push-pull SMPS datasheet mistake ? [TL494CN] [SOLVED]

[Mate_Well]

For a project i will need to step up voltage from 12v battery (to respectively +17 v and - 17v rail) i'm left with creating smps becouse of this, found an ic that has all features (if not more) that i need.
The ic is TL494CN  (datasheet : https://www.gme.cz/data/attachments/dsh.322-001.1.pdf ) i have tryed to understand the basic functions, calculated most of the component values, transformer etc....

Either i understood it's function wrong or there is mistake in the datasheet, i feel like the bases of the stransistors should not be conected together (since it switches only 1 of them at time, that part is also "highlighted" in light red)
in attachments there should be 2 pictures (scheme from datasheet, and one that i used for understanding it)   

any help would be appreciated <3, don't know if i should trust datasheet or my gut feeling in this case

[T3sl4co1l]

Yeah, obvious goof.

Much better anyway to use N-MOSFETs; even the contemporary IRF540 will perform better than the TIP32s shown.  The gates will need a bit of a driver; the TL598 can be used directly, or a dual gate driver IC added.

Here's a basic scheme:



Notable improvements that should be made:
- Note that the output is stacked on top of (i.e., in series with) the supply; this saves on capacity, but also means the output can never be unpowered (even when the converter is disabled -- if you need that as an option -- 12V flows out through the diodes).  This connection is easily enough changed to a regular common-ground secondary, like shown above.
- And of course, for +/-17V, you'll need a FWB and two chokes, instead of a FWCT rectifier and one choke.  (The secondary CT will be ground, no choke there.)
- The snubber is awfully aggressive (the 50 ohm 5W), probably that much won't be necessary.  Depends on the transformer's leakage inductance, which depends on how it's wound.
- Current feedback can be compensated by bringing an R+C from pin 3 to pin 15, and reducing the capacitance on pin 16.  Probably as shown, it would oscillate or chatter when current reaches the threshold; probably not destructive, but not ideal.
- Even more preferable, current feedback should be used as a primary feedback option, and a separate error amp used for voltage feedback.  This changes the controller to average-current-mode control, a dramatic improvement over the voltage-mode control shown.
- The capacitance between pins 2 and 3 needs to be an R+C as well.  Typical starting value 100k + 10nF and adjust up or down from there until good stability and load step response is found.
- An R+C from "VFB" to pin 1 may also be desirable, particularly for voltage-mode control (combined with the R+C between pins 2-3, this is called lead-lag compensation).
- Note this shows how to use a TL494 with a discrete gate driver.  Using a TL598, the 1k pulldowns on pins 9 and 10 are not necessary; the drive diodes and 2N4403s may not be necessary either.
- If isolated output is required, that "VFB" connection won't be possible; instead, you need an optoisolator with a TL431 or similar voltage reference / error amplifier.
- Depending on your source, protection may be desirable: transient protection, additional filtering, UVLO/OVLO, "hiccup" current limiting, overtemp, etc.  Some of these can be added on easily (e.g., TVS diodes, filter caps and chokes), others require additional circuitry; the main downside is the circuit simply grows as you tack on more features.  At some point you're probably better off using a more modern controller that contains all these features.
- And for that matter, if you're not drawing much current from it after all, you may be better off with a flyback type supply, even a ready-made off-the-shelf converter/isolator module.

As for transformer winding, it's worthy of its own topic, really.

Tim

[Mate_Well]

Huge thanks <3 
the part behind secondary will have to be redone, i will use 2 FWB with both chokes (and "big" capacitors), i will use the overcurrent preatty much as only hiccup mode (if the amplifier part that will be behind, get's shorted or something like that  ) the plan for the project is learn a lot, and ideally also make 25~40W (x2, one for left and one for right chanel) car amplifier, primary will have to handle somewhere around 5~6A (worst case) 

[T3sl4co1l]

Only need one FWB, no need for extra voltage drop or extra transformer windings. FYI.

Tim