Hi guys!
I've been trying to modify an ATX computer PSU into a lab bench PSU,with a fixed +24V output at the beginning.I spent days with it,firstly removed most of the secondary components and the output voltage got increased to over 20V,but not for long as it dropped immediately with some load and the power transistors started to switch randomly without deadtime.I thought that bad feedback at pin 3 of the TL494 causes this and tried different negative feedback capacitors.
Right now I have a weird gate waveform and the power transistors are heating slowly but no output voltage.Changing the TL494 doesn't solve the problem.Has anybody got an idea why I have this bad gate waveform?


Schematic which I made:


Modified ATX PSU:

 

Is the Push-Pull Power Output section - the original design or is it your design?
It appears that Q1 and Q2 have a Pull-Up resistor ( R14 & R15 ) that cause them both to be ON, during dead time?
When one output of TL494 is ON ( Active Low ) then only that corresponding transistor should be ON?

 

The output stage is original as you can see here:


I think during deadtime both transistors should be ON to short out the gate drive transformer in order to turn OFF both power transistors on the primary side.My problem doesn't happen during deadtime,but while switching.It looks like something is bad and the waveform can't be stable and starts to fall back to zero.(see scopeshot)

 

Q1 & Q2 do not "short out" the primary of T3
Q1 and Q2 allow current to flow to their respective half of T3 Primary
This causes "Magnetic Reset" of the Gate Transformer.

During Dead Time, it appears both halves of T3 Primary are ON ( Current limited by R17 ) and
then one side turns OFF to generate the Gate Pulse in the Secondary to activate the Power Transistor?

At what point in your posted schematic is that "Scope Shot" taken from?
Can you show the circuit between T3 ( Gate Drive ) and T1 ( Power Output ) Transformers?

 

The posted waveform comes from the base of one primary switching power BJT.Here is a very similar complete schematic of an ATX PSU:

 

The posted waveform comes from the base of one primary switching power BJT

What was the Ground Reference point?

 

Remember if using a PC PS for a lab supply that the DC common is usually referenced to earth ground.
Important to know when using grounded probes such as a 'scope' etc.
Max.

 

What was the Ground Reference point?

It was the negative output of the full bridge rectifier(E of Q2 on the complete ATX schematic).I get the same waveform across a single external turn around the base drive transformer.

Remember if using a PC PS for a lab supply that the DC common is usually referenced to earth ground.
Important to know when using grounded probes such as a 'scope' etc.
Max.

Yes,it is referenced to earth with the screws,which are removed.

 

Yes,it is referenced to earth with the screws,which are removed.

The schematic shows the earth ground symbol throughout!
Max.