Michele's New Ideas - Solid State Preamp

[m_buzzi]

Hi all,

Since I have at home a lot BF245C, BC550, BC560 and lots of led I decided to design a preamp with these chips.

This is the best I managed to take out. Spice says that THD on 10k load at full output is about 4*10^-4%. Risetime is about 1uS.

The output stage is highly biased (40mA) since I wanna have the opportunity of connect an headphone and use this preamp as an headphone amp.

Many of the LEDS seems to be reverse biased but this is only because in the simulation I substituted them with a zener.

50 Ohm Resistors on BF245 gate stands for a trimmer which will be used to adjust balance in the diffamp.

I was wondering if a DC-Servo would be a important improvement. Is it?

Please feel free to express your advice on the design I made.

Thanks,
Michele

[Giaime]

Hola Michele!

Good design. A couple of notes:

- try it in reality, not just in SPICE

- you'll need pretty matched transistors. I suggest replacing at least R14 (or others with the same placement) with a trimmer, to adjust quiescent current.

- BF245? Use japan semiconductors!!!

I'd surely want to hear how does it sound

[m_buzzi]

Hi Giaime,

Thanks for replying!!

Now I'm really busy due to exams at university, so I will try this design as soon as possible, probably this weekend I can prototype one channel. I do know that with Japanese transistors will work better but I have a lot of those transistor so I can match them really well. If I can find some LSK389 I will try them out. Spice says that THD at full load reduces to 1.3*10^-4.....

I forget to specify it but the bias current in the output stage will be regulated via a 22Ohm fixed and 50 Ohm trimmer.

Michele

[scott wurcer]

[i]

- you'll need pretty matched transistors. I suggest replacing at least R14 (or others with the same placement) with a trimmer, to adjust quiescent current.

- BF245? Use japan semiconductors!!!

[/B]

I have found the Philips "knock-offs" of some of the Toshiba JFETS to be quite cost effective (ie. really cheap) and very available. I have no problems with the performance either. Hand matching SOT23's is another matter.

[m_buzzi]

Hi,

You are really right. There are SMD transistors that are much better than the usual small signal TO-92 with very high transition frequency and low input capacitance. I was thinking infact to try out a discrete SMD opamp but the little spare time I have blocks me.

Would you improve the schematic with a push pull output stage or I can leave it with the single ended CFP output?

Michele

[anatech]

Hi Scott,
I made a couple of adapters for SOT type packages for a couple meters I like to use for matching. Believe it or not, I use a Heathkit IT-18 to match surface mount transistors when I service a Cyrus.

The adapter is simply a SOT pattern on a PCB that plugs into a normal transistor socket. Just don't sneeze when matching parts.

Don't laugh, because I have sneezed - twice.

-Chris

[scott wurcer]

Originally posted by m_buzzi
Hi,

You are really right. There are SMD transistors that are much better than the usual small signal TO-92 with very high transition frequency and low input capacitance. I was thinking infact to try out a discrete SMD opamp but the little spare time I have blocks me.

Would you improve the schematic with a push pull output stage or I can leave it with the single ended CFP output?

Michele

Not necessarily much better but roughly equivalent and the die in many cases is the same in SOT or TO. For instance the BF862 specs almost lie on top of the 2SK170. The few I measured looked good for low frequncy noise too. At 15 cents each I bought whole tubes.