Another boring FET bias thread (azabache)

[thermionix]

yeah i suppose any dc blocking is going to happen regardless of value

Unless it leaks.  But in that position, I think a leaky cap would lower your source voltage, not raise it, and only measurably when the bright switch was on, shunting the 1M.  And leaky caps just aren't that common these days.  Seems like the problem almost has to be a faulty FET or a bad connection at either end of the 470R.

[reddesert]

I think it would be helpful to measure the Vp and Idss of the FETs to rule out the possibility that the FETs are out of spec. You can measure them with ROG's simple circuit in Fig 10 of http://runoffgroove.com/fetzervalve.html

The Vsource of 2.7V is unusually large and likely outside the range the circuit is intended to operate. If the FET, drain trimmer, and source resistor are all connected correctly, then you should have had a reasonable DC value for Vsource even if there is a weird component after Q1.

I have an ROG Ginger vero build here, whose Q1 and Q2 stages are fairly similar to the Azabache's Q1, and on the Ginger Q1 and Q2 I measure Vgate = 0, Vsource =0.8-0.9V, Vdrain = 6.0V. (This is with 2N5458 FETs not 2N5457, but I measured the Vp and Idss to be reasonably close to the spec.)

[diydave]

"meten is weten".
Before messing with fet's (in any circuit), it's always a good thing to measure the Vp and Idss of each fet you are going to use.
Keep in mind that it's not the Drain-resistor which biases the Fet (eventhough you always see those damn drain-trimpots), it's the Source-resistor.
Rd just gives load and gain, not bias.

If you know Vp and Idss, you can calculate any Vgs and appropriate Rs.
Id = Idss x {1-(vgs/vp)}²
Rs = Vgs / Id

I use spreadsheet in which 10 values of Vgs and Rs are calculated (every 1/10th of the Vp). This gives me a nice list, and even a line-chart, in which I can choose my biaspoint. In doing so, I can nearly predict if some value will work or not. Once you know your Id and Vgs, it's simple to calculate an aprioriate Rd to give load and gain to your fet.

It's possible that the 470R Source-resistor for the fet you are going to use just biases the fet to much near Vp (cut-off). This can make the fet sputter. Swapping out fets (2n5457 by mpf102) will not cut  it, because of the big spec differences between fets, and surely between fets from another "brand".

If it's a quicky, I sometimes use a rule of thumb:
half Vp gives about 1/4th of Idss.
Example (same numbers of runoffgrooves 2n5457)
VP = - 1.58 V
Idss = 3.27 mA

For Vgs of -0.79 V will give me an Id of 0.817 mA, which in return ask's for a Rs of 966 Ohm (1k).
To get a voltage drop over Rd off 3 volts (so Vd is about 6 volts), I would need a Rd of about 3.6 kOhm.
If I really would want to use 470R as source-resistor for this specific fet, I would know that this will push the fet more into cut-off. I also would need a higher Rd  to maintain the same drainvoltage Vd.

If I would want to use another fet (j201, mpf102, etc...), I would do the measuring and math all over again.
Up untill now, this hasn't failed me yet.

[Frank_NH]

If one really wants to explore this further, then putting the input stage on a breadboard is the way to go.  I may do that for the entire Azabache circuit as I've been curious about it for some time.  Last night I even went as far as to sim the entire Azabache in LTSpice!     I confirmed one mod I read about for people that complained about it being too bright.  Reduce the value if the 47K resistor at the tone control to 4.7K or 10K - that seems to help make the response less treble heavy and the scoop switch work better (note that the tone control is simply a modded Big Muff tone control, so you can play around with it using the Duncan Tone Stack calculator).  I also came across something weird, namely that the T-notch filter at the end of the circuit seems to affect the output wave form so that it looks more like a fuzz (using a 0.1 V sine wave @ 800 Hz as the input signal).  If you replace it with the simple low pass filter arrangement of the Professor Tweed, the wave forms look more amp-like (to me).  I can post some images for those who are interested.

[kdmr]

swapped out Q1 with a 2n5457 from a working circuit...same voltages all around. going point by point now...its gotta be SOMETHING.

[thermionix]

With Q1 pulled out of its socket, measure resistance from the source socket pin to ground.  Do you get 470 ohms?  I know that's been covered already in general, but have you measured from the actual socket pin?  I'm assuming you're using sockets, with all the swapping you've mentioned.

[kdmr]

Sure do. I'm halfway considering just building the first stage again separately just to see if it will work on its own.

[kdmr]

i'm seeing a variety of biasing methods for 2n5457 on other ROG designs (from the fetzer valve stage to full amp-in-a-box circuits). seeing everything from 390R with 6v drain to 1.5k with 4.5v drain. before i start building extra stuff maybe i'll socket Rs and see if i can do this the old fashioned way.

[kdmr]

well, i ended up finding a bag of 2n5458s under the rubble and popped one in q1. swapped in a 10k trimmer...HUGE difference...sourced voltage at something like .2v, 6v on drain. not ideal but its starting to sound like something. just have to tame the high gain squeal now. thanks for dropping some knowledge on me, everyone. i wish this was more climactic.

[Frank_NH]

Looks like you have it sorted now.  I had a go with the Azabache on my breadboard tonight, and it worked fine for me, but as you said it's somewhat bright and brash.  I found that using the simple 10K/2.2nF LPF at the output of the ROG Professor Tweed (see link below) really improved the sound (for me). There are other mods you can try, like modding the Big Muff tone circuit or putting a 10uF cap to ground across Q4's source resistor for a boost (could be made switchable).

http://runoffgroove.com/professor.png

[kdmr]

it's very annoying to me that Q1 doesn't behave as specified. Probably going to haunt me forever. it's very curious that every 2n5457 or mpf102 that i tried yielded almost identical wonky source voltages yet the very first 5458 that i threw in resulted in such a low source voltage and produced so much amplification that i'm now dealing with a squealing issue. even breadboarded up ROGs fetzer valve to test FETS and every 5457 or 102 i tried resulted in the same 2.5-2.7v source voltage. still pretty certain that something is incorrect...maybe its in a parallel dimension.

[kdmr]

i guess i should get a few fets to sample from a reputable retailer...wondering if i got burned on some out of spec units on the ol fleabay

[rankot]

If you know Vp and Idss, you can calculate any Vgs and appropriate Rs.
Id = Idss x {1-(vgs/vp)}²
Rs = Vgs / Id

Dave,

Could you please be so kind to explain what are Vp and Vgs in those equations? For some strange reason, Runoffgroove use Vp to describe what is usually called Vgs(off) in datasheets, now, I am not sure what are you reffering to.

[robthequiet]

Pinchoff?

[reddesert]

it's very annoying to me that Q1 doesn't behave as specified. Probably going to haunt me forever. it's very curious that every 2n5457 or mpf102 that i tried yielded almost identical wonky source voltages yet the very first 5458 that i threw in resulted in such a low source voltage and produced so much amplification that i'm now dealing with a squealing issue. even breadboarded up ROGs fetzer valve to test FETS and every 5457 or 102 i tried resulted in the same 2.5-2.7v source voltage. still pretty certain that something is incorrect...maybe its in a parallel dimension.

But didn't you also change the trimmer at the same time?

If you just want to get a working circuit, then you may be done, but if you want to understand what happened, then I'd suggest at least two things: 1. measure trimmer resistance and voltages at drain, gate, and source, then swap a 2N5457 back into the circuit, re-bias and measure the resistance and voltages.  2. Measure Vp and Idss of the 2N5458 that works and the 2N5457, using the measuring circuit from the Fetzer Valve page (it's really simple), so that you can get an idea of how different they are and what Vgs is needed to make the circuit work with each one.

[diydave]

If you know Vp and Idss, you can calculate any Vgs and appropriate Rs.
Id = Idss x {1-(vgs/vp)}²
Rs = Vgs / Id

Dave,

Could you please be so kind to explain what are Vp and Vgs in those equations? For some strange reason, Runoffgroove use Vp to describe what is usually called Vgs(off) in datasheets, now, I am not sure what are you reffering to.

Vp is - as robthequiet stated - the Vgs(off), or Pinch-off voltage. It's the voltage in where there is no current going through the fet.
Vgs is a voltage between Vp and 0 volts.

Do you have the values (Idss and VP - by measuring, not datasheet) off your 2n5457?

[rankot]

Vp is - as robthequiet stated - the Vgs(off), or Pinch-off voltage. It's the voltage in where there is no current going through the fet.
Vgs is a voltage between Vp and 0 volts.
Do you have the values (Idss and VP - by measuring, not datasheet) off your 2n5457?

Thanks a lot! So you try your circuit with different Vgs values and listen which one is best? I have bought 50 pcs of 2N5457 from China and they are all around the low Vgs(off) threshold, 0.5V±0.1V (usually under it), so I presume they sent me some out of class transistors. I have measured Vgs(off)/Vp and Idss with that simple circuit form ROG.

I tried different preamp circuits with them and they really have problems with clipping. So I have ordered few 2N5458 hoping they would be better.

[diydave]

So you try your circuit with different Vgs values and listen which one is best?

I calculate first, and listen afterwards.

If your fet has a Vp of 0.5 volts of less (which is odd for a 2n5457, you're more likely to see this with a j201), it won't take a lot to start clipping at the input. Anything with a swing of more than 0.25 volts will clip.

Not much room to experiment with Vgs as well. Midpoint bias (vgs = -0.25 volts) would be the way to go. So I would select / calculate a Source-resistor which gives me about 0.25 volts on Source. But it wouldn't take much to get clipping at the input, that's for sure. There's simply little or no margin.

A fet with higher spec's (higher Vp and Idss) would be the way to go if it's more of a clean stage you're after.

[duck_arse]

from the fetzer page ......

Example: A J201 with Vp=-0.75V will accept up to +/- 0.75V before entering into the clipping region.

so Vp of -0V5 allows signal of +/- 0V5 before clipping [I_think] (1V peak to peak) [/I_think].

[diydave]

from the fetzer page ......

Example: A J201 with Vp=-0.75V will accept up to +/- 0.75V before entering into the clipping region.

so Vp of -0V5 allows signal of +/- 0V5 before clipping [I_think] (1V peak to peak) [/I_think].

If Vp = -0.5 V, there's room to wiggle between -0.5v and 0 V (being -0.5v = cut-off / no current and 0v max current) around -0.25V as symmetrical Qpiont.
So I'm entitled to think that runoffgroove's example is really saying: A J201 with Vp=-0.75V will accept up to 0.75V peak-to-peak before entering into the clipping region. As long as Vgs = -0.375V - midway the Vp value.