Simple I/V for TDA1541

[rbroer]

Built this the last week; it's a variation of Jocko's simple I/V, using all generic unmatched parts, larger bias currents, and the R4 upstream of C3.
This is used in a non-oversampling cd player with Kwak-Clock version 3 and TDA1541A-S1.
I have adjusted Vc(Q3) with pot R6 to 10.0 Vdc
Ve(Q3) is adjusted with pot R3 to 0 Vdc, and readjusted after a few hours. It doesn't change much over time, within +/-1.0mV.
Current source Q1 is about 23mA, of which 6mA flows into R4, 17mA through Q3 and 15mA through current sink Q4.

All bjt's have hfe of ca. 300-400 measured with multimeter.

Since I didn't measure Vbe on Q3 and Q2 I ended up with different R3 values; ca. 5.6k and 22k.
I suppose one could select to get the lowest R3 value, that is, have some moderate current flowing into Q2.

The sound ?
Well my previous I/V was a LM6172 opamp driving a J309 jfet biased by few mA, Riv between +V and fet's drain.
Later I replaced the J309 by a BS170 mosfet which has higher transconductance, hence less "correction" to be done by opamp on fet's gate to keep output voltage compliance of 0V.
The latter sounded more sweet, treble got more transparant.
Just last week I replaced the BS170 with a BC547B and it sounded a little better again, so I decided to built this circuit.
Got an improvement again; sounds like treble is more free and transparent and basses seem to be driven by some "mass".

This circuit can be universally used for other current output DAC's;
use pot R6 to get the wanted Vce on Q3.
To use it with Philips DACs like the TDA1543 and TDA1545A with DC voltage on output, replace Q2 with LED's, zener, or diodes to get stay within the valid DC compliance range.

I do have some SA970's and SC2240's to try, which might sound better...but even with generic parts it sounds pretty good.

Have fun,

[Nicke]

Nice to see someone else tried Jocko´s circuit.

I have tried a similar in a DAC with PCM63´s,but with 2SA970´s & 2SC2240´s.
I have put a capacitator in parallell to R4 as a filter and all is followed with a simple JFET follower(2SK170).

It sounds really good,better then the previous analog stage which was the usual opamp style.

Nicke

[Kuei Yang Wang]

Hi,

Originally posted by rbroer
I do have some SA970's and SC2240's to try, which might sound better...but even with generic parts it sounds pretty good.
[/B]

You may wish to consider cascoding both the current source (load)transistor AND the current conveyor transistor and to use a lot of voltage in the "+B"....

BTW, the same principle of course works great with valves too....

Sayonara

[Ric Schultz]

Try replacing the LEDs with three signal diodes. Try fixed resistors instead of pots once you have figured out the value needed. Try changing the 10K bias resistors (currently used on LEDs) to 20K and putting them to the opposite power supply rail instead of to ground. If you do any of these mods, let us know what you hear. To me LEDs and pots are extremely non-linear (noisy and distorted).

Ric Schultz

[rbroer]

Cascoding:
Yes, I've been thinking to make the constant current source a little better by cascoding, but, higher B+ are hard to find in cd players (this is not a standalone DAC). Dunno if there will be audiable gains though.
A next version shall use a SA970 as CB, so I can cascode the current sink with a J111 fet.
What is the benefit of cascoding the current conveyor transistor ?
Then the cascode transistor will have the "problem" of a variable collector voltage.

LEDs & pots:
Really ? In fact I did some measurements last week, comparing how different LEDs and three 4148 change their voltage with increasing current. Red was best, green next, yellow changes a lot, like the three diodes. As a reference for a BJT CCS I suppose the steadier, the better.
Many other inmates seem to be using them.

But I'll compare LED/diodes on the pcb version I'll be using on a cd player with the TDA1543 (this was just a quick prototype).
A hunch tells me you already know the outcome Rick ?

Indeed those resistors on the LEDs can be connected to the opposite rail. One can even save a resistor by connecting the upper LED in series with 18k and next in series with the lower LED.

Happy new year to y'all.

[Kuei Yang Wang]

Hi,

Originally posted by rbroer
Cascoding:
Yes, I've been thinking to make the constant current source a little better by cascoding, but, higher B+ are hard to find in cd players (this is not a standalone DAC).

For a higher +B simply build an external supply and while you are at it also make seperate suppies in the same case for clock and DAC, leaving only servo and uP powered from the internal supplies.

What is the benefit of cascoding the current conveyor transistor ?
Then the cascode transistor will have the "problem" of a variable collector voltage.

The changes in collector voltage change the transfer curve of the Transistor. If you "nail down" this voltage via a cascode you much reduce this effect. Same goes for FET's and Valves.

If you then use a suitably high voltage cascode transistor (or better FET, Valve!?) you can use a very high Supply Voltage which increases headroom and helps to keep the modulation of the Collector/Emitter or Drain/Source Voltage in percent of the absolute voltage small.

My Final Gambit would then be a suitable output transformer driven by the cascode and directly into a suitably low load...

LEDs & pots:

I do agree on the pot's (but that should be self evident), on the LED Front, why not use a TL/LM431 instead, it costs more than LED's or diodes but is best in the context and for a DIY Unit....

Also, feed references NOT via resistors but via current sources, J-Fet "constant current diodes" or selected J-Fets tend to work fine...

Later T

[tschrama]

Hee.....this IV-stuff just keeps coming back.. great ... I use a very simple IV stage with 1 BJT, 1 diode, 1 cap and 3 resistors.. and it works just fine..... better than passive IV with just a resistor I think.. but how about this one .. DC to 10MHz.... better put some LP in it somewhere..nice output of 2Vrms too .. and low distortion.. but propably needs very good supply... all simulated offcourse (if you know me )...

best of new years greetings,
Thijs

[OliverD]

@tschrama:

Your I/V-stage seems to be fed by a voltage source w/ series resistor... Hm. Why don't you use a current source in the simulation?

@Nicke:

Could we have a look at your circuit?

Happy new year!

[Nicke]

Originally posted by AMT-freak
[B

@Nicke:

Could we have a look at your circuit?

[/B]

It is almost the same as Jocko published but without the coupling cap.
And it is followed with a JFET source follower.
I don´t want to publish yet,because I got some clues from Jocko in a private email to him and it is his decision to publish it.
If you read the whole thread you could figure it out myself.
Why don´t you try the published circuit to start with?
I am very satisfied with my I-V stage.

A last hint:Look at Nelson´s SU-SY patent.

[jkeny]

Rudolf,
How does the sound compare to your similar circuit with irf610 on 36v rails?

Can the CCS be done with simple jfet & resistor?

John