simple ccs under 5-10mA ?

[bembel]

I'm about to build an ECC81/12AT7 differential phase-splitter and I am looking for a simple and cheap (and european aviable) CCS like the LM317, but in the range of a few mA.

Any ideas ?

(This with or without a negative supply, better, reuse the planned DC filament supply 6.3 or 12.6v)

[Giaime]

Here it is:

http://www.giaime.altervista.org/CCS.gif

feel free to use mostly any transistor you have at hand (that will survive the voltage and current needed, of course).

-16V can be obtained from 6.3VAC filaments with a voltage doubler.

[rdf]

http://www.ixys.com/pdhvcr01.html

See Pete Millet's site for perfomance numbers. Easily available mail order.

[bembel]

Thank you very much, it seems to answer so quicky to what I was lookink for (I'm a few hours' fresh member of the forum).

But Giaime, it's too bad that your link is not working at now, (either the one to RDH4).I got a big bag of BC337 will it fit to that circuit ? (that I haven't seen)

I saw also some data about tl431 (I got some of them) but I don't know if it realy needs a transistor (as said in the datasheet) or if it can be directly used like the lm317 for current up to 4mA? (3.52mA where simulated with ECC81(differential common cathode B+=292 Ra=82k Ua=175V), any thoughts ?

thx again

[Miles Prower]

Originally posted by bembel
I'm about to build an ECC81/12AT7 differential phase-splitter and I am looking for a simple and cheap (and european aviable) CCS like the LM317, but in the range of a few mA.

Any ideas ?

(This with or without a negative supply, better, reuse the planned DC filament supply 6.3 or 12.6v)

For something like this, just cascode a couple of BJTs. That's what I did (shown here). The two BJTs are MPSA42 high voltage, video amp, types to avoid possible poofage by overvolting. This works with two sections of a 6SL7 to form a differential phase splitter. Simple, and it works just great in maintaining both amplitude and harmonic balanve between sections. A 12AT7 will require a bit more current than a 6SL7, but still under 10mA. Keep the grids near ground, and add a negative rail voltage of about 20Vdc, and just about any small signal NPNs will get the job done.

[bembel]

I got a big bunch of BC337, otherwise, I was wondering what was the difference between: bipolar , FETs , & ICs
I mean, how does the impedance of the CCS affects the circuit ? and then, which technology of component offert the best ratio
simplicity(laziness)/impedance,PSRR(good sound) ?
RDF suggestion about IXYS IC is a very pleasing idea too, but I was initialy thinking of a CCSink (in the cathode circuit). but it seems that this kind of component can work with low voltages too, any experiences about ?
Also it would be a nice thing from the CCS to be satisfied whit a negative rail of (-6.3 or -12.6V as I planned a DC filament supply for this phase-splitter stage)

[tubelab.com]

Consider the LM334Z. I think that it does 10 mA max, but works very well with low voltage. Negative voltage is not always required. I tried it with a 12AT7 in LTP. I think that it sounded better tied to -5volts instead of ground. I get -5volts by rectifying the filament voltage and using a negative regulator.

Morgan Jones uses this chip in some of his designs. If you have the third edition look on pages 476, 477, 478, and 484.

The IXYS chip needs a little more voltage to work well, something like -12 volts.

[Giaime]

Why my links always don't work

However I was suggesting something identical to Miles' schematic.

[Miles Prower]

Originally posted by bembel
I got a big bunch of BC337, otherwise, I was wondering what was the difference between: bipolar , FETs , & ICs

BJTs work better as a constant current source than do most FETs. The current stays more constant with voltage variations. Also, since the BJT is a much higher gain device, the impedance of the BJT CCS is also higher. An IC would have even higher gain, since it's usually composed of several BJT gain stages. Cascoding the BJTs is helpful in getting the impedance up, and helps to avoid high frequency AC from flowing around the CCS. An IC is a good way to go, but I didn't have one, and did have the necessary transistors, so that's what I did.

I mean, how does the impedance of the CCS affects the circuit ?

The higher the impedance, the better the balance, both AC as well as harmonic distortion, and the higher the CMRR, which is important for rejecting whatever noise may be coming from the CCS in the tail.

and then, which technology of component offert the best ratio
simplicity(laziness)/impedance,PSRR(good sound) ?
RDF suggestion about IXYS IC is a very pleasing idea too, but I was initialy thinking of a CCSink (in the cathode circuit). but it seems that this kind of component can work with low voltages too, any experiences about ?
Also it would be a nice thing from the CCS to be satisfied whit a negative rail of (-6.3 or -12.6V as I planned a DC filament supply for this phase-splitter stage)

No easy answers there. If you need something with a low drop out voltage, then some of the aforementioned ICs would probably be the best way to go.

[ray_moth]

bembel,

I take it that your LTP splitter is not directly coupled to a previous stage, as in Miles's circuit? If it is, then you clearly don't need a negative supply.

What you could do, to avoid having a negative supply and also to give yourself a wide choice of CCS types, is to raise the grids (and coupled cathodes) above ground potential.

You can do this by arranging a potential divider from B+ to ground, using high value resistors of several hundred kilohms. Connect the "grounded grid" of the long tail pair (LTP) to the tap on the potential divider and also a capacitor (say 0.33uF) from that grid to ground (so that it is still grounded as far as the signal is concerned).

Then you need to have a 1Meg resistor between the two grids of the LTP, so they are at the same potential. The existing resistor betweeen the input grid of the LTP and ground must be removed. You will also need to use a coupling capacitor (say 0.047uF) as input to the LTP, since its grids are now elevated above ground potential.

Choose the resistors in the potential divider to give enough voltage at the grids and cathodes for your choice of CCS. You could even use a pentode CCS if you wanted to.