[Sigurd Ruschkowski]

Hi,


I want to have a very low noise Bugle. I have two of the PCBs and are planning to build one with very low resistor values to make the
resistor noise as low as possible.

Of course one can use lower noise OPamps than the OPA134, but I want to build a antiRIAA that is a lowest noise version by just lowering and removeing resistors.

Lowering resistor values means increasing some capacitor values. I have increased most capacitors by a factor 10. See schematics below.
RIAA correction resistor values are lowerd by about a factor of 10. Not exactly 10 as I
wanted a better RIAAaccuracy than is standrad in the Bugle.

Resistors R29 and R27 are removed (1k standard).
The input series resistor is lowered from 1k to 39 Ohm.

Feedback resistors are lowered about a factor 10, too.

I am lucky to have 220n, and 100n polystyrene caps. C3 will be a 2.2u WIMA MKP10 (or MKP10) and is paralleled with 47n PS for best sound quality.

My simulations in SPICE show good promise for my modificatons. WIth Hagerman's RIAA reference (see schematics below) I get about +-0.025 dB RIAA inaccurace from 20Hz - 20kHz!

Can there be any problems / disadvantage with my modifications?


Schematics:


RIAA inaccuracy from 20Hz - 20kHz:


RIAA inaccuracy from 1Hz - 100kHz:


Hagerman's mathematical RIAA:



Sigurd

[hagtech]

Looks very interesting.  Let us know how it comes out.  

jh

[peranders]

If current consumption is an issue, only the first stage needs 220 ohms or less as feedback and the other stages can surely manage 1 kohms or so.

Don't forget Sigurd that low impedance also increases distortion but it may not be so important.

You may notice how much gain you'll need but I think you have a bit much now.

[Sigurd Ruschkowski]

Per-Anders -
no current consumption is not an issue.
Distorsion is a big issue - shoule be very low.

Gain is too high - thanks!


Sigurd

If current consumption is an issue, only the first stage needs 220 ohms or less as feedback and the other stages can surely manage 1 kohms or so.

Don't forget Sigurd that low impedance also increases distortion but it may not be so important.

You may notice how much gain you'll need but I think you have a bit much now.

[Sigurd Ruschkowski]

I have now measured capacitors and finally found some that work well as pairs.

Schematic is now updated with resistors that match the measured cap values:








Sigurd

[Sigurd Ruschkowski]

Jim -

here is the assembly that is only missing some resistors.
Silver wire is used to replace the input resistors of stage 2 and 3.

I have measured all caps and paired them. Resistor values are then matched to these resistors.

The opaque light blue caps are RIFA426.

The transparent light blue caps are RIFA polystyene.

Red caps are WIMA MKP10 polypropylene.

Electrolytic caps are ELNA Cerafine 47u.

One can also see that there is a cap under the PCB. It is a polystyrene cap that is i parallel with the large WIMAs.

I added a blue LED, too.













Sigurd

Looks very interesting.  Let us know how it comes out.  

jh

[/img]

[Sigurd Ruschkowski]

I got the special value resistors today. I have now tested the my ultra low noise Bugle with extreme antiRIAA inaccuracy,
and  only find one small issue with it: DCoffset. Offset is 8 and 18 mV in left and right channels. That is a bit high for a all out DCcoupled stereo system like mine. I will try some other OPA2134s and hope they have lower input offset values. Or even some other OPamps.

Anyway, here are some oscilloscope measurements. One at 20Hz, one at 2kHz and the last one at 20KHz.
Chanel 1 is input and ch2 is output voltage.

At 2.2  kHz I have 43.6 dB gain,
at 20 Hz 65.5 dB gain, and
at 20 kHz 27.7 dB (shouldn't that be lower, around 24dB?)










Sigurd

Looks very interesting.  Let us know how it comes out.  

jh

[Sigurd Ruschkowski]

With the AD8620 dual OPamp I have solved the DCoffset problem.

I now have below 1 mV DCoffset. Really amazing performance of the AD8620!
I am using one of BrownDog's SMD-DIP adapters. See below.




Sigurd

and  only find one small issue with it: DCoffset. Offset is 8 and 18 mV in left and right channels. That is a bit high for a all out DCcoupled stereo system like mine. I will try some other OPA2134s and hope they have lower input offset values. Or even some other OPamps. quote]

[Sigurd Ruschkowski]

I have now changed the first stage OPamp into AD745s. A single OPamp that I have managed to squeeze into a Brown Dog dual single opamps to dual opamp with IC sockets. It is a 16-pin IC, so I cut of 4 pins.

The AD745 is a super low noise OPamp designed hor higher gains like the Hagerman RIAA uses. FETinput, and super low distorsion.














Sigurd

[Sigurd Ruschkowski]

The Hagerman/Sigurd RIAA amp is now installed in a box.



Power supply voltage is +- 12V DC. Max voltage for the AD8610 is +- 13V.

I use Per-Ander Sjöström's Jung-Superregulators to create the +- 12 V
from the unregulated +- 27VDC.
http://home.swipnet.se/~w-50719/hifi/jsr03/index.html

Unregulated voltage is created from Per-Anders Sjöströms RFB02 rectifyers. I use two the RFB02 with in total 8 Schottky diodes in a dual bridge.
http://home.swipnet.se/~w-50719/hifi/rfb02/index.html
22 000 uF per rail is used together with 2x 10UF polypropylene.
Snubbers used.

AC mains is filted with two caps.

Transformer AC is also filtered with two caps.

80VA toroid transformer.

2.7k bleeder resistors.





Sigurd

[peranders]

Those AD745 looks cool but it's possible that you could achieve even better results if the opamps were tighter to the pcb.

[Sigurd Ruschkowski]

AD745s are excellent choices for low level applications with high gain such as a RIAA amp.

Yes, Per-Anders, they could be mounted much better. It was planned to be just a prototype test, but ended up being the finished product  



Sigurd

Those AD745 looks cool but it's possible that you could achieve even better results if the opamps were tighter to the pcb.