His Master's Noise: A Thoroughly Modern Tube Phono Preamp

Jump to Latest
#562
Read through this whole thread and really enjoyed it.. Not to mention the HMN design itself! 🙂 Boy a few times I did read that first post.. Thanks a mill Sy for your contribution!
Inspiration implanted, parts are on the way (using Siemens E280F second generation + Tesla E88CC no.32), hope to build it in 1-2 weeks. Curently having an Ortofon X1-MC to start with which is a HOMC at about 2mV.

Now I have a question: Is it mandatory to use the DN2540 N channel depletion MOSFET? These buggers are nowhere to be found in my vicinity and I was thinking about simply using either a BJT-NMOS or a BJT-BJT CCS of course both implemented with low noise transistors matching the operating voltages and currents.
Would this make any significant difference..?

Thanks & regards,

Zsolt
 
#563
jerrybenko said:
Now I have a question: Is it mandatory to use the DN2540 N channel depletion MOSFET?
Click to expand...

The depletion MOS has impedance in the tens of megOhms, but you could just as use a resistor. The advantage of DMOS is that it offers tremendous PSRR. If you use a nicely regulated power supply you can achieve much the same result.

I think SY was trying to make a point, but he has left us wandering in the desert.
 
#564
I just happened to run into Stuart ;-). He says:

Yes, it helps the PSR, but it also makes the biasing automatic. Since Merlin's paper came out on noise optimizing, if I were to do it again, I'd drop the current of the first stage to 10mA and the second stage to 5mA. The cathode follower could continue at 10mA.

The use of depletion mode MOS makes setup of the CCS easier, but if one wants to work out a new biasing scheme, other CCS configurations can work, there's nothing magic about the depletion mode cascode CCS other than simplicity. FWIW, Mouser has DN2540 at about a buck each, in stock.

Jan
 
Last edited:
#565
Many thanks Jack, Jan (and Stuart - remotely 🙂 )

For some reason it's difficult to find N-channel depletion mode mosfets around here (Hungary) and I try to avoid shipping from abroad as that takes time and I'm impatient having all esle in my possession.. 😀 But I'm gonna check again for DMOS transistors here and - duhh - order online if need be. I mean this design deserves to make no compromise. 🙂

In the relevant datasheets I see that the internal plate resistance of the E280F is 1k8 and the D3a has 1k9 (both triode strapped). Should this difference worry me in terms of driving the the RIAA network?
 
#570
Let's say you want to use the 20k resistor, but have a tube with anode resistance of a different value. Here are some fundamental relations.

R1/R2= 6.877
C1/C2= 2.916
C1= T4/R2 = 318e-6/R2

If you have a tube with Rp of 10k, R1 has become Rp + 20k = 30k, thus:
R2 4.36k,
C1 72.9nF,
C2 25nF
 
#580
jerrybenko said:
Many thanks Jack, Jan (and Stuart - remotely 🙂 )

For some reason it's difficult to find N-channel depletion mode mosfets around here (Hungary) and I try to avoid shipping from abroad as that takes time and I'm impatient having all esle in my possession.. 😀 But I'm gonna check again for DMOS transistors here and - duhh - order online if need be. I mean this design deserves to make no compromise. 🙂

In the relevant datasheets I see that the internal plate resistance of the E280F is 1k8 and the D3a has 1k9 (both triode strapped). Should this difference worry me in terms of driving the the RIAA network?
Click to expand...


But with CCS low mu Total Ri = Ri / 4 more or less Real Ri E280F for RIAA 1k9 / 4 to match R1/R2= 6.877
 
You must log in or register to reply here.
Top Bottom