Question about biasing

[nee]

Greetings!

I built a distortion unit using J201s, breadboarded it, set the biasing with trimmers so the drains were 4.5V and replaced the trimmers with resistors. I built it into an enclosure to try live. Sounded good, but I had some ideas for a few improvements.

I breadboarded the original circuit again and it was immediately apparent that biasing of one (or more) of the J201s was not right. I put in a trimmer and where originally I had a 47K drain resistor, now I needed 22K.

Is it normal for J201s to be so "all over the place"? A 50% reduction in the value of the drain resistor seems mighty large.


IanG

[aron]

I would reset the biasing when it's on the board - not only on the breadboard. I've found the breadboards to be somewhat unstable in a variety of ways.

[John Lyons]

Jfets vary a a good bit.
Thats why most circuits use trim pots.
Either that or they use a self biasing shceme.
Search for STM and Jfet and you'll fnd a monster thread.

john

[demonstar]

As John said JFETs do vary a lot from device to device. What this means is if you used different devices each time you breadboarded and didn't design for the variation in devices you can experience problems such as significant variation in operating point, such as you did.

The topic of biasing JFETs comes up a lot here and often provides a lot of dicussion and disputes over the 'correct' method of doing so. This can lead to a lot of confusion and misunderstandings in biasing these devices to account for variations. If you search the forum you are bound to find a lot of these discussions on this topic. One article that nearly always gets mentioned is the one below so you may want to take a look at it.

"http://www.datasheetcatalog.org/datasheet/vishay/70595.pdf"

I've found sometimes there's no substitute for a good textbook and a lot of spare time when it comes to learning about things like this.

[kurtlives]

I would reset the biasing when it's on the board - not only on the breadboard. I've found the breadboards to be somewhat unstable in a variety of ways.

Could you expand on that please?

[frank_p]

I would reset the biasing when it's on the board - not only on the breadboard. I've found the breadboards to be somewhat unstable in a variety of ways.

Could you expand on that please?

I don't want to talk for Aron.  But could it be that a plugboard or any type of connector might have more resistance (contact resistance) than solder.  So when biasing some amp stage, you can (un-or)amplify these current (and volages) losses.

Anyway, I've built some springboards (as connector) and had this problem.

[aron]

Thanks Frank. Yes, sometimes resistance, sometimes weird capacitance issues, unstable - parts falling out, touching etc....

I would bias (if possible) in circuit and then replace with a fixed resistor if possible.

[demonstar]

I was just reading what I posted late lastnight and now I'm awake I thought I should point something else out. Remember when John and I said with JFETs one has to design for variation in device to device well that's why you chose to use the trimpot so you can alter the operating point in the desired range even when device characteristics are changing.

[DougH]

Some things to think about-

1. 4.5v is not the ideal bias voltage. You want the drain voltage half way between the supply voltage and the source voltage for maximum headroom. Source voltage will depend on the size of the source resistor and the drain resistor. Ideal drain voltage will be > 4.5v.

2. When you change the value of the drain resistor (to account for device variation), you are also changing the cutoff frequency of the high pass filter formed by the RC combination of the drain resistor and the coupling cap it's attached to. Depending on how much the resistor changes (2x is significant), the difference in the filter frequency will be audible.

3. When you change the value of the drain resistor (to account for device variation), you are also changing the overall gain of the stage. Depending on how much the resistor changes, the difference in the gain will be audible  (gain change of 2 is 6db which is noticeable, but gain change will also depend on the load of the next stage so 2x resistance change may not equal 2x gain change).

Different trim settings to set bias for varying devices can produce significant sound changes for two seemingly identical circuits.

[nee]

Some things to think about-

1. 4.5v is not the ideal bias voltage. You want the drain voltage half way between the supply voltage and the source voltage for maximum headroom. Source voltage will depend on the size of the source resistor and the drain resistor. Ideal drain voltage will be > 4.5v.

2. When you change the value of the drain resistor (to account for device variation), you are also changing the cutoff frequency of the high pass filter formed by the RC combination of the drain resistor and the coupling cap it's attached to. Depending on how much the resistor changes (2x is significant), the difference in the filter frequency will be audible.

3. When you change the value of the drain resistor (to account for device variation), you are also changing the overall gain of the stage. Depending on how much the resistor changes, the difference in the gain will be audible  (gain change of 2 is 6db which is noticeable, but gain change will also depend on the load of the next stage so 2x resistance change may not equal 2x gain change).

Different trim settings to set bias for varying devices can produce significant sound changes for two seemingly identical circuits.

Thank you all for your replies. Doug, your point number 1 caused a penny to drop, thanks, and points 2 and 3 made me realise that twirling a drain trimmer willy-nilly is not the answer. This made me seek further information in the large "Source v Drain trimmers" thread, and the AN102 datasheet. Gad! The more one knows the more one needs to know!!

[DougH]

Gad! The more one knows the more one needs to know!!

Yeah well, I'm lazy and got away from fooling with JFETs a long time ago due to all the complications. And after reading AN102 I haven't really mustered up the energy to start messing with them again. That bias selection method just looked like too much work to me. IMO, there's nothing really special about the sound of JFET distortion. The little "ampy" circuits are fun but I suspect a lot of that has to do with EQ and gain staging more than anything.

The way I look at it is there are a bazillion ways of distorting a signal- pick one. Then it's basically the same game- gain, EQ, headroom, etc... I suspect the easiest way to work with JFETs is probably just to preselect them, as was covered in one of the other threads.