Not OT: BC850 as a diode clamp?

It might just be that Vbe is more tightly specified than forward voltage of a regular diode.

It is shown as min 580mV, typ 660mV and max 700mV at 2mA and 25C, which is pretty tight, and unusually detailed. there's no minimum voltage at 10mA.

Low leakage and lower dynamic resistance due to transistor action.

piglet

Someone at NIST built a complete ring mixer around that structure.

< >

Gerhard

The be-junction is in parallel with two BAV99 (? illegible) diode junctions, so there is no leakage advantage.

That's a really weird circuit.

it sorta works like a zener diode

I never said it was appropriately used in that circuit! Probably designed by a physics student who thought it looked clever?

piglet

It was probably intended to protect the person or animal to which the electrodes are connected from being injured by excessive dc in the event of a failure of one of the other components. I'm assuming it was intended as an ECG or EMG amplifier.

John

I thought of that, but this is preceded by a 36 kohm/470 pf filter and followed by a 5.1 kohm resistor and a pair of BAV99 diodes to the power rails and on to differential inputs of an op amp, INA321. I don't get the need to clamp both inputs to ground on the negative excursion and not the positive. Oh, wait, I may have found it. The INA321 is not connected to V-, rather the circuit ground. The circuit on the input side of the INA321 is referenced to ground, while the output side is reference to a midpoint reference formed by an op amp. The input side has a pair of op amps in a "bootstrap" arrangement with a separate virtual ground slightly below that midpoint. There's no DC blocking caps on the input, so I guess it's ok to provide a low voltage DC to the test subject through the input electrodes.

This author circuit has a number of circuits for amplifying "biosignals" meaning very low amplitude. I'm not clear on the advantages of many of these circuits.

Not exactly the same circuit. The schematic in question grounds the collector/base connection, so the two transistors work independently. I suppose it does however, give you the same effect, by clamping like a diode at 0.6V below ground, and some zener-like effect above ground. I didn't think of that, but it suits the circuits this guy designs I suppose. I need to remember this one. I don't think I can use it in my current design though. That is clamped to the power rails on the analog inputs, ±12V. I may need to rethink that circuit for other reasons though. The input switches are hard to find.

The problem with diodes to the power rails, is that it injects current into those rails. This circuit dumps the bulk of the energy into ground and the diode clamps deal with the rest. Looks like a good circuit to me. Amplifier inputs are not always about minimizing input capacitance. Only optimize what needs to be optimized. You save a lot of work and the design is more widely useful.

But why? The BAV99 already clamps up and down. I assume the supply is

5 volts, split-rail at 2.5, so the BAV clamps at +-3 roughly.

Of course the split supply is soft.

Very weird circuit.

$40 to see the PDF. No thanks.

You have the same guy, but this is the paper the circuit I posted came from. I was able to download the PDF file.

You may have to select it before you try to open the link since it is a bit long.

The opamps run single-ended so don't need much cmrr. CM will be dominated by resistor and capacitor tolerances.

Horrible mess. Why not buy one good diffamp?

Only to someone who doesn't understand it. The BAV diodes clamp to the power rails, which means a heavy surge can overload every part on the board. In this design the BC850 acts as a higher current path to ground for the bulk of the overvoltage. Grounds can typically handle that sort of thing better than power rails.

If you read a few of his articles, you might learn something. The guy is not just another pretty face.

søndag den 29. januar 2023 kl. 19.15.22 UTC+1 skrev John Larkin:

until you push too much current in to the supply

the BAV99s are not really needed the opamp is rated for +/-10mA into the input

I typically use Schottky diodes for this, to get the lower Vf.