Good things need improvement

Part 2: From Prototype to Production

Last modified: 01-Jul-2003

Basic Construction

The phonoclone will be made on a eurocard type board with soldering islands on one side of the board only. In the picture one can see how the signal goes from input (front) to the output section (back end) and how both channels are separated completely.

All resistors are metalfilm 1% types of either 0.6W or 0.25W, and every OpAmp does have 2 Elna caps of 2200uF for additional filtering as well as two Tantalum caps close to the pins 4 and 7 with values of 4.7uF and 0,1uF.

For each channel the two left most dip switch positions select the amplification factor (R0) and the rightmost 4 positions select the input impedance (100, 220, 1000 or 47k Ohms).

All additional options (impedance selection etc.) will take more room, and thus this design needs more space than the gainpre. Still, it fits in one of this flat Conrad housings (the ones I use for Tweety, Loekie, Granny etc.)

Starting Point

As explained on the first page, I started this project with the same design as the GainPre, but with the intention of making it less noisy and make input impedance and gain setting adjustable.

After reading the datasheets of the AD797 again I decided to use the AD797 again for the second stage but use lower resistor values for gain setting.

Apart from the layout on the Eurocard board, nothing really changed, and therefore the next step was to make some changes ...

First Proto

As explained in the datasheets, it is best to keep the noise floor as low as possible by choosing using lower values for the resistors in the feedback loop.

I decided to start with the 2nd stage by replacing the 160K feedback resistor and the 1k resistor on the inverted input by a 2.4k and 20 Ohms pair.

In the figure below the OpAmp in the first stage is a AD797, but I must say that I used the OPA637 Opamps of GainPre during these initial stages instead (after all that's the advantage of using IC sockets).

I'm ordering some LT1037 and LT1115 with my next orders for Conrad and DIL so it may take a while before the amp takes its final form.

2nd Version

The first modifications to the original design were OK but did not improve the noise in a huge way. Therefore it was time to rework the first stage as well, and replace the high-value resistors with lower and sometimes less noisy ones.

Also I replaced the feedback capacitor C1 and the coupling capacitor C4 by Audyn 5% MKP types. Should I build this amp ever again I would use their premium capacitors with 2% tolerance. Of course it is possible to use 1% caps but these are available as polypropylene and styroflex; expensive and only available in smaller values.

After building version 2 according to the schematic above I was surprised by the much lower noise levels. And although it is not so quiet as Loekie, I decided to give it a break-in before judging the qualities.

The finished Amplifier, one LED per channel indicating power status

The board of version 2

The backside with connectors

Version 3

The next step is to make the phonoclone really usable for MM cartridges.

In order to make the PhonoClone suitable for MM cartridges we need to limit the gain to less than 50 dB. In version 2 this can be done using the dip switches in the input section of the first stage. Looking back I should have used the second stage for gain management instead of the first stage. Reason is the feedback loop that contains a capacitor parallel with a resistor and both are in turn in series with a second resistor.

The most obvious way to limit the gain of a non-inverted OpAmp design is using a higher value for R0. However, doing this in a RIAA filter in the first stage will result in less than perfect bass response.

In any case, when using a higher value for R0 will help reducing gain, but I should have implemented a better approach. For active feedback stages it may be best to keep the gain fixed and adjust the total gain by changing the second stage only. Just my luck: I located the dipswitches at the input side only and therefore I'm limited to the crude method of reducing gain.

OK, now we know what the best way is, in version 3, I fixed the gain of the first stage to 36 (about 1+(3,300+330)/20) and rework the second stage such that gain can be set to two values: One for MC input and a second with 10 times less gain for MM input.

Therefore I changed the second stage such that the resistor R5 with a value of 2400 Ohms got a little brother R5b which has a value of only 240 Ohms. A small dipswitch S2 allows me to set the gain in the second stage to either 1+(2400/20) = 121 times or to 13 times for MC cartridges.

YEEES, this sound a lot better and for MM cartridges I have a very acceptable noise floor now which gives me even more pleasure listening to this amp.

Surprise: The design is very stable for both the OPA637 and the AD797 in the first and second stage. Given the better noise values for the AD797 obviously I decided to put that opamp in the PhonoClone. The results are better than I expected.

Conclusions & ToDo

I'm getting more satisfied about this project by the minute. The amp is soooo fast. But noise can be difficult to control with OpAmps (like humm is difficult to control with tube amps)

The following list is a placeholder for future improvement:

Well, probably most of these -wannabee's- will have to wait until I start the last OpAmp project, the PhonoCard.

<< Page 1: Intro

Back to the DIY page

Page 3: Listening >>


©, April 2003, 2006