| salas |
Hello
Can I use a current source heater PSU (LM317 R=1.25/I type) for a cathode follower and still bias it with voltage so to keep the cathode no more than 100V DC positive in respect to the heater?
I want to use 0.3A current source for 12AU7.
Do you prefer the sound of a current sourced heater? |
|
|
| kmtang |
I don't think you need constant current source for 12AU7 - indirectly heated filament. You can simply use DC regulator instead.
Johnny |
|
|
| salas |
I have read many times that a CS heater sounds better. That is why I want to try it.
Why you think its not needed? |
|
|
| planet10 |
| quote: | Originally posted by salas
I have read many times that a CS heater sounds better. That is why I want to try it. |
That's what i've heard... i've not built anything with DC heaters yet so i can't talk from experience.
It is pretty easy to implement why not just try it.
dave |
|
|
| salas |
I have got the parts already (he he) but I was mainly interested in floating CT with voltage to avoid exceeding over 100VDC difference between heater and cathode. So I asked if they do this and how on a current sourced heater. I will listen to it anyway and let you know p10.
PS did you see that?:
http://www.diyaudio.com/forums/show...&threadid=57210
you are right to be a proponent |
|
|
| kmtang |
The constant current source for filament supply is mainly used for Direct Heated Tubes like the 300B and 2A3. The reason is that the filament forming part of the cathode circuit which would affect the sound.
However, the indirectly heated tubes like the 12AX7 having the filament separated from the cathode, I don't see what CCS could do anything at all.
If you want to float the filament voltage, I think you can connect the -ve side of the CCS supply to the B+'s divider.
Just my 2 cents.
Johnny |
|
|
| SY |
| quote: | | I don't see what CCS could do anything at all. |
The advantages are a gentler startup (heaters have lower resistance when cold) and a somewhat better potential efficiency. There's a nice treatment of these issues in Morgan Jones's book. The main thing to avoid is CCS operation of parallel heaters. |
|
|
| gingertube |
The Allen Wright "Tube Preamp CookBook" recommends exactly this method - example circuits of current sourced heater supplies at elevated DC voltage abound. A RECOMMENDED book for any serious DIY Audio Nutter.
Cheers,
Ian |
|
|
| thoriated |
| I've used a slightly different approach - a voltage regulated heater circuit that is current limited only a little higher than the required heater rating for soft start up. |
|
|
| salas |
| I have prepared this. Its ready. Is it ok to apply it? |
|
|
| Circlotron |
| quote: | Originally posted by SY
The advantages are *snip* a somewhat better potential efficiency. |
Efficiency of what? The heater heating the cathode, or plate cct efficiency? ...or efficiency of the snake-oil cycle? :dodgy:
Sounds a bit strange to me. |
|
|
| SY |
| Efficiency of the heater regulator. |
|
|
| SY |
| quote: | Originally posted by salas
I have prepared this. Its ready. Is it ok to apply it? |
The one glaring problem is that you're using a CCS to feed a parallel pair of heaters. A CCS is best used on heaters in series- in a sense, it forces matching between tubes (or, in your case, between sections of tubes).
You probably do not need that reverse-voltage diode since this is not a HV application. |
|
|
| salas |
| I lack more than 12V rectified supply, else I would do it in series, its a plus. So I will apply it since nobody sees circuit error. |
|
|
| rdf |
| When driving indirect heaters with a current source is there any benefit to bypassing the coils with capacitors? A small film cap driven by a very high impedance sounds like a cheap and cheerful way to minimize the potential of mains-bourne junk entering the circuit via the heater. Another consideration is that a solid state current source will only act as one to a certain frequency (well below RF?) Anyone played with this combination? |
|
|
| EC8010 |
| It would indeed be an excellent filter, except. Except that most CCSs for this application are not only not very high impedance, but they're distinctly iffy above a few hundred Hz. I've been doing some preliminary investigations on using power FETs or bipolar audio transistors (such as BD131) in a cascode. |
|
|
| salas |
Did it....Its great! It sounds fantastic! More focus less grunge, more depth, faster. Bassist woke up.
*I use 10uf tantalum on pins 4,5-9 |
|
|
| rdf |
| quote: | Originally posted by EC8010
...not only not very high impedance, but they're distinctly iffy above a few hundred Hz. |
Thanks for the head's up! Heater CCS's were in consideration for my first self design but I hadn't considered the possibility of such poor behaviour from typical implementations. When you say 'most', you mean the the usual IC voltage regulator restrapped as a current source? |
|
|
| EC8010 |
| That's right, if you look at the curves of output impedance against frequency or ripple rejection against frequency for an adjustable regulator you soon realise that what you're looking at is a graph produced by the falling open loop gain of an op-amp. And not a very good one at that. A 741. |
|
|
| salas |
| Considering the fine sonic upgrade I got by using LM317 current source I would think its worthy to investigate a superior descrete CCS. Do you have any promissing circuit in mind that is fairly developed and tested? |
|
|
| EC8010 |
| Been away... Not yet. Although uprating a cascode to take account of the current or simply using a DN2540 depletion-ode JFET would work, it wouldn't necessarily be very temperature stable, so a little more thought is probably required. Watch this space... |
|
|
|
