2M2/2M2 voltage divider to bias follower dropping 3/4 the voltage instead of 1/2

[SnickSound]

Pulling my hair over this... Got a feed from the B+ line to a 2M2 resistor in series with another 2M2 resistor to ground. This is to bias a MOSFET follower stage. B+ also goes to the Drain, mid-point goes to Gate (along with the signal through a DC-blocking cap), source to 100K then to ground.

But even without the MOSFET in the circuit, I'm getting 80V at the mid point and 340V at the top. I've reflowed the joints, measured the resistors, changed both resistors... still 80V at the mid point.

And with the MOSFET in the circuit, I still get 80V, but the Source is standing at around 140V?!?! Ergo, Source is 60V higher than Gate (it's a big MOSFET that is internally zenered, shouldn't be able to do that).

And mostly... it works. 

There must a glaring mistake I'm making, or my multimeter is screaming for a replacement. I've measured from every probing point I could find, used different ranges on the DMM, changes testing probes, the MOSFET definitely has its Gate at 80V and Source at 140V. It's one thing to occasionally get a false reading, but to get it over and over? And as I said... it works, sound is passing through and I can't hear any odd artefacts.

[2deaf]

There must a glaring mistake I'm making, or my multimeter is screaming for a replacement.

Both.  Meter has a 1M impedance when measuring voltage.  I would expect 81V under those circumstances.  When you take the meter off of the gate, the voltage is more like 170V.

[SnickSound]

There must a glaring mistake I'm making, or my multimeter is screaming for a replacement.

Both.  Meter has a 1M impedance when measuring voltage.  I would expect 81V under those circumstances.  When you take the meter off of the gate, the voltage is more like 170V.

Of course! At such high resistances, the DMM's internal resistance becomes a factor.

Thank you so much! I'll be able to sleep tonight.

[jjasilli]

There must a glaring mistake I'm making, or my multimeter is screaming for a replacement.

VTVM     11M input impedance.  Some Eico's have 25M.

[2deaf]

I take that back about replacing your DMM.  As long as you are aware that you are connecting a 1M resistor when using it, it's just a matter of using a calculator, slide-rule, abacus, etc. to interpret the results.  Even with a 10M probe, I would have expected 153V instead of 170V. 

[PRR]

Going way back.....

We used to use passive meters. The movement could be a 1mA, a 100uA, or a 50uA.

To read volts there is a resistor. 1K/V, 10K/V, 20K/V.

Since 300V is around we would use the 500V scale. These meters would be 50K, 500K, 1Meg.

Then we stuck tubes in front of the meter. Small tubes at low current can be stable with 10Meg grid resistors. VTVMs were either 10Meg on all ranges, or featured a 1Meg in the DC probe for less AF/RF loading, hence 11Meg. Some brands ran to 22Meg or higher. FET-input "VTVMs" usually hewed to these higher resistances.

The early DMMs kinda sorta followed these loadings. But real high impedance circuits are now rare. I'm seeing more popular-price DMMs with 1Meg inputs. This confuses pedal builders who look at 1 Meg bias resistor to a bias supply, it reads "low". It would still read low with 11Meg or 50Meg, but not so dubiously low.

Just remember your meter is like a 1Meg resistor you are poking at random points in circuits that may be high impedance.