Single rail, active I/V for TDA1543, TDA1545A

Status
Not open for further replies.
Jump to Latest
#43
Michael,

Maybe you've got some minor oscillation problems.
I included SMD base stopper resistors (optional) for most on the super pair pcb's in case anybody needs them.
What transistors do you use for the I/V ?
What kind of supply regulator/caps do you use to feed the TDA1545A with ?
 
#45
rbroer said:
Michael,

Maybe you've got some minor oscillation problems.
I included SMD base stopper resistors (optional) for most on the super pair pcb's in case anybody needs them.
What transistors do you use for the I/V ?
What kind of supply regulator/caps do you use to feed the TDA1545A with ?
Click to expand...

Did I hear the word PCB?
Is that PCB for the 1543/5 or for the 1541? (if it could be good for both I'd love it...)
Any details ?

Cheers

Andrea
 
#46
PCB's

Andrea,

I've had numerous requests for pcb's so I've ordered one final batch, most of 'm sent out, just a few left, basically the TDA1543 version from the super pair thread and a generic super pair I/V for e.g. TDA1541
But it's a lot of work supporting everyone, helping people source components and in return I just asked for some feedback on this forum, like Peter Daniel did.
A few boards left of each, contact me offline if interested. But please do read the "super pair I/V thread for TDA1543, it mentions the "special" parts you'll need and the circuit which is almost the same in the latest incarnation...
 
#47
Rudolf,

I use BC547B and BC557B transistors as shown in your schematic. Supply for the TDA1545A is a TL431 regulator and 470µF electrolytic, bypassed with 100nF Wima cap. I don´t think the hiss comes from the DAC supply, because sound is quite good when I use the opamp output. Definitely better than the original DAC supply, and absolutely no hiss. (the I/V opamp is in a socket, when I remove it I take Iout from the opamp socket and feed it to the SE I/V stage, so changing between the two is easy)

What values should I use for base stopper resistors? Do I need these base stoppers on every transistor or only on some of them?

Michael
 
#49
mihu said:
Rudolf,

I use BC547B and BC557B transistors as shown in your schematic. Supply for the TDA1545A is a TL431 regulator and 470µF electrolytic, bypassed with 100nF Wima cap. I don´t think the hiss comes from the DAC supply, because sound is quite good when I use the opamp output. Definitely better than the original DAC supply, and absolutely no hiss. (the I/V opamp is in a socket, when I remove it I take Iout from the opamp socket and feed it to the SE I/V stage, so changing between the two is easy)

What values should I use for base stopper resistors? Do I need these base stoppers on every transistor or only on some of them?

Michael
Click to expand...

Hi Michael,
I've seen values for base stoppers ranging from 22 ~ 100 ohms or so.
I would 1st make a quick test; disconnect the DAC output from the I/V and listen to it (the I/V that is). This will help you determine where to look next...
 
#50
Hi Rudolf,

I checked that yesterday and the results are:

Ranking for lowest noise while not playing music is
- SE stage disconnected from DAC, no cap between GND and V+
- SE stage disconnected from DAC, 2200µF between GND and V+
- SE stage connected, but not playing
- AD712 Opamp output is noisiest

When I play music, things change. Now the Opamp stage has lowest background noise, SE stage has more noise and distortion with or without 2200µF between V+ and GND.

I have no clue how to interpret these findings...
Michael
 
#51
upgrade with follower?

rbroer said:
Michael,

I built it a while ago on vero-board, it's quite easy.
IMHO it sounds better than the active I/V posted in the "simple I/V" threads, but not as smooth as the folded cascode. YMMV.
I used just the raw supply for the positive rail, no need for regulator because of the constant current for the LED/diodes references.
Click to expand...


Dear Rudolf!
The literature says that transistor circuit made as common base has high output impedance. So why do not to connect directly output of your circiut to the brave follower (that you have introduced in a different place of diyaudio) in which the fet is working with a constant power?
The simulation of that follower shows that he is "almost distortionless" in a wide band of input voltages!
After this we may to load it with very low impedances making it independent of amplifier impedance input and from the influence of the weak interconnects.
The only one minus of this is that EF might not sound too good...
 
#52
Re: upgrade with follower?

padamiecki said:

Dear Rudolf!
The literature says that transistor circuit made as common base has high output impedance. So why do not to connect directly output of your circiut to the brave follower (that you have introduced in a different place of diyaudio) in which the fet is working with a constant power?
The simulation of that follower shows that he is "almost distortionless" in a wide band of input voltages!
After this we may to load it with very low impedances making it independent of amplifier impedance input and from the influence of the weak interconnects.
The only one minus of this is that EF might not sound too good...
Click to expand...

Sure, that's something you can try. We're eager to hear how it sounds and compares to a bufferless circuit.
I wouldn't put all my money on simulator THD results 😉
 
#53
Re: upgrade with follower?

padamiecki said:



Dear Rudolf!
The literature says that transistor circuit made as common base has high output impedance. So why do not to connect directly output of your circiut to the brave follower (that you have introduced in a different place of diyaudio) in which the fet is working with a constant power?
The simulation of that follower shows that he is "almost distortionless" in a wide band of input voltages!
After this we may to load it with very low impedances making it independent of amplifier impedance input and from the influence of the weak interconnects.
The only one minus of this is that EF might not sound too good...
Click to expand...

Tell me if I'm wrong, but the common base itself has obviously a high impedance, since it's a current source.
The I/V is done with the output resistor.
 
#54
mihu said:
Hi Rudolf,

I checked that yesterday and the results are:

Ranking for lowest noise while not playing music is
- SE stage disconnected from DAC, no cap between GND and V+
- SE stage disconnected from DAC, 2200µF between GND and V+
- SE stage connected, but not playing
- AD712 Opamp output is noisiest

When I play music, things change. Now the Opamp stage has lowest background noise, SE stage has more noise and distortion with or without 2200µF between V+ and GND.

I have no clue how to interpret these findings...
Michael
Click to expand...

I'm running out of suggestions, you might wanna check this thread:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=31279
 
#55
Re: Re: upgrade with follower?

Bricolo said:


Tell me if I'm wrong, but the common base itself has obviously a high impedance, since it's a current source.
The I/V is done with the output resistor.
Click to expand...


yes, i think so
but anyway EF would make more comfortable working of the output transistor,
do you think that without this resistor (dc coupling to the FET's gate) there will be no I/V convert because of extreme high impedance of EF?
 
#56
This has been discussed before; and I think that the conclusion was "you must have a resistor somewhere to do the I/V conversion"

What else can you use for I/V? I don't see any other component that has such a linear current/voltage caracteristic


I/V, either active or passive have always be made with resistors, AFAIK

I see Rudolf's circuit, and the other ones (opamps, common base... even transformers) as something to help the DAC to drive the I/V resistor. But the last one is always needed.


But I might be wrong. Maybe you have something new in mind. I'm curious to see a pure semiconductor based I/V 🙂



Your first suggestion (to add a buffer, and to let the resistors in place) is interesting. Rudolf's circuit has a Zout of 1.2kR, but my gainclone has a 10kR input impedance. A buffer may be welcome
 
Status
Not open for further replies.
Top Bottom