Author Topic: High power transistor  (Read 19694 times)

0 Members and 1 Guest are viewing this topic.

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 751
  • Country: gb
  • Life's too short.
High power transistor
« on: November 13, 2016, 11:22:03 am »
I was looking for a pnp transistor with more current capabilities than an MJ2955. I've found a 2N5684 and from what i can see it would be a higher current transistor, my intended use was for a pass element in a voltage regulator circuit. I was wandering if anyone can confirm this 2N5684 pnp transistor as a higher power alternative to the MJ2955 transistor. Thanks for reading any advice much appreciated.
 

Offline 2N3055

  • Super Contributor
  • ***
  • Posts: 6451
  • Country: hr
Re: High power transistor
« Reply #1 on: November 13, 2016, 12:31:18 pm »
They look fine.. What currents do you expect.. These guys need serious driving currents, low hFE... Also make sure you check SOA, as in linear series regulators most of the time it's power that kills you...
 

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 751
  • Country: gb
  • Life's too short.
Re: High power transistor
« Reply #2 on: November 13, 2016, 12:48:27 pm »
Thanks for your reply. Would they be suitable as a higher current alternative in these circuits below. Thanks for your help.
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14076
  • Country: de
Re: High power transistor
« Reply #3 on: November 13, 2016, 01:05:47 pm »
The TIP36 is a little more powerful than the TIP2955, but not much.

The circuit with an normal regulator driving external PNPs cause quite some voltage drop, and is not very stable against oscillations. So for a high power regulator I would prefer a different circuit.
 

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 751
  • Country: gb
  • Life's too short.
Re: High power transistor
« Reply #4 on: November 13, 2016, 01:15:54 pm »
I have tried it with a mj2955 and it worked well. I went on to try it with a high power darlington, and that did oscillate. I was looking for a higher current handling pnp transistor than the mj2955. But its not so easy to find one, that's why i asked about the use of the 2N5684 pnp transistor, and if it would function in these circuits.
 

Offline Cliff Matthews

  • Supporter
  • ****
  • Posts: 1910
  • Country: ca
    • General Repair and Support
Re: High power transistor
« Reply #5 on: November 13, 2016, 01:38:02 pm »
FWIW, look for a design that isn't so much of a room heater.. Maybe try £54 and get some quiet adjustable efficiency:

If you sell the huge transformer, rectifier and filter caps it will need, you can save double on electricity cost and buy 4 of these:
http://www.ebay.co.uk/itm/400-Watt-9-15-Volts-DC-36-Amp-HAM-CB-Power-Supply-12-Volt-Real-MegaWatt-/180777043955
 

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 751
  • Country: gb
  • Life's too short.
Re: High power transistor
« Reply #6 on: November 13, 2016, 01:44:24 pm »
Thanks, I've done this with 1200 watt server power supplys. Its building this linear circuits I'm interested in. The TIP36 looks promising, although a transistor that had a collector current of 50 Amps would be better. If i could find a suitable component. Preferably in a TO3 package. Do you think the 2N5684 is a viable option ?
 

Offline Cliff Matthews

  • Supporter
  • ****
  • Posts: 1910
  • Country: ca
    • General Repair and Support
Re: High power transistor
« Reply #7 on: November 13, 2016, 02:00:24 pm »
Dropping 20V at 50a is serious biz.. and 2N5684 is 300w. Unless you expect a lot of brownouts, even dropping down to an 18V transformer would help. Is the final objective learning and/or to swap the reg to an LM317? Is that the reason for 24V?
 

Offline 2N3055

  • Super Contributor
  • ***
  • Posts: 6451
  • Country: hr
Re: High power transistor
« Reply #8 on: November 13, 2016, 02:48:20 pm »
Thanks for your reply. Would they be suitable as a higher current alternative in these circuits below. Thanks for your help.

Search internet for HAM PSU schematics 12-13.2 V  20-30A
Plenty of them, those based on old LM723 are actually quite good.. You will get all kind of practical explanations that go with it too..

This was the topic few days ago.. Although making a simple PSU is very easy, making good, stable, high power linear PSU is not simple. Plenty of details to go wrong...
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16547
  • Country: us
  • DavidH
Re: High power transistor
« Reply #9 on: November 13, 2016, 06:31:12 pm »
For a linear power supply this large, I would also want foldback current limiting which is common in 723 designs.  Some improvement can be made in this design by adding diodes to compensate for transistor Vbe allowing the integrated regulator to better protect the transistors as shown below from the 1980 National Semiconductor Voltage Regulator Handbook.  If the integrated regulator is mounted to the same heat sink and the thermal resistances are matched, then it can also protect the transistors thermally.

In high power designs, one advantage of using a greater number of lower power transistors is that heat is spread over a larger area effectively lowering thermal resistance over using fewer higher power transistors.  I usually end up selecting parts based on power and thermal characteristics rather than current capability which also means that they operate at a higher part of the hfe curve.

I have never had a problem with stability when PNP boosting an integrated 3 terminal regulator.
« Last Edit: November 13, 2016, 06:35:13 pm by David Hess »
 
The following users thanked this post: Cliff Matthews

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16272
  • Country: za
Re: High power transistor
« Reply #10 on: November 13, 2016, 06:35:07 pm »
Simplest is just to add more TO247 PNP transistors and emitter resistors in parallel, as the drop in gain with increased current is going to really ruin your day. That way you run each individual transistor at a safe SOA point ( lower current, the 15A current is as a saturated switch with gain of 2in most cases, not what you want as a voltage regulator) with lower current, and they will also be able to dissipate the heat properly, and as a bonus the gain will be a lit higher per device. giving a better performance.

TO247 over TO3, as the heatsink space is smaller, you only need one hole and insulator ( unless you use a ISOTOP version in which case you only need the hole and thermal compound), they have the same thermal resistance junction to heatsink ( surprise, but comes from not having a steel case but Kovar instead) and are a lot cheaper.

6 for a 30A supply is barely able, giving each a 5A share, but better is to drop it to around 3A using 8, and then for higher current use 8 per 30A block, though there will need to be another transistor then as a driver, your 78xx regulator is only going to be able to supply 0.7A or so of drive at most, even flat out.

Over 20A I would personally recommend using a different topology, using a 723 regulator instead, as you can take this and scale it to a 100A power supply and have it still be stable, and still have current limiting that works. You then also add a power transformer that has multiple voltage taps on it, so you can select a range to minimise differential somehow.
 
The following users thanked this post: Cliff Matthews

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 751
  • Country: gb
  • Life's too short.
Re: High power transistor
« Reply #11 on: November 13, 2016, 06:56:05 pm »
Thanks for the replys. To many to quote..
Here in the uk brown outs are very rare, if you've ever seen photonicinduction on youtube you're understand why. I don't endorse his stunts though. I'm  all for the 78xx and the Lm 317 and similar regulators. The 24 Volts is to high on the input, i plan to close that gap, where each transistor shares 5 of the 30 Amps in the big circuit, that's to much for me from each mj2955.

This is the main reason i asked about the viability of the larger transistor, or one that can carry more current. I would add input and output protection to that big circuit. I have built it with 4 x TIP2955 transistors, but at 10 Amps (powering a linear amplifier ) It ran rather hot for my liking. And that was air cooled. I'm  after a transistor, preferably TO3 package that's pnp that will give far greater head room in that big 6 x transistor circuit. So does anyone know if my finding that larger current capabilities transistor above will do the job ? Or if not, do you know a transistor with greater collector current and power rating that will. Thank again all.
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16272
  • Country: za
Re: High power transistor
« Reply #12 on: November 13, 2016, 07:10:24 pm »
There is such a transistor, but the price is such that using 50 regular ones will be cheaper, and you will still need a 20kg heatsink on it. You are still limited by input output differential voltage times current which is heat in watts, and that has to be dissipated somehow, and the 2 ways are convection cooling with a great big finned heatsink, or a slightly smaller one with forced cooling.

If your unregulated input is 20VDC, your output is 12VDC and you are wanting the full 10A then you will have to dissipate 80W, and this will need a big chunk of metal to keep the transistor junction temperatures below 70C for best reliability, or you will have long term failures with then going short circuit and applying 20V to the radio, most of which really do not like having 20V shoved into the power connection instead of 13V8.
 

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 751
  • Country: gb
  • Life's too short.
Re: High power transistor
« Reply #13 on: November 13, 2016, 07:22:30 pm »
I'm  using a secondary voltage of 15 to 18 volts AC. I've  just had a wander round again and come across the mj4502 ! Its collector current is 30 Amps. Can anyone tell me if that's a viable transistor in place of the mj2955, it again looks promising.  I will get round to an LM723 based psu circuit, but first want to try this minimalist circuit above. Its attraction is low components count and cost. I can get the mj4502 reasonably cheap, is that one a viable unit ?
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16272
  • Country: za
Re: High power transistor
« Reply #14 on: November 13, 2016, 07:43:22 pm »
Not going to work, unless you can find a heatsink under 0.5C/W to mount it on, and your regulator would not drive it anyway. At that spec for the supply you have the gain for a "typical' device is 5, and thus your supply ability will be 5A.

You will still need at least 4 in parallel on the same 0.5C/W heatsink to get to the 10A mark, and it will still need to dissipate 110W of power in the pass element. Still a room heater, nothing aside from a 300x300x40mm slab of heatsink, RS 490-7040 will do, or RS 692-0629 with a fan.
 

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 751
  • Country: gb
  • Life's too short.
Re: High power transistor
« Reply #15 on: November 13, 2016, 07:55:55 pm »
I have used a 50 Amp darlington, but it was unstable and oscillated. Its gain was 1000,the mj 4502 has a gain of 100 maximum. Would it not work in place of an mj2955.
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16272
  • Country: za
Re: High power transistor
« Reply #16 on: November 13, 2016, 07:58:17 pm »
It will work in place, but you will still need to keep it at a case temperature under 70C to keep the transistor alive.
 

Offline 2N3055

  • Super Contributor
  • ***
  • Posts: 6451
  • Country: hr
Re: High power transistor
« Reply #17 on: November 13, 2016, 08:00:19 pm »
Like Sean and everybody else say,  max Pd of transistor is not MaxIc * MaxUce...   It is defined by thermal design, or choice of transistor+whole cooling solution, that includes ambient temperature (and humidity, even above the sea height ...)

Ptot in datasheets are defined as keeping case of transistor on a heatsink that has const temp of 25C, and no thermal resistance whatsoever.. Also they keep the junction at max. temp allowed (sometimes at 175C)... I real life you can count on 1/5 of that figure... 

For starters look at this :
http://www.designworldonline.com/how-to-select-a-suitable-heat-sink/#_
http://www.swarthmore.edu/NatSci/echeeve1/Class/e12Code/HEAT-NOTE.pdf
http://nvlpubs.nist.gov/nistpubs/Legacy/SP/nbsspecialpublication400-14.pdf

And plenty more..
 
The following users thanked this post: salbayeng, Cliff Matthews

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 751
  • Country: gb
  • Life's too short.
Re: High power transistor
« Reply #18 on: November 13, 2016, 08:02:40 pm »
Ok thanks, looking at the data sheet again the mj4502 has a minimum gain of 25 and a maximum of 100, other parameters don't  look a million miles of the mj2955 specs. Other than a greater collector current and a higher power unit.
 

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 751
  • Country: gb
  • Life's too short.
Re: High power transistor
« Reply #19 on: November 13, 2016, 08:09:33 pm »
If i take the boost regulator circuit above..
Would the mj4502 a single transistor supply current up to a maximum of 3 to 5 Amps on a suitable heatsink ? Its gain is 25 minimum and 100 maximum. Would the mj4502 work as a single current boost for a 78xx or lm317 circuit.
 

Offline Cliff Matthews

  • Supporter
  • ****
  • Posts: 1910
  • Country: ca
    • General Repair and Support
Re: High power transistor
« Reply #20 on: November 13, 2016, 08:43:24 pm »
With the 24v transformer, you were dropping 20v (~32v to 12v) so 100W looks like it'll be toasty warm..
I see SOA is OK at Vce 20v is 9 amps, but the Hfe graph above it has me confused.. Would that read .33 x Hfe (at temp)?
 (guess that's why we're learning here..)
« Last Edit: November 13, 2016, 08:45:56 pm by Cliff Matthews »
 

Offline 2N3055

  • Super Contributor
  • ***
  • Posts: 6451
  • Country: hr
Re: High power transistor
« Reply #21 on: November 13, 2016, 08:46:21 pm »
If i take the boost regulator circuit above..
Would the mj4502 a single transistor supply current up to a maximum of 3 to 5 Amps on a suitable heatsink ? Its gain is 25 minimum and 100 maximum. Would the mj4502 work as a single current boost for a 78xx or lm317 circuit.

As I said, in this case you go in opposite direction... you calculate max power dissipated, divide that by power per transistor (35W, very good reserve, 50W possible ) and than see..

Funny thing is that since thermal resistance of transistor is summed with that of heatsink.. Putting two transistors in parallel will allow you to make heatsink smaller, even in cases where on transistor would be good enough..
Sometimes that is cheaper, heatsinks are expensive... That is reason where I saw 13.2V/30-40A PSU with 8 2N3055 in parallel.. 4 would have been quite enough in theory, 8 gave greater thermal reserve and slightly smaller heatsink.. Not to mention that it made PSU practically indestructible...
 

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 751
  • Country: gb
  • Life's too short.
Re: High power transistor
« Reply #22 on: November 13, 2016, 09:02:52 pm »
this is not "how far can a push a transistor" its how much more head room can i put between what i need and the transistors maximum output. So for example, if i need 15 Amps continuous, its handy to know i've got a maximum of 30 Amps.

The mj2955 does the job well, but i'm looking for a transistor that has less voltage sag, is rated to dissipate more power (not to max it out) but for greater head room and more efficiency.

when i tried a high power darlington, when it wasn't oscillating, its efficiency was good, it ran cooler, no voltage sag under load etc. But it did oscillate periodically. That was a 50 Amp collector darlington, so all i'm after is better head room for the same current use. The two i've found are the 2N5684 pnp and the MJ4502 pnp transistor. all i'm looking or, is are they viable options for better head room than the MJ2955.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16547
  • Country: us
  • DavidH
Re: High power transistor
« Reply #23 on: November 13, 2016, 09:06:30 pm »
The current rating in this case is less important that the power rating and heat sinking.  Once the later are taken care of, the total current rating will be plenty.

Instead of looking at the current ratings, consider the physical layout and heat sinking requirements.  More than 4 TO-3 transistors on a single heat sink will be difficult.  2 TO-3 transistors on two separate heat sinks may be more convenient.  In the past when doing high power designs for cheap, I started with the heat sink I had available and worked from there to get the best performance out of it.  For what it is worth, the cylinder head from a Volkswagen bug as seen below has good power handling capability.

I have used PNP Darlingtons in this circuit without problems but it takes more care in design to prevent oscillation.  The Sziklai pair shown in the application note I posted solves the gain problem and allows for NPN output transistors but adds complexity; that was more important in the past before good and inexpensive PNP power transistors were available.

The PNP MJ4502 (and NPN MJ802) is a fine choice and probably what I would use but only because I have a drawer full of them.

Consider what happens with an output short.  A more complex design which includes foldback current limiting will not self destruct with an extended output short.  To achieve the same without foldback current limiting requires a design capable of handling higher power.
 

Offline 2N3055

  • Super Contributor
  • ***
  • Posts: 6451
  • Country: hr
Re: High power transistor
« Reply #24 on: November 13, 2016, 09:09:22 pm »
I hate to state the obvious with 8 transistors in parallel you have 8x the reserve... No single transistor can do that, and as I said, thermally, there is NO single transistor that can dissipate as much as a few in a parallel...

Best case to junction numbers I've seen was maybe 0.5W/K, with few TO220 transistor in parallel you can do better than that...
 
The following users thanked this post: salbayeng


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf