Noise measurement amplifier - interested?

[jackinnj]

Last evening I finished the PCB for the replacement amplifier board for my HP465A test amplifier. The 465A had gone wonky, with a lot of noise and a DC offset problem.

The circuit which is attached uses the SSM2019 programmable gain amplifier for 20, 40 and 60dB gain (trimmable), and follows an idea which Jung published in AudioXpress in 1995 and similar designs found on the Analog Devices website. It has "belt and suspenders" input protection -- perhaps unnecessary since the SSM2019 seems to have this incorporated.

The chip Walt used is no longer in production (or at least I couldn't find it) and almost all devices are SMT. If there is interest I could send the boards out for production and distribute at cost. The proto-type has the option for using the Linear Tech LT1964 and LT1761 ultra-quiet regulators (shown on the schematic) AND a TL431 type regulator. The proto's fit exactly into the space allotted in the Hewlett Packard 465A box. The power coming off the HP transformer is high so this had to be tamed with an active zener. If anyone is interested in testing power supplies, this is a handy instrument to have, and the HP 465A power supply, chasis etc. makes it convenient for DIYrs.

[janneman]

Hi Jack,

I would be interested. I *may* still have a few 2019's around, else I use one of your recommended ones. Are you thinking about a (semi) kit including say the regulators and opamp?

Jan Didden

[jackinnj]

I use the "W" version of the SSM2019 -- easier for me to hand-solder. I was just thinking of just having the boards made -- I used the SSM2019 since I had them on hand. Using this chip does "bandwidth limit" the project a bit (compared to the HP465A which is flat to 1MHz).

I was wondering whether to do some active filtering on the output. so that the noise can be referenced to a specific bandwidth.

Right now I am having an arguement with my PCB software -- it isn't allowing me to place the PowerPlane where I want it so I have had to "pour copper".

[jackinnj]

I added a 12.5 Hz Chebyshev low-pass filter to the circuit -- you can jumper it in or out with a couple of molex connectors (i.e. if you want to measure power supply noise, the filter is wired in, if you want to measure to 1 MHz, the filter is out.) I also changed the input caps to "back to back" 1000 uF electrolytics -- thus the response should be 0.1 Hz to 10Hz. This necessitated going from surface mount to through-hole.

I also removed the protection diodes -- and will rely on the diodes on the SSM2019.

Jack

[jackinnj]

here's a link to the article --

http://www.tech-diy.com/hp465a_ampli...ernization.htm - for purposes of clarity the ground plane isn't shown in the PCB design.

the jury is out on the LTC1761 and LTC1964 ultra-low noise regulators -- they may be unnecessary (and as I state, just add to the cost and complexity.)

C1, C4 and C7 have been omitted. if there is an RF or EMI problem they can be soldered in.

i will add a clipping indicator, am thinking of changing the resistor arrangement so that a trimpot can be inserted for the 20, 40 and 60dB ranges -- for the moment I have been soldering in different SMT resistors.

the SSM2019 does run "hot" -- the initial tests show that it is extremely quiet. I did find a cookie tin !

[ergo]

Count me in for 2 PCB-s. I know my friend will want one too

Ergo

[jackinnj]

I updated the PCB -- now have a dual-comparator set up as a clipping indicator (over=>red, under=>green).
I tested the Linear Tech low noise LDO regulators -- they tame what was already a pretty quiet supply with the TL431 pre-regulator , but I don't know that they add to performance in any meaningful way.

Since I have just a little more space on the board I was thinking of gilding the lilly and adding an RMS detector...or should I put in a DPDT relay for DC-coupling (and by shorting the 1000uF input caps.)

http://www.tech-diy.com/hp465a_ampli...ernization.htm

[jackinnj]

The amp in its current incarnation doesn't seem to add any distortion or noise -- here's a pic -- without the clipping indicators installed yet:



works just fine at 20db and 40db, but the connector for 60db isn't functioning quite right yet:

[cocolino]

Hello Jack,

just recently I stumbled over a very similar project as Yours (unless that it works with a ground referenced input only and lacks the differential input) .
It is well documented and I thought maybe You`ll find it interesting (but might well be that You know it already) :

http://tangentsoft.net/elec/meas-preamp/

BTW: as the SSM2019 is an obsolete and hard to find part, it appears to me it should be also possible to substitude it with an INA-103 (Burr-Brown) while getting very similar performance. What do you think?

[jackinnj]

Originally posted by cocolino
BTW: as the SSM2019 is an obsolete and hard to find part, it appears to me it should be also possible to substitude it with an INA-103 (Burr-Brown) while getting very similar performance. What do you think?

The SSM2019 was chosen as the replacement for the SSM2017 which is obsolete. Last I checked the '2019 was current production. WJ used the '2017 in his Audio Amateur article about 8 years ago. If you wanted greater bandwidth an even better preamp would be a good idea. The '2019 here is used in the wide SMT package.

I used dual BNC jacks on the input -- important to ground one of the inputs if you are going single-ended.