Topology Federmann, HQQF-55 ...

[PMA]

Originally Posted by BV

Overcurrent "protection" is one mistake, showed frequency compensation second.. It will be for sure power oscilator (for short time before explosion), slow output stage and "fast" input and VAS without local compensation closed in global NFB..Unrealistic. We have to wait for measured results.

Yes, BV, we are still talking about the worst mistakes. Also, absence of the MOSFET driver circuit is a big issue.

[Juergen Knoop]

thank you Federman for posting the picture and telling us about the current status of your project.
Regards

[stinius]

Juergen

One positive thing is that if you build this thing and turn it on it will be completely dark.
You know ladies like candle lights.

Sorry for being of topic, but here is a link.

Cheers

[Juergen Knoop]

Stinius,
OT:
ouuh, but they certainly don't like campfires indoors!

Back to topic...I don't care very much about simulated circuits and virtual projects. If Federmann presents something readily build, I would be curious about it and would happily add my two cents why it works or fails.
Regards

[Federmann]

Originally Posted by megajocke

Here is why it will not work well without source resistors. Let's look at Vgs = 4.5V and Vds > 10V. In that case, the drain current is:

For temperature = 25 Celsius, it is Ids = 0.6A.
For temperature = 150 Celsius, it is Ids = 25A.

That is not a reduction in current with increased temperature!

All the parallelled transistors have the same Vgs so the hottest one will work harder than the others. Drawing a load line and pointing at it is nonsense. It is current sharing that is the problem, nothing to do with the load line.

However, the BUZ9** transistors you mention on your page do not have this problem. They can be parallelled just fine without resistors.

Transistors are next to each other, it is not possible that the first had a temperature of 25°C and the second 150°C. Forget the positive temperature coefficient RDS. I do not recommend temperatures above 80°C. Chips are selected for BUZ what the problem is to choose four transistors? Why look for problems where not?

Nice night

[PMA]

Juergen,

simulation, if properly used, is a great help and saves lot of time to a designer. However, its is the aid, not the final goal. Every professional company use simulators, please read Scott Wurcer's comment on that. But, if let in hands of pure theorists with no practical background, simulation may be misleading. Also, it does not substitute lack of professional background.

So please be careful before you disapprove simulation, rather consider the content.

[PMA]

Originally Posted by Federmann

Transistors are next to each other, it is not possible that the first had a temperature of 25°C and the second 150°C. Forget the positive temperature coefficient RDS. I do not recommend temperatures above 80°C. Chips are selected for BUZ what the problem is to choose four transistors? Why look for problems where not?

Nice night

No two IRF240 or IRF9240 have same charecteristics!!!!!

You consider one datasheet plot example as to be same for all devices, and this is a huge lack of understanding. The difference of characteristics leads to different current load to paralleled devices and therefore to different chip temperature. These things happen instantly and you do not see them in plots you are permanently displaying. Why do not you study publications on breakdown that danhard has linked here??

[Federmann]

Originally Posted by PMA

No two IRF240 or IRF9240 have same charecteristics!!!!!

Once again, good night

[PMA]

Federmann, please take a soldering iron in your hand, and build a working sample. Then you can defend your circuit ideas.

[Upupa Epops]

I am wailing, Federmann....try to make simply experiment: setup Vgs at 100mA Ids with IRFP240 and let it be. After a few seconds current will rise up to ampere region and after while device will be in silicon heaven...Now do the same with 2SK134 ( lateral type ) - here is Ids quite stable, or little fall with rising of temperature....It is reason, for what we can connect lateral ones in parallel, because they have in this working point zero Tc, but vertical ones not....Is not easy to understand ?