TSL 2N3771 fakes or not?

[sv3ora]

Hi a few years ago I have purchased three TSL 2N3771 transistors. They are exactly like the ones shown in the photo (markings, metal finish etc). Mine have dates 01 96 and 04 95.
Looking at the photo do they look fake to you?

I need to make a new 41A linear PSU, using my transformer and the rest of the components from my recently failed 40A 9x 2n3055 PSU

[wraper]

Regardless if they are genuine or not, your PSU is broken by design if you think passing 41A through 3x 2N3771 in linear operation.

recently failed 40A 9x 2n3055 PSU

Even 4.5A per 2N3055 is way too much unless you are dropping like 7V volts maximum on transistors which are mounted on a large heatsink.

[sv3ora]

Regardless if they are genuine or not, your PSU is broken by design if you think passing 41A through 3x 2N3771 in linear operation.

Why is that? They are rated for 30A max and I will drive them  at less than half their max power. There is such a project here https://www.hobby-machinist.com/threads/a-complete-13-8-volt-50-amp-power-supply.63178/ with 4 of these though.
By the way, my transformer measures 14.2x11.4x13 centimetres in size and the secondary wiring in it is 4.14mm. Would this be capable of 41A?

Even 4.5A per 2N3055 is way too much unless you are dropping like 7V volts maximum on transistors which are mounted on a large heatsink.

The old PSU was marked as RMS (made in Italy company) K1540, which might mean 15v max 40A, or maybe not.
I cannot measure the output voltage of the transformer right now, but I could tune the PSU for 14.2v (I did not try further) if this tells anything to you. It was working ok for months, then suddenly the fuse was blown. I have not yet figured out if any transistor is fused or not btw. 41A were drawn in peak, allow me to say (SSB operation) so yes, maybe on CW it would not have the guts. However, I never left the two heatsinks to go very hot, I installed a blower to it and they were always cool. Also, when the PSU was broken, the PSU had just been switched on, so I had not drawn excessive current before failing. these are the hints on this problem basically.

[Kleinstein]

The relvant limitation is usually the cooling. The 2N3055 is specified to allow some 115 W if the case is cooled to 25°C (maybe 20° C in some versions). Real life outside Greenland one would more like get some 50-80 C at maximum power.  This reduces the maximum temperature difference from junction to case from some 150 K to some 125 - 90 K, or some 16-40% reduction in permissible power. Better cooling would need a larger fan or heat sink -  to day better use a few more transistors.  Current sharing between the trasistors would not be perfect so that the worst transistor could get some 20% more current than average.  So one can more like use something like 1/2 the nominal P_tot - so more like 50-60 W per 2N3055.  In a design without fold back curent limit and some 15 V raw voltage this would be 3.5-4 A per transistor maximum. With a higher votlage the current per transistor would be lower.

15 V is still relatively low. With more than some 25 V one has to also look at the SOA curve that may allow less power than the nominal P_tot when the voltage is high.

Without PFC the transformer would need to deliver something like 1.6 times the DC current. So for 40 A DC the transformer would need to be good for some 60-70 A.  It maybe slightly better with a more soft transformer and slightly high with a large ring core. So chances are the transformer is too small to deliver the current for a longer time.

Only some 30 W per transistor is rather low / conservative, but better safe than sorry.

A 2N3771 is less powerfull than 2x 2n3055.

The design shown has a relativly high drop out. So it is not really suited for high power and may need a relatively high transformer voltage. By reducing the drop out ( e.g. use a PNP to drive the 2N3055) and use a seprate fitler / rectifier for the 723). one may reduce the needed transformer voltage by some 2-3 V and thus get less heat.

[sv3ora]

If that's the case why thy are making them so powerful? It sounds like cheating to me.

[Kleinstein]

The bold letter specs are for marketing and a quick first view only.  Giving the power rating for a fixed 25 C case temperature is kind a standard for power transistors  (low power ones use 25 C environment, so a little more realistic).  It still helps to get comparable data, just don't expect to really use that muchpower, unless very well cooled with cool water or similar.

In the early times when the transistors were more expensive a larger heat sink may have been more ecconimic. Even than P_tot was allways some (theortical) upper limit nothing for real life.

It gets even worse with modern MOSFET ratings: somehow they claim 150 W for a TO220 case (e.g. IRLZ44), while real life TO220 is no good for more than some 30-40 W and in the old days usually got a power rating of < 90 W.  The SOA is another special problem there:often not there and if there ususlly often not reliable / applicable as the irgnore the relevant part why they have a SOA curve and not just a P_tot number. 

[jmelson]

If that's the case why thy are making them so powerful? It sounds like cheating to me.

No.  Power transistors can switch lots of current for a short time, or pass large current with small voltage drop, or small current at a large voltage drop, continuously.  But, when you ask for large current at the SAME TIME as large voltage drop, that will cause failure of most transistors.  It is called secondary breakdown, and is caused by localized heating of the transistor die.  This is all shown in the detailed data sheets in the "safe operating area" graph.

Jon

[sv3ora]

If that's the case why thy are making them so powerful? It sounds like cheating to me.

No.  Power transistors can switch lots of current for a short time, or pass large current with small voltage drop, or small current at a large voltage drop, continuously.  But, when you ask for large current at the SAME TIME as large voltage drop, that will cause failure of most transistors.  It is called secondary breakdown, and is caused by localized heating of the transistor die.  This is all shown in the detailed data sheets in the "safe operating area" graph.

Jon

All right, so basically this tells me to measure the AC output voltage of my power transformer.
No I know that the previous PSU was able to do 14.2v at least, so if I measure an AC voltage as close to it as possible this should give me an idea of how much current I can draw out of the transistors. Small difference=large current and vice versa.

[SilverSolder]

Here is an example of a good quality fanless commercial power supply, rated at +/- 15V at 4.5A (135W) at up to 70C ambient.



It uses a total of 12 pass transistors in Darlington configuration (six on each side) to do that.

Basically, it takes a surprisingly large amount of silicon to make a reliable linear power supply...