Better current source than LM317 on a 12B4 line amp

[sparkle]

better current source than LM317 on a 12b4 line amp

Hi guys,
I have been using for few weeks LM317 in cathode arranged as current source in my 12B4 grounded cathode line amplifier. Plate current is arround 25mA and voltage on LM317 current source is arround 10-11V depending on a tube caracteristics.

I would like to use something that is better than LM317 that will give me arround 25mA and that will work better on that voltage across ccs. A friend of mine suggested TL431 - I would like to hear some other options that might be even better ....
Maybe something discrete .....
best regards

 

[sparkle]

i was thinking about something that is better than this tl431 and LM317 ccs.....
Vi=10V
R1= 1K
Rs=100R
Io=25mA
hfe=the bigger the better

 

[SY]

Could you explain further? You say that this is a grounded cathode circuit but you have a current source on the cathode?

 

[sparkle]

oups..... made mistake... picture of the current arrangement (almost current-but You will get the idea)...
((100uF is 220uF in fact and it is bypassed with 1uF MCap; also voltaage on anode of the tube is not 65V but 130V and high voltage on plate res. is arround 260V))

 

[poobah]

Why do you not like the 317 circuit? What are you wishing to improve or change?

😀

 

[ilimzn]

In this circuit, it is the cathode capacitor that is the most critical component in the cathode circuit. It's whole point is to efficiently bypass the CCS anyway, so the CCS only sets DC condition. Last I checked, humans can't hear DC 🙂 so if something is wrong, your cap is not doing it's job.

 

[Mark A. Gulbrandsen]

I have to agree.... whats wrong with the 317.... they can be had for 50 cents? Any noise that it might introduce would be bypassed with the caps anyway. There are other regulators one could use by IXYS and others however the 317 is pretty cost effective and you'll still need the boutique capacitors in the cathode with the others as well..... I doubt there would be any major audible difference between cathode based current sources due to the major effect the bypass caps have on overall sound.

Mark

 

[Onra]

The datasheet says the tube needs about -15 to -20V gate voltage to deal with 130V plate voltage and 25mA anode current.
Your 10V-Cap will not withstand this voltage for a long time.
Onra

 

[Tubelab_com]

I am partial to the IXYS 10M45 chip myself, but I have not tried it in this circuit........yet.

In this circuit the cap would eliminate all artifacts caused by the CCS, IF THE CAP IS PERFECT. We all know that caps are far from perfect.

I want to build this circuit soon, I have all of the parts, a lab grade power supply to run it (Fluke 407D) and both kinds of CCS chips. I am just short on time right now.

 

[mach1]

I'm just putting the finishing touches to a new 12B4 linestage, the details of which I will post shortly.

I tried a 317 on the cathode with a couple of good quality bypasses (silmic II, cerafine) plus a generic for good measure. However, I preferred a 10M45S cascode CCS on the anode with no bypass cap on the cathode resistor.

The sound with the 317 CCS changes markedly depending on the cap used: I think tubelab hit the nail on the head - this type of circuit is great, assuming you have the perfect cap. Otherwise you are listening to capacitor artefacts, which to my ears are worse than those of the 10M45S.

Doubtless there will be those who contend that their latest iridium / boron laminate caps with powdered unicorn horn dielectric can achieve sonic nirvana in this application. Me - I'd rather not have a cap there in the first place...

 

[sparkle]

🙂 Guys I did not want to argue about how good is LM317 here - .... it is working fine in my circuit that is posted up there in this thread.... I was wondering what will happen if I try a better ccs than LM317 .... that's all ....I want to try everything I can - and I see that this ccs in the cathode might be a good thing but I have also perceived that it's sound is very dependent of quality of the cathode cap bypass - so why not try something better before I say that this project is finished - why not... why we have to stick with the LM317 - when we know that it is not perfect and that cathode cap bypass has problems to deal with it and that sound is dependent on how good that cap is really (I know that it is cheap and everything - but I would like to try better).....that is why I have turned here to ask the question..... 🙂

I have read the thread that is running arround this part of the forum about LM317 in the cathode where our dear Mark A. Gulbrandsen can be seen often - so I am familiar with all the problems (I hope)........... 😀
to be precise... one of my tubes has 10.5V on the cathode and another 11.5V (well not that precise). I have said that I have not 100uF cap bypass in the cathode but 220uF Black Gate and 1uF MCap from Mundorf in paralel.....So it should be good enough I think....😀

 

[auriontoledo]

Hi mach 1,

I also want to build a preamp using 12B4's. But haven't decided on a design yet.... just too bad that I bought a box of 5687's.

You don't like caps, so I am wondering whether you built your 12B4 preamp without?

I am looking forward to more details on your 12B4 as soon as you are finished!!

 

[sparkle]

auriontoledo - You can try Aikido preamplifier with 5687 al through the way.... I think You will like it and it will give You the total gain of arround 8.....
best regard

 

[auriontoledo]

Sparkle: about your Aikido suggestion

Thanks for the tip. I've been busy looking up info about this design. There is one schematic on the web by the designer that uses the 5687 for the second tube, but a 6N1P for the first.

Do you know whether I have to change any other parts if I use only the 5687's?

Also for the first tube, other tubes are recommended: 6N1P, 6DJ8, 6922, 6SN7. Is it worth the effort to locate one of these tubes on Ebay and use it for the first tube rather than a 5687?

 

[sparkle]

there was a good thread on the diy forum regarding aikido.....try to find opinions there also.....
6N1P tube is a very, very good tube ... the problem is that it has higher mu than 5687 so if You need total gain than is lower than 10 than 5687 might be better because You will not have that low total gain with 6n1p (with 6n1p you will have arround 17 and that is too much if You have solid transistor/mos or similar power amplifier). The option is to use 6n6p which is even better than 5687 as people say but this tube has a very low cathode/heater voltage so You might find it slightly difficult to implement....... as You have a box of 5687 - it would be loggical to me that this would be Your first choice since this is a very nice tube and You will have very reasonable total gain. You can make the pre with 5687 all through the way and try it - if You find it interesting and worth the modifications and worth buying a new set of front tubes than You can do it very simply by changing the resistors that are in the cathode of the both front tubes (so that You optimize the working point of those tubes - remember the resistor must be the same for the lower and for the upper tube) and this is it.... For 5687 You can use front part of the pre the same as the rear pair of the triodes.... use the same value in the front as at the rear.... it will work o.k. with it - no problem......hope this helps
regards
daniel

 

[auriontoledo]

Sparkle's instructions on using 5687's

Thank you! Without even referring to the schematic again, I get a very concrete idea of how to do it. You broke it down into a very easy to grasp picture for me.

So I plan to try it out with my 5687's first, then if I find some other nice tubes, I can easily substitute.

BYW, my power amp is a Pass Zen DIY in the Aleph current version. It is supposed to need a preamp with lots of gain. I have the Balanced Bride of Zen as a preamp now, but it just doesn't sound good: too many sharp highs. I've fiddled around, but still not good.

Sometime I'll build a tube power amp as well.

 

[sparkle]

If I would be in Your place I would use 6n6p or 5687 in the second stage and I would run a lot of current through them.... You will need it to drive mosfets in Your power amplifier.... I don't know about the total gain You need but I think that it will be sufficient to use 5687 or similar... You will probably have more than enough gain (again I do not know how much You need - i am to lazy to go and see how much the Bride of Zen has).... Aikido is a VERY,VERY good preamplifier .... it sounds very good and can drive very difficult loads if You have something nice at the output like 6n6p or 5687 in Your case......

best regards



btw.... i am working arround TL431 connected as a shunt in the cathode or as a ccs - don't know which one to use yet.... time and ears will tell......

 

[Mark A. Gulbrandsen]

The sound with the 317 CCS changes markedly depending on the cap used: I think tubelab hit the nail on the head - this type of circuit is great, assuming you have the perfect cap.

Has anyone built and listened to both versions yet? Of course it still has yet to be determined if the CSS sounds best in the plate or in the cathode.... or weather the caps in the cathode sound better than the plate side CSS does and so on. I don't feel that the cathode side CSS is all that cap fussy either... as long as you use good quality multiple parallel caps across the 317 the effect should be about the same. While I use pretty darn good caps I don't use caps that I would consider very exotic.... just rel-caps PPMFX. I ordered some 10M45S's today... perhaps I get to be the first to compare.

Mark

 

[kathodyne]

Hi,

i used a lm317 ccs also on my amp (5965 in ltp driving 6l6gc)
it was said that the lm317 lost its properties above 1-2khz.

so i tried something else, 2xbc547b, biased by two led's.
(see attachment)

The sound did change a bit, but not staggering.

next ccs will be 2xbc547b with the bases connected together....
don't know if it helps.....

 

[ilimzn]

Yes but your topology is completely different so cannot really be compared. In yours, the CCS works at DC and AC, and in his, it is completely bypassed for all usable AC signals, and sets only the DC conditions.