That's actually exactly what I need for work. I was actually considering building a slightly modified clone of your version, but at 300 CHF this device seems like a no brainer. Lets hope that procurement lets me buy from a Swiss startup...
honestly the way we Swiss love our bureaucracy, I'm willing to bet the house that they will be ready to comply with whatever paper trail your procurement requires.
One very simple method of achieving similar fast-settling LNA performance is to split the overall gain into two stages. An ultralow noise first stage gain of about 250x is plenty to get the signal of interest well above the noise floor, but is low enough to prevent the first stage from railing due to input electrolytic leakage. If you follow that stage up with a high-pass filter, the DC offset is removed and the signal can be safely amplified further. I'm glad that you mentioned input voltage safety, since with such a large input electrolytic it is very easy to overload the input, even with the input clamp diodes. Many opamps don't allow more than 5mA, so either higher than desired input resistors, or JFET current limiters are required. My design uses the OPA2182, which does not have input back-to-back clamp diodes, so it can handle differential inputs up to the VDD/VSS span. I also included a series mosfet voltage limiting circuit, so the input is fully safe for hot plugging into anything up to 30V, regardless if the LNA is powered on! I'm on rev 2 of such a design, and have gotten a 72nVpp / 10.6nV RMS shorted input noise floor! Current consumption is only 7mA total, so it gives plenty of life from a 9V battery. If you search the eevblog metrology page for "DIY 0.1 to 10Hz Noise Amplifier", posted June 12th 2023, that's the build log of my initial design, which I will be updating with the second revision very soon! *The EEVblog post is updated with my Rev X2 results! 78nVp-p typical and 10.54nV RMS typical noise floor with a shorted input
As an audio guy, I know exactly where that "DC Servo" moniker comes from. Glad to hear there's actually something interesting happening with it, and it wasn't solely marketing wankery.
We are engineers through and through and despise marketing wankery too 😉. Nevertheless I must say that good engineering marketing is very nice, like it was in the past.
AVX makes Wet tantalum capacitors with ultra low DC leakage (less than 1uA). I believe Mr. Williams used a few of them for some ultra low noise preamplifier (AN 124 July 2009).
@@DanielHeineck yeah its about 60-150 euros per piece. Ive been on excursion in AVX where they make it but they didnt want to give me a samples :( I would gladly accomodate some nice 4700uF wet tantalum cap :)
Since there is an audio servo used the question of audio grade caps from nichicon (muse) comes to mind. Those are rather lower grade caps . The effect of voltage rating also is of interest in that application . The cost difference is small in case .
@@johnwilliamson467 Exactly. It means they are rated to work within specifications for a given amount of time at that temperature. In the case of this amplifier 105oC caps are going to work for a much longer time at lower temperatures.
@@zsigmondkara For these caps the rule of thumb is 2x per 10c work for about a 40 c range thus 120c over 85 c is aprox 16 times the life given the test I have watched.
0:30 The "BNC Polarity" is just a bit 'funny'(being an AC coupled device), but I guess they're referring to the battery charging function. Poor label layout, that BNC Polarity advice should have been printed directly adjacent to the battery charging note for it to make immediate sense.
I personally enjoy my own naturally occurring noise sources. You have an approximate 2kHz noise source whenever unmuting your mic. This may be from test equipment active in a lab, not studio environment.
So cool. I wonder how the guys at welectron handle your humor. I'm in Canada a purchased a bm869s recently and they were about as helpful to me as the dolphins were to the fish population before they left earth haha
There's me expecting a huge InterFET FET - I've seen they have some running at >50 euros a piece but low noise is the name of the game there. I'll stick to the LSK389 and similar.
To measure how noisy is the DC volatge reference. The less noise it has the more exactly you can measure its voltage, nowadays possible to 7 places which is 0.1 ppm.
I love this stuff. However, I'm not a fan of the DC cancelation circuit. My approach was to build a composit LNA with minimal gain and use a very low noise SAR like the AD4630-24. I put a brick wall on the .1 to 10hz bins and ifft to get the RMS. No settling is really needed since there are no AC caps. Also, it's nice that the SAR has a little bit of over rang, so you can measure the same reference noise.
With 10V Input Voltage and 24 bit resolution you will have 600nV LSB resolution. How do you want to measure 10nV or less? You need a gain of 100 or more.
Swiss startup? Hmm, let's see where they from ... and of course, they are two villages over from where I live. 😅 Even though it's mostly villages here just at the border, for some reason this is tech central of Switzerland.
BNC charging is a bit ... silly ... however i am reminded that the end user is an idiot and would gladly keep a device like this plugged in while using it so its a nice cheap lockout mechanism
Ich fange erst zu staunen an, wenn der Verstärker von 10 Hz bis 10 Ghz einen Amplitudengang von +/- 0,1 dB hat und der Rauschpegel unter der Messgrenze liegt. 😂
That's actually exactly what I need for work. I was actually considering building a slightly modified clone of your version, but at 300 CHF this device seems like a no brainer. Lets hope that procurement lets me buy from a Swiss startup...
honestly the way we Swiss love our bureaucracy, I'm willing to bet the house that they will be ready to comply with whatever paper trail your procurement requires.
One very simple method of achieving similar fast-settling LNA performance is to split the overall gain into two stages. An ultralow noise first stage gain of about 250x is plenty to get the signal of interest well above the noise floor, but is low enough to prevent the first stage from railing due to input electrolytic leakage. If you follow that stage up with a high-pass filter, the DC offset is removed and the signal can be safely amplified further.
I'm glad that you mentioned input voltage safety, since with such a large input electrolytic it is very easy to overload the input, even with the input clamp diodes. Many opamps don't allow more than 5mA, so either higher than desired input resistors, or JFET current limiters are required. My design uses the OPA2182, which does not have input back-to-back clamp diodes, so it can handle differential inputs up to the VDD/VSS span. I also included a series mosfet voltage limiting circuit, so the input is fully safe for hot plugging into anything up to 30V, regardless if the LNA is powered on!
I'm on rev 2 of such a design, and have gotten a 72nVpp / 10.6nV RMS shorted input noise floor! Current consumption is only 7mA total, so it gives plenty of life from a 9V battery. If you search the eevblog metrology page for "DIY 0.1 to 10Hz Noise Amplifier", posted June 12th 2023, that's the build log of my initial design, which I will be updating with the second revision very soon!
*The EEVblog post is updated with my Rev X2 results! 78nVp-p typical and 10.54nV RMS typical noise floor with a shorted input
Good job! Thank you for your community support.
Congratulations to Euler Precision.
If you see an order from Italy, it might be mine!
Edit: it is definitely mine. Just ordered!
I feel so spoiled, so much great content from Marco. Thank you so much for giving us the best. The nerd in me is giggling with delight. 😂
Absolutely top engineering and a very good video about this product! Great and very entertaining, as usual.
A return to form! Delightful, and such an interesting device. Many thanks.
As an audio guy, I know exactly where that "DC Servo" moniker comes from. Glad to hear there's actually something interesting happening with it, and it wasn't solely marketing wankery.
We are engineers through and through and despise marketing wankery too 😉. Nevertheless I must say that good engineering marketing is very nice, like it was in the past.
I understand about 10-12% of your videos but I watch every one all the way through.
I don't fully grasp all the nuances, but I enjoy your enthusiasm for it.
I mean that LFLNA is sweet but that MXO 4 is.... Just something else.
AVX makes Wet tantalum capacitors with ultra low DC leakage (less than 1uA). I believe Mr. Williams used a few of them for some ultra low noise preamplifier (AN 124 July 2009).
And those lovely caps are about as much as the whole preamp! :D
@@DanielHeineck yeah its about 60-150 euros per piece. Ive been on excursion in AVX where they make it but they didnt want to give me a samples :( I would gladly accomodate some nice 4700uF wet tantalum cap :)
Nice bent tweezer, I personally use mine whenever pcb pointing is called for.
it would be so cool to see a video on the fluke 731B
Since there is an audio servo used the question of audio grade caps from nichicon (muse) comes to mind. Those are rather lower grade caps . The effect of voltage rating also is of interest in that application . The cost difference is small in case .
It was probably a longevity decision. Most audio grade caps are only rated for 85oC.
@@zsigmondkara Good point are those 125 C then the longest lasting ?
@@johnwilliamson467 Exactly. It means they are rated to work within specifications for a given amount of time at that temperature. In the case of this amplifier 105oC caps are going to work for a much longer time at lower temperatures.
@@zsigmondkara For these caps the rule of thumb is 2x per 10c work for about a 40 c range thus 120c over 85 c is aprox 16 times the life given the test I have watched.
Marco, please do a teardown of the Fluke 731a!!
Pretty impressive equipment indeed, Marco! 😊
Stay safe there with your family! 🖖😊
Sounds like a threat.
@@bismuth7730 Absolutely no. "Enjoy your next 24hs", on the other hand... 🤨
Great content, Marco! Im playing currently with rtl-sdr's so this was interesting for me as well.
That: okay, bye got me off guard xD
I didn't know a Low Frequency Ultralow Noise Preamplifier could be so sexy...
Impressive.
Very nice.
Lets see Dave Jones' teardown.
Noice!
I see hat you did there, and what a clever design it is.
0:30 The "BNC Polarity" is just a bit 'funny'(being an AC coupled device), but I guess they're referring to the battery charging function. Poor label layout, that BNC Polarity advice should have been printed directly adjacent to the battery charging note for it to make immediate sense.
Hey, we liked your feedback and found it very valid, so we updated the design of the top sticker 👍
This is literally the first time in my 51 ( so far) life I've ever heard the term "millihertz". Crazy!
Google "HP3325 generator" if you want to play with mHz. Got one in my shack.
Hi, anyone knows where to buy metal can shown at 8:55 and it's connector sold on to PCB?
ahahaha singular 1 "Hert" 3:44
3:42 "...one hert... ...one hert..." The singular of 'hertz' is still 'hertz'. 🙂!!!
what is the software you use to simulate the circuit? thank you very much, as always excelsior the content of the channel. greetings
I would like to see the noise measurement of a 9v battery using this setup. It would be a good follow-up to the noise measured in the previous video.
You know, I used to not be concerned with all this precision stuff. Now I'm seeking out and hanging on to every ppm.
Yes its the reps effect
I personally enjoy my own naturally occurring noise sources. You have an approximate 2kHz noise source whenever unmuting your mic. This may be from test equipment active in a lab, not studio environment.
At 8:07, what circuit simulation software are you using please?
falstad
So cool.
I wonder how the guys at welectron handle your humor. I'm in Canada a purchased a bm869s recently and they were about as helpful to me as the dolphins were to the fish population before they left earth haha
I wonder if this could be coerced into some close in phase noise measurement?
Which simulation software Marco use in its videos?
There's me expecting a huge InterFET FET - I've seen they have some running at >50 euros a piece but low noise is the name of the game there. I'll stick to the LSK389 and similar.
Wow I might actually pester my boss into buying this
8K oscilloscope. NICE
In what area is this device used? How is it applied?
Nice relatively simple design - but my brain is still searching for a practical application 😅
Make guitar more louder obviously. Maybe? I hope!
To measure how noisy is the DC volatge reference. The less noise it has the more exactly you can measure its voltage, nowadays possible to 7 places which is 0.1 ppm.
you can tell, by the content and quality of a video, if a man is married and has kids, or not.
But does it have a Swiss accent?
Nice.
I don't understand a bit of sorcery he performs here, just come here for his soothing voice 😅
Jim Williams feeling the joy in his grave
I love this stuff. However, I'm not a fan of the DC cancelation circuit. My approach was to build a composit LNA with minimal gain and use a very low noise SAR like the AD4630-24. I put a brick wall on the .1 to 10hz bins and ifft to get the RMS. No settling is really needed since there are no AC caps. Also, it's nice that the SAR has a little bit of over rang, so you can measure the same reference noise.
With 10V Input Voltage and 24 bit resolution you will have 600nV LSB resolution. How do you want to measure 10nV or less? You need a gain of 100 or more.
needs more doggo
Aw, I thought this was LF in the radio sense (30 kHz - 300 kHz), but this is ELF :)
Swiss startup? Hmm, let's see where they from ... and of course, they are two villages over from where I live. 😅 Even though it's mostly villages here just at the border, for some reason this is tech central of Switzerland.
Switzerland is such a tiny country 😆
BNC charging is a bit ... silly ... however i am reminded that the end user is an idiot and would gladly keep a device like this plugged in while using it so its a nice cheap lockout mechanism
My RUclips bugged out and didn't play any sound, I thought I was the victim of an elaborate shitpost, that everyone but me was on 😭
ok bye
First, yey!
Thörd
Second
Ich fange erst zu staunen an, wenn der Verstärker von 10 Hz bis 10 Ghz einen Amplitudengang von +/- 0,1 dB hat und der Rauschpegel unter der Messgrenze liegt. 😂