diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Analogue Source (http://www.diyaudio.com/forums/analogue-source/)
-   -   more gain please (http://www.diyaudio.com/forums/analogue-source/183457-more-gain-please.html)

martingagnon 19th February 2011 01:10 AM

more gain please
 
i'm amplifying a Denon DL-160 Phono Cartridge

here is the spec:
Output: 1.6mV
Stylus: .1 x .2mm special elliptical diamond
Cantilever: aluminum
Frequency range: 20-50,000Hz
Compliance: 14x10-6cm/dyne
Tracking force: 1.3-1.9g
Weight: 4.8g


with a Burr Brown INA103 and im getting a gain of about 55dB and i need 75dB
so my question is how do i achieve a range of 70 to 80dB with a buffer, what kind etc...

here is the schematic:

http://martin_gagnon.tripod.com/images/INA103.jpg

and my hack:

http://martin_gagnon.tripod.com/images/pre.jpg


by the way it sound wonderful with a 64-Bit RIAA Vinyl Correction Curve. (pure vinyl)

thanks

Martin Gagnon

indianajo 20th February 2011 04:35 PM

Look at U1B in the second picture of this thread:http://www.diyaudio.com/forums/analo...-dj-mixer.html It is set up as a unity gain buffer with input resistor of 10k and feedback resistor of 10k. You want about a gain of 4, so the feedback resistor should be about 39k or 47k. The feedback resistor is the one coming from the point of the triangle (the output) to the minus input. Metal film resistors make the least noise. Put the gain 4 buffer between the point of your device and the output. Your power supply looks okay for 5 more milliamps. The drawing shows 5532 op amp. I like ST33078 or NJM4560 for better slew rate & noise specs. Both of the latter numbers you have to put .01 to .1 uf disc caps on the power supply within an inch of the IC, and a 20 pf or so disc cap across the feedback resistor, to prevent oscillation. Some people say 4562 is even better than 4560, but I haven't tried it or the 5532 either. Watch on 5532, the most common package I have found is 14 pin DIP which is different pinout than the 8 pin dip package of 33078, 4560, 4562. You have to check the datasheet (from datasheetcatalog.com) of what you buy against the pinout shown on the +in, -in, and out, of the schematic on the link. Watch buying SIP package: they are made, but hard to buy IC sockets for. For a board to put it on, look at Small IC Prototype Board | 21-110 (21110) | Distributed By MCM
Don't forget to buy 10 uf electrolytic caps to put on the op amp PCB, sockets, standoffs and screws to screw it to something, and hookup wire. Get it all on one $6 shipping fee. Farnell.com operates in Canada, also somebody recommended future in Vancouver. Crossing the border runs freight up to $12 or more UPS; maybe cheaper USPS/RMail.

martingagnon 20th February 2011 09:35 PM

is this good
 
http://martin_gagnon.tripod.com/images/NE5534.jpg

thanks for the guidance

Martin

indianajo 21st February 2011 03:29 AM

Close. The 20 pf cap goes in parallel with (across) the 47K resistor not in series with it. It shorts out the 47k at very high frequencies, causing the op amp to have low gain at those frequencies (which you can't hear anyway). The 10 uf just need to be on the PCB somewhere near the second op amp. The fast op amps also need a .01 or .1 disc cap really close, I put mine right between pins 8 & 4 of the socket on the back. You may be able to get away without the disc caps on a 5534, I've never used it but it's slew rate is slower. You also need a 10 k between the output of your old IC and the minus in of the new one, and a 1k from the output of the old amp to signal ground.
I don't have a datasheet on a 5534, I think the pin numbers you are using are for an obsolete TO5 round package or something. On the 8 pin dip package that nearly everybody stocks, 8 is plus supply, 4 is -supply, 5 is amp plus in, 6 is amp minus in, 7 is amp out, other amp (2 to a package) 3 is plus in, 2 is minus in, 1 is output. when you get them, pin 1 has a dot on it, or if there is only a cutout on one end, holding that end up pin 1 is top left from the top side. Legs point down at the socket. Count pins counterclockwise from the topside. To be really specific I have ST33078N's from farnell.com (newark.com in NC, USA) with legs that plug into tyco phosphor bronze sockets I soldered into my disco mixer. The D package has the legs cut off and bent for surface mount package, which package you only use with $50000 wave solder machines (you don't have). Put the op amp in the socket after the soldering is all over, and inspect for bridging, it is hard not to bridge those tiny pins. (I did it twice in 4 op amps, had to suck the solder off with old stripped wire).

martingagnon 23rd February 2011 12:46 AM

and like that
 
http://martin_gagnon.tripod.com/imag...03-ne5534a.jpg

thanks again

Martin

indianajo 23rd February 2011 01:01 PM

yes.

AndrewT 23rd February 2011 01:27 PM

the inverting opamp has a gain set by the 47k and the total source impedance seen at the -IN terminal.
If INA103 has Zout = zero ohms then the gain is -4.7 (+13.44dB inverted).
The opamp also multiplies the INA103 noise by the same factor, +13.44dB.

You asked for 10times (=+20dB). you need a ratio of 1:10 for those gain setting resistors.

I cannot see the extra 1k loading the output of the INA103 as having any advantage. My view is that it reduces the performance of the INA for no benefits at all.

kevinkr 23rd February 2011 01:54 PM

I noticed that the input gain setting configuration is not per the recommended configuration in the TI/BB application notes. Should be 15 & 13 tied together, and 6 & 2 tied together with the gain setting resistor (or rheostat connected pot) placed between them. The gain equation as implemented currently will be incorrect. This is actually a pretty serious error and also has implications for CMRR performance if you are running a balanced connection. (Not to mention DC offset) To fix just jumper pins 2 & 6..

Note also from a noise perspective it would be better to realize as much of the gain as possible in the front end of the INA103 as opposed to an external op-amp as that device will also amplify all of the noise generated by the INA. 60dB is possible with an INA103 with an rg of about 6 ohms.

The differential input impedance needs to be 47K to properly load the cartridge, I'd change those 10K input resistors to 24K which would be close enough. I doubt this is having a huge effect on cartridge performance since the generator impedance is pretty low.

I'm also puzzled by why you think you need so much more gain with a DL-160, with "just" 60dB of gain you'll get 1.6Vrms @5cm/sec which should be more than enough for most applications, and if it isn't you'd need only a couple of dB more. Your proposed gain would result in 4.8Vrms (70dB) and 16Vrms (80dB) which is clearly a bit over the top.. :D

Note also that the inverting mode of operation in an op-amp has an inherent noise disadvantage because the (noise) gain at the non inverting input is 1+Rf/Rin and at the inverting input Rf/Rin. This is very significant at low gains. (Worst case is a unity gain inverter where the noise penalty is 6dB) Note that most modern op-amps have sufficiently good common mode performance that inverting operation is not going to cause meaningful errors in audio applications.

Personally I think you'd be better off applying the EQ in the analog domain and then all of the frequency dependent headroom vs SNR issues go away at the input of your ADC...

AndrewT 23rd February 2011 02:10 PM

Quote:

Originally Posted by kevinkr (Post 2480045)
Note also that the inverting mode of operation in an op-amp has an inherent noise disadvantage because the (noise) gain at the non inverting input is 1+Rf/Rin and at the inverting input Rf/Rin. This is very significant at low gains. (Worst case is a unity gain inverter where the noise penalty is 6dB)

putting the inverting opamp second in the amplifying chain avoids this added noise problem, particularly when the previous stage has lots of gain and the inverting stage uses a low noise configuration (combination of opamp choice and resistor values).

kevinkr 23rd February 2011 02:23 PM

Quote:

Originally Posted by AndrewT (Post 2480067)
putting the inverting opamp second in the amplifying chain avoids this added noise problem, particularly when the previous stage has lots of gain and the inverting stage uses a low noise configuration (combination of opamp choice and resistor values).

True, it is not a big contribution given the huge amount of gain ahead of it, but I was trying however to suggest that perhaps that op-amp is not needed at all. (And raise the OP's awareness of noise issues in general) It seems to me that the OP really doesn't want the kind of output voltages implied by 70 - 80dB of gain. At 80dB an op-amp providing the additional gain would have to be running on +/-25V rails minimum and there are few that can actually do that. (OPA445 is one that can)

I use less gain with a 350uV LOMC (65dB with passive EQ) and get sufficient output level even for the ADCs in my media server.


All times are GMT. The time now is 01:26 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2