L'Fake - A Cheap- A** Expedient to Replace the Elusive SIT

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I finally got around to fixing up the gain-phase in the amplifier shown in post #131. Most of the corrections consisted of snipping out unnecessary compensation components. The two channels don't match all that well, though the gain and phase margins are quite adequate on both.

As per my findings so far, a Zobel RC network across the output terminals is absolutely essential for keeping the gain below 0dB after crossover. I'll look into the reasons for the difference between channels at some later date. For now, I'll hook the thing up to some speakers and see if I actually enjoy it. I'll post an updated schematic when I get around to it.
 

Attachments

  • 159_Fake_G-P.GIF
    159_Fake_G-P.GIF
    43.2 KB · Views: 428
One of my New Year's resolutions is to tie up a lot of the unfinished projects I have lying about (too many). Tidying up the control loop of the L'Fake was a start. I haven't forgotten about L'Fake Lite and its switching regulator/current limiter. I finally have come up with a regulator circuit that satisfies me and provides the control of start-up current that is needed to tame the savage light bulb and prevent my line fuses from exploding.

There might be a chip that does all the things I need, but I didn't bother to go looking. Instead, I used the TLC555 as the nucleus of a switching regulator (CMOS version of 555, much faster than the original bipolar 555). The circuit runs in variable frequency constant on-time mode, which allows me to really run down down the effective duty cycle by decreasing the operating frequency. This allows one to shed volt-seconds accumulated during the fixed on-time and maintain an iron control over the output current limit, even during an effective output short circuit. The TLC555 floats around the input voltage. A ground-referred LM392 is used to sense output current (it talks to the TLC555 via an optocoupler), and the extra comparator in the package is used to implement a frequency soft-start scheme that gives the current sense amplifier time to wake up and assert control of things. I verified basic functioning over the holidays, and the only item left on the plate is to stabilize the current control loop - the average value of the output current is controlled, but the loop oscillates in current mode. That'll get fixed. I also need to make cure that the hand-off from voltage mode to current mode is smooth, always a concern with CVCC circuits. A circuit will follow when I happy with things. It's something that the average Joe might actually be able to build, as the output inductor is based on a standard Micrometals powdered iron toroid (mix 52) that can be found on Ebay or bought from folks like Amidon Associates.
 
I did my testing of the switching regulator/ current limiting circuit with an electronic load (Kikusui) that cuts out at around 1.5V. It very well could be that this is the cause of the instability I see when the regulator turns on in current limit. Before I mess around with the circuit any more, I think it would be worthwhile setting up the regulator to run my L'Fake Lite, light bulbs and all - I may be pleasantly surprised. I'm also using a small 5A circuit breaker to protect the input (it's nice looking and doubles as on/off switch and pilot light) so at least I won't have to worry about replacing any blown-up fuses.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.