Yes, I did make them like that for a while, maybe more than twelve years ago! But I found that for many constructors, myself included, the single-point connexion to ground (or Rk) usually wins the listening tests - especially if you have loudspeakers that can take advantage of the wider and taller stereo space.
Would a Keantoken circuitt be useful @ 2A rating as CM? It floats but is high impedance too. The basic is a darlingnot instead of a darlington, with a special buffer circuit.
The cap-multiplier could be a CFP (Szlikai) circuit, if you have the time to build and measure.
I did not find any advantages (with the transistors tested) at these higher currents.
CFPs also need care to be sure there is no instability, which is no advantage at all! At lower current, which IIRC the Keantoken addresses, the control- and the pass- transistors are dynamically close in behaviour, and this makes stability less troublesome.
I did not find any advantages (with the transistors tested) at these higher currents.
CFPs also need care to be sure there is no instability, which is no advantage at all! At lower current, which IIRC the Keantoken addresses, the control- and the pass- transistors are dynamically close in behaviour, and this makes stability less troublesome.
Yes I remember that Kean described such difficulties and some have been adding a cap between emitter and collector - strange having to to remedy it that way.
Nevertheless, in amplifiers like Le Monstre by Jean Hiraga the CFP works flawlessly, I had it running at > 1.2 Amps (within a closed loop of about 500KHz).
Rod
I have a question about connecting your DHT filament regulators.
I have build a Tubelab SE amp, using 45 and/or 46 DHT's. It uses DC heaters, heater power supplied by a high current linear reg.
The schematic for the TSE amp can be found here
Schematic | Tubelab
I am not sure how to connect your reg to my amp?
Your general installation guide shows a 1-10 ohm current sense resistor for a fixed bias amp. Since you don't specify a value, seems like this should be part of the amp, but I don't have one.
Should I add one? What value? or does it matter?
Also not sure if I should isolate both heater pins, 3 and 4, at the 45, and connect those pins to fil-pos and fil-neg? IOW, do I disconnect TSE ground from the 45, and connect to fil-neg instead. Or does fil-neg get connected to TSE ground?
TIA for your help here.
Randy
I have a question about connecting your DHT filament regulators.
I have build a Tubelab SE amp, using 45 and/or 46 DHT's. It uses DC heaters, heater power supplied by a high current linear reg.
The schematic for the TSE amp can be found here
Schematic | Tubelab
I am not sure how to connect your reg to my amp?
Your general installation guide shows a 1-10 ohm current sense resistor for a fixed bias amp. Since you don't specify a value, seems like this should be part of the amp, but I don't have one.
Should I add one? What value? or does it matter?
Also not sure if I should isolate both heater pins, 3 and 4, at the 45, and connect those pins to fil-pos and fil-neg? IOW, do I disconnect TSE ground from the 45, and connect to fil-neg instead. Or does fil-neg get connected to TSE ground?
TIA for your help here.
Randy
Hi Randy,
In the Tubelab-SE, the 10Ω resistor R29 does the same job, but it's in the plate circuit.
you can ignore the 1-10Ω resistor in my instructions, and instead just follow the "Setting Up the Bias" paragraphs in the Tubelab manual.
There are a couple of other changes - I will write them out and post later.
Last edited:
Great thanks, looking forward to the other changes you have.
I think my only other question is if I would connect the TSE ground to either the Fil Pos or Fil Neg outputs from your reg.
Randy
Please connect FIL+ to Sytem Ground. The PCB track to the power-regulator should be cut, or in some other way make full isolation between U1 R2 and C3 from the tube socket.
The other precautions:
- Please connect the two pads of F-JP together, once the new regulator is wired in. This will reference the 5842's heater to ground, instead of the DHT filament.
- check that the 5842 heater is still running at 6.2 - 6.3V, once everything is set up correctly. The power transformer will have lighter loading without the 45 filament (a good thing!) but R3 may need to be increased to keep the heater happy.
The other precautions:
- Please connect the two pads of F-JP together, once the new regulator is wired in. This will reference the 5842's heater to ground, instead of the DHT filament.
- check that the 5842 heater is still running at 6.2 - 6.3V, once everything is set up correctly. The power transformer will have lighter loading without the 45 filament (a good thing!) but R3 may need to be increased to keep the heater happy.
I think I should move a jumper.
Right now, since I have a 45 tube based amp, so I jumper from PAD1 to PAD4, grounding the center tap of the filament winding.
If I move that jumper to PAD3 to PAD4, then it connects 5842 F- to amp ground. I'll also cut the filament pos traces at the 45's, which isolates the output of U1, so I'll no longer be using U1.
Randy
Preamp tubes and tubes used in pre driver and driver stages can benifit by been driven by dc voltage supplies as it could reduce hum. Coleman reg is also a specialized dc supply intended to be used for different purposes. The current regulated Coleman supply will definifely look after the filament of the tube, but would be an overkill at a high cost.
So finally getting to checkout and installation of my coleman filament heater.
I used some big power resistors to check out my raw dc supply, and then the coleman reg. So far have one side completely checked out, installed, and powered a tube. Tube is a 47 at 2.5V and 1.75A.
Based on my measurements of the raw supply, seems like my 47 isn't drawing 1.75A, maybe less then 1.5A. I need to check again, I'll be more careful when I bring up the 2nd channel, and record a few voltage/current points with different loads on it.
The 47 I'm using has always been on the weak side, need to crank up its bias pot to get the right bias setting, and seem to glow less then its partner.
But wondering how real life the data sheet numbers for heaters are?
Is 1.75 supposed to be a max number? Or should it have some tolerance?
Randy
I used some big power resistors to check out my raw dc supply, and then the coleman reg. So far have one side completely checked out, installed, and powered a tube. Tube is a 47 at 2.5V and 1.75A.
Based on my measurements of the raw supply, seems like my 47 isn't drawing 1.75A, maybe less then 1.5A. I need to check again, I'll be more careful when I bring up the 2nd channel, and record a few voltage/current points with different loads on it.
The 47 I'm using has always been on the weak side, need to crank up its bias pot to get the right bias setting, and seem to glow less then its partner.
But wondering how real life the data sheet numbers for heaters are?
Is 1.75 supposed to be a max number? Or should it have some tolerance?
Randy
Hi Randy,
For "Voltage" specified filaments, which is most types, and includes the 47, the 2.5V filament voltage is fixed, and although you are allowed ±5% tolerance in the data sheet, it is best to keep as near to the Nominal as you can.
The current drawn at the nominal Voltage is only for guidance, and the range is not always specified. It is usually within ±10% of the guide value (1.75A), but older tubes (before 1950,say), and those with many different manufacturers, and many different physical structures, the variation may be more.
If the current of a given sample is >20% down, I would regard it as suspect, and possibly it has run many more hours than it was able to support.
But never increase the voltage to get the current you are expecting - this will lead to damage, for sure. (With current specified filaments like the WE 10xx series, and STC 3A/10xx series: this is allowed, however - but these are exceptions).
What you can do if the regulator won't adjust all the way down to 2.5V: change the current-program resistor. I can guide you, if you think this might be needed - please send some email.
For "Voltage" specified filaments, which is most types, and includes the 47, the 2.5V filament voltage is fixed, and although you are allowed ±5% tolerance in the data sheet, it is best to keep as near to the Nominal as you can.
The current drawn at the nominal Voltage is only for guidance, and the range is not always specified. It is usually within ±10% of the guide value (1.75A), but older tubes (before 1950,say), and those with many different manufacturers, and many different physical structures, the variation may be more.
If the current of a given sample is >20% down, I would regard it as suspect, and possibly it has run many more hours than it was able to support.
But never increase the voltage to get the current you are expecting - this will lead to damage, for sure. (With current specified filaments like the WE 10xx series, and STC 3A/10xx series: this is allowed, however - but these are exceptions).
What you can do if the regulator won't adjust all the way down to 2.5V: change the current-program resistor. I can guide you, if you think this might be needed - please send some email.
Thanks for that explanation Rod, that all makes sense.
I brought up the other channel yesterday, and that tube seems dead on. The voltage from my raw supply when I power the tube is the same as I get when I power my resistor load.
I think the bad tube is about 20% off.
I have a pair of 46's in the mail to me now that I bought from a real dealer, so I expect them to act better. These 47's were ebay specials.
I was wondering if I could increase the voltage a bit on the bad 47, won't try that now.
Randy
I brought up the other channel yesterday, and that tube seems dead on. The voltage from my raw supply when I power the tube is the same as I get when I power my resistor load.
I think the bad tube is about 20% off.
I have a pair of 46's in the mail to me now that I bought from a real dealer, so I expect them to act better. These 47's were ebay specials.
I was wondering if I could increase the voltage a bit on the bad 47, won't try that now.
Randy