AMZ Mosfet Booster Biasing?

[nessbass]

Hey guys,

first post here. I have some questions regarding Jack Orman's Mosfet Booster.

After I've built one this weekend, I was digging around in the archives here today and read of several people having bias problems with the circuit. Since I had noticed a slight distortion of the attack portion of my bass's signal, I thought I could try something in this regard.

In all of the cases in the archives, it was recommended to use a pot for R2 to find the right resistance value so that the voltage at ref point A would be 4.5-5.5 volts. So I unsoldered the 62k resistor that was there and connected a 100k pot instead. With the resistor still in place, the voltage had been 5.6 volts. With the pot, voltages from 4.7 to almost 7 could be dialled in.

In order to find the proper voltage, I connected the Booster to my computer so that a 1kHz sine wave was fed to the input and the output fed the computer's soundcard. Using a software scope, I noticed that with the stock value, the negative-going cycle was distorted (not clipped, but bent slightly), so I twisted the pot until the signal looked most symmetrical.

The value I determined this way was 81k, the resulting voltage at the reference point 6.33 volts (supply was 9.34). So why is it that these values deviate so much from what is spec'd?

I then went on to trust my ears more than the scope and plugged my bass in. On one end of the rotation of the pot, the signal died (voltage at A: 4.72), before that, it was quite heavily distorted. Going the opposite direction, the signal became more solid, less distorted and louder. And that went on till the end of the pot (which is 90.2, not exactly 100k)! It didn't get worse....?

I thought that was strange enough to post a thread here. So, can anybody make sense of this? I used a BS170, BTW.

Thanks in advance,

David

[nessbass]

BUMP

[davebungo]

Hello,
you didn't say wether or not you had the gain switch (SW1) in or out so I'll assume it was out.  

Consider the DC conditions of this circuit.  You have between the 9V rail and 0V, three components: a 2K7, the drain-source of Q1 and another 2K7.  The same current flows through all three components.  So, if you measure 5.6V at Q1 drain (VD) you must have 9-5.6=3.4V at the source (VS) as the same current flows through both 2K7 resistors generating the same voltage across each.  As VD decreases on the output negative going cycle of the signal, VS will increase.  But, and here is the catch, what happens to VDS i.e. VD-VS the voltage across Q1 drain-source?  

In the quiescent state VDS = 5.6 - 3.4 = 2.2V.  For every mV VD decreases, VS increases by 1mV and VDS decreases by 2mV.  If you examine the output characteristic of a MOSFET, you will see that there is no drain current flow when VDS falls below a certain level (typically less than a volt or so I'm guessing).  So what you are experiencing is Q1 cutting off due to a lack of VDS.  By increasing VD (by bias adjustment) you reduce VS and increase VDS giving the circuit more leg room.  Obviously, the opposite is also true i.e. VD cannot increase beyond 9V so there will be a happy medium value to be had which allows the maximum signal swing before the onset of either cut-off in either direction.  I would have thought this is going to be around VD=6.7V (VS=2.3V and VDS=4.4V).

[nessbass]

Thanks for your reply, davebungo.

Your explanation really makes sense. And my measured 6.33 are quite close to your estimated 6.7, so we are on the same track here. I also read your post about the circuit simulation you did with it. Did you try it out yet? Oh, and it doesn't matter where the gain knob is set, I get these slight distortions even at unity. But I would really need the knob to sit at approx three o'clock (more distortion there, of course).

But why is it that all this is not dealt with in the original article, why is the biasing describes as it is? Maybe amz-fx can chime in here...

[davebungo]

No I never did try it out.  I'm not sure why the original article specified the bias that way but perhaps it sounds OK for a 6 string guitar and is more noticeably distorted on a bass especially as you would normally be playing more single notes (and perhaps the bass guitar output is higher?).

I think there is a common misconception about biasing in a lot of circuits where the bias will be adjusted to achieve 4.5V or half rail at the drain or collector assuming that this means that the output can then swing either way to the supply limits.  This would be true for op-amp circuits but not for most discrete transistor circuits.  Note that Jack specified 4.5V to 5.5V and not 3.5V to 4.5V i.e. he was heading in the right direction and provided the signal at the output is not too large it would probably sound nice and maybe "warm".  With high gain settings, the output would be hitting the end stops anyway so the asymmetry is maybe not a problem and may even be desirable.  I wouldn't get too hung up on this question - just adjust it to your own personal taste and as long as you understand what is going on with the circuit you will be happy.

[nessbass]

Alright, thanks a lot!

David

[amz-fx]

Simulations and and theoretical calculations are all well and good, and I use them in designing a circuit.  However, all of  my circuits are based on a real-life prototype of the project and not just estimated values.

The threshold voltage of mosfets will vary just as they do for jfets, so a calculated value may or may not be correct.  Also if you substitute a BS170 from a different maker, or you use a 2N7000 or other substitute, the Vth will likely be different from what I got in the prototype.  

The value of the bias resistor R2 may have to be tweaked to get the drain voltage in the correct range...  R2 will likely fall in the range of 62k to 100k depending on your particular mosfet.

If your drain volatge is 4.5v to 5.5v then it is okay to use without further tweaking.

regards, Jack

[davebungo]

If the drain is at 4.5V then you will have a nice half wave rectifier.  If the bias is increased further then you will get less and less of the peak of the signal until a point where you get nothing.  The DC drain voltage will limit at nominally 4.5V.  It can't physically go any lower.