RELIC HAWK

[KingJL]

Hi,

I think the coil is pretty critical component here much more than in other designs... and I think changing op amp will not solve this.

I agree about coil being critical component. My simulations show that after 2.5 ohms, the coil current starts to decrease rapidly. Anything over 3 ohms stops oscillations. This is one coil where Litz wire may be very beneficial.

The op-amp does have an effect (probably due to the slew rates), although I agree that it will not fix it. The TL074 so far works best in my simulations. All others have a slower oscillator startup or less coil current.

Regards,
J. L. King

[Max]

Found SPICE models for TLC074, TLC27M4, and TL074. TLC074 causes low amplitude (4vpp) oscillations, slow oscillator startup. TL074 seemed to work best (see atachment). Coil series resistance set to 2.2 ohm in this run. With 2.2 ohm coil, peak coil current is 400mA.

Hi,
thanks for these informations... my problem is that I've used TL084 and it doesn't work with the 2.5ohm coil.

Troubles are still there I think. I used BC557 and BC547 but no way... if I replace the coil with one with lower series resistance it works (now I changed tx coil and get 6.8Khz at oscillator, still far from 5Khz).

I think the coil is pretty critical component here much more than in other designs... and I think changing op amp will not solve this.

Kind regards,
Max

[KingJL]

Found SPICE models for TLC074, TLC27M4, and TL074. TLC074 causes low amplitude (4vpp) oscillations, slow oscillator startup. TL074 seemed to work best (see atachment). Coil series resistance set to 2.2 ohm in this run. With 2.2 ohm coil, peak coil current is 400mA.

[KingJL]

Max,
If you change the components in LTSpice to the correct models, the circuit works with 2.2 - 2.5 ohm coil (attached LTSpice output was run with coil series resistance at 2.5 ohm). The only model that I do not have that correspond with schematic is U1. I am currently btrying to find a model for a TLC074 that works wit LTSPice. The TLC074 has better noise specs that the TL074 and much much better than the 27M4.

[Max]

Ok, I've tested with an old big coil 700uH and 0.7ohm resistance... used in past for PIs... and oscillator is up and running.

Sure frequency is higher but is right using 700uH stuff... so 2.5ohm is sure a wrong value.

Kind regards,
Max

[Max]

Hi,
I have bad news for those who already built coils...

I made yesterday a coil with resistance = 2.5ohm like described... BUT it doesn't work on this oscillator at all.

The most of my PCB is populated so I tested it and oscillator doesn't work and checks confirms all is fine about circuit... so error is probably at coil only.

I assume who made measure had bad instrument, cause from simulation I see clear that coil series resistance must be under 1ohm, probably between 0.5ohm and 0.8ohm to work cool.

So... 1mH seems right to match 5Khz with 1uF BUT keep resistance lower... otherwise you'll get no oscillations!

I've still not tested but seems realistic to me that resistance is wrong...

If you use 35turns on 30cm diameter (94.248cm circ.) you'll get about 33meters of wire... for a resistance of (about 70ohm/1000m) 2.31 ohm total... but this just for exact 0.55mm diameter... and simulator say it will never work.

Kind regards,
Max

[VUJO]

Hi Max
Thank you for the information, I have just took a look at it after installing it and I already like it.

[Max]

Hi,
this is revision 1.6 , corrected bad voltage pins at U10.

PCB is already right way.

Kind regards,
Max

[Max]

Hi,
here's simulation with 3v3.

There's about 25Hz modulation of the 5Khz carrier...

Kind regards,
Max

[Max]

Hello Max
I am interested in which program is made oscillator simulation?
Thank you in advance.

Hi,
it's SwitcherCad (LTspice) by linear technology, you can download for free.

Kind regards,
Max

[VUJO]

Hello Max
I am interested in which program is made oscillator simulation?
Thank you in advance.

[Max]

Hi,
now the question is: can we trust coinseeker schematic about value of that zener diode ?

I don't know... but I see change few if you use 3v6 or even 6v2 zeners there, at least I cannot notice any difference.

It works also with a regular diode (but reversed respect to original schematic that's wrong way), but I think the zener solution is right cause a reverse polarized normal diode make no sense to me there.

I don't remember if I have 3v3 (I think yes) but otherwise will use 3v6.

Kind regards,
Max

[Max]

Hi,
Here's revision 1.5 of schematic.

Kind regards,
Max

[Max]

It likes this even better!

Hi,
yes, right... in RH schematic we have it's reported as a regular diode... but it cannot work this way so I think too there's a zener diode there...and now that I saw your simulations I'm sure of that, thanks.

I'll made another revision with that error corrected.

Kind regards,
Max

[KingJL]

It likes this even better!