Sinus PSU Prototype

[sinus trouble]

Hello Necks
I have another PA and exciter built and ready to go! So I wanted to design a basic linear PSU specifically for its application!
Afterall the PSU is a very important part of any transmitter and this time round, I wanted to go full regulated!

Sinus PSU.PNG

The circuit above uses 3 adjustable independent outputs controlled by LM317s! One to power a fan, One to power the exciter and the last driving two big fets for the PA!
Based on a single sided board and an aluminium plate will be mounted underneath the semiconductors secured with insulators and bolts!
I also added a small transistor for external protection? But haven't decided yet on current or SWR based?
If anyone can see any possible problems? Advice would be much appreciated!

[Analyser]

You need to add some bypass caps around the LM317s or they might go unstable.

[sinus trouble]

Cheers Mr Analyser!!
I have added some 0.1uf ceramics to the inputs of each LM317! The datasheet suggests adding caps to adjust and output pins yet optional? Maybe it would be good practice to include these?
Anyways here is the new version!

Sinus PSU Ver2.PNG

[Albert H]

Small bypass caps at the inputs and outputs are sensible, and I have had problems with hum modulation on the supply rails if you don't bypass the "adj" pin too!

It would also be sensible to space the '317s further apart, to better distribute the heat across their common heatsink. I'd move your shut down circuit to "below" the '317s, and spread out the regulators. You also need space for local 10 - 47µF caps on the regulator outputs - usually a good idea. Also - filter the common incoming supply: A big smoothing cap locally to the board can't hurt, with some RF filtering on the incoming supply too.

[sinus trouble]

Nice One Mr Albert!
I guess more filtering would be best, Especially in an RF environment!
The protection circuit was a tricky one? As is? It shuts down the gates of the power fets yet could be arranged to pull down the adjust pin to ground??
The PCB as stands is less than 5" in length so extra spacing could make a huge difference? Afterall I know the fets are gonna get very warm! lol
Il have a tinker and see what progress I can make?

[sinus trouble]

Hello Again Necks!
I have finally built a test prototype and its looking promising!

20170428_200217.png

Without the unit bolted to a suitable heatsink? I cant push it yet!
So far ive managed just under 3amp @ 18volt from the FETS! Warm to the touch but not too bad lol

[radio-berlin]

Looks neat, welldone

[sinus trouble]

Cheers Mr Berlin!
The K2690 datasheet is somewhat ambitious? But only time will tell if it survives! Lol!
More testing to be done!

[sinus trouble]

Hello Again Necks!
I have been doing some further tests and something I did not expect is happening with the FETS?
My first test was with an input of 24v and output set at 10v consuming 3A? Over time, the FETs reached around 80/90 degrees ish?
Second test was again 24v input but output 20v consuming 3A?? Now I assumed that the temperature would rise considerably? Yet this is not the case? Temperature seems to be the same??
My only guess is that there is some relation between the supply voltage and output difference?
Nether the less, its good news so far!

[Analyser]

Sinus, with all linear voltage regulators the voltage you need to drop is wasted in heat, therefore the more voltage you drop the greater the heat generated.
An example-
In case 1 your output voltage is 10V. Your input voltage is 24V. That means you're dropping 14V in the regulators (FETs). If you multiply this by the current you get watts dissipated in the circuit. In this case 42W.

In case 2 your output voltage is now higher at 20V, meaning you're only dropping 4V from an input of 24V. With another 3A load the total power dissipated is 12W.

So, the 20V output should make the FETs get hot less quickly. This all depends on your heatsink! Try not to run them at 80/90 degrees though, it sounds like you need more cooling.

[sinus trouble]

Yess Mr Analyser! It all makes sense now!
I deffo need a proper heatsink for continuous use! I did have an old fan connected to one of the 317s to blow gently over the board during testing? However this only added more heat emanating from the 317 lol