Considering these two phono stages
Adam Stouffer
these two phono stages?
http://i88.photobucket.com/albums/k162/astouffer/pre_amp_6dj8_srpp.gif
http://margo.student.utwente.nl/el/tubes/schema/preamp/fb-e180f.gif
I am partial to the second link because of its simplicity and how long
it seems to be around on the internet. The only part that looks odd is
the 33m resistor in the feedback loop. Would leaving it out even cause a
measurable difference :)
Adam
Eeyore
Adam Stouffer wrote:
God Yes ! It's part of the RIAA eq.
The second one is likely to be quite noisy too. The configuration is rubbish.
Graham
Chris Hornbeck
<adam_s...@hotmail.com> wrote:
>http://margo.student.utwente.nl/el/tubes/schema/preamp/fb-e180f.gif
>
>I am partial to the second link because of its simplicity and how long
>it seems to be around on the internet. The only part that looks odd is
>the 33m resistor in the feedback loop. Would leaving it out even cause a
>measurable difference :)
Looks like a typo; 3.3M is more likely. Note that anode
follower feedback interacts with source impedance. Unless
the phono cartridge is low or resistive impedance it
will affect frequency response. And the 51K will make
noise.
All good fortune,
Chris Hornbeck
Eeyore
> On Sun, 29 Oct 2006 20:40:19 GMT, Adam Stouffer
> <adam_s...@hotmail.com> wrote:
>
>
>> http://margo.student.utwente.nl/el/tubes/schema/preamp/fb-e180f.gif
>>
>> I am partial to the second link because of its simplicity and how long
>> it seems to be around on the internet. The only part that looks odd is
>> the 33m resistor in the feedback loop. Would leaving it out even cause a
>> measurable difference :)
>>
>
> Looks like a typo; 3.3M is more likely.
> Note that anode
> follower feedback interacts with source impedance. Unless
> the phono cartridge is low or resistive impedance it
> will affect frequency response. And the 51K will make
> noise.
The 51k is needed - it's part of the RIAA eq.
Graham
Chris Hornbeck
<rabbitsfriend...@hotmail.com> wrote:
>> Looks like a typo; 3.3M is more likely.
>No. It should indeed be 33M
Doh! You're right. For infinite open-loop gain, it'd
be about 11.8M. Operating this close to open loop, I'd just
leave it out; the real value needed will be critically
dependent on actual open-loop gain. It sets the 50 Hz pole.
I'd also reduce mid-band gain from about 46dB to something
more managable, maybe 35 to 40dB.
>> Note that anode
>> follower feedback interacts with source impedance. Unless
>> the phono cartridge is low or resistive impedance it
>> will affect frequency response. And the 51K will make
>> noise.
>The 51k is needed - it's part of the RIAA eq.
It's part of the feedback network, but it's in series
with the input for noise purposes.
FWIW, this circuit, requiring an open loop gain of 66dB
(a factor of 2000!) in a single stage, is whack. It also
requires a device that probably *cannot* be made anymore.
Much thanks, as always,
Chris Hornbeck
Eeyore
Chris Hornbeck wrote:
> On Mon, 30 Oct 2006 00:23:39 +0000, Eeyore
> <rabbitsfriend...@hotmail.com> wrote:
>
> >> Looks like a typo; 3.3M is more likely.
> >No. It should indeed be 33M
>
> Doh! You're right. For infinite open-loop gain, it'd
> be about 11.8M. Operating this close to open loop, I'd just
> leave it out; the real value needed will be critically
> dependent on actual open-loop gain. It sets the 50 Hz pole.
>
> I'd also reduce mid-band gain from about 46dB to something
> more managable, maybe 35 to 40dB.
>
> >> Note that anode
> >> follower feedback interacts with source impedance. Unless
> >> the phono cartridge is low or resistive impedance it
> >> will affect frequency response. And the 51K will make
> >> noise.
> >The 51k is needed - it's part of the RIAA eq.
>
> It's part of the feedback network, but it's in series
> with the input for noise purposes.
It is which is rather bad news.
> FWIW, this circuit, requiring an open loop gain of 66dB
> (a factor of 2000!) in a single stage, is whack.
I only make it 666x ( 34M / 51k )
> It also requires a device that probably *cannot* be made anymore.
Certainnly no pentode is going to manage that. It would be bass light.
Graham
Adam Stouffer
> On Mon, 30 Oct 2006 00:23:39 +0000, Eeyore
> <rabbitsfriend...@hotmail.com> wrote:
>
>
>>>Looks like a typo; 3.3M is more likely.
>>
>>No. It should indeed be 33M
>
>
> Doh! You're right. For infinite open-loop gain, it'd
> be about 11.8M. Operating this close to open loop, I'd just
> leave it out; the real value needed will be critically
> dependent on actual open-loop gain. It sets the 50 Hz pole.
>
My thinking is that whatever voltage is across that 33m resistor is
going to be so low that leaving it out would have no audible effect. Try
even locating a 33m resistor thats not a glass high voltage type. Sure
you could probably use a 20m and 13m.
> I'd also reduce mid-band gain from about 46dB to something
> more managable, maybe 35 to 40dB.
>
Whats the best way to go about that?
>
>>> Note that anode
>>>follower feedback interacts with source impedance. Unless
>>>the phono cartridge is low or resistive impedance it
>>>will affect frequency response. And the 51K will make
>>>noise.
>>
>>The 51k is needed - it's part of the RIAA eq.
>
>
> It's part of the feedback network, but it's in series
> with the input for noise purposes.
Doesn't it also keep the input impedance up?
>
> FWIW, this circuit, requiring an open loop gain of 66dB
> (a factor of 2000!) in a single stage, is whack. It also
> requires a device that probably *cannot* be made anymore.
>
> Much thanks, as always,
>
> Chris Hornbeck
Heres a link to the E180F specs:
http://www.mif.pg.gda.pl/homepages/frank/sheets/009/e/E180F.pdf
I just bought two Tesla E180F on ebay for $6 each. Hey its a cheap
gamble and better than the phono stage I have now (nothing)
Adam
Chris Hornbeck
<adam_s...@hotmail.com> wrote:
>> I'd also reduce mid-band gain from about 46dB to something
>> more managable, maybe 35 to 40dB.
>>
>
>Whats the best way to go about that?
Doh! AGAIN! You'd probably do best to just ignore me;
I've been making so many stupid blunders.
Mid-band (between the 500 Hz zero and the 2122 Hz pole)
gain is set by the ratio of the 1M feedback resistor and
the sum of the 51K resistor and source impedance. IOW, about
26dB with a very small source impedance. Ignore, as
possible, my factor-of-10 errors.
You'll probably need to select the "33M" resistor value
for a particular valve and loading, to make the 50 Hz
pole 17dB up from mid-band. Loading cannot be ignored
here, especially with a 100K ohm plate load. And
connection to typical solid-state impedances will be
a real problem.
>> It's part of the feedback network, but it's in series
>> with the input for noise purposes.
>
>Doesn't it also keep the input impedance up?
It *is* the input impedance. But it's also in series
with the signal and contributes noise proportionally.
>I just bought two Tesla E180F on ebay for $6 each. Hey its a cheap
>gamble and better than the phono stage I have now (nothing)
Tesla, as in modern manufacture? I'm quite impressed;
these high Gm bottles require a high level of precision
to make. Very cool that somebody can still do it.
Very good fortune,
Chris Hornbeck
Ian Bell
> Anyone care to stop the name calling for a few minutes and comment on
> these two phono stages?
>
> http://i88.photobucket.com/albums/k162/astouffer/pre_amp_6dj8_srpp.gif
>
> http://margo.student.utwente.nl/el/tubes/schema/preamp/fb-e180f.gif
>
The first design will probably be seriously coloured. The second, using a
single stage with feedback may well run out of open loop gain which will
again lead to colouration. If you like that sort of thing I guess it is not
an issue.
IAn
Eeyore
Adam Stouffer wrote:
> Chris Hornbeck wrote:
> > On Mon, 30 Oct 2006 00:23:39 +0000, Eeyore
> > <rabbitsfriend...@hotmail.com> wrote:
> >
> >>>Looks like a typo; 3.3M is more likely.
> >>
> >>No. It should indeed be 33M
> >
> > Doh! You're right. For infinite open-loop gain, it'd
> > be about 11.8M. Operating this close to open loop, I'd just
> > leave it out; the real value needed will be critically
> > dependent on actual open-loop gain. It sets the 50 Hz pole.
>
> My thinking is that whatever voltage is across that 33m resistor is
> going to be so low that leaving it out would have no audible effect.
That's not exactly very scientific thinking ! It's part of the RIAA EQ !
> Try
> even locating a 33m resistor thats not a glass high voltage type. Sure
> you could probably use a 20m and 13m.
>
> > I'd also reduce mid-band gain from about 46dB to something
> > more managable, maybe 35 to 40dB.
> >
>
> Whats the best way to go about that?
Scale the feedback component values.
> >>> Note that anode
> >>>follower feedback interacts with source impedance. Unless
> >>>the phono cartridge is low or resistive impedance it
> >>>will affect frequency response. And the 51K will make
> >>>noise.
> >>
> >>The 51k is needed - it's part of the RIAA eq.
> >
> > It's part of the feedback network, but it's in series
> > with the input for noise purposes.
>
> Doesn't it also keep the input impedance up?
It actually sets the input impedance ! At least it's the major factor in it.
Graham
Eeyore
flipper wrote:
> Since we're looking at tube RIAA preamps, what about this one?
>
> http://www.geocities.com/rjm003.geo/rjmaudio/diy_pho3.html
>
> http://www.geocities.com/rjm003.geo/rjmaudio/images/p6dj8_cir.gif
That looks a heck of a lot better.
Graham
Patrick Turner
Adam Stouffer wrote:
The first schema will be noisy but have plenty of gain and suit
high output MM only, and definately not MC carts.
The second schematic with E180F is likely to be also noisy and harder to
hand trim the values of R&C to get a correct RIAA response, but
the gain should be sufficient because the tube s much higher gm than an
EF86.
Quad and Leak and others used single EF86 for their phono stages with NFB
and
since the output voltage required was less than 100mV for following amps all
was OK
if the cart was high output, say above 3mV at 1 kHz.
Neither of the two schematics or anything by Leak or Quad 22 etc
is any good for MC as is. For MC with such amps you need a 1:10 step up
tranny or much lower input noise
and for far better phono circuits check out my preamp pages at
http://www.turneraudio.com.au
The step up tranny of 1:10 gives a 100 impedance change but since the Zout
of many MC is
less than 20ohms, then the Rout at the secondary = 2,000 ohms and low enough
to drive into the
51k of input resistance of the EF180 circuit without having to worry about
the
effects of the higher Zout of many MM carts which will be interactive with a
shunt NFB
input circuit.
And after the EF180, a direct coupled triode Cathode follower buffer would
work well to maintain the
pentode gain uniformly so the FB values don't load down the pentode at HF.
For the FB to work more predictably around a given tube gain stage the use
of a pair of 12AX7 cascaded triodes is better and with a buffer CF after the
second AX7 triode,
or use a ľ follower stage like I do in various versions of amps at my site.
Using the E180F as a triode in cascode with 2SK369 j-fet as a cascode driver
input device
with all passive RIAA and ľ-follower second gain stage is an excellent low
noise
solution; the more effort you put in the better the recortds will sound.
Patrick Turner.
Patrick Turner
flipper wrote:
> On Sun, 29 Oct 2006 20:40:19 GMT, Adam Stouffer
> <adam_s...@hotmail.com> wrote:
>
> Since we're looking at tube RIAA preamps, what about this one?
>
> http://www.geocities.com/rjm003.geo/rjmaudio/diy_pho3.html
>
> http://www.geocities.com/rjm003.geo/rjmaudio/images/p6dj8_cir.gif
These would be easier to get working correctly than the pentode circuits,
and also quieter.
Patrick Turner.
Eeyore
flipper wrote:
> On Mon, 30 Oct 2006 09:53:14 +0000, Eeyore
> <rabbitsfriend...@hotmail.com> wrote:
> Great. Thanks.
>
> I kind of like the passive filter approach.
There was a neat circuit Quad used that had a bit of both actually. It avoids
the interaction common in most implementations.
Of course I find myself wondering the lathe manufacturers used. Were their RIAA
curves that accurate ?
Graham
Patrick Turner
flipper wrote:
> Thanks for the comment. How much noise would you expect?
> Objectionable? How about compared to a shure m65 12AX7 version?
With most MM carts with 3mV of rated output at 1kHz, a non modulated groove of
an LP
will produce noise that is usually just above the noticeable
noise of an average triode preamp. Engineers usually seek to have the noise of
the background
concert hall or venue as low as possible so one lone quietly played violin is
well above the noise, but in practice vinyl rarely allows the playback
experience of
low level signals to sound as quiet as the original concert/recording studio
experience if you were present.
Either the amp hums and/or hisses a bit or the vinyl is noisy.
But with rock and jazz it matters not.
Vinyl is good when the SNR is -55dB below a quietly played violin.
Noise concerns are all about its noticeability with vinyl and tape and other
analog sources.
A moving magnet phono amp with gain of 100x or 40db at 1khz
should be tested with a shorted input grid to ground so that the typical loading
of 47k at the input has its noise shunted to 0V.
The noise at the output should be less than 1/1,000 of output level, unweighted.
So that if you have a cart that lists its output at 5mV at 1kHz, then output
voltage
at 1khz if gain = 100 at 1kHz = 500mV = 0.5Vrms, and noise should be < 0.5mV.
The SNR is then said to be -60dB.
With MC, typical cart output may be only 0.3mV, so if gain = 100x,
then output is only 30mV, but the tube noise will be the same at 0.5mV and
SNR is a poor -35dB.
So with such an MC, the SNR of the input device which is ALWAYS what determines
the
SNR in most microphone and phono amps must be better, and using a 2SK369 or
2SK147
j-fet will achieve a dramatic improvement in SNR with MC low level signals.
A typical well chosen 12AX7 with its grid shunted to 0V will manage to
produce about 2uV of noise at its input and this is unavoidable with either NFB
or passive
preamps because this noise which is called grid input noise is outside the
the FB loop and in series with whatever follows.
The gain of a pair of cascaded 12AX7 halves can be say 60 x 70 = 4,200,
and 2uV of noise across a 40Hz to 20kHz bandwidth woukld be amplified
to 8.4mV were it not for the effect of the RIAA filter which reduces the noise
to acceptable levels of less than a mV at the output.
However, the noise of the input triode is often dominated more by low frequency
noises and the gain of a phono amp is highest at LF so you can get
lots of rumbly noise with tube phono amps as well as hum if you don't
choose the tubes carefully and you have a low output cartridge.
>
>
> I've been seriously thinking of building that one as I've got the
> 6DJ8s (not by accident ;))
>
Well then stop being lazy, build the darn amp and find out all about noise and
good vinyl replay.
>
> Spice simulations shoving an RIAA into it gives a bell curvish
> response down about 0.6dB at both 20Hz and 20kHz. And looks pretty
> much the same even at 140V B+.
Forget Spice. Get away from the PC and out to your
workbench. Make a premphasis cutting head amp filter.
Use that between the output from your sig gene and any phono amp you build.
The response of sine waves at the output after RIAA eq can then be forced to be
flat by
careful hand trimming of the values.
Check your work to proove its right to yourself.
Then and only then will you have good vinyl replay.
Its surprising how many old amps I fix that have +/- 3db along the band
and with one channel different to the other.
>
>
> I would have though reverse RIAA would be rather 'textbook' but
> looking at different approaches I'm running into all kinds of
> variations with the 'new' curve being down 10dB at 20Hz on some
> 'application note' circuits, like AND8177-D (an opamp design) from ON
> Semiconductor, essentially a duplicate of 1988 AN142 from Philips.
Reverse EQ and practical hands on is the ONLY way.
Good phono stages cannot be achieved by designing and building
without testing as you go.
Capacitance value variations and your inexperience will trip you up
every time.
The only dependable thing would be resistance values which can easily
be attained within 1%, except that RIAA values are rarely all standard values
and have to be made up by the DIYer with parallel/series arrays of R.
>
> And an opamp quickie I built from a web site was down 4dB at 20Hz but
> up 1.8dB at 20kHz. I've since modified it from at hand parts to about
> -2dB at 20Hz and then essentially flat to 20kHz.
The CRO can spot about 0.5dB, but a good wide band millivolt meter
will help.
>
>
> Anyone have an idea what those 15 buck el cheapos you can buy do? Not
> that I'm going to buy one but my bother in law did and he thinks it
> stinks, especially the low end, so I'm curious if they do the 'new'
> RIAA.
>
> I'm thinking of giving him my opamp version for Christmas and make the
> 6DJ8 version for myself.
You have a tolerant BIL.
Or he is partially deaf.....
Patrick Turner.
>
>
> >Patrick Turner.
> >
Patrick Turner
Eeyore wrote:
RIAA eq probably was NOT exact amoung all the recording companies.
If one collects all the available reverse and inverse RIAA networks and tests them,
one will find considerable response variation and inconsistencies.
Many networks are seen, there is scant information which
gives the exact tested guranteed response of the networks shown, so you are on your
own as a DIYer
and you must test to make sure any filtering one does is correct
and never assume anything.
If the response was +/-2db due to such analomies in cutting head amps it mattered
little
to most studios because most music is pop so who fukkin cares?
The eq imposed upon the music during post recording and processing
will be hugely greater than the tolerance variations of RIAA cutting amp filters
but one wouldn't want to make matters worse by having ones own phono amp
boosting, cutting, and phase shifting unecessarily, and with channel in-balance
which will
make imaging poor.
Patrick Turner.
Patrick Turner
flipper wrote:
> On Wed, 01 Nov 2006 10:16:22 +0000, Eeyore
> <rabbitsfriend...@hotmail.com> wrote:
>
> >
> >
> >flipper wrote:
> >
> >> On Mon, 30 Oct 2006 09:53:14 +0000, Eeyore
> >> <rabbitsfriend...@hotmail.com> wrote:
> >> >flipper wrote:
> >> >> <adam_s...@hotmail.com> wrote:
> >> >>
> >> >> >Anyone care to stop the name calling for a few minutes and comment on
> >> >> >these two phono stages?
> >> >> >
> >> >> >http://i88.photobucket.com/albums/k162/astouffer/pre_amp_6dj8_srpp.gif
> >> >> >
> >> >> >http://margo.student.utwente.nl/el/tubes/schema/preamp/fb-e180f.gif
> >> >> >
> >> >> >I am partial to the second link because of its simplicity and how long
> >> >> >it seems to be around on the internet. The only part that looks odd is
> >> >> >the 33m resistor in the feedback loop. Would leaving it out even cause a
> >> >> >measurable difference :)
> >> >> >
> >> >> >
> >> >> >Adam
> >> >>
> >> >> Since we're looking at tube RIAA preamps, what about this one?
> >> >>
> >> >> http://www.geocities.com/rjm003.geo/rjmaudio/diy_pho3.html
> >> >>
> >> >> http://www.geocities.com/rjm003.geo/rjmaudio/images/p6dj8_cir.gif
> >> >
> >> >That looks a heck of a lot better.
> >> >
> >> >Graham
> >> >
> >>
> >> Great. Thanks.
> >>
> >> I kind of like the passive filter approach.
> >
> >There was a neat circuit Quad used that had a bit of both actually. It avoids
> >the interaction common in most implementations.
>
> Yes, both approaches have their advantages and disadvantages. I wonder
> if mixing them gives you the best or worst of both ;)
>
> Got a schematic?
>
> >Of course I find myself wondering the lathe manufacturers used. Were their RIAA
> >curves that accurate ?
>
> I don't know but I'd imagine it's like most things, some do it well
> and some don't.
>
> Doesn't really matter to a preamp, though, does it? I mean, botching
> it on the tail end wouldn't 'fix' more than the wildest stroke of luck
> coincidental botch on the other but make the rest, especially whoever
> did it well, worse. Or so it seems to me.
>
> But it's kind of hard to tell as there seems to be a pretty wide
> variation in the various implementations.
>
> And then you've got sources talking about the 'missing' corner at
> 50kHz.
The cutting head amps often use NFB loops to boost the frequencies above 50Hz
thus limiting the space needed to contain the wide amplitudes of bass waves cut into
records.
This gave records a lon play time. The "emphasizing" process also
assisted the SNR on replay so that vinyl was able to be a great medium for so long.
The HF boost done with FB could not infinitely boost HF at 6dB/octave above 20kHz.
If that were possible, any stray HF or oscillations asociated with over zealous
HF gain could cause cutting head burn out or overload the cutting amp
of stuff a good cut.
So people cutting records more or less agreed that boosting so that there was a pole
at 50kHz
was a sensible thing and placing a plateau of cutting head amp gain at the 50kHz
pole
wasn't likely to be audible of cause noticeable audio F roll off.
So when you have an RIAA filter in a preamp there is no need to have an infinitely
rolling
off HF cut; a shelf beginning at a pole at 50kHz is OK. Where NFB
is used for the HF cut, such a shelf which is simply created by placing a resistance
in series with the RC network giving 75uS so that an additional resistor giving a
time constant of about 3uS
will aid the stability of the amp circuit involved, and avoid having a final
capacitance between output
and input of the feedback amp. In RIAA amps where NFB is used for the RIAA eq, the
RIAA
eq network LOADS THE AMPLIFIER and must be considered carefully
as to what effect it may have on open loop gain and HF stability.
Some would say that the additional 50kHz pole in the phono amp
makes the sound better because the phase shift due to the extra TC
better complements the emphasizing process used in the cutting amp.
In other words fidelity is better above 5kHz.
I always include the extra TC, even in passive RIAA networks which are all i
ever seem to need to use.
Shunting out the extra R needed for the 3uS TC doesn't make a huge difference
to the sound I have to say but then my ears are older than more critical listeners.
With passive networks there is no worry about HF stability caused by FB and
capacitor loading phase shifts.
Nevertheless, for anyone who takes my advices about using j-fets and triodes in
hybrid
phono input stages that ARE VERY DARN GOOD had better get used having leads short
and all bypassable things including heater wiring all treated like the circuit
designed for RF up to 500MHz, because with high gm j-fets and triodes
they do try to oscillate at RF unless carefully designed NOT to oscillate.
And if they do oscillate, sometimes they squeg, and an audio tone is mysteriously
generated because the amp is oscillating with RF bursts that start and stop due to
short time saturation
that have an audio frequency.
So using ceramic bypass caps and short leads is a must in such circuits.
And one must KNOW for sure that your nice phono stage is NOT oscilating
because some oscillate a bit while also amplifying the audio.
What can oscillate does oscillate, every time, in an electronic circuit
because what may appear to be an audio circuit may turn into an RF circuit
due to stray inductances in wires and caps at a high enough frequency.
Schematics at my site of hybrid phono input stages do not
convey the means by which layout of the input stage is confined to a
board area 50mm x 40mm, and where leads are short thick copper wires.
Using a 2SK369 with a 6EJ7 strapped in triode took some clever
layout work to prevent HF oscillations.
I may add that i tried a cascode circuit with 2SK369 powering the cathode
of the 6EJ7 while trying to keep it working in pentode mode rather than triode
and the circuit was grossly microphonic, and it oscillated at RF and squegged
viciously
despite all attempts to stop it.
There are many ways to make a phono amp.
Good luck with whatever you try.
Patrick Turner.
>
>
> >Graham