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.
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.
Attachments
Thank you Rudolf, I´ll try that version. I guess for a TDA1545 I have to replace D5 with a 1N4148, anything else to change? What type of capacitor do you recommend for C1? 100µF is a pretty high value that makes a a "better quality" cap pretty expensive.
How about driving a passive preamp from this stage?
Michael
How about driving a passive preamp from this stage?
Michael
- read my earlier posts for MY opinionsAndypairo said:
- I think I used something like raw 18V after the rectifiers
- depends on voltage of battery and value of Riv
- green
- don't think any change needed for TDA1545A, except Riv/Civmihu said:
- use whatever cap YOU like. I just used 100u for simulations, 2.2u Audyn KpSn in real life.
- well..., try, and let us know
The intend was to keep it simple, active stage with low parts count, just one rail in stead of two, using raw supply.
For really low parts count, one can even ommit casode transistor in the current sink (Q2, D2) and the common base (Q3, D3) as well.
Bricolo, at a first glance what you ask is not trivial.
The emitter of Q4 is at a fixed voltage (2Vled - 0,7V), and you should bring that node to a voltage suitable for the DC compliance of the 1541, which is much smaller as I recall.
However, if you bring the Idac node too low the current sink tied to Gnd can't work anymore, hence a negative supply is needed...
Cheers
Andrea
The emitter of Q4 is at a fixed voltage (2Vled - 0,7V), and you should bring that node to a voltage suitable for the DC compliance of the 1541, which is much smaller as I recall.
However, if you bring the Idac node too low the current sink tied to Gnd can't work anymore, hence a negative supply is needed...
Cheers
Andrea
I meant 2*Vled - 0,7V which is the Vbe of the Bjt.
Vled it the forward drop of the leds, and depends on the colour and (slightly) on the current that flows in them.
If you use them at appropriate currents the voltage can be considered constant.
Check some datasheets for nominal voltage drops.
Cheers
Andrea
Vled it the forward drop of the leds, and depends on the colour and (slightly) on the current that flows in them.
If you use them at appropriate currents the voltage can be considered constant.
Check some datasheets for nominal voltage drops.
Cheers
Andrea
I built one channel of the SE output stage this evening and connected it to my Marantz CD38 (TDA1545A DAC).
It works, but there is a humming noise I can´t get rid of. The humm is there no matter whether the player is playing or not. GND is connected near the DAC chip, V+ (12,5V) and I-dac is taken from the I/V opamp socket (opamp removed of course..). Cables are approx. 20cm from Player to output stage and 20cm from output stage to cinch connector. Civ, C1 and R23 are soldered to the cinch connector, the rest is built on a small veroboard. Values for Riv and Civ are as shown in Rudolf´s schematic, what values should I use for the TDA1545? Any ideas where the humm could come from? I had to use a BF245B instead of the BF245A, could this make a difference?
Michael
It works, but there is a humming noise I can´t get rid of. The humm is there no matter whether the player is playing or not. GND is connected near the DAC chip, V+ (12,5V) and I-dac is taken from the I/V opamp socket (opamp removed of course..). Cables are approx. 20cm from Player to output stage and 20cm from output stage to cinch connector. Civ, C1 and R23 are soldered to the cinch connector, the rest is built on a small veroboard. Values for Riv and Civ are as shown in Rudolf´s schematic, what values should I use for the TDA1545? Any ideas where the humm could come from? I had to use a BF245B instead of the BF245A, could this make a difference?
Michael
@Bricolo:
Yes, the hum is there with output connected or not, frequency could be around 100Hz (that´s just a guess...). I tried 10µF between GND and V+ on the veroboard, didn´t change anything.
@Andrea
Could you explain what this difference in Idss means for the whole output stage? Sorry, I´m a mechanical guy and don´t know a lot about electronics.
A revised schematic wouldn´t be useful, I built it according to Rudolf´s schematic, only difference is the BF245"B" instead of the "A" (couldn´t get the "A" type) and a 2,2µF MKT output cap instead of the 100µF in the schematic. I can´t measure right now, but voltage drop over the lower 2 LEDs (D4/D5) is 2V/LED and 0,7V for the 1N4148 (D3) giving 4,7V total. D1 seems to be not as bright as D4/D5, and voltage drop is a llittle lower. Power supply is 12,5V from the I/V opamp socket. I´m using green LEDs, as Rudolf suggested.
@Moderator
There seems to be a lot of interest in the SE output stage, so if Rudolf agrees it could make sense to move post #47 and up to a new thread?
Yes, the hum is there with output connected or not, frequency could be around 100Hz (that´s just a guess...). I tried 10µF between GND and V+ on the veroboard, didn´t change anything.
@Andrea
Could you explain what this difference in Idss means for the whole output stage? Sorry, I´m a mechanical guy and don´t know a lot about electronics.
A revised schematic wouldn´t be useful, I built it according to Rudolf´s schematic, only difference is the BF245"B" instead of the "A" (couldn´t get the "A" type) and a 2,2µF MKT output cap instead of the 100µF in the schematic. I can´t measure right now, but voltage drop over the lower 2 LEDs (D4/D5) is 2V/LED and 0,7V for the 1N4148 (D3) giving 4,7V total. D1 seems to be not as bright as D4/D5, and voltage drop is a llittle lower. Power supply is 12,5V from the I/V opamp socket. I´m using green LEDs, as Rudolf suggested.
@Moderator
There seems to be a lot of interest in the SE output stage, so if Rudolf agrees it could make sense to move post #47 and up to a new thread?
Michael,
I think you're hitting a rail, your reference current is too big I think.
You will need a multimeter.
The BF245A together with R22 forms a constant current course.
Measure and choose R22 such that you get about 1mA
When using the BF245B, you need a higher value for R22 (measure with 2.2k for R22 as a starting point) to get the same 1mA. You can also use a LM334 (see its datasheet).
I hope you guys read the other threads as well and understand RSO1A is a potentiometer. Michael, when using the same Riv and Civ, I would suggest a 200 ohms potentiometer, set right in the middle (105 ohms) is good starting point prior to adjusting it.
When playing no music, I would adjust RSO1A such, that voltage OUT1 is right in the middle between VB(Q2) and VB(Q3). This gives the same clipping margin for positive and negative output swings.
R23 is there for simulator only, I never installed it.
When using the 3k3/220pF from the original CD723 opamp I/V,
you just need to adjust RSOA1 higher, say 150 ohms.
Here are some voltages for your application, hope it helps.
I think you're hitting a rail, your reference current is too big I think.
You will need a multimeter.
The BF245A together with R22 forms a constant current course.
Measure and choose R22 such that you get about 1mA
When using the BF245B, you need a higher value for R22 (measure with 2.2k for R22 as a starting point) to get the same 1mA. You can also use a LM334 (see its datasheet).
I hope you guys read the other threads as well and understand RSO1A is a potentiometer. Michael, when using the same Riv and Civ, I would suggest a 200 ohms potentiometer, set right in the middle (105 ohms) is good starting point prior to adjusting it.
When playing no music, I would adjust RSO1A such, that voltage OUT1 is right in the middle between VB(Q2) and VB(Q3). This gives the same clipping margin for positive and negative output swings.
R23 is there for simulator only, I never installed it.
When using the 3k3/220pF from the original CD723 opamp I/V,
you just need to adjust RSOA1 higher, say 150 ohms.
Here are some voltages for your application, hope it helps.
Attachments
Well, let's see if I understand the circuit:
The CCS of the JFet is used to give the right polarization to the LED/Bipolar Current sources/sinks.
In Riv flows the difference between the upper CCS current and the current sunk by the lower BJT (4,5 mA?) plus the DAC current, say about 1mA at idle.
Changing The upper CCS current we have to get the voltage on Riv at about half ail voltage.
Correct me if I'm wrong, I just gave it a quick look..
Cheers
Andrea
The CCS of the JFet is used to give the right polarization to the LED/Bipolar Current sources/sinks.
In Riv flows the difference between the upper CCS current and the current sunk by the lower BJT (4,5 mA?) plus the DAC current, say about 1mA at idle.
Changing The upper CCS current we have to get the voltage on Riv at about half ail voltage.
Correct me if I'm wrong, I just gave it a quick look..
Cheers
Andrea
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.