Step-up transformer to connect to an AWG output (for a curve tracer)

[RoGeorge]

Would like to connect a transformer directly to the 50\$\Omega\$ output of a generator, to step-up the voltage in order to build a curve tracer from an AWG and a DSO in XY mode.

I have 3 small transformers at hand, 2 from former wall wart power supplies, and a 3'rd of an unknown provenience:



- T1:  220Vac/12Vac,  450mA,  N ratio=18,  Rprimary=1670\$\Omega\$,  Rsecondary=5.5\$\Omega\$
- T2:  110Vac/10Vac,  200mA,  N ratio=10,  Rprimary=570\$\Omega\$,  Rsecondary=4.7\$\Omega\$
- T3:  ?Vac/?Vac, ?mA,  N ratio=30,  Rprimary=257\$\Omega\$,  Rsecondary=1.0\$\Omega\$

The AC voltages that the AWG (a Rigol DG4202) can generate on a 50\$\Omega\$ resistive load is max 10Vpp between 0-20MHz.  Would probably never use more than 1kHz anyway for curve tracing.  However, max 10Vpp (as in +/-5V) is too little to test certain components with a curve tracer (e.g. for a more than 5V Zener diode, or to see the breakdown voltage in transistors, etc), so I want to increase the max AC available with a step-up transformer.

I'm tempted to use the transformer with the biggest transforming ratio, 30, though that would mean the 50\$\Omega\$ output impedance from the AWG (neglecting the resistance of the transformer windings) will be seen as a 30*30*50\$\Omega\$=45k\$\Omega\$ on the other side of the transformer.

45k\$\Omega\$ in series with the DUT seems pretty big to me, and I've never had a proper curve tracer, so I don't know if I should go for a x30 step-up voltage (+/-150V max) at 45k\$\Omega\$ output impedance, or for a x10 transformer (+/-50V) at an output impedance of 5k.

What other aspects to consider for a curve tracer?
Which transformer to choose?

[moffy]

Give yourself a choice and use both, easily switch selectable, there could be applications where one or the other would be more suitable.

[RoGeorge]

T3 of unknown provenience turned out to saturate very easily.  The voltage out on a 10k load resistor drops 2-3 times at less than 500Hz.  I suspect it might be an audio transformer from a vintage radio with tubes, thus the very small resistance on one of the windings (1\$\Omega\$).

Was 30:1 a usual ratio for an output transformer in audio amplifiers with tubes?


As for the other 2 transformers, would probably use both of them for the curve tracer.

Either in series, to increase the total voltage available for the DUT (though, T1 and T2 are quite different), or even better, use the transformer with a higher voltage output to drive the DUT, and the other one as a current transformer, to read the current through the DUT, like in the draft schematic. 

[coppercone2]

why not get a small power amp? The frequency is low so its easy to develop

I am not hopefully about getting a transformer that will work correctly on a 50 ohm input impedance when they are all designed for mains in mind.


A simple voltage follower or unity gain differential amp with a power op amp would boost your capabilities significantly. You basically need to do zero engineering to ensure proper behavior under a few KHz. This is so slow it might even be viable with a discrete amplifier that is not super well engineered for response.

IMO a 45k impedance source is too high. Won't it normally be say 600 ohms? I don't think its even near realistic component values at 45k.


The form factor of a transformer sucks anyway. You can't make a nice module that plugs into the FG anyway, because its clunky already with three transformers. making it into a box that has its own power and sits under the FG would basically be a requirement anyway.

[RoGeorge]

why not get a small power amp? The frequency is low so its easy to develop

Was trying to keep it dead simple, all passive.

Indeed, a low impedance repeater before the transformer is a very good idea, thank you. 
Very tempting.  That will need yet another wire for power, but it will also allow to use a bigger transformer (which I might have already if I dig through the scrap boxes), and that will make possible to test power components at higher currents than only ~5mA (like it is now).

[mawyatt]

Here's a link to something that might be useful, note the links to AWG HV Buffer Amps in #13.

https://www.eevblog.com/forum/testgear/fooln-around-with-dso-awg/

Best,

[RoGeorge]

https://www.eevblog.com/forum/testgear/fooln-around-with-dso-awg/

You may find the coincidence amusing, but that's exactly the link that, no longer than 2 days ago, triggered again my need for a curve tracer.  More precisely, the swapped CE transistor trick you shared there, and it's region of negative resistance.

Last week I was looking into a type of LC oscillator said to be very stable (the Franklin oscillator) ---> which led me to a book stating that negative-resistance based oscillators are particularly stable (which I didn't get why, but wanted to experiment with) ---> which reminded me about your trick with the reversed CE transistor and its negative resistance ---> which made me wanting again a curve tracer, because just like back in 2022 when I've tried with a SF128 transistor, I still couldn't find any negative resistance region (this time for BCY59).

Now, in a Franklin oscillator I suspect the main factor for its frequency stability is the very low coupling with the LC tank, and wanted to experiment with this idea of a Franklin-like low-coupled LC tank + a genuine negative resistance device, with the hope to get a new funny oscillator.

As usual, any intermediary step turns into a project in itself, with its own intermediary steps, and so on forever. 

I remember your HV buffer for AWG, is based on an HV APEX opamp which I don't have.  That's on my project list since long ago, just that right now I was looking for a quick improvisation, so to curve-trace more transistor types, and at a higher voltage than my AWG can do (+/-5V peak in 50\$\Omega\$), in order to finally find a negative resistance reversed BJT, so to advance with the Franklin-like oscillator experiments.

[mawyatt]

In this link there is a HV AWG Buffer Amp based upon the OPA462 as well as the APEX PA443. The OPA462 chip can only support lower voltages as noted. APEX has a single channel version, believe it's the PA441, that might be more available. There's also a version based upon the LM3866 which supports a lower voltage but much higher current use (maybe good for step up transformer use).

https://www.eevblog.com/forum/testgear/hv-buffer-amp-for-awg/

Best,

[Mortymore]

If it helps, the use of an AWG for a curve tracer was also addressed here:

https://www.eevblog.com/forum/beginners/can-i-build-this-curve-tracer-circuit-with-an-awg/