|Equipment & Tools From test equipment to hand tools|
Please consider donating to help us continue to serve you.
Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
||Thread Tools||Search this Thread|
|7th February 2011, 04:24 PM||#42|
|10th February 2011, 03:40 PM||#44|
The Hammond enclosure is fully drilled and the board fit very well.
I have also designed a laser printable sticker as a silkscreen for front and rear panel. It's easy to do and result is pretty good.
I will add it in the next design folder of this project that i prepare.
You can see the result on pictures below.
Next step now, is to test it !
I'll do that this saturday. To soon.
|20th February 2011, 08:06 AM||#45|
ERMSDCV2-- All measurments results.
After soldering all parts, do the functional tests and drill it's enclosure,
i have run many measurements on my ERMSDCV2 to qualify it's practical specs.
Test concern the three functions of the ERMSDCV2 design :
_The DC..1MHz RMS to DC converter
_The LOG AC voltmeter.
_The 80dB 10Hz..100kHz LNA
As you will see below, results are very good. The instruments could be really
useful for audio/electronics investigations.
I have write a test and trimming procedure allowing very simple adjustment
to get a fully functional and calibrated device. I will send new documents to all PCB buyers.
Note that i had need to do some little modifications on the design to correct mistakes.
(Please read the wiki main page about that). This will be also included in new files.
You can download the latest schematic version (including all modifications)
here : ERMSDCV2.3_sch.pdf .
Otherwise specified, all tests are made using 1X probe (a coaxial cable) with the
jumper W3 closed, and front panel switch in DC position (down).
The device is powered by a +/-9VDC linear PSU.(+/-6v to +/-15v maxi)
1/ The DC..1MHz RMS to DC converter.
1.A/ Linearity test.(DC output level vs AC input level for 1kHz sine wave).
As you see, the RMS converter work very good from about tenth mV to more than 2.5V.
The usable lower limit is about 20mV (<10% error).
1.B/ Measurement error test.(Error in % of input level for 1kHz sine wave).
1.C/ Frequency response measurement test.(Error in % of input vs frequency for 1Vrms sine wave).
The error versus frequency show very good linearity up to 400kHz and fall at about 600kHz.
The frequency flatness depend largely on the AD817 low-pass filter Butterworth conformance (set to 1.8MHz cutoff).
1.D/ 5Vpk-pk 10kHz Triangular wave RMS measurement test. (Yellow=signal input , Blue trace= DC output).
Here i test the RMS conversion of triangular wave (crest factor of 1.7).
1.E/ 2.5Vpk-pk 5% duty-cycle rectangular wave RMS measurement test. (Yellow=signal input , Blue trace= DC output).
Test with high crest factor signal (here 4.5), the RMS conversion stay very good despite this.
1.F/ Digital pink noise RMS measurement test. (Yellow=signal input , Blue trace= DC output).
Test with random digital pink noise, again very good RMS conversion.
I have also try measurements with x10 and x100 scope probe (W3 jumper open),
allowing the extend of measurement range, and it work fine.
You just need to trim your probe compensation for frequency flatness and then you can measure up to 300Vrms!
(with a 100x probe)
2/ The LOG AC voltmeter.
Note: To test the log voltmeter output, the input signal must be very very low,
so i use precision shielded passive voltage divider to get the right input level.
2.A/ Linearity test.(DC output level normalized as dB vs AC input level for 1kHz sine wave).
2.B/ Measurement error test.(Error in dB of input level for 1kHz sine wave).
2.C/ Log output frequency response test.(Vin=1Vrms).
2.D/ Log output frequency response test, zoomed view of the previous graph.(Vin=1Vrms).
The front-end of the LOG voltmeter(U16-U12)) use an OPA627 followed by an AD817.
The noise floor of the log voltmeter depend largely of the noise of these two IC.
On my design, with 0dBV calibrated to 1Vrms, i get a noise floor of about
-84.5dBV 60ÁV).The AD8307 noise floor is about -90dBV.
As you see the frequency flatness is about +/-0.1dB in 100Hz-1MHz range !
3/ The 80dB 10Hz..100kHz Low Noise Amplifier.
3.A/ LNA frequency response test (normalized).(Vin=100ÁVrms red= Gain , blue=Phase).
3.B/ LNA frequency response test, zoomed view (normalized).(Vin=100ÁVrms red= Gain , blue=Phase).
3.C/ LNA maximum output voltage before saturate.(Vin=310ÁVrms Yellow=LNA output).
3.D/ LNA output voltage for 100ÁV 1kHz sinewave input.(Vin=310ÁVrms Yellow=LNA output).
3.E/ LNA pulse response.(Vin=+/-100ÁV 10kHz square wave Green=signal input before attenuator Yellow=LNA output).
This test show clean response of the Butterworth fifth order low pass filter.
3.F/ LNA noise floor.(No signal, 50 Ohms plug on LNA input, Yellow=LNA output).
This signal show the own output noise of the LNA.This represent about 520nV RTI (Referred To Input).
It's mainly the noise of input OPAMPs, this is why they must have very good noise specs.
3.G/ LNA low level signal reconstruction.(Vin=500nV 1kHz sinewave Green=signal input before attenuator
Using the scope digital averaging to eliminate the uncorrelated noise, you can show clearly a sinwave signal
as small as 500nVrms.
The 50mV (green trace) is attenuated by 100dB (1/100000) before to go at LNA input.
3.H/ LNA Ultra low level signal reconstruction.(Vin=100nV 1kHz square wave Yellow=LNA output).
Here, it's same as previous screenshot, with 100nV pk-pk square wave signal and scope digital averaging.
3.F/ LNA noise floor spectral view. (No signal, 50 Ohms plug on LNA input, Blue=LNA output spectrum).
Here, the full output spectrum of the LNA output noise.The noise is very flat in all bandwidth, showing white noise type.
No visible harmonics of main (50/60Hz and multiples) and no spurious.
The ADC show a RMS level of -47.27dBV that correspond to about 4.33mVrms. (433nV RTI)
Roughly the same measurement value done with scope on section 3.F.
Ouf ! I'm sorry this long reading...But if it's any consolation, it was even longer for me to write that ! ;-)
I hope all builders will be happy with their devices. For me, that will be a new useful "DIY" instrument in my lab.
Don't hesitate if you have any question, i will always try to answer you.
Many others info and files for download can be find on the dedicated wiki page HERE . Thank's !
(Note that i still have some PCB available for this design, i you're interested just send me a PM).
|20th February 2011, 09:55 AM||#46|
Join Date: Jul 2004
Location: Scottish Borders
An enormous body of work, just to show us test results.
Thank you !
regards Andrew T.
|20th February 2011, 10:47 AM||#47|
Join Date: Dec 2003
Location: Frankfurt area
The results are very impressive, including your very extensive and professional testing. So first of all many thanks for making this happen. The questions that I have right now are:
1. The LTC1968 is really tiny. How did you manage to solder it to the PC board - are there any special tricks? I had already ordered a sample (and of 1967) a long time ago but did not venture using them until your project came along.
2. The OPA627 is very expensive, what I saw so far is well above EUR 20,--. Are there any equivalent OpAmps with the same specs that we might substitute? I can go and file through spec sheets but maybe you know some already.
|20th February 2011, 11:50 AM||#49|
i soldered the LT1968 (MSOP package) using a small iron tip, a good light and a magnifying glass. Of this work require very good eyes !
You can also to try a method that consisting in place the IC, and then doing a big solder along each sides and remove excess of tin with a desoldering braid.
You can find many video showing how to do that, on Youtube for example.
You can replace the OPA627 OPAMP by others, with FET input(low Ibias), high frequency bandwidth and low noise. The AD711 can be good low cost choice. Many others can work too. Note that if you need to use the LOG output, the measurements noise floor depend on this IC (and the AD817).
I'm sorry but i don't want to organize that, it will take a time that i don't have.
But of course, you can do it !
|22nd February 2011, 04:28 PM||#50|
I'm a little surprised about the little readers comments.
I have made today new measurements that i don't have done before.
RMS to DC converter settling time.
The input signal is a 1Vrms 100Hz sine-wave (blue), the output in yellow
show the response time.
As you could see on the picture, the output settle in about 500ms.
(front panel switch in DC mode).
RMS to DC converter DC linearity .
In these tests i use a DC scignal to be measured by the ERMDCV2.
(front panel switch in DC mode).
A/ Linearity vs DC input level 0 to 5VDC .
B/ Linearity vs DC input level zoomed view, 0 to 250mVDC.
C/ % Error vs DC input level 0 to 5VDC .
C/ % Error vs DC input level 0 to 250mVDC .
Note that all measurements i have done has been made with precision calibrated DC voltmeter and DC generator.
To summarize, i'm very happy about all results obtain.
The LTC1968 and the AD8307 really keep their promise and offer very good solution.
My next DIY project that i will post, will change a little of previous.
I design a small laboratory SMPS based on resonant fly-back converter.
It's specs will be ; 75w output power, 2 symmetric variable outputs (+/-5..+/-15V) and another output of +3 to +9V.Each output could handle 2Amp, small standard Hammond enclosure, low output noise, easy to build.
Of course, i will do new post when i will be ready.
(The project have already seriously progressed)
|Thread Tools||Search this Thread|
|Thread||Thread Starter||Forum||Replies||Last Post|
|Simple 60dB discrete low noise amplifier (lna)||iko||Solid State||440||19th November 2010 04:18 PM|
|New To Site?||Need Help?|