DoSuSy - LTspiced project. Comments?

Status
Not open for further replies.
Jump to Latest
#21
mosfets selection

Is it hard work to dissipate about 50W with TO-220AB? There are 4 mosfets
in the project with constant dissipation about 50W*4 and 2 mosfets with 25W*2.
Now, irf9540/640 are in the model. Probably, these mosfets selection is not sutable
at dissipation point of view. Are there (not very expensive) alternatives?

To Grey:

1. I have looked at a thread you share your experience concerning
TO3, TO247 an so on. What about TO-220?

2. Have you looked at the last schematics?

Andrew
 
#22
Andrew - 50W for one TO220, is too much, it is dangerous!

The bestest is TO-3, due to a big surface, connecting to the heatsink.
But TO-247, even with dielectric (nomax or something else) betwen the fet and the heatsink - is also good.

Don't afraid high tempareture of the heatsink, I said more - you don't need to measure it !

Check the fet's case temperature (especially if You'll use a dielectric), it is more important!

Keep in mind, than in the worstest case (hot summer + bad airflow around the amplifier + possible closing of the vent-holes in the amplifier's case) - the mosfets case temperature should not be more then 130C !
Depending of the dielectric between the fet and the heatsink (or it's absence) - heatsink' temp maybe 10-30C less.

If your final construction ("in metal") is running very closely to the limits - use thermoprotection.
 
#23
Alex,

OK, I have replaced output mosfets with irfp9140n/irfp140n (TO247AC in
datasheet).

BTW, what is average open threshold voltage difference for these and
9610/610 mosfets? I mean mosfet pairs from the same production cycle.

You see, I try to shift the model closer and closer to reality 🙂

Andrew
 
#27
Hi, Anli,

Very nice schematic. You even don't have to put Ground Resistor at differential mosfets, instead you put simple common DC servo at CCS for output. I like that smart idea.😀

But in your Version 2.0, I see a possible problem. Since you don't use current mirror in Version 2.0, the voltage drop in R13 won't be the same as the voltage drop at R14. The same goes to voltage drop at R15 won't be the same as voltage drop at R16. This is because in basic transistor equation IE=IC+IB, so even if it is very slight, I in collector won't be 100% equal to I in emitors. From this point of view, I think your Version 1.0 will be better.

What if you use current mirrors instead?
 
#28
Lumanauw,

I have tried to use a current mirror (to be more strictly, a "current mirror
with current amplification") and have got about 1db worse result (I mean
3d harmonic). Probably, the reason is curremt mirror transistors works with
very different collector potentials.

I think, preamp-repeaters-inverters part can be better. At least, I'll try.
The main problem is to make the project as simple as possible 🙂

Thanks for your feedback! Any other thoughts?
Andrew
 
#29
altor said:
The bestest is TO-3, due to a big surface, connecting to the heatsink.
But TO-247, even with dielectric (nomax or something else) betwen the fet and the heatsink - is also good.

Keep in mind, than in the worstest case (hot summer + bad airflow around the amplifier + possible closing of the vent-holes in the amplifier's case) - the mosfets case temperature should not be more then 130C !
Depending of the dielectric between the fet and the heatsink (or it's absence) - heatsink' temp maybe 10-30C less.

If your final construction ("in metal") is running very closely to the limits - use thermoprotection.
Click to expand...

Dear Alex,

in fact the TO-247 has better thermal conductivity than TO-3, although from visual point this seems not logical.

In older Pass amps there are tons of TO-3 cans used. If you compare to more advanced IR fets like IRFP244 or similar you will notice that the thermal resistance from junction to case and from case to heatsink is lower with TO-247. This is one of the reasons why TO-3 fets seldom can handle more than 125W (IRF240)power dissipation whereas TO-247 fets easily handle >150W (IRFP244).

A case temperature of 130° C is FAR to high! In case of TO-3 this means 150° C junction temperature at 20W dissipation! A typical TO-3 fet like IRF240 has a thermal resistance of 1° C/W from junction to case. A typical TO-247 fet like IRFP244 has a thermal resistance of about 0,8 from junction to case.

For safety and reliability reasons the junction temperature should NEVER exceed 100° C e.g. ~80° C case temperature.

Uli

:nod: :nod: :nod:
 
#30
For safety and reliability reasons the junction temperature should NEVER exceed 100° C e.g. ~80° C case temperature.
Click to expand...

I have a very good experience using mosfets (moslty in TO220, D2Pack and D-Pack) in my SMPS designs, with 100-110C case temparature in mass production units (>30K/year) with lifetime 60-80K hours (and this lifetime depend only of the El.Caps' temperature).
 
#31
difference

As a silicium junction normally breaks at more than 150° C
it is recommended not only to care for case temperature. The more important issue is the junction temperature. In fact case temperature is only important for the near surrounding (-> Elko).
The reliability of a semiconductor device depends heavily on the junction temperature.

An advice to drive a mosfet junction in an amplifier to 150° C is not a good advice!

As this is passlabsforum I assume that fets are driven at least at 15-20% of their capacity and driving a mosfet like IRF240 to 130° case temperature means 150° (20W) - 160° (30W) junction temperature - > fet dies. R. I. P. :bawling: :bawling: :bawling:

Of course driving a fet to 130° case t. with 5W power is "safe" but this leads to Aleph3 with 20 fets per channel😀

Uli

:nod: :nod: :nod:
 
#32
I don't remember of the IRF240, but IRF150 has max junction temp - 170C.

I'd not adviced running 150C junc.temp at the amplifier (130C case), I'd adviced that this is absolute maximim!
Of course, its better to have lower temp.

But (!) - I'm not agree with people, who recommend to keep case temp not more then 60-70C !!!
80-100C - this is a good compromize between the reliability and the heatsink size.
 
#33
Hi,

Next version is published:

The project page: http://freedns.ru/andrew/dosusy/index_en.html
Bode diagram: http://freedns.ru/andrew/dosusy/bode.pdf
FFT diagram, 1W: http://freedns.ru/andrew/dosusy/20KHz-R37-1W.pdf
FFT diagram, 10W: http://freedns.ru/andrew/dosusy/20KHz-R37-10W.pdf
FFT diagram, 33W: http://freedns.ru/andrew/dosusy/20KHz-R37-33W.pdf

Changelog:

- IRFP9140N/IRFP140N are in use,
- preamp-driver mosfets were replaced with bipolar transistors,
- repeaters/inverters were replaced with emiter followers,
- frequency/phase correction done!!

Hey! As always, I'm waiting for your critiques!

Andrew
 
#37
Re: Re: Dead Links

anli said:

Elso,

I have rejected this project. Now I'm DIY-ing another one. But I have uploaded last DoSuSy schematics for you especially 🙂 :

http://gaydenko.com/archive/DoSuSy-0.31.png

Andrew
Click to expand...
Hello Andrew,
Many thanks!🙂
The cascoded inputstage is interesting though I will do it with FETs.
http://www.diyaudio.com/forums/showthread.php?postid=507432#post507432
I need to have a word with that IDIOT Quirck though.
Looking forward to your next project Andrew.

😉
 
Status
Not open for further replies.
Top Bottom