Has somebody any idea? I tried some things as lowering the bleed resistor, (gain and noise lowered slightly) bypassing lower ccs with capacitor ( hum increased dramaticly), but to no avail.I mean, it should be possible to amplify 50mV without to much noise with a d3a.
People are using it for mc phono stages, with less than 1mV input...
The most frustrating thing is after i changed the coupling cap from silver mica to teflon, the sound changed from great to superb... If it werent for the noise I would be in audionirvana....
Last edited:
That was only for the photo; I had them nearby. (very perceptive though)I have four original siemens d3a in there now; I wil try using them for a few days, if thats the cure, it is an easy one.
Hmm.. I have 2 d3a's in my power amplifier for a few years; I could swap them....
thanks!
Paul
Last edited:
Ok I just found confirmation, ltp with active load, being a ccs or current mirror, has high output impedance. I'm dont think I can use my schematic in this form. I think it explaines the slight hum,and would be very hard to get rid of.
I think i keep the pcb, because it would be a nice phase splitter... Maybe i will use it in a balanced poweramp.
Meanwile i'm going to re-examine the normal Ltp, and check how that performs.
I think i keep the pcb, because it would be a nice phase splitter... Maybe i will use it in a balanced poweramp.
Meanwile i'm going to re-examine the normal Ltp, and check how that performs.
Last edited:
Something that bothers me though. I tried to find formula's for calculations on ltp's with ccs in tail.
I couldnt find any! Can someone help me with that?
I'm interested in output impedance calculations. With a non ccs ltp its Ra parallel to ra.
They describe it as plate load resistor parallel to the resistance you see as you look into the plate. But what would that be with the ccs below cathode?
I couldnt find any! Can someone help me with that?
I'm interested in output impedance calculations. With a non ccs ltp its Ra parallel to ra.
They describe it as plate load resistor parallel to the resistance you see as you look into the plate. But what would that be with the ccs below cathode?
Hello again.
No, 30k is too high and surely not correct.
See Morgan Jones Valve Amplifiers 3rd Edition pp 129/130 - where he points out that in a diff pair when one output is loaded "output resistance rises considerably'
Is the anode resistance of your D3As actually 1.8k at your operating point? If so then these resistances are still in parallel with the load resistances.
Finally if all else fails, a nice cathode follower will solve your problems; with Rin of say 20 meg who cares about output resistance!
Best o luck
No, 30k is too high and surely not correct.
See Morgan Jones Valve Amplifiers 3rd Edition pp 129/130 - where he points out that in a diff pair when one output is loaded "output resistance rises considerably'
Is the anode resistance of your D3As actually 1.8k at your operating point? If so then these resistances are still in parallel with the load resistances.
Finally if all else fails, a nice cathode follower will solve your problems; with Rin of say 20 meg who cares about output resistance!
Best o luck
Absolutely! As mentioned above, a follower, albeit source, is already in the circuit
Hmm, 7N7 asleep at the switch as usual - whoops...
Yes, I really don't see that the CCS should affect the anode resistance in a differential pair - consider that the perfect CCS presents an infinite impedance to the joined cathodes and thus with a diff pair that is operating correctly anode resistance will not be affected.
Output resistance though as I suggested above wil be affected if the stage is overloaded.
P.
Thanks all you guys for helping me!
I tried the output above the sourceresistor of the mosfet already; in fact I tried it even before your post but forgot the put it on DIYaudio.
I measured the outputimpedance yesterday, and it was a lot better, but still not as low as I would like. The outputimpedance is ~10K. (isnt this called a mu follower?)
It has to drive my line-in from my tube power amp, (100k input impedance)
I will remove the 100k and the 0.1 uF; it put it there to copy the input load of the power amplifier, so the ltp gets "symmetrically" loaded on both branches. I understand now that the load gets distributed across both tubes anyway.
@7N7 and Oshifis: I know that a cathode follower could be a possible solution, but I want to keep the schematic as simple as possible; I dont want another tube if its not absolutely neccesary. Ony of the reasons I choose a d3a is because of its very low plate resistance of 1.9K. If a normal LTP output impedance is Ra parallel ra then I thought that I would get an output impedance of ~2k.
The problem is this I think: (maybe I'm completely wrong though)
Output impedance is ra // Ra:
load resistor // plate resistance if you look into the plate.
And its that last wich I fear is to ground; and in series with the plate is the ccs wich has a very large resistance. If you have the normal LTP with only a ccs in the cathode, and the plateload resistor is connected to the powersupply, wich is a low impance path for ac/signal to ground.
So in a normal LTP, its high impedance plate to ground resistance(100's of Kohm) in parallel with plateload resistor (10Kohm in my case) is ~ 9K output resistance.
In my active loaded LTP there is a CCS wich behaves as a very large resistance below AND above the tube.
Here it is 100k's // 100k's; so it leaves me with a very large output impedance and a linestage very sensitive to hum. The testleads of the multimeter alone were enough to inject a lot of 50 Hz rubbish. With low output impedance that would be much better.
But its all theory based on (possibly wrong) assumptions.
Thats why I wanted to see some formulas of LTP's with CCS's in the tail; they would give me some guidance if i'm on the right track.
I'm gonna try it anyway; convert it to ccs in the tail only, and measure output impedance.
The output impedance was measured as follows:
By scope; first open circuit, then a potmeter connected between output and ground, and adjust till output is half. Then measure the potmeter.
Value of potmeter=output impedance.
Last edited:
They describe it as plate load resistor parallel to the resistance you see as you look into the plate. But what would that be with the ccs below cathode?
Are you shure? So a small resistor (100 ohm or so) in the cathode line or a ccs (100Kohm or so) dont change output impedance? Or to say the same in another way: with a ccs in the cathode circuit the "plate to ground resistance" doesnt change? What am I missing?
Are you shure? So a small resistor (100 ohm or so) in the cathode line or a ccs (100Kohm or so) dont change output impedance? Or to say the same in another way: with a ccs in the cathode circuit the "plate to ground resistance" doesnt change? What am I missing?
Distortion looks good to; had a strange "pumping" effect in the FFT, but it measured almost the same for the ccs side as the normal side. (I converted the left channel to "normal"LTP operation, and left the right side with the 3 ccs's so I could easy compare them.)
I'm going for a quick listening test now...
I'm going for a quick listening test now...
I'm listening right now...the first listening test showed a huge differnce between the normal and the triple ccs version. No noise, much less hum, perfect. Good sound too, so far as you can hear with two different amps for left and right.
I then rushed upstairs to convert the other channel too: and back downstairs...
Last of the hum is now dissapeared too. No noise audible at all.. It sounds incredible....
I'm listening to old records like they are new. Voices are so incedible....
Harry Connick jr. From the cd blue light red light, the song "if i could give you more" never sounded like this. Instruments are really 3d separate from each other. Depth and stereo landscape is amazing...
I then rushed upstairs to convert the other channel too: and back downstairs...
Last of the hum is now dissapeared too. No noise audible at all.. It sounds incredible....
I'm listening to old records like they are new. Voices are so incedible....
Harry Connick jr. From the cd blue light red light, the song "if i could give you more" never sounded like this. Instruments are really 3d separate from each other. Depth and stereo landscape is amazing...
Well done, though I must admit I find this quite baffling. If the anode load was say 1 meg and the valve's anode resistance at your operating point was say 2k then the resultant must surely have been less?
As far as I can remember, I had no such trouble with the prototype I built; in those days I had a 'scope and would have seen any extraneous rubbish.
Anyway, happy listening.
P.
As far as I can remember, I had no such trouble with the prototype I built; in those days I had a 'scope and would have seen any extraneous rubbish.
Anyway, happy listening.
P.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.