Krill - The little amp that might...

Status
Not open for further replies.
Jump to Latest
#561
Hi John, another question:
in the 50W version the VAS BJTs are 2SC2240/2SA970 while in the 100W version you specified 2SC1845/2SA992.

I didn't pay attention to this difference (was thinking about using just the power section) until I decided to try the VAS section too.

I guess the diference lies in the different power handling of the devices (300mW vs 500mW), that make the 2sC2240 unsuitable for the 100W version (where each device of the LTP has to dissipate about 150mW).

I searched on various online shops but most don't carry either the 2SC1845 or the 2SA992...and I don't want to pay more in shipping than for the devices....
Do you have any suggestion for (cheap) online sources and/or suitable substitutes?

Last question, should I look for the highest Hfe grade?

Ciao

Andrea
 
#565
Andypairo said:
Hi John, another question:
in the 50W version the VAS BJTs are 2SC2240/2SA970 while in the 100W version you specified 2SC1845/2SA992.

I didn't pay attention to this difference (was thinking about using just the power section) until I decided to try the VAS section too.

I guess the diference lies in the different power handling of the devices (300mW vs 500mW), that make the 2sC2240 unsuitable for the 100W version (where each device of the LTP has to dissipate about 150mW).

I searched on various online shops but most don't carry either the 2SC1845 or the 2SA992...and I don't want to pay more in shipping than for the devices....
Do you have any suggestion for (cheap) online sources and/or suitable substitutes?

Last question, should I look for the highest Hfe grade?

Ciao

Andrea
Click to expand...


Hi Andrea,

Who is John?

I changed to the newer transistors because they have very similar specs, are easier to find and cheaper.

Q1 and Q2 will dissipate about 0.056W. Q3 about 0.12W so either transistor is suitable.

I always buy the highest Hfe grade. I see no reason not to in this application.

All the transistors are available from Mouser. I do not know what shipping would be to Italy. If shipping is prohibitive from suppliers I can send you the parts you need. Those transistors should mail in an envelope, which would be fairly cheap.
 
#566
Hi Steve,
sorry for calling you John 😀

I made a rough calculation and actually Q3 passes nearly 3mA, resulting in about 120-150mW
Q1 and Q2 pass half the current so the power is about halved.

If the transistor model change is not caused by technical reasons I think I will use the 2SC2240/2SA970 I have.... actually I might also use the dual 2SC3067 for the diff pair, what do you think?

http://pdf1.alldatasheet.com/datasheet-pdf/view/108588/SANYO/2SC3067.html

Ciao

Andrea:smash:
 
#567
Andypairo said:
Hi Steve,
sorry for calling you John 😀

I made a rough calculation and actually Q3 passes nearly 3mA, resulting in about 120-150mW
Q1 and Q2 pass half the current so the power is about halved.

If the transistor model change is not caused by technical reasons I think I will use the 2SC2240/2SA970 I have.... actually I might also use the dual 2SC3067 for the diff pair, what do you think?

http://pdf1.alldatasheet.com/datasheet-pdf/view/108588/SANYO/2SC3067.html

Ciao

Andrea:smash:
Click to expand...


The 2SC3067 should work without any problems as long as you can fit it to the board. It looks like one transistor in the pair will have the pins reversed.

R5 has 2mA of current (1.995ma). How do you get Q3 at 3mA?
 
#569
Andypairo said:
Hi Steve,
I assumed 0.7V per diode, so 0.7/240 = almost 3mA

Probably the drop on the diode is pretty lower at the actual working current, so I think I overstimated the current and consequently power dissipation.

Yes the dual BJT will need some adapting but I think it will also provide better matching and thermal tracking.

Ciao

Andrea
Click to expand...


OK. I see where we have a small discrepancy. I am taking into account the temperature of the components. At start up, the current through R5 is 2.18mA and the voltage drop across D1 or D2 is 0.6V. As the circuit stabilizes the temperature of the components rises causing a lower voltage drop across the diodes and less current in R5.

This is where actual measurements come in handy. I must admit it is a pain epoxying thermocouples to all the components you want data on.
 
#570
Component Layout

Hi c2cthomas,

In your latest version of the component layout (post #554), C4 seems to have disappeared. Since it does exist on the board, I assume this is an error?

Thanks for all your hard work on this!

Best wishes
Dave
 
#571
Hi Steve,
assembly of the first board is almost done - basically I'm missing the 56V zeners and a couple of trimmers.

A couple of questions: on the board there is provision for a cap named C2 between base and collector of Q4.
I guess the cap needs to be installed only in case of oscillations, right?

I also noticed the space for a resistor (labeled R28 I think, silkscreen is blurred) that seems to allow balanced input (toghether with C22 if I get it).

In case one wants to use the Krill in a balanced fashion, which values should have R28 and C22?

Ciao

Andrea
 
#572
Andypairo said:
Hi Steve,
assembly of the first board is almost done - basically I'm missing the 56V zeners and a couple of trimmers.

A couple of questions: on the board there is provision for a cap named C2 between base and collector of Q4.
I guess the cap needs to be installed only in case of oscillations, right?

I also noticed the space for a resistor (labeled R28 I think, silkscreen is blurred) that seems to allow balanced input (toghether with C22 if I get it).

In case one wants to use the Krill in a balanced fashion, which values should have R28 and C22?

Ciao

Andrea
Click to expand...


You are correct about C2. I left a place for it if it was needed. The value would be 30pF or less. I never had a problem but this makes it easy to correct if there was one.

A place for R28 was put on the board so I could access the inverting input of the amp. You could use a pair balanced as you say, but the intent at the time was to use the row of holes there (1-6) to mount a daughter board to use the amp in a bi or tri amp system.

The value of R28 is selected to set the inverting gain equal to the non inverting gain. For the 100W amp the gain is 21. For an inverting amp, R10 is replaced by R28. The value of R28 would be 952.4. Use the closest 1% value of 953.
 
#575
Hi,

i did a 50W version of the Krill Amp using alex mm early pcb.
There are some errors on this pcb which represent errors in the early schematic of the 50w version. I think its quite interesting that the combination of sc3298 and sa1306 from chinese vendor isc in position q8 q11 is not working. Bias could not be adjusted. A replaclement with bd139/bd140 cured this. In my build, R12 had to reduced to 132Ohm to get rid of the static bias (Carlos had the same Problem i think)
Thank you Steve for this nice amp ...
 
#576
heinz said:
Hi,

i did a 50W version of the Krill Amp using alex mm early pcb.
There are some errors on this pcb which represent errors in the early schematic of the 50w version. I think its quite interesting that the combination of sc3298 and sa1306 from chinese vendor isc in position q8 q11 is not working. Bias could not be adjusted. A replaclement with bd139/bd140 cured this. In my build, R12 had to reduced to 132Ohm to get rid of the static bias (Carlos had the same Problem i think)
Thank you Steve for this nice amp ...
Click to expand...

Hello Heinz - one of the mistakes that I discovered on the PWB's manufactured by Steve versus the schematic for the 50W version is that R9 actually goes to GND (not -36V regulated). I don't know if this error is on the "alex" PWB or not. You might want to investigate this. According to Carlos correction of this mistake ran much better in his simulations.

Congratulations on your construction!! Any photo's?
 
#578
What does that mean, Stinius? If you are trying to start another Class A/ not Class A debate then forget it!!

Edit: OK it was pointed out to me that you were just pointing to the origination of the Class A confusion - apologies if I was blunt but I'm sick of hearing this Class A S**T
 
Status
Not open for further replies.
Top Bottom