OP tube grids. That will ensure that the negative rail is up before the positive rail, so the OP tubes are safely biased .
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.