A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

[JR.]

well my personal interest started from explorations into MM carts with the conclusion that almost none of the commercial solutions are fit for purpose. Holman was onto something back in the 70s. For MM you can make a very cheap phantom powered headamp that you can put inside the turntable. Worth knocking up for a bit of fun*. Plus the look of horror on audiophiles faces when you mention phantom power and Vinyl in the same sentence .

*Other's definition of fun may vary. Not sure JR can have fun unless an angle grinder is involved these days

I enjoy reengineering my solar lamps

JR

[mediatechnology]

I wanted to let our readers know that the ZTX851-based moving coil preamp board will be available for purchase in about 10 days.

The initial prototype run sold out and I've ordered production quantities.

Construction information, a bill-of-materials and circuit description are available in the "Build" section of the forum. viewtopic.php?f=7&t=1179

Wayne

[mediatechnology]

Flat Moving Coil Phono Preamp PC boards are available now.

https://ka-electronics.com/shop/index.p ... uct_id=105

[mediatechnology]

I built a MC preamp board today and before installing the ZTX851 transistors I decided to test the ZTX689B "in-socket."

The MC preamp, with the ZTX851, shows that the rbb' of the ZTX851 is around 1.75Ω.
(Ein about -141 dBu, gain = 62 dB, Rsource 10Ω, Rg 1Ω produces a total Rnv of about 14.5Ω. Ic = 5.5 mA nominal)

Given the low noise of the MC preamp it makes a good test-bed for the ZTX689B.

The ZTX689B, and its PNP complement the ZTX789A, have much higher gain.
The ZTX689 has an hfe of 500 minimum. Cob is lower at 16 pF typical.
The maximum collector voltage is 20V compared to the 60V of the ZTX851.

https://proaudiodesignforum.com/images/pdf/ZTX689B.pdf

I didn't expect that the ZTX689B would perform as well as the ZTX851 but the higher gain, and lower Cob, might make it a good choice for mic preamps with their higher source impedance in the 50-300Ω range.

The ZTX689B has an estimated rbb' of about 5.5Ω based on the same test conditions.
(Ein about -139 dBu, gain = 62 dB, Rsource = 10Ω, Rg 1Ω for a total Rnv of about 22Ω.)
This works out to be about 600pV√Hz.

The ZTX689B is a strong contender for mic preamps.
It's 1/f performance appears just as good as the ZTX851.
With higher source impedances the lower rbb' of the ZTX851 may be less of an advantage compared to the higher gain of the ZTX689B.

[terkio]

The ZTX689B is a strong contender for mic preamps

Great.
Another interesting point is.
When using 2 x 10K bias base resistors, the high hfe makes it easier on hfe matching.
This is because, hfe mismatch induces a Vb mismatch from ib x Rb.

[mediatechnology]

Yes, Ib and Ios should be much smaller.

[mediatechnology]

I started looking for vendors for the ZTX851 matched pair's acorn nut since I had exhausted Home Depot's supply and they're over $1 each.
The nut is almost as expensive as the two transistors inside it.

There's a local Fastenol on the other side of the airport from me so they were my first choice.
As I shopped Fastenol's website I discovered aluminum acorn nuts. https://www.fastenal.com/products/details/0176228
Then it occurred to me: Chrome-plated stainless steel look very cool but when compared to aluminum stainless sucks at thermal conductivity.

The 2011-T3 aluminum nuts have about 10X the thermal conductivity of steel or stainless steel.
Even more beautiful is that they're about half the cost and in stock.



I may wind up being in the low noise matched transistor pair business.
Matching and gluing these things is kinda fun.

Speaking of matching transistors Textool's 14 pin ZIF socket has a divider to accept, in addition to 14 pin devices, 8 pin and 6 pin parts such as dual op amps and opto-isolators. https://www.digikey.com/products/en?mpa ... 0602J&v=19
Stray contact resistance in the emitter terminals can give false Vbe readings.
I've been using Protoboard contacts, which I've cleaned, and have considered machined pin IC sockets as test sockets which would likely were out quickly.
Textool's ZIF socket seems much better and the 6 pin section would be great for the transistors.

[JR.]

Somewhere I have a 14p zero insertion force socket that I used back in the 70s when I 100% tested ICs I sold through my kit company. I found it useful to know that the ICs worked before the customers received them, so I could say with authority the parts were good.

In the early days I would have single digit cull outs from 1,000 piece batch of say TL074s... by the 80s the failures were so low it wasn't worth continuing 100% testing.

JR

[terkio]

The 2011-T3 aluminum nuts have about 10X the thermal conductivity of steel or stainless steel.
Even more beautiful is that they're about half the cost and in stock.



I may wind up being in the low noise matched transistor pair business.
Matching and gluing these things is kinda fun.

Another advantage of aluminum. It is softer, it will be easier to remove the thread inside the nut.
BTW, while you are at this, look at best thermal conductivity glues.

[mediatechnology]

I've updated this design with a new PC board.

Hans Polak convinced me that I should use emitter servo injection to reduce ZTX851 matching requirements. In the process we discovered that huge improvements in IMD2 distortion could be realized by matching resistors and C4/C5.

The 19/20 kHz 1:1 f2-f1 IMD2 product at 1 kHz at +10 dBu output and 62 dB gain drops from around -83 dBc to the Analyzer A/D residual around -117 dBc.


KA Electronics FMCP Flat Balanced Input Balanced Output Moving Coil Preamp Using the ZTX851


Schematic of the KA Electronics FMCP Flat Balanced Input Balanced Output Moving Coil Preamp Using the ZTX851

Large schematic: https://proaudiodesignforum.com/images/ ... amp_18.png

Original FLIP PCB with unmatched components.
The spur to the right of 1 kHz is environmental.
The IMD percentages shown by ARTA are incorrect due to an improper setting.


The actual f2-f1 IMD at 1 kHz is 0.0075% in the above image.


New FMCP PCB with matched components:

The actual f2-f1 IMD at 1 kHz is 0.00015% in the above image.