mosfet follies - cathode follower

[MarkusBass]

Since we're on the subject here's a practical question from few days ago. On the pic below is an LND150 based FX loop also know as Metro Zero Loss FX loop. There's no CF there but the first stage is wired as inverting feedback amplifier which has a very low output impedance assuming the role of a CF.
So after firing this thing up I measured 2 Volts at the send stage drain which is not normal. At the other end of the 100k resistor voltage was 254V which indicated ~2.5mA of current draw which is strange. After replacing components and scratching my head for couple of hours without any result I added a 1M resistor from gate to ground and very slowly the voltage at the LND drain in question went to 220V while at the other resistor's end to 305V. The only thing missing from my build is that R4 220k resistor.
So I was wondering what happened and why? Shouldn't the loop work without that resistor to ground as per schematic?

[ATTACH=CONFIG]42716[/ATTACH]

I have a question regarding this schematic. What is the purpose of the R2 pot? It shorts the input of the first stage of the loop with the output. Is it correct, or it's a mistake on the schematic?

Mark

[Dave H]

but, as emosms says, even if the transistor is shorted, you cannot get higher current in this circuit than 11 mA. So the information about 25 mA limit is "general" and it applies to other values of R5 resistor and/or other power supply voltages.

Isn't 11mA is just the max current you can get in the 22k source to ground resistor (R5). The FET can source more current to a 2k2 load than that.

[Dave H]

I have a question regarding this schematic. What is the purpose of the R2 pot? It shorts the input of the first stage of the loop with the output. Is it correct, or it's a mistake on the schematic?

It's adjustable negative feedback to set the send level.

[emosms]

5vRms/2k2 is 2.27mA.

So, how does the value of R5 (22k) affects the driving capability ??

For the gate threshold voltage Vgs, values are 2 to 4 v indeed.

So, 12 (zenner)-4 for the Vgs is 8/330, or abt. 25mA.
Eventhough, I do not understand - is that 4v dropped in the gate-source, so we get only the difference of 8 volts?

Further on, if R4 dissipates that current, should'nt we add the output current to that 25mA? So, consumption is over 25mA?

Finally, If R4 eats that current, what if I substitute it with 620 ohms. Then we have 8/620 = 0,13mA.
The "nominal" impedance is 620 ohms, like standard, and the output driving capability is not impaired.
(Whatever the reason that it depends on R5 is...)

p.s. Also, I hope, that the G-S zenner of 12v does not clip the tube output swing to 12v.....

[Dave H]

So, how does the value of R5 (22k) affects the driving capability ??

It sets the negative clipping point. The current through R5 is 119/22 = 5.4mA. When the FET shuts off (negative clipping point) there's no current through the FET so 5.5mA peak is pulled from the load. If the load is 2k2 that's 2.2 x 5.4 ~ 11V peak.

So, 12 (zenner)-4 for the Vgs is 8/330, or abt. 25mA.
Eventhough, I do not understand - is that 4v dropped in the gate-source, so we get only the difference of 8 volts?

Further on, if R4 dissipates that current, should'nt we add the output current to that 25mA? So, consumption is over 25mA?

You've got it in the first line The voltage between the gate and bottom of R4 is limited to 12V by the zener and the FET needs 4V to turn on so there's 8V max across R4 or 25mA.

25mA is the fault current if you like. It's the current you'd measure if you shorted across R5 with an ammeter. The current consumed from the power supply is the no signal (quiescent) current through R5 i.e. 5.4mA. The current through R5 (and the load) will vary with signal but the average from the power supply will be 5.4mA.

Finally, If R4 eats that current, what if I substitute it with 620 ohms. Then we have 8/620 = 13mA.
The "nominal" impedance is 620 ohms, like standard, and the output driving capability is not impaired.
(Whatever the reason that it depends on R5 is...)

Yes you could do that but all of that 13mA isn't available to drive the load some of it is the current through R5. I'll try and work out where the positive clipping point is with R4 at 620R to see if it's OK.

[emosms]

Hi, I think I am getting it, thanks .
The fault consumption of 25 ma is 25 * 8volts drop = 0,2 VA of the PSU. This is the case when we have cold cathode and almost HT at mosfet's gain. Lets leave R4 as it is.
This is not to be considered uner normal operation. There would be no voltage exceeding 12v across gate-source, because the source "follows"

---
Trying to calc some worst case scenarious. If the valve is 6n2p, 100k anode load, 300v DC - the load line shows that anode voltage could go over 240v DC...
It is not that much for 6n1p, so taking the value of 220v as worst case.
Disregarding the drop in R4 and the voltage needed for mosfet-on.

220v between R4 and R5, 220/2k2 - we have 10mA consumption - "worst case scenario"
If the voltage is higher, I can rise R5 - 25k, 27k, 33k... (more than worst case scenario)

- Here a side question. When setting the load line, all the grid voltage curves - is that RMS voltage ???? Needed to more precisely do the math.
F.ex. if I measure the guitar output with multimeter, I can get the RMS AC voltage (I guess it is not peak to peak).
Then, I can use that to get the max grid volgate and the max possible anode voltage at given HT/anode load.

---
Further on.
Suppose I order transformer 250v AC 40mA. This is 250 * 0,04 = 10VA. Bridge rectifier and caps resutls in 250*1.414 = 353.5v DC.
10VA/353.5 = ~28mA expected possible current draw without sag.
28mA is 11(mosfet CF)+11(mosfet CF)+3 (tube load)+3(tube load). Provided the tube will operate lower than 3mA, I will have > 11mA per mosfet.

More considerations - In another forum, people want me to get even bigger transformer. Suggesting ratio of 2/2.2 to 1 - AC to DC current.
I did the math above and the result is 40mA/ 250v AC vs 28mA/ 353v DC. This is a ratio of 1.42 to 1.

The only thing I am missing - how much extra current/power would eat 2 electrolytic capacitors in the CRC chain ??? Is it that much? Is it something to consider at all ? .

[eschertron]

Hi, I think I am getting it, thanks .
The fault consumption of 25 ma is 25 * 8volts drop = 0,2 VA of the PSU. This is the case when we have cold cathode and almost HT at mosfet's gain. Lets leave R4 as it is.
This is not to be considered uner normal operation. There would be no voltage exceeding 12v across gate-source, because the source "follows"

---
Trying to calc some worst case scenarious. If the valve is 6n2p, 100k anode load, 300v DC - the load line shows that anode voltage could go over 240v DC...
It is not that much for 6n1p, so taking the value of 220v as worst case.
Disregarding the drop in R4 and the voltage needed for mosfet-on.

220v between R4 and R5, 220/2k2 - we have 10mA consumption - "worst case scenario"
If the voltage is higher, I can rise R5 - 25k, 27k, 33k... (more than worst case scenario)

Thoughtfully considered. Well done!

Here a side question. When setting the load line, all the grid voltage curves - is that RMS voltage ???? Needed to more precisely do the math.
F.ex. if I measure the guitar output with multimeter, I can get the RMS AC voltage (I guess it is not peak to peak).
Then, I can use that to get the max grid volgate and the max possible anode voltage at given HT/anode load.

Data points on the load line plot are instantaneous voltage. So if you impose a signal on the bias, use peak-to-peak values to see how the anode voltage will track, positive and negative from the bias 'point'.

how much extra current/power would eat 2 electrolytic capacitors in the CRC chain ??? Is it that much? Is it something to consider at all ? .

The only time the caps are 'eating' current is when they are charging. Initial inrush figures into things like sizing caps to prevent destroying tube rectifiers; normal half-cycle charging under load figures into ripple. Under a steady-state condition, the caps do not consume any power, so do not 'eat' any current.

[emosms]

^ seems almost ok (still I don't know the max peak to peak input levels)

But I am still confused. The signal after the CF is going to feed opamp EQ, +-15v supply.
The signal needs to be attenuated. I attached all tube circuit (matchless hotbox) with CF. The attenuation is achieved with 100k resistor.
- How to calc the attenuation resistor?
- How does that resistor affect the ouptut impedance??

Next, the clipping diodes (2 x 10v) set the max output swing (peak to peak??). I guess, for a +-15v circuit is better to set 2 x ~6v diodes.
The output clipping diodes must be after the attenuator resistor ^.
All possible signal clipping is not going to be tube clipping, but zenner clipping.. .

[eschertron]

That 100k 'distortion output' control looks funny. Do you have another drawing, or an actual unit, to compare?
The attenuation will be = (wiper resistance)/(wiper resistance + 100k). The part of the volume knob that goes to the output might actually be connected (or jumpered to) the wiper.
If you want to sub a mosfet for the CF with the 56k on the source, just 'plug it in' following Merlin's guidance on gate resistor and zener Vgs protection (and use an 'enhancement mode' device!). It really is that simple. A tube CF or Mosfet SF work almost identically in terms of gain and impedance buffering.

edit: I see your other comments. The input impedance for an OP amp should be high enough that your interstage load doesn't matter for any tube type circuit. A couple dozen kOhms to a couple hundred kOhms won't matter. You are feeding the interstage divider into an OP amp, right? select your diodes to clip before you hit the +/-15vdc power rails. However, (train of thought thinking here) depending on the feedback in an active Op amp-based EQ, you may still clip the rails even with a lowish input signal.

[emosms]

edit: I see your other comments. The input impedance for an OP amp should be high enough that your interstage load doesn't matter for any tube type circuit. A couple dozen kOhms to a couple hundred kOhms won't matter. You are feeding the interstage divider into an OP amp, right? select your diodes to clip before you hit the +/-15vdc power rails. However, (train of thought thinking here) depending on the feedback in an active Op amp-based EQ, you may still clip the rails even with a lowish input signal.

Leave the matchless hotbox ..

Does the above mean, I can feed the opamp EQ without CF whatsoever? No loss of clarity (especially the highs)?
Anyway, I think I know what transformer I need, with or without mosfet CF's.
CF should be solved... I hope


What I am digging now is another story (full train of thoughts).

The EQ is suggested to have input voltage of ... 0.775 volts (I guess RMS).
If I calc 0,775*2*SQRT(2) - we have 2,19v peak to peak. 2,19v * 15db band boost could result in ... ~12v peak to peak. Near the power rails.
There might be something more, I have to analyze the circuit. I am not much better with opamps, than with tubes

My idea is to put the eq in between two triode stages. I just imagine, a heavily boosted EQ band would yiled plenty of nice harmonics and clip softly, if necessary. The whole amp is going to be hybrid stuff, solid state power amp.
What do you think?

[eschertron]

Look around at some designs. One of the designer's first decisions is whether they want a tube in a SS amp, or some SS components in a tube amp.

Just thinking out loud here, but if I were to want a SS EQ in a 'tube' amp, I'd choose between on the one hand putting the SS after all the tube drive components (including a SS power stage) and on the other hand use the SS for getting a clean and clear signal through the (SS) preamp and EQ to feed the tube distortion stages. I see two advantages: 1) I've chosen a topology that lends itself well to either a high-gain 'sculpted' sound, or a classic clean/light crunch. 2) Minimize transitions from SS to tube and you've minimized the headache or trying to match level and impedance without adding tons of extraneous components.

If you want versatility, make everything rackmount, matching standard line level and impedance, and patch it together as desired for the occasion.

edit: .775vrms is specified to be 0dBu