[enemigo]

I'm bulilding a mint biased to class A with the JFET Cascode. I'm having trouble figuring out how the transistors are oriented, but I assume Q2 drain is connected to the OPA output, and the Q2 source is connected to the Q1 drain. Am I correct?

I'm using 2N5486 transistors for both Q1 and Q2, rather than 2N5484, will that be a problem?

Knut

 

[steinchen]

http://tangentsoft.net/audio/opamp-bias.html

you should measure IDSS of the transsistors and/or use a resistor with them

 

[enemigo]

Yepp, done that, but I can't figure out how the JFET's are wired.

Knut

 

[Daroid]

Do you mean that you are not sure what is drain, source and gate on the schematic ? Then this will maybe help: http://www.rason.org/Projects/jfetamp/design1.gif

 

[steinchen]

2n5484 datasheet

jfet schematic symbol

 

[enemigo]

I've looked at the documentation! At the tangentsoft site, there is a picture of a JFET cascode, the Q2 is "flipped". What I am wondering is if the drain of the Q2 goes to OPA output, and the source of the Q2 goes to Q1 drain. I can't see any indications of what is source and drain on that schema. I'm assuming this is correct, but just want to be sure.

Knut

 

[__redruM]

Take a look at the pimeta:
http://tangentsoft.net/audio/pimeta/...1.1a-hires.png
You're answer should be in the picture. Digikey should have the pinouts on the opamp in a data sheet.

BTW are you etching, or making your own PCB? Otherwise the CRD is the easiest way to go...

 

[enemigo]

No etching for this amp, back to protoboard now. Couldn't find the CRD at my distributor, so I'll try the cascode JFET's, space is not an issue this time. I'll study the pimeta pcb and schematic then, thanks!

Knut

 

[amb]

Look at the Pimeta schematic diagram as an example for the cascoded JFET current source, the drain is at the top for both JFETs. If you use 2N5486 for both JFETs, then you should insert a resistor (or better yet, a 1K ohm trimpot) to adjust the current to exactly what you want. If you use a fixed resistor you would have to experiment with the value to achieve the desired current. Start from about 470 ohm and go from there. See the M³ amplifier schematic for an example of two 2N5486s and a trimpot.

 

[stadams]

Enemigo,

I quickly drew these and scanned them for you. They are a top view representation of the JFET that you are intending to use in your cascode from a pin connection or footprint representation. The first drawing is for a fixed current source whose current will be dependent upon the V- that you are using or you can adjust the current by trying different values in place of the 1k resistor. The second schematic is the version that amb refers to, and is much more adjustable and allows for the use of 5486's that have much closer Idss's.

Hope this helps.

Later,

 

[enemigo]

Thanks guys!

The JFETs' current are reading
Q1 10.6 and 9.8
Q2 11.5 and 10.9

I'll try calculating a good value for the resistor for Q1.

Knut

 

[enemigo]

The BUF634 has an input impedance of 100ohm. According to tangentsoft.net, the max peak voltage will be 0.7v for 32 ohm phones. So max current level between opamp and buffer will be:
(1.414*0.7v)/100ohm=9.9mA

So to limit Q1 to 8mA:
(1.414*0.7v)/X=8mA
X=124 ohm

So chosing an Rs of 120 ohm will be correct??

Knut

 

[amb]

Quote:

Originally Posted by enemigo The BUF634 has an input impedance of 100ohm.

You read the datasheet wrong. The 100 ohm is the output load impedance condition used for the spec. The input impedance is 80M ohms in low bandwidth mode, or 8M ohms in wideband mode (each in parallel with 8pF capacitance).

So, with such a high input impedance it takes very little current through the JFET cascode to keep the opamp in class A. Even 0.5mA is more than sufficient.

 

[enemigo]

Thanks for noticing.

So for high bandwidth mode:
(1,414*0,7v)/8M ohm = 0,124uA
Down to 0,0124uA in low bandwidth.

So to bias the circuit to about 0.5mA is acheived by chosing an Rs of 1,8k ohm?
(1,414*0,7v)/1.8k ohm = 0,55mA

But since you suggest starting out with a 470 ohm resistor, or better yet, chosing a 1k pot, I take it I will be needing a bias > 0,5mA.
(1,414*0,7v)/470 ohm = 2,1mA
(1,414*0,7v)/1k ohm = 0,99mA

Just to make sure I understand this correctly. Figuring out a value for Rs is all I need to set the bias?

Knut

 

[amb]

Yep, you can set the JFET cascode current to a couple of milliamps and it'll work just fine. It's low enough not to cause any duress on the opamps or become excessive battery draw, yet deep enough in class A to feel good about.