Why these trimmer pots? Bias adjustment?

[dbanner]

Can anyone shed light on the use of such trimmer pots as indicated in this schematic. This is an old design, but it is an interesting schematic.
Could these trimmers be eliminated altogether with a redesign of this section? This part of the schematic is using non-inverting inputs throughout.

Also the power supply (+6V +3V 0V-battery negative)is different than the traditional positive and negative rail voltages, with +3V serving as a virtual ground.

[Carl-NC]

The first 2 trim out the offsets of U3d & U5b. So if you use opamps with really low Vos those can go. The other, I'm not sure. It's an offset adjust, but not sure why. U5a appears to be a rectifier.

That type of power supply is very common in VLF designs. Many detectors are now +5V with a rail splitter to create a +2.5V "analog ground."

[dbanner]

Thanks for explaining that. So it's an offset adjustment which trims out the input offset voltage error (LF444 datasheet indicates a Vos of 3mV minimum to 10mV) The TL074 is just about the same.

I like to go over these old schematics, they contain a lot of tricks the young players like myself (maybe not so young) can learn from, like the two series led's in the power supply.

[dbanner]

Question, if a non inverting op amp is followed by an inverting op amp, do the offset errors cancel? I see other schematics using this topology for this section of the circuit, but without offset trimmers. They are using common TL072 chips, nothing special.

[Carl-NC]

Opamp offset is statistical, sometimes it adds, sometimes it subtracts. So you always assume it adds. The gain of these stages (they are the motion filter stages) is around 120, other designs may have filters with low gain. For this design, you need both low Vos and low Ib. The AD8609 would work, probably there are others.

[dbanner]

Thanks Carl. I'm going to work on this circuit some more, hopefully I can make a working machine out of it.

[dbanner]

Well I tried ltspice sim of the tx oscillator, the thing won't oscillate. Looking at the schematic, there's no dc bias resistor on the base of the npn transistor.

Luckily I have some photos of the PCB, so I'll try to see what's up with the schematic.

This circuit s a 13.77kHz machine.

[Carl-NC]

Oscillators often need a "kick" in Spice. Put a pulse current source in parallel with the coil; pulse, say, 1mA of current for a few microseconds, then off for the rest of the sim.

[dbanner]

When I added a 68k resistor from the base of NPN transistor to the supply rail, the thing runs beautifully.

But now the two phase shifted outputs look bad.

[dbanner]

My laptop has crashed, so I'm using my desktop PC, as soon as I can figure out how to get it online, I'll post the spice file and the results.

I'm using my phone to connect to the forum. Lord have mercy!

[dbanner]

I added a 68k resistor from the base of NPN to the 6volt supply rail after the 10ohm resistor. The circuit now works. But the shifter outputs to the gates look wrong.
I will also try what you suggested.

[Qiaozhi]

Oscillators often need a "kick" in Spice. Put a pulse current source in parallel with the coil; pulse, say, 1mA of current for a few microseconds, then off for the rest of the sim.

Since Spice solves for the operating point prior to the start of a time domain simulation, you can also try ramping the power supply. This will sometimes give the oscillator the necessary kick. Another trick is to set an initial condition on one of the capacitors. This is particularly useful if the oscillator is in a balanced state when the simulation starts.

[Carl-NC]

Since Spice solves for the operating point prior to the start of a time domain simulation, you can also try ramping the power supply. This will sometimes give the oscillator the necessary kick. Another trick is to set an initial condition on one of the capacitors. This is particularly useful if the oscillator is in a balanced state when the simulation starts.

True & true. Another trick that works with either the current pulse or the I.C. on the main tank cap is that you can usually determine what is needed to kick the thing into a nearly-settled running condition. That way, you don't have to wait for it settle out.

[dbanner]

I got the oscillator running, turns out it was missing a resistor on the schematic. This was confirmed by photos of the PCB.

What's more, I have phase shifted sine waves going to the cntrl pins of the CMOS gates.

Quite a difference compared to the use of comparator chips in other circuits, which outputs a more rectangular phase shifted pulse.

I don't know how this is supposed to work, but I guess it does.

The PCB is a two sides board, and I'm not in the mood for tracing the schematic for errors.

[waltr]

Ok, good news.

Guess there is an error in the schematic you posted in post #1.

If so you may want to recheck the transistor from pin 1 to 2 of opamp U5.
This may be miss drawn and should be acting as a Log Amp.