I'm in trouble with designing and testing tl494 halfbridge converter (schematic based on atx power supply and this one: https://www.hobbyelektro.eu/spinany-zdroj-0-30v-0-6a/) Converter wasnt quite working, so I decided to make it part-by-part. Present configuratoin is: 
So i'm providing wariable pwm signal from tl494, and output is loaded, so i will not saturate TR2 core.
I'm also looking at oscilloscope in points A and B, referenced to low voltage ground. when i turn the potentiometer i can smoothly regulate ducy cycle, and see it on A channel, but when i'm in some narrow range of regulation it makes horrible sound like buzzing or squeek, and when i'm outside of it it's completly fine. When it makes this sound the scope also shows strange behavior on channel A. The best think is, when i disconnect high voltage i can see these "strange thinks" dissapearing and after 6 seconds i have got pure suare on the scope. Every measurment is made using short probe leads (this springy ones) so i don't think it's any kind of induced signal.
SCOPE SIGNALS:
Singal on points A(pink) B-completly flat and good looking. These somethinks after falling edge corresponds with the buzzing.
What can cause this behavior? Thanks in advance for your answers!
EDIT:
I did one think, added a buffer gate (made of two cmos nands) between TL494 and transistors:
And this not solves the problem... Waveforms on bases are ok, but circuit is still buzzing in narrow width of regulation. What can oscillate or something like that in such a simple circuit?
