Low Noise Transistors: Comparing the ZTX1053A and ZDT1053 (Dual) vs. the ZTX689B and ZTX851

[mediatechnology]

I was looking for a dual low noise transistor in the Zetex line and came across the ZDT1053 dual and it's through-hole single the ZTX1053.

I decided to compare the ZTX851, ZTX689B and ZTX1053 in through hole before I test the ZTD1053.

I setup a simple test circuit similar to the one here:



In order to maximize the transistor's noise contribution I made Rsource and Rgain 1Ω.
Rfb is 316Ω. (Not the 499Ω shown in the above drawing.)
Rg is 1Ω. (Not 2.2Ω as shown.)
Gain = 61.6 dB
Ic is about 7.5 mA.

Here are the results in this circuit:

ZTX851 Ein -143 dBu Rnv 9.4Ω Noise Density 390 pV√Hz.
ZTX689B Ein -137.9 dBu Rnv 30Ω Noise Density 695 pV√Hz.
ZTX1053 Ein -139.9 dBu Rnv 19Ω Noise Density 553 pV√Hz.

The ZTX1053 looks ideal for mic preamps.
hFE for the ZTX1053 I tested ranged from 300-500 at low test currents.
Cobo is also low-ish around 21 pF.

With the ZTX1053 single looking so promising I'll be checking the dual ZDT1053.
I bought 6 - the Vos spread of the pairs was <<500 µV on all 6.

[JR.]

Nice .... put some of these in the time capsule to send back to 1975....

JR

[mediatechnology]

Nice .... put some of these in the time capsule to send back to 1975....

JR

I'll leave some in my basement.

A visual comparison:


ZTX Low Noise Transistor Comparison ZTX689B (top) ZTX851 ZTX1053A (middle) ZTX851 (bottom)

ZTX689B Ein -137.9 dBu Rnv 30Ω Noise Density 695 pV√Hz.
ZTX1053 Ein -139.9 dBu Rnv 19Ω Noise Density 553 pV√Hz.
ZTX851 Ein -143 dBu Rnv 9.4Ω Noise Density 390 pV√Hz.

Zetex Low Noise NPN Datasheets:

ZTX689B: https://proaudiodesignforum.com/images/pdf/ZTX689B.pdf
ZTX851: https://proaudiodesignforum.com/images/pdf/ZTX851.pdf
ZTX1053: https://proaudiodesignforum.com/images/pdf/ZTX1053A.pdf
ZDT1053 Dual: https://proaudiodesignforum.com/images/pdf/ZDT1053.pdf

[terkio]

Great.

From the ZTD1053TA datasheet.
There is no information about transistor matching.

Foot notes at thermal characteristics, indicate the two transistors are on separate dies.
It is not clear wether we are better with a dual ZTD1053 rather than with two singles ZTX1053 we can match.

The case is plastic, most heat dissipates via the collector leads.
Temperatures of the PCB tracks at the collectors are the key factors about junction temperatures hence V be matching.

[mediatechnology]

My thought was the dual SMT could have a metal slug glued onto it to equalize the temperatures (and add thermal mass) as well as a PCB pour under it.

[mediatechnology]

There may be some Zetex duals that share a common die.
We may have quite a few devices Zetex designed for low Vce saturation which are also low noise.

[terkio]

My thought was the dual SMT could have a metal slug glued onto it to equalize the temperatures (and add thermal mass) as well as a PCB pour under it.

What do you mean by a PCB pour under it ? Some thermal pad.

I was struck by the thermal resistances mentioned in the ZTD1053 datasheet.
The Rth from junction to collector leads: 30 K/W
The Rth from junction to ambient: 50 K/W
I imagine the junction heat mostly dissipate via the collector leads, this is why there are two pins for each collector. Some heat dissipate to ambient by the case, but much less than by the collector pins.
I understand, they count on the collector tracks to act as heatsinks. They better have equal thermal resistance to ambient. Same area same air flow.
Plus as an improvement to a best thermal equalization, a thermal bridge between the collector tracks should be good to have. Hence adding some PCB thermal pad under the chip.
[Edited]
Heath sinking by the PCB tracks.
From the data sheet I understand:
25mm x 25mm 2 oz copper at each collector pin pairs. That is quite a lot of PCB real estate for a tiny SMD circuit 6mm x 7mm. I wonder about the layout. May be the SMD on a piggy back circuit ?
[/Edited]

[mediatechnology]

I was visualizing a thermal pad.

The Vos may be purely the luck of the draw with two separate die.
Of the 6 I tested all would have been usable in terms of Vos. (delta-Vbe.)

We don't need the big thermal areas since we're not switching amps of collector current.

[mediatechnology]

I got bored and decided to measure some low noise PNP transistors by inverting the polarity of the test circuit.

The top trace is the 2SB737; the bottom trace the ZTX951.


Low noise transistor comparison: The 2SB737 (top) versus the ZTX951 (bottom).

The PNP ZTX951 is about 0.4 dB quieter than the NPN ZTX851.

The 2SB737 is 3-4dB quieter than the high gain ZTX689B and ZTX1053A NPNs so, for low impedance sources, the 2SB737 remains the best choice when high hFE is needed.

[terkio]

2SB737 has become unobtainium.
No hit on octopart.
High prices on unreliable sources, risk of fakes.

With low cost in mind. ZTX1053 or ZTD1053 paralleling can give a lowest noise and better Vbe Hfe matching.