#41 and neutralizing C_Miller
Hi Arne K,
thanks for your reply. I am using the thin satellite dish coax 50pF/m.
In my old RIAA pre I put a resistor and a capacitor in a cinch plug
and put it to an additional cinch input. Some RIAAs have switchable
resistors and caps at the input. Since I did only use the 100pF 50K
I soldered this value into my new RIAA pre. link
In the active RIAA topology discussed here in this thread, it is possible
to neutralize link (comes from link ) the Miller capacitance Cga.
See attached zip below.
There is no need for a cascode to reduce the input capacitance.
The exact values depend in the layout, the better the layout the closer
you come to the calculated values. I think making it without a
printed circuit board provides the best results.
Kind regards,
Darius
Hi Arne K,
thanks for your reply. I am using the thin satellite dish coax 50pF/m.
In my old RIAA pre I put a resistor and a capacitor in a cinch plug
and put it to an additional cinch input. Some RIAAs have switchable
resistors and caps at the input. Since I did only use the 100pF 50K
I soldered this value into my new RIAA pre. link
In the active RIAA topology discussed here in this thread, it is possible
to neutralize link (comes from link ) the Miller capacitance Cga.
See attached zip below.
There is no need for a cascode to reduce the input capacitance.
The exact values depend in the layout, the better the layout the closer
you come to the calculated values. I think making it without a
printed circuit board provides the best results.
Kind regards,
Darius
Hi Darius,
I saw your article on the stereo decoder. Looks sort of interesting, also looked at your Cmiller canceling RIAA circuit - how effective is this in your opinion - how much reduction of Cmiller would you expect, and how is this affected by tube aging induced loop gain reductions? Are you using a gimmick style capacitor (basically twisted wires) or are you actually able to source very small high quality mica caps or similar?
I have seen similar techniques applied in oscilloscope designs, but never considered applying them in a phono stage. I've pretty much always preferred passive EQ. I currently use a single triode connected D3A in the input stage and the Cmiller present is not a problem with my Grado.
Clever idea, and perhaps I owe you an apology for my skepticism. Truthfully I am not a big fan of the 12AX7A, (I think that is obvious from my previous postings ) although at one time I really was, and used them extensively in my own designs. My current favorites are the D3A, 5842, 5687, and a slew of dht - that is basically where I am coming from.
Would you consider starting a new thread relating to your stereo decoder project. I think this might be of considerable interest to others here. I don't read German, and your English is certainly good enough - any possibility of an eventual translation of your article?
I saw your article on the stereo decoder. Looks sort of interesting, also looked at your Cmiller canceling RIAA circuit - how effective is this in your opinion - how much reduction of Cmiller would you expect, and how is this affected by tube aging induced loop gain reductions? Are you using a gimmick style capacitor (basically twisted wires) or are you actually able to source very small high quality mica caps or similar?
I have seen similar techniques applied in oscilloscope designs, but never considered applying them in a phono stage. I've pretty much always preferred passive EQ. I currently use a single triode connected D3A in the input stage and the Cmiller present is not a problem with my Grado.
Clever idea, and perhaps I owe you an apology for my skepticism. Truthfully I am not a big fan of the 12AX7A, (I think that is obvious from my previous postings
) although at one time I really was, and used them extensively in my own designs. My current favorites are the D3A, 5842, 5687, and a slew of dht - that is basically where I am coming from.Would you consider starting a new thread relating to your stereo decoder project. I think this might be of considerable interest to others here. I don't read German, and your English is certainly good enough - any possibility of an eventual translation of your article?
SY said:
SY is right about keeping the electrical character of the 'X7 and similar types firmly in mind, when designing for them. The high RP/ low gm/high mu types present 2 obvious difficulties: a large Miller capacitance and wretched drive capability. The drive capability problem can be dealt with easily enough by DC coupling a ZVN0545A source follower to the triode's anode. I don't have such a quick solution to the Miller capacitance issue. Perhaps Darius could provide additional info. about his "neutrode" concept.
BTW, the ZVN0545A exhibits a tiny reverse transfer capacitance, which makes it well suited to mating with high RP types. ZVN0545A data sheet here.
I'm attaching a schematic for a tweaked version of the "classic" passively EQed RCA phono preamp, in the hope that Darius will apply the neutralizing method to the 1st gain block.
Attachments
SY said:
Hi Sy,
My measurements confirm your comments, I was always able to get distortion numbers of 0.05% thd or better at fairly large voltage swings, (10Vrms or better) and I am well versed in their limitations as well. I've just found most modern ones to be rather inconsistent, and vintage ones to be rather expensive, also in passive phono designs I have found the much higher transconductance of types like the D3A to be advantageous from a noise perspective. (Not that they don't have a whole other set of application issues as you know.) I used to parallel up to 4 sections of 12AX7A or 5751 in order to the get the benefits of higher effective transconductance, as you can imagine the effect on Cmiller was not too good which since I was primarily using high output moving coil cartridges at the time was not a major concern at least with the low capacitance cables I used at the time. The D3A design is at least 20dB quieter than my best 12AX7A based design and has moderate Cmiller on the input.
Very interesting discussions. Thanks all for your contributions. But frankly, I was not expecting such vigorous argument over such a simple little circuit. I'm planning to build a EAR834P replica and Mr. Wright circuit and try them into the Copland. I'll just build other pcb's and use the Copland own supplies. I prefer not to touch the original PCB, to be able to put it back to its original state if needed.
I'm planning to build a EAR834P replica and Mr. Wright circuit and try them into the Copland. I'll just build other pcb's and use the Copland own supplies. I prefer not to touch the original PCB, to be able to put it back to its original state if needed.Hello Eli Duttman,
in your topology it is not possible to neutralize in the way I showed it
in the copland topology. See #42. In your topology the input capacitance
is low, ( 6n2p: Cg1a ≈ 0p7F
; ECC83: ≈ 1p6F) and it remains constantwith the frequency. Thus there is no need for a neutralisation.
Kind regards,
Darius
#43 neutralisation
Hello Kevin,
thanks for your reply. For the neutralisation you need an inverted signal.
The phase and the gain of the inverter should be constant.
This neutralisation is not intended to prevent tgtp oscillation.
Thus there is no need for a perfect neutralisation and a difficult alignment.
This neutralisation is intended to reduce the input capacitance to say 20pF,
stray capacitance not added.
I would recommend to take the calculated values, this prevents an over
neutralisation.
There are three things that set the gain of the triode inverter.
First the transconductance,
second the µ of the triode
and third the negative (current) feedback.
The transconductance has an aging. Thus make sure that the transconductance
has a low influence in the gain.
I cut a piece ot of double side copper plated epoxy PCB as a neutralizing cap.
Kind regards,
Darius
TGTP: Tuned Grid Tuned Plate, Huth-Kühn
PCB: Printed Circuit Board
Hello Kevin,
thanks for your reply. For the neutralisation you need an inverted signal.
The phase and the gain of the inverter should be constant.
This neutralisation is not intended to prevent tgtp oscillation.
Thus there is no need for a perfect neutralisation and a difficult alignment.
This neutralisation is intended to reduce the input capacitance to say 20pF,
stray capacitance not added.
I would recommend to take the calculated values, this prevents an over
neutralisation.
There are three things that set the gain of the triode inverter.
First the transconductance,
second the µ of the triode
and third the negative (current) feedback.
The transconductance has an aging. Thus make sure that the transconductance
has a low influence in the gain.
I cut a piece ot of double side copper plated epoxy PCB as a neutralizing cap.
Kind regards,
Darius
TGTP: Tuned Grid Tuned Plate, Huth-Kühn
PCB: Printed Circuit Board
# 43 OT
Hi Kevin,
I never build this so called Pass topology. In my RIAA 2001 I used an EF40 triode connected
and an ECC81. The EF40 is a linear preamp. The first section of the ECC81 makes RIAA eq.
and the second section is used as a cathode follower and bootstraps the first section of the ECC81.
It depends in the circuit which tube I use. My preamp favourites are AU/AT/AX ECC82/81/83 and
similar tubes.
Do you help me to translate german english to oxford english in the decoder blog, please?
Kind regards,
Darius
Originally #43 posted by kevinkr
Hi Darius,
...
I have seen similar techniques applied in oscilloscope designs, but never considered applying them in a phono stage. I've pretty much always preferred passive EQ. I currently use a single triode connected D3A in the input stage and the Cmiller present is not a problem with my Grado.
Clever idea, and perhaps I owe you an apology for my skepticism. Truthfully I am not a big fan of the 12AX7A, (I think that is obvious from my previous postings
Would you consider starting a new thread relating to your stereo decoder project. I think this might be of considerable interest to others here. I don't read German, and your English is certainly good enough - any possibility of an eventual translation of your article?
Hi Kevin,
I never build this so called Pass topology. In my RIAA 2001 I used an EF40 triode connected
and an ECC81. The EF40 is a linear preamp. The first section of the ECC81 makes RIAA eq.
and the second section is used as a cathode follower and bootstraps the first section of the ECC81.
It depends in the circuit which tube I use. My preamp favourites are AU/AT/AX ECC82/81/83 and
similar tubes.
Do you help me to translate german english to oxford english in the decoder blog, please?
Kind regards,
Darius
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.