Simple I/V for TDA1541

[rbroer]

jkeny:
Dunno, never used the IRF610 since inmate Dave from downunder compared it to BJT, and latter sounded and measured better.

After all suggestions I've come up with a version I'll try next (still enjoying the 1st version though);
I added a PNP cascode on the current conveyor to keep it's Vce pretty constant and as a bonus it creates more Vce voltage headroom on Q5 since it's referenced to negative rail.
I have cascoded current sink Q7 with Q6 since this one will be "seeing" the output voltage swing.
I1 ~ I4 are constant current sources made of jfets with resistor.

I'll be using 2SA970´s & 2SC2240´s
bias Q3, Q5 = 15mA
Vce(Q5) = 19V,
Ve(Q5) = 11V,
Vc(Q5) = -8V (+/-3V signal swing)

Any suggestions for a "bigger" Q5 ?
I may add the Borbely follower with two 2SK170 BL's

Regards,

[sonnya]

When using 2SC2240/SA970 at currents above 7mA you will operate them outside there linear range.

Pretty good transistor with good hfe linearity is parts like BC807/817 SMBT06/56(1 - 70mA) or BC846/847/848/849/850/856/857/858/859/860 (100uA - 8mA)

Sonny

[sonnya]

i did post some wrong info about 2SC2240 and 2SA970. The 2240 can easyli be used up to around 30mA but the complementary 970 starts rolling off at around 10mA.

I have a suggestion :
Design a complementary I/V stage. Does not have to use more than 4 BJT's in the active stage.
pro's :
Lower input impedance the half of a single BJT.
DAC see's a constant impedance.

[Nicke]

It works great with 5 mA through each stage also...
Perhaps it sounds better with more,but I haven´t tried that yet.
I think that 2SA970 & 2SC2240 is rated at 0,3 W so about 10 mA should work without problems with 15 V rails.

Nicke

[tschrama]

Sonnya,

check the other I/V thread ... there you can the my efforts in designing complementary I/V stage. .. simulated great!


That other guy found a way optimize the canceling of the 2nd harmonic distortion, allthough he says it not a slightly different way and makes it look like something bad...

[ThorstenL]

Hi,

I have a suggestion :
Design a complementary I/V stage. Does not have to use more than 4 BJT's in the active stage.
pro's :
Lower input impedance the half of a single BJT.
DAC see's a constant impedance.

Well, in that case, why not use a transconductance cell on a chip ("diamond transistor") like the OPA660? There you have already perfectly matched transistors on one chip, a buffer and can be run open loop. I think AD also has a transconductance Op-Amp (743? not sure) where the transconductance node is available externally and that can be used open loop thusly....

Of course, that replaces all that lovely discrete stuff with a chip and a few resistors but it will sure sound better AND measure better than all that cludge, but it would go against the DIY Spirit. Anyway, these days it is VERY HARD to make discrete designs that are even competetive against competent implementations of the right type of chip, but that's just me.

Sayonara

[tschrama]

Hi Kuei Yang Wang,

very interesting ... please could you make a simple example schematic...

gr,
Thijs

[sonnya]

Originally posted by tschrama
Sonnya,

check the other I/V thread ... there you can the my efforts in designing complementary I/V stage. .. simulated great!


That other guy found a way optimize the canceling of the 2nd harmonic distortion, allthough he says it not a slightly different way and makes it look like something bad...

Yes i have read it. Add a folded cascode as you mentioned instead of your caps in the collectors. When modifyed this circuit could easely work with PCM1738... I know i know this is not part of this thread.

By the way. Use Switchercad III from linear tech. with this simulator you won't get fantastic distortion levels.

My circuit measure around -92dB based on the modified Jocko circuit first posted from Rbroer but with theoreticly current sinks and sources!

if i interchange them with fet's or BJT the numbers get 10dB higher.

[ThorstenL]

Hi,

please could you make a simple example schematic...

Not off-hand, but it really should be obvious from the Datasheet. Here the basics for the OPA660.

You connect a 270 Ohm resistor for biasing from pin 1 to the negative rail.

You decouple the negative and positive rail (pin 4 negative and Pin 7 positive) well and according to RF principles to ground, supply +/-5V.

You connect pin 3 via 100 Ohm to ground and pin 2 to the current input from the DAC. You place the required I/V conversion resistor on pin 8 and link pin 8 via a 100 Ohm resistor to pin 5.

Buffered output is from pin 6 via a 51 Ohm Build out resistor.

This has no lowpass, I would recommend a 62nF capacitor to ground on the output, this forms with the 51 ohm build out resistor a 50KHz 1st order lowpass, sufficient if the DAC is operated with 4 Times oversampling or more. You could make the Filter 2nd Order by paralleling the I/V resistor with a suitable value capacitor. The requirements for filtering for non oversampling operation are controversial, many people have their own ideas, as do I.

With -4mA Peak current from a TDA1541 the I/V conversion resistor should be sized so that the typhical output Voltage compliance is not exceeded or a +2mA compensating current should be injected into the TDA1541 Output. This is done easiest by using a 2k4 Resistor from a Well filtered +5V supply via a further 100ohm/1,000uF Filter section.

In this case the compensating current is used we have +/-2mA full scale peak current out of pin 8. To get 2.8V Peak (2V RMS) we thus need a 1k4 resistor, nearest standard value is 1k5 so we use that from pin 8 to ground. For a 50KHz 2nd Order lowpass we use a 33nF capacitor on the output and a 1nF capacitor in parallel to our 1k5 I/V conversion resistor.

I would feel that shunt regulating the +5V and -5V supplies via TL431/LM431 Shunt regulators is highly desirable.

L8er T

[tschrama]

Got it, thanks!

Nice interesting datasheet... I always love those simplified circuit diagrams.. Maybe use BUF634 instead of the 'onboard buffer' and connect direct to your favorite headphone


gr,
Thijs

PS
I'm still trying to make a perfect, ultra low noise, super fast V/I convertor (1-2mA max)at here at my work (Phd student). Wouldn't this opa660, with emmitor degeneration do this job?