Quick review and teardown of low cost AlphaLab LNA10 low noise amplifier

[maxwell3e10]

AlphaLab LNA 10 preamplifier was mentioned in a few posts, for example here https://www.eevblog.com/forum/projects/some-information-on-the-lna-10-low-noise-oscilloscope-amplifier/msg2682060/#msg2682060 and here https://www.eevblog.com/forum/testgear/preamplifier-for-noise-measurements/msg3073534/#msg3073534.
It's primary advantage is low cost ($270) and rough similarity to Stanford Research SR560. So I got one for general lab use.

As shown in the picture below it uses a quad OPA4209 op-amp for the input stage. The performance is decent, but since it uses a quad op-amp, there is limited possibility to upgrade to a different chip. The power supply to the op-amps is +/- 4.5 V (derived from 12 V wall wart), so the output voltage is limited to about +/- 3V. The maximum 3dB bandwidth is about 800 kHz and can be adjusted down to 1 Hz. I was particularly interested in the noise spectrum for gains lower than maximum. Indeed, for gain of 100 the noise level remains  the same as for gain of 1000, about 4 nV/Hz^(1/2). For gain of 10 the noise level increases to about 10 nV/Hz^(1/2). This is about the same as for SR560.



The current noise of the amplifier is however larger than for SR560. Shorting the input with a 1 MOhm resistor (instead of a dead short) gives the noise of about 200 nV/Hz^(1/2) vs. Johnson noise of 127 nV/Hz^(1/2). In contrast, according to specs SR560 has a noise figure close to 0 dB for 1 MOhm source impedance. The current noise of LNA10 can be estimated to be about 0.2 pA/Hz^(1/2), which actually seems somewhat smaller than the spec of OPA4209 op-amp of 0.5 pA/Hz^(1/2).

The noise level and the amount of 60 Hz pick-up did not change appreciably when I switched from wall power to a 12V battery to power the amplifier. The amplifier output data were acquired with MAX11410 evaluation board (a free giveaway from Maxim a few years ago).

The build quality seems pretty good, BNC connectors are of decent quality and the knobs seem solid. However, my unit seemed to occasionally produce glitches, which consistently went away when I turned the box face down. This was the case both before and after I opened the amplifier.

So overall the performance of LNA10 is pretty good, matching most features of SR560, except the input current noise, at a 10th of the cost. It is perhaps not state-of-the art, but a very good deal for a general purpose pre-amplifier.

[sixtimesseven]

The performance is decent, but since it uses a quad op-amp, there is limited possibility to upgrade to a different chip.

Looks to me like a pretty simple two layer board. So rather than "upgrade" the opamp I would be more interested in an improved version with the LNA10 as a basis.

Would you provide me with some measurements if I start drawing a schematic from your pictures?

[thm_w]

Thanks for following up from the other thread.

Wonder if shielded leads going to the BNC would help much. The white wire on the back is grounding the case?

[maxwell3e10]

I was considering tracing the signal path through the circuit, but it doesn't seem so interesting. The GBW of the op-amps is only about 20 MHz, so to change gain it probably switches in additional stages. The voltage regulator generates about 9 V, so the ground level is generated by one of the op-amps. Then you have a switchable differential amplifier and a filter with switched capacitors and a variable resistor. So that accounts for most of the op-amps. I can make a few measurements, but as my grandfather used to joke, you save more money by running behind a taxi instead of the bus.

The metal case provides some shielding, but the blue ends are plastic. The white wire gets jammed between the case clamp shells to ground it.

[sixtimesseven]

I'm more interesting in improving (particularly bandwith, maybe current noise as you mentioned). Particularly in the opamp department there seems to be a lot of room for improvment.

I had a project recently where the low sensitivity of 8bit scopes really bugged me. It was, however more in the 1-10MHz range so the stanford reasarch as well as this lna are useless.

It bugs me that in order to look at small signals in more details one has to go to high resolution scopes. 8bit would be plenty if one could preampliefy the signal. Kind of how SA's get preamps to increase the dynamic range, oscilloscopes with a preamp would be nice.

[maxwell3e10]

Some scopes have better input noise level than others, see https://www.eevblog.com/forum/testgear/oscilloscope-input-noise-comparison/. Generally input noise spectral density in 1-10 MHz range is between 2 and 10 nV/Hz^(1/2) on the lowest gain setting. You can also look at Owon XDS series scopes, they have 12 bit resolution and good input noise level at a very reasonable price.

For an amplifier in MHz range it depends on your desired input impedance. For 50 Ohm impedance there are many inexpensive options with noise figure of 1-3 dB. High impedance and wide bandwidth are a rare combination. NFCorp and Femto have some options but they are relatively pricey. So maybe there is room for a DIY design.

[sixtimesseven]

Some scopes have better input noise level than others, see https://www.eevblog.com/forum/testgear/oscilloscope-input-noise-comparison/. Generally input noise spectral density in 1-10 MHz range is between 2 and 10 nV/Hz^(1/2) on the lowest gain setting. You can also look at Owon XDS series scopes, they have 12 bit resolution and good input noise level at a very reasonable price.

For an amplifier in MHz range it depends on your desired input impedance. For 50 Ohm impedance there are many inexpensive options with noise figure of 1-3 dB. High impedance and wide bandwidth are a rare combination. NFCorp and Femto have some options but they are relatively pricey. So maybe there is room for a DIY design.

Yes, a 12b scope, that would be scope number 4 
Yes, scopes are noisy and the dynamic range is limited with 8bit scopes. But if one wan't to look at small signals with lowish dr, then why not amplify it before feeding it to the front end to improve snr relative to the scope noise? Yes it adds the amplifier noise but it's kind of the name (lna10).

50 ohm amplifiers are another way. However they would need an "active probe" discretely or integrated to match to a high impedance test source if that would be required.

[CChin254]

IC Identification: