I have not posted much here because i was busy to do other things.
One is to support the Paradise builders. Many thanks to Frans, Alfred and Michael to sort out the remaining issues. The first boards are now playing and i am dedicated to help the other ones play too.
Some time ago we started to design a single ended stage called Cro Magnon.
This is not forgotten and i have now a concept that can work. Maybe in the next days i publish the concept. The main problem is the " current fighting " we experienced in the input stage. To make a one stage single ended non negative feedback RIAA the input stage needs to have a very high output impedance working on a very high output impedance CCS. The thermal drift in that arrangement can set up the DC conditions.
I think that can be solved. Preferably the whole stage should be DC coupled so we need a reliable Servo.
On another thread i am designing a monitor speaker for SB Acoustics. So when you are interested how i design speakers you can go to the Verdors Forum. Under Verdors Bazar you can find the thread " SB Acoustics Satori Monitor ". I am very close now to have a pair playing. I plan to bring that to the Mini Frickelfest on 9th of february.
One is to support the Paradise builders. Many thanks to Frans, Alfred and Michael to sort out the remaining issues. The first boards are now playing and i am dedicated to help the other ones play too.
Some time ago we started to design a single ended stage called Cro Magnon.
This is not forgotten and i have now a concept that can work. Maybe in the next days i publish the concept. The main problem is the " current fighting " we experienced in the input stage. To make a one stage single ended non negative feedback RIAA the input stage needs to have a very high output impedance working on a very high output impedance CCS. The thermal drift in that arrangement can set up the DC conditions.
I think that can be solved. Preferably the whole stage should be DC coupled so we need a reliable Servo.
On another thread i am designing a monitor speaker for SB Acoustics. So when you are interested how i design speakers you can go to the Verdors Forum. Under Verdors Bazar you can find the thread " SB Acoustics Satori Monitor ". I am very close now to have a pair playing. I plan to bring that to the Mini Frickelfest on 9th of february.
I now got the Peak Atlas DCA75 Pro curve tracer. It is really cool and will allow me to measure the early voltage in BJT transistors. I also started to build a measurement setup for Rbb`. This two measurements will allow me to identify BJT transistors that can be used in low noise, high speed MC input stages. The reason for this is that the super low voltage noise Japanese BJTs from the 70th and 80th are no more. After ROHM and Toshiba, Hitachi Renesas has also given up to produce them. We can of cause parallel several medium noise ones like our BC337 and BC327 but some think that only one pair sounds better. With the curve traces i can also find out if NPN and PNP are truly complimentary. I have designed a simple MC head where i can try the candidates in praxis.
A lot of medium power and even high power BJts have low Rbb´.
Some of them have being designed for RF or Video use.
Most of them can also stand a lot of voltage and have a low saturation voltage.
Your BD115 has quite a high saturation voltage but that does not mean that it is not usable.
Preferably i would like to find complimentary pairs because i design parallel symmetric mostly.
A problem in medium and high power BJTs can be low Hfe. When a common base stage is used that is not much of a problem because a common base stage has zero current gain. In common emitter though high Hfe plays a role to give the circuit a high enough input impedance and minimize base current or we get DC offset into the cartridge.
Hfe linearity will then also be important to make the circuit low distortion. High early voltage is good too to make the circuit fast and low distortion in the treble. Also Cob makes a contribution so output capacitance should be low. That may be the reason that i heard from seasoned designers that often transistors in a small case sound better then their bigger brothers. Anyway, i have drawn a simple parallel symmetric MC head circuit without any current source or cascode that will bring out any flaws in the chosen transistors. I will use the same circuit and only exchange the transistors. i will then measure bandwidth, distortion and noise.
Some of them have being designed for RF or Video use.
Most of them can also stand a lot of voltage and have a low saturation voltage.
Your BD115 has quite a high saturation voltage but that does not mean that it is not usable.
Preferably i would like to find complimentary pairs because i design parallel symmetric mostly.
A problem in medium and high power BJTs can be low Hfe. When a common base stage is used that is not much of a problem because a common base stage has zero current gain. In common emitter though high Hfe plays a role to give the circuit a high enough input impedance and minimize base current or we get DC offset into the cartridge.
Hfe linearity will then also be important to make the circuit low distortion. High early voltage is good too to make the circuit fast and low distortion in the treble. Also Cob makes a contribution so output capacitance should be low. That may be the reason that i heard from seasoned designers that often transistors in a small case sound better then their bigger brothers. Anyway, i have drawn a simple parallel symmetric MC head circuit without any current source or cascode that will bring out any flaws in the chosen transistors. I will use the same circuit and only exchange the transistors. i will then measure bandwidth, distortion and noise.
Low noise, low Rbb transistors
Hola.
Toshiba 2sc3329, Comp: 2sa1316
"For Low Noise Audio Amplifier Applications and
Recommended for The First Stages of MC Head
Amplifiers"
Best regards,
Francisco
Hola.
Toshiba 2sc3329, Comp: 2sa1316
"For Low Noise Audio Amplifier Applications and
Recommended for The First Stages of MC Head
Amplifiers"
Best regards,
Francisco
First tests have shown interesting results. Our classic BC550C, BC560C pair shows very good linearity of Hfe and a high early voltage up to 2.5mA. More then that the PNP and NPN go astray and the transistors get slower and dissimilar. I would say a decent MC head can be made of 4 pairs running on a total idle of 10mA.
The now for secret Sanyo medium power video transistors show very high early voltage up to high currents. I can not measure more then 12mA though. Unfortunately they do not match well in Hfe but that can be selected. Lets see how they fare on the noise and bandwidth test.
The super low voltage noise Hitachi-Renesas do great up to 10mA but there i also experienced a Hfe mismatch. Again that can be selected.
The now for secret Sanyo medium power video transistors show very high early voltage up to high currents. I can not measure more then 12mA though. Unfortunately they do not match well in Hfe but that can be selected. Lets see how they fare on the noise and bandwidth test.
The super low voltage noise Hitachi-Renesas do great up to 10mA but there i also experienced a Hfe mismatch. Again that can be selected.
The Toshibas should do great too. I have some but they are now even harder to get then the Renesas. Also the ROHMs are more or less extinct. This is the main reason that i am searching for alternatives.
Here are the ROHMs.
Yes, Salas, the new Altas gives curves of idle current over CE voltage over base current.
Maximum idle is 12mA and maximum voltage is 12V.
The straighter the curves, the higher the early voltage and the better is the BJT as a current source so it is also a measure of output impedance.
Unfortunately the current software does not allow printing. The content can be paisted into the clipboard but i have no spread sheet program on that particular computer.
I buy this week a new laptop that will allow me to do that.
I could do a screen shot with camera.
Yes, Salas, the new Altas gives curves of idle current over CE voltage over base current.
Maximum idle is 12mA and maximum voltage is 12V.
The straighter the curves, the higher the early voltage and the better is the BJT as a current source so it is also a measure of output impedance.
Unfortunately the current software does not allow printing. The content can be paisted into the clipboard but i have no spread sheet program on that particular computer.
I buy this week a new laptop that will allow me to do that.
I could do a screen shot with camera.
It also gives Hfe, Ube and leakage, it is also very easy and convenient to use.
There is a small graphics bug when you make the second measurement on a pair.
The curves get cluttered so the second measurement has to be erased and repeated.
There is a small graphics bug when you make the second measurement on a pair.
The curves get cluttered so the second measurement has to be erased and repeated.
Yes, i have.
Here you find more information :Peak Electronic Design Limited - Atlas DCA Pro - Advanced Semiconductor Analyser - Model DCA75
Provided that the measurements are accurate this little Gem is really helpful.
Here you find more information :Peak Electronic Design Limited - Atlas DCA Pro - Advanced Semiconductor Analyser - Model DCA75
Provided that the measurements are accurate this little Gem is really helpful.
As i said before i have also build the Rbb´jig that Vogel shows in his book " The Sound of Silence ". I may be able to make a wide band measurement of noise and a spectral analysis too in the near future.
Its good to know more about what you got in your semis drawers. Do you find the Atlas verifies the curves in the datasheets from the few tests you did, or there are serious deviations? (up to 12mA, 12V of course).
Well, i did not compare this to the nth degree but i will.
From the session tonight i know that each type of transistor needs an optimal bias current.
For example the BC550C, BC560C combo really detoriorates with higher idle then 2.5mA-
This is actually no new news but the curve traces allows for better fine tuning and matching.
From the session tonight i know that each type of transistor needs an optimal bias current.
For example the BC550C, BC560C combo really detoriorates with higher idle then 2.5mA-
This is actually no new news but the curve traces allows for better fine tuning and matching.
Here you see what we call in German " Kennlinienfeld".
It shows that the Vce curve against idle gets flatter when the idle goes down.
My BC550C and BC560C samples did even better.
http://www.htl-rankweil.at/grass/ftkl/datenblatt_BC547.pdf
It shows that the Vce curve against idle gets flatter when the idle goes down.
My BC550C and BC560C samples did even better.
http://www.htl-rankweil.at/grass/ftkl/datenblatt_BC547.pdf
I think 12V Vce is sufficient in a low voltage noise BJT design because with higher voltage there is a risk of avalanche noise. The bad thing with avalanche noise is that ones the BJT is over driven the noise does not go down again.