NEW !!! Full differential PI front end.

Well...

Well... Your result is much worse now. Opamp stays saturated for ca 30us and is still unbalanced...

[moodz]

Oh .... substituted 1N4148 for above test instead of schottky ( thanks Aziz ) Also the blocking diode is the 1N4004 only. ( no shottky ).
The 4.7 pF cap removed across the 1M resistor. 30 pF cap added across pins 1 & 8.

Moodz.

[moodz]

U ,unregistered guests ,need to buy yourselfs a cro ,an get your act together!Ha!an do some work of yur own! nitpicks!then post it here! so we can pick!Ha!Trolls!
We enjoy moodz test an experiment,I do,so register an share your knowledge,I want to learn more.

thanks Rov ...

Progress continues ....

Here is a scope pic showing the end of the tx pulse WHITE ( mosfet gate drive ) and the ouput of the amp GREEN ( no target ) ... the inverting leg of the amp is fed with 4.58 volts for balance. Note this is not as per schematic.

The Tx pulse is set for 100 uS in this case.
Note the fly pulse has decayed in 10 uS this is better than my previous efforts.

NO TARGET



BIG STEEL TARGET ( my what big steel targets you have grandma ! )



Note this is the SINGLE ENDED performance ..... tacked a pot to the inverting side. The differential is significantly different. More in my next post.

I have made further improvements to noise without dropping gain.

cheers from moodz.

[Rov]

U ,unregistered guests ,need to buy yourselfs a cro ,an get your act together!Ha!an do some work of yur own! nitpicks!then post it here! so we can pick!Ha!Trolls!
We enjoy moodz test an experiment,I do,so register an share your knowledge,I want to learn more.

well...

you need to get your glasses checked ... the scope is set for 5 uS / div ... 20 us is nearly off the display.

Boy, have a look at the small hf artefact amplitude - it is back "normal" at 20us...

[moodz]

Thanks Aziz .... I had the impression the voltage drop across the schottky diodes would be lower ... I will give the 1n4148 a go. ( they will be cheaper too. )

I have been tweaking the circuit some and have got the decay rate twice as fast with no ringing that I can see. Peak voltage is now 450 volts ... I feel I could go higher but the front end of my CRO is dear to me.

Will post pics when I rebuild the circuit.

cheers moodz.

[Aziz]

Hi moodz,

don't use the schottky diodes. They have a quite long recovery time. The 1N4148 has a much lower capacitance and they recovery from clamping very fast.

This can be seen on the circuit simulation.

Aziz

[moodz]

Thanks Carl .... yeh the damping is a bit out by the looks of that dip ... I will post a shot of the scope showing the voltage on each side of the coil ... there is a high flyback voltage on the mosfet side as shown already but there is also a pulse of lower voltage on the other input ... but kind of delayed or something else.

I have noticed that with the addition of the so called blocking diode my flyback voltage is much increased ... my "traditional" single ended version of this circuit using the same coil, mosfet, damping and 5 volt TX supply peaks at 140 volts. The sample point is about 20uS. ( with schottky diode clamping ). With this "balanced" version the clamping time has decreased below 10uS and the peak flyback voltage is approx 260 volts ... though I suppose if I increase the damping this will fall.

moodz.

[Carl-NC]

Paul,

This looks a bit better, but I'm not sure that it will offer any real benefits. Would need to give it a try. A while back Aziz posted a coil method that resembles an autotransformer, and it has the potential for true differential operation.

In your "No target" pic the opamp output has a single oscillation, suggesting that the damping needs adjustment.

- Carl

[moodz]

Thanks, Moodz. I would like to suggest that it would make a whole lot of sense if you also posted the corrected circuit with the balun, on the off-chance that some of us uneducated dummies might get an inkling. Please?

hiya P ...
the balun idea in version 0 was to reduce EMI ( RF noise ) entering the amp. Baluns dont work at DC so the biasing arrangements that 'float' the coil are to some extent unaffected by the presence of the balun. So I will proceed without the balun for now as I really need to get this front end going .... promised the guys at work

PS I am sure you're not all dummies.

Well... all I can see from the scope is a poorly balanced opamp output and blocking (or at least decreased ampli) for ca 20us...

you need to get your glasses checked ... the scope is set for 5 uS / div ... 20 us is nearly off the display.

well...

Well... all I can see from the scope is a poorly balanced opamp output and blocking (or at least decreased ampli) for ca 20us...

[porkluvr]

Thanks, Moodz. I would like to suggest that it would make a whole lot of sense if you also posted the corrected circuit with the balun, on the off-chance that some of us uneducated dummies might get an inkling. Please?

[moodz]

balun revised to floating coil

moodz,

I'm traveling this week and cannot do much more than look at your circuit, but I'm failing to see how this is any different than the traditional front-end method. Can you post the scope signals at A and B (at the same scale)?

- Carl

Ah yes .... actually I fail to see it also as I have left a blocking diode out of the +V feed .... so I have remodded the circuit again to try to demonstrate the benefit of a differential coil over a single ended / grounded coil.
With no balun this time as it probably confuses things unless you are pulse engineer.
Here it is .... explanation below picture.




NO TARGET



10 cm copper target at 25 cm.

Note that this is the output directly from the amp in the schematic above.

Circuit operation : what I am trying to do here is to present the amplifier with a differential input from the coil. In order to do this we must disconnect the coil from signal ground at the time when we want to amplify the signal from the coil ( ie straight after the tx pulse + back emf). On one side this is achieved when Q1 turns off.
On the other side D3 .. the blocking diode .. is reverse biased by the coupled EMF from the flyback pulse. Note that in this circuit there is a voltage on both sides of the coil ... ( A B ) however they are not equal. Note also that in this configuration that even using less than 5 volts to drive the coil we are getting a 250 volt plus peak with a decay rate that lets the sample begin at less than 10 uS.
As you can see from the scope shots above the decay is very good ( well it is for me :-) )
My noise problems with this ampifier have now gone ( fingers crossed ) I could not really get this to work with the 'traditional' input as Carl mentioned above.

The scope shots above were taken directly from the sig out of the amplifier from a bread boarded version of the schematic above. No additional components or tweaking (variable pots, caps etc ) were used ... and the breadboarding is really rough .... believe me.

I am hopefull of more performance tweaks when I can layout a PCB for the front end ( hopefullly this weekend. )

Anyways thanks for reading ...

moodz.

[Carl-NC]

moodz,

I'm traveling this week and cannot do much more than look at your circuit, but I'm failing to see how this is any different than the traditional front-end method. Can you post the scope signals at A and B (at the same scale)?

- Carl