Nikko preamp with some questions

Hi Pat,
If you hear noise, have you checked with an oscilloscope to see if you have RF oscillation? If you're using a DSO, you may have to play with the sampling rate to see it.

Check the power supply for oscillation as well.
 
That would be a nice reference to compare Pat's noise measurement with.

The noise floor is around -136 dBV with 192000/65536 Hz bin width, so roughly 100 nV/√Hz of output noise, at a gain setting of -10 dB, if I see it correctly. It increases below 2 kHz or so. I didn't correct for the effect of the windowing, but it's orders of magnitude that matter.
 
Last edited:
Hi Marcel,
He shouldn't be more than 10 dB higher in noise, less probably. As you can see, performance is pretty darned good.

As I mentioned before, the 60Hz peak is due to the test setup. The analyzer is connected to the computer, as is the bench amplifier system and some equipment. That and a LAN, so many injection points. Better these days, and I use audio isolation transformers when needed for line noise isolation, but they add distortion to the audio measurements. They aren't cheap transformers either.
 
Hi Pat,
If you hear noise, have you checked with an oscilloscope to see if you have RF oscillation? If you're using a DSO, you may have to play with the sampling rate to see it.

Check the power supply for oscillation as well.
with a scope i dont see any oscillations either on the power supply or at the output. it is trully hiss noise associated with high gain stage. almost like a phono stage with full volume on with open inputs. well it is not that comparable but the hiss is clearly audible and it drives me nuts. i can bypass the nikko with just a pot volume in/out from the source to the cross over and noise them become null. the nikko is the only item in my audio chain loop that i can point at like "its him". maybe it is within spec i cannot tell. if there is no way to reduce the hiss then i prefer my black window pot volume alone in a hammond enclosure. fwiw my scope is an Agilent 54622D.
 
also just to make sure its understood, the nikko is not full of random modifications and stupid errors. i only replaced the opamp with Burson audio VI5D and did try other such as the LM4262 and NE5532, the Burson gives the best. Only recently i replaced some resistors with RN55 metal films. The best modification i made recently is adding a shield in between the transforer and switch area, but that was a slight 60hz humm issue that is now corrected.
Maybe the nikko is ok as it is and i am just to fussy with my ears.
 
Hi Pat,
Yes, I got that from you. The noise is not normal for this unit at all. You are not being fussy, or equipment is defective.

Are you using an analogue scope? What make and model is it?
Agilent 54622D, i have a TDS220 also. Both channels hiss the same level. Whats drive me nust the most is that it hiss at volume=0, and thats because the volume is before the final stage. if i bypass the outputs stage it become null. if the mute relay is engaged the noise is null as well so it really comes from the final stage.
 
Yes, something is wrong.

The DSO's can be noisy, I still use analogue scopes as well. The Philips PM3070 and PM3365A are my favorites, TEK 2465B CT not so much. I have an Agilent 54642D, similar to your scope, and the newer MSOX3104T. They aren't as good for low noise stuff. In my view, a DSO is for large signal type stuff. Even in 12 bit mode, the MSOX isn't as sensitive as my older analogue oscilloscopes.
 
my scope uses a CRT so its a mix match, its called a mixed signal scope. I am at work for now so i cannot make any tests. i will do some experiments this weekend, like installing back the original opamps to make sure something not got missed right from the start.
 
Well, no.
A mixed signal oscilloscope has digital (logic) channels in and analogue channels. The 54642D has a CRT, analogue channels and 16 digital logic input channels. The MSOX3104T has 4 analogue channels and 16 digital input channels. It is also classified as a mixed signal oscilloscope. Your 54622D is a nice scope, good for about 20 MHz analogue input (using the 5X highest frequency criterion).

My new scope has an OLED display instead of a CRT. The CRT is limited by the drive electronics and digital resolution. Early LCD displays could be limited by the actual display compared to CRT displays.

In any event, you have a nice scope. Just not the best for low level noise. For that, a spectrum analyzer is far superior. A THD meter would be better.

Yes, let us know Pat. You have an interesting fault.
 
Hi Marcel,
He shouldn't be more than 10 dB higher in noise, less probably. As you can see, performance is pretty darned good.

It might just be audible over a high gain amplifier and sensitive loudspeakers in a very quiet room:

-136 dBV in 192000/65536 Hz
92.6 nV/√Hz
about 10.5 uV A-weighted (noise bandwidth of A-weighting is around 12.5 kHz to 13.5 kHz)

With a gain-of-50 amplifier into an 8 ohm load:
about -74.6 dBW A-weighted

Assuming stereo, a 92 dB at 1 W, 1 m loudspeaker sensitivity and a 1 m reverberation radius:
about 20.4 dB(A) SPL
 
lol!
I've been watching Vunce.

Hi Marcel,
Normal amplifier gain ranges from +27 to +34 or so dB. Typical room noise is a * little * over 20 dB(A) SPL. You would stick your ear near the tweeter to hear it. Not from a sitting position. In North America our homes tend to be larger, with larger listening rooms. My test bench speakers are 99 dB/watt (Klipsch THX 6000) and I would absolutely hear hiss, zero doubt. Bench amp is a Marantz 300DC (forget the exact gain, close to the spec in the manual). I used the Beta 30 for a bench preamp for years, replaced with a Yamaha C-50.

From what Pat is describing, he has a problem. If anyone would have complained about hiss, it would have been me.