No Signal out of FET - Troubleshooting Help

[BuddyPrince]

I'm working on adding an active 2-channel mixer to a pedal so I can have independent volume controls for a clean and dirty channel.  That thread is here  http://www.diystompboxes.com/smfforum/index.php?topic=119886.0

I'm doing a 2-channel version of the JFET Parallel Looper over on Tag Board Effects.  http://tagboardeffects.blogspot.com/2010/11/jfet-parallel-looper.html

I've got the circuit built on stripboard (removing the third channel, crossed out in blue in the attached pic).  I've quadruple checked it against the stripboard layout and the schematic.  It all looks good.

I've used my audio signal tracer and discovered that the signal goes into the FETs at the Gate, but isn't coming through the Source.  I've checked the pinout (DGS) and it is fine.  I'm not sure what's going on.

One thing I'm not sure about is that I think that the electronics shop sold me MOSFETs and not JFETs.  I'm using 2N7000 transistor, which when I look online tells me it's a small signal MOSFET.  I don't know why they gave me that one, it's not what I asked for, but what do I know, it looks like the JFETs I've used, it had the right pinout...

This is about where my limited knowledge stops.  I'm assuming that a MOSFET should work in place of a JFET in this situation, but I don't know what the functional difference is between the two, nor how to troubleshoot what's going on.  On top of it, I think my multimeter isn't working properly, I don't seem to get any steady readings when I measure voltage or resistance.  The values keep dropping and are inconsistent, so I don't accurately know what the voltage drop is across different components or parts of the transistor.

So, how do I troubleshoot this?  Should I just get one of the recommended JFETs?

Here's the layout I'm using with the unused components crossed out in blue.


Here's the schematic

[idy]

They sold you those because through-hole jfets are "collectors's items". Try googling"AMZ basic buffers" for the mosfet configuration.

[MaxPower]

For future reference: Tayda has through-hole jfets (2n5485 and 2n5458 I think).

I don't  think mosfets are biased the same way as jfets.

[PRR]

MOSFET is not JFET.

The original plan has dubious bias even for JFET.

Given a fistful of 2N7000, do this.

[BuddyPrince]

Thank you guys!

Paul, thanks for those changes to the schematic!  I ought to be able to figure out how to change the stripboard layout.

I've been reading a bunch online about transistors and trying to understand how they work, some of the differences between them, biasing, etc. It's slow going,  but coming together.

[R.G.]

Something that helped me with understanding transistors of several kinds and tubes as amplifying devices was the notion that they're all the same in some ways. After that, you can concentrate on the differences in understanding them.

All common amplifying devices have a main current path where all or nearly all of the current travels. For tubes, this is the plate-to-cathode path. For bipolars, it's the collector to emitter. For FETs it's the drain to source. Bipolars (NPN and PNP) are the ones that require that "nearly all" term, as a tiny fraction of the current in the emitter lead comes from the base, but it is tiny.

So there's a current path through two of the three pins. It lets more or less current through depending on what's happening between the third pin and the emitter/cathode/source. The current paths have different limits on how much current they'll let trough, how much voltage they will hod off, etc, but the principle is the same.

The big variations are in what you have to do to that third pin to make current flow or prevent it from flowing. For vacuum tubes and JFETs, current flows through if you don't do anything to the third pin, just leave it open. You have to apply a negative voltage to the grid/gate to turn it off. These are called by the name "depletion devices" because you have to do something to deplete the normally full-on flow of current in the current path. As a side/foot note to this, there are some specialized MOSFETs that are depletion devices, but they're the exception, not the rule.

Bipolars and MOSFETs (in general) are enhancement devices. That means that if you don't do something to the third/control pin to make current flow, nothing happens. They'll hold off current flow as long as they can. For enhancement devices, you have to put in an increased voltage on the control pin to make them conduct.

Both bipolars and MOSFETs have a threshold voltage on their control pin. For bipolars, as you increase the base to emitter voltage from zero, nothing much happens until you get to maybe 0.45 to 0.5V, at which point the current increases exponentially with increasing base-emitter voltage. A current flows through the base to the emitter, and the current in the collector to base increase approximately linearly with base-emitter current.

For MOSFETs, there is also a threshold voltage below which nothing much happens, but this gate threshold voltage is not set by the material properties as in bipolars, but by the internal design of the channel material. So you can have low threshold MOSFETs where the turn on voltage between gate to source is as low as a volt or so, up to needing as much as ten volts before current starts to flow in the drain to source path. Once current starts to flow, it's approximately a linear relationship of gate voltage increase to drain current increase.

The two types of FETs are very different when looked at this way. To bias a JFET into it's linear amplifying region, you have to turn it partially off by introducing a back bias on the gate to throttle the current down, as it's a depletion device. To bias a MOSFET on, you have to turn it partially on, introducing a forward bias on the gate.

It helped me to think of how the devices were similar, then add the minor differences to the control pin., Hope that helps you some as well.

[BuddyPrince]

RG, That's really helpful!  Let me check my understanding with regards to troubleshooting why this circuit that was designed with JFETs isn't working with MOSFETs, it essentially boils down to how a JFET gets turned on vs. how a MOSFET gets turned on.

The JFET is normally on, allowing full current to flow from drain to source (Ids).  In order to get the JFET to work it's magic, you introduce a back bias, which if I understand correctly means that the gate voltage (Vg) is less than the supply voltage (Vdd).  R2 and R3 accomplish this in this circuit.  They're acting as a voltage divider setting Vg at 4.5v when Vdd is 9v.

The MOSFET, on the other hand, is normally off and no current flows until we apply a voltage across gate to source (Vgs) that is higher than the threshold voltage (Vth) of the MOSFET.  We want Vgs to be greater than Vth.  I'm not clear on how we achieve that.  When I look at the changes that Paul (aka PRR) made to the circuit, it looks like Vbias is 4.5v, which is connected to the gate, and therefore Vg =1/2 Vdd as in the JFET case.  Am I missing something?

[R.G.]

You got it right.

All amplifier biasing is about manipulating the input side so the output side is amplifying the way you want it to. For the simple circuits most often looked at here, with a single output resistor (ie. collector resistor, drain resistor or plate resistor) you try to get the amplifying device to split the available power supply so that a bigger version of the signal can swing as widely as you need it to.

MOSFETs need "help" to turn on, JFETs need "help" to turn off enough, a characteristic they share with tubes.

Congratulations! It is sad but some people never get to where you are. Keep digging in!!

[PRR]

As a rough dart-toss, you want the output pin to sit near half of the available supply.

"Near half" is not optimum, but in most cases is not far off the optimum.

In a simple Follower, the output follows the input, except usually with some DC offset.

On a JFET you bias the gate maybe -1V from where you want the source.

On a MOSFET you bias the gate maybe +2V from where you want the source.

When supply voltage is considerably more than signal voltage (often true in 9V pedal work), it is expedient to just use equal resistors to derive a half-supply bias. The JFET may sit at +5.5V output. The MOSFET may sit at +2.5V output. Either will work.

If you don't need "optimum output", a JFET follower may be biased with gate at zero and source self-lifting on a several-K resistor. Real darn simple. Will clip above about a Volt or so.

Your JFET plan did work through the input stages even with MOSFETs (assuming you IDed the leads right. Your output stage was sitting there utterly dead.

In retrospect, for MOSFETs, I might use 47K+100K or so divider to get bias near 6V and source near 4V.

[BuddyPrince]

My head's hurting...I've been at this all day, like most of the last 12 hours...reading, re-reading, and re-reading some more.  Looking at schematics, looking at strip board layouts, re-organizing them, checking them, testing them, measuring things, etc. etc. etc.

I don't have it working yet...and probably need to call it a night.

The easy thing would have been to just wait and get some JFETs, but then I wouldn't have the opportunity to learn about how these things work.  The biasing piece of this puzzle, both in theory and in practice, is still a little out of my reach, but I'm getting there.

I'll have more questions after a good night's sleep and a fresh mind...

Thanks for helping me understand this stuff!

[antonis]

My head's hurting...I've been at this all day, like most of the last 12 hours...reading, re-reading, and re-reading some more.

Maybe because R.G. "forgot" to make discrimination between BJTs & FETs "linear" regions.. 
(meaning what's "saturation" for FETs is "linear" for BJTs..)

P.S.
Just kidding of course, but sometimes that's an elementary confusing parameter/behavior.. 

[BuddyPrince]

OK...getting closer, here's where I'm at this morning...

I know my goal is to have a gate voltage that is about 2v greater than the source voltage.  As PRR mentioned, I'm shooting for a gate voltage of about 6v and a source voltage of about 4v.  I'm using a 47k and 100k resistor to get a bias voltage of 6.3v.  Here are the voltage readings I'm getting:

Vdd = 9.39v
Vbias = 6.3v
Vgate = 5.07v
Vsource = 8.79v

A couple of questions:
1. When R2 = 2.2M, Vgate = 5.07v.  When I put in a 10k resistor Vgate = 6.3v, which is where I want it.  I just want to confirm that this is one change I want to make.

2. How do I lower the source voltage?

Here's the schematic that PRR provided.  The only change is that I replaced the first biasing resistor with a 47k.

[PRR]

> about 2v greater than the source voltage.
> Vbias = 6.3v
> Vgate = 5.07v
> Vsource = 8.79v

First: that Vgate voltage is bogus. There should be zero drop across the gate resistor. The gate is at 6.3V.

UNTIL you put the meter on the gate. The meter loading on the large gate resistor loads-down the voltage while you are measuring it. Assuming a 10Meg meter on a 2Meg resistor, you do get about those voltages.

"About 2V" Vgs is very approximate, for reasons R.G. touched upon. We see your MOSFET is 6.3V-8.79V or 2.5V g-s.

I may have got mixed-up. Assume we want source near half-supply, which we round to 4.5V. Gate must be 2.5V lower or 2V. I think 100K on top bottom and 47K to ground supply gets closer. 100K and 27K gets super close. 100K and 33K also works. Get the 'Bias' around 6.5V and it should work good. (Do not know why yours doesn't.)

[PRR]

2N7000 is specified Vgs 0.8V-3.0V at 1mA. Our operating point will be near 0.5mA (half of 9V, divided by 10K). So the 1mA spec is close though not spot-on. 2.5V at 0.5mA is entirely on spec. The 0.8V-3.0V shows the great variability of MOSFET Vgs.

[R.G.]

I may have got mixed-up. Assume we want source near half-supply, which we round to 4.5V. Gate must be 2.5V lower or 2V. I think 100K on top and 47K to ground gets closer. 100K and 27K gets super close. 100K and 33K also works. Get the 'Bias' around 2V and it should work good.

Close. It's easy to get the gate-source direction messed up between JFETs and MOSFETs. For a MOSFET, an enhancement device, the gate must be Vt higher than the source before conduction begins. So for a 2N7000, with Vt of 0.8 to 3.0 Vgs at 1ma, the source will settle 0.8 to 3.0V below wherever the gate is being held.

So if you want the source at 4.5V, the gate bias voltage needs to be 5.3 to 7.5V.

I remembered an article on MOSFET biasing for boosters at geofex:

http://www.geofex.com/Article_Folders/mosboost/mosboost.htm

It's for boosters, but it gets the gate voltage and source voltage offset right, I think.

[BuddyPrince]

Ok. So the Vgate measurement is inaccurate due to the loading from the multimeter, and it's safe to assume that the actual gate voltage is the same as the bias voltage.

Based on what RG stated, it looks like my gate voltage is in the correct range, at 6.3v.  however, it looks like my source voltage is 2.5v higher than the gate at 8.79v. (Assuming I've measured everything correctly).

Still, I'm not getting any signal out of the MOSFET...

[antonis]

Are we talking about JFEts or MosFets..?? 

[R.G.]

At least we're not talking about GaAsFETS.

Which, by the way, are a real thing.   

Other FET technologies are advancing. There exist FETs made from materials other than silicon that will have their own odd characteristics. GaAsFETs are made from gallium arsenide. Ti is now pushing its line of gallium nitride. FETs made from silicon carbide are able to withstand massively higher temperatures than silicon. There's even work on FETs made from odder materials like graphene. People desperately want to make FETs out of vapor-deposited diamond, but it is very,very difficult to dope a diamond. Someday.

Of course, it will be a long time before we can think of gallium nitride FETs as "vintage" or "NOS".   

[antonis]

Judging from OP values posted above, he probably uses some Galenium-Selenium_FET..

Otherwise., +2.5V Source-Gate should work for almost any general purpose n-channel FET and couldn't be obtained for none of MosFets..

[BuddyPrince]

Thanks for that article RG.  It helped me understand MOSFETs better, but there are still some gaps.  I followed along with pencil and paper and worked out calculations.  However, I'm still left with the question, how do I get the source voltage down to 4.5v?  The bias voltage seems to be int he right range...

At this point, I think I'm bailing on these MOSFETs and waiting for the JFETs that I initially wanted to come in.

It's been a really good learning experience, and I've gotten a lot out of it!  Just not my desired current.