monica usb jfet mod

[kkerwin]

hi all,
i did the jfet buffer mod on the monica usb using the 2sk170's.  i've run it for a good period of time and along with noticing a much higher level of detail, more life and music forward, i am also hearing very strong ssss's and even ffff's have some "fuzz" about them.  i too have the white noise at higher volumes. 
based on what i've read in previous posts, it sound like the silver mica (430pf) capacitor across the jfet buffer should relieve this issue. agreed?  and visually, the capacitor connects to the outermost legs of the two jfets that are joined together? 

thanks
k

[traskutis]

Hi, i have white noise too.. would like to try silver mica's but where do i connect it?

I have usb monica as in this picture.

cheers

[ole]

I'd suggest a 1k stopper resistor between ss i/v output and gate of the buffer stage. This was part of the original Nelson Pass circuit, and intended to prevent parasitic oscillations as much as I understand.

Ole

[assendor]

Hello you guy's...

Yes it did help in my circuit with the 430pF Silver Mica. But both on the I/V resistor and the output after the Buffer stage.
It does seem like even when this oscillating is removed before the buffer, it is reappearing after/IN the Buffer stage.
"I do think" that this may vary, depending on how much noise your computer sends out.
It did look like these problems where not always' there in my circuit.


It is a better idea to try what Ole here is suggesting, of course it is always best to resolve the problem, before it occur.

I've been so busy with other stuff and work lately, Sadly for me and maybe others  The progress have been NONE.
Really will try the 1K resistor asap, but if anyone else try this and find this beneficial, please post it here.

Good luck !!

[Paul N]

Hi

Try the gate stopper! Adding extra caps is a dangerous thing to do. Every cap intended for rf bypass can also form a resonant circuit with the stray inductance of the circuit, possibly making things worse! (And, in aprticualr application, a by pas cap with a highish esr may work better than ione with low esr- the former damping any oscillation

Last weekend, feeling dissatisfied with the sound of my modified Monica 3 (possibly worse than my unmodified Monica 3!) I built new/V boards- all BC550/560, with a BF245A in the LED’s CCS, and LSK170’s in the buffer. I used a 1k carbon film as a gate stopper. (I’m a fan of stoppers in general- they can introduce noise in low level circuits if too big, otherwise harmless or beneficial in my experience).  After a couple of hours run in, I stayed up far too late listening to a wide range of music. Excellent sound!. Problem piano pieces sounded much better.

The I/V stage they replaced had a grid stopper (a SMD 1k metal film) but also had 2C2547E in place of BC550 at Q3 and Q4 (see original circuit). I fitted these in the days of  Monica 2 and reckoned they improved the sound. They should be more linear. But they are more susceptble to rf problems, and at 24V the I/V stage became unstable- whining on output and or hiss- which required ferrites on connections to and from I/V boards to cure (and was incurable if all the transistors were changed to 2SC2547E/2SA1085. Even now, this circuit change won’t work with the onboard I/V circuit of Monica 3, even when I broke the tracks from the DAC and put in ferrite SMDs. Which I why I use separate I/V boards with Monica 3- well that and  wanting to try separate supplies on all stages)

I think the addition of  the buffer stage to this modified circuit pushed things into mild rf instability, A NOS DAC + wide band I/V stage + wide band buffer is a potentially dangerous combination in rf terms! I for one won’t be playing any more with wider bandwidth transistors, despite their supposed advantages in other respects

So, to summarise, avoid high bandwidth transistors; use a carbon film 1k gate stopper, soldered as close as possible to the gate of the buffer; use ferrites if possible on the output and inputs on the I/V stage (a 100R -1k resistor on the I/V output may help); use a preamp with low rf susceptibility;  try moving the wiring about; and, if all else fails try lowering the supply voltage to the I/V stage -and I'd suggest a maximum of 24V.

BTW I 've no experience of USB Monica, just the "regular" versions

Hope this helps

Paul

[ole]

I just followed my own advice below and added gate stoppers, eventhough I thought I don't have a problem about oscillation. And, it's really better! Complex recordings of large orchestras and choirs sound less congested. Without the stoppers, it was sometimes a little muddy. Certainly worth a try!

Ole

[Nuuk]

Is there any chance that somebody could post a diagram showing exactly where the grid-stopper resistor should go please? Perhaps somebody could modify the digram that shows how to add the LSK170's. 

[kkerwin]

yes, adding to nuuk's sketch would be very helpful.  i'm really "following instructions" with this.  i don't know enough electronics to translate application between the nelson pass schematic and the monica usb.
thanks
k

[ole]

On Nuuk's picture, you would have to cut the track below where the transistor pin marked with a single "2" is soldered in, so that this "2" pin is not connected anywhere anymore. Then you would need to bridge this cut with the stopper resistor. On both channels, of course.

So to say, the stopper resistor connects the output of the i/v stage to the gate of the jfet buffer.

Ole

[Nuuk]

I can't claim credit for that original picture!

Is this correct?

[Paul N]

Hi

The diagram's correct!

Paul

[kkerwin]

thanks,
i'll try this and  post my observations.
k

[ole]

Nuuk's picture is correct. That is the input of the buffer.

Hi

No! Lead 2 on the other transistor- i.e. the input of the buffer

Paul

[assendor]

Hello you investigators

This is getting exciting 

This NOS DAC with no Oversampling nor any filters, lets it output analogue "like" Square wave levels.
Everything is almost totally analog, but the form of the signal waves have steps in total Squares.
These things explains like Paul N notes, why the wide band I/V stage + wide band buffer gets into some trouble.
Square waves, or a analog level change that is in square form, really is a level change that is a Sine Wave
containing "all" frequencies. (The Flat part is 0Hz, the up is Infinite Hz)

Despite all these possible issues,
It seems that this is the absolute best way to obtain the most analog and natural sound from a digital source, if compared to any other way of doing this.
More bits, more Sampling, more filters, just anything will destroy the fantastic Naturalness of this NOS
DAC.

I've been thinking of trying a Transistor 2N5416 and 2N3439 instead of the BC550/BC560/BC557B/BC547B.
Reason this is a good transistor is that it is supposed to be very linear and it has a rather low transition frequency. min=15MHz. I've seen these used in places where stability and linearity is essential.
They really are switching transistors, but performs flawlessly also in audio circuits because of their
frequency/gain linearity.
This should, if it at all works in this circuit resolve the oscillation problems in this part of the Circuit.

As I do not know if this transistor really is within working range with the Currents and voltages present here in this Gainstage circuit, maybe someone here can help out with this?
Maybe some values somewhere must be changed?

2C2547E in place of BC550 at Q3 and Q4, maybe these can be used with even more stability when the other transistors have a lower Transition Frequency?

Something else, As it is rather useful having a Grid/Gate resistor on the buffer stage, could it also be beneficial to do the same thing at the I/V stage Jfet 2N3819/BF245A??
OR Would it be possible to build a Constant current source with a lower bandwidth??

I do understand this was a lot of ideas in one post, but just want to hear what others think.
It is very possible all of these suggestions of mine are totally useless 

Thanks you guy's for all the superior tips so far 

-Jonas

[milen007]

hi. can someone verify the connection of the grid stopper resistor in pictures? thx heaps