Square wave into a transformer, low impedance output needed

[gnurf]

I need to connect the transformer to an 5V open-collector source as shown below, but the load must not see a larger impedance than 50Ω. The 5V source is an internal source to the driver and not available outside of it. As it stands, the seen impedance will be 5k when the bjt is off, so this will not work. I need something in between and was thinking of an emitter follower (or maybe a opamp buffer), but my simulations went a bit awry (struggling with my transformer model) so I thought I'd throw this out here in case there is an obvious solution (which there probably is) that I've missed. Any ideas are welcome.

 

[berkeman]

I need to connect the transformer to an 5V open-collector source as shown below, but the load must not see a larger impedance than 50Ω. The 5V source is an internal source to the driver and not available outside of it. As it stands, the seen impedance will be 5k when the bjt is off, so this will not work. I need something in between and was thinking of an emitter follower (or maybe a opamp buffer), but my simulations went a bit awry (struggling with my transformer model) so I thought I'd throw this out here in case there is an obvious solution (which there probably is) that I've missed. Any ideas are welcome.

Yes, I'd use an emitter follower off of the collector. And you need to move the DC blocking capacitor from the output/secondary to the input/primary...

 

[berkeman]

On second thought, the CE output can only pull up on the output, so there is nothing to pull down for the other half-cycle. You should probably use a complementary output stage with NPN and PNP transistors...

 

[gnurf]

Thanks for the input. I've played around in spice and this is what I got so far:

As the circuit stands the impedance the transformer (1:1) sees will be < 50 Ω and thus the load on the secondary side will be satisfied on that account. Additional load requirements are transition time < 100ns and TTL voltage levels. I'm not happy with how delicate the balance between sharp edges (low values of R1 and R2) and enough voltage headroom on the output (large R1 R2 give amplitude closer to 5V than low values of R1, R2). Any quick fixes for this problem, or is this a tuning problem I have to live with?

Any favorite work-horse transistors/diodes that could work well under the above conditions?

Are there any standard improvements to this circuit?

What kind of load is the 1:1 transformer? Will it affect performance (assuming it has a bandwidth of 400MHz)?