Ok, so I’m about to ask weird / odd questions randomly, but please, help is needed…

Q1.) Does anyone know of a website with a search engine where I can just enter a Transistor part number & it will output a list of all the Equivalents, containing different brands, etc? Datasheet archive had this feature, but for some reason they took it away! So, where can I search for transistor equivalents now?

Q2.) What EXACTLY is the “Collector Power Dissipation”, “Turn-On Time”, “Storage Time”, “Rise Time” & “Fall Time” of a NPN Power Transistor & what does it do & are higher or lower values better?

Q3.) My Computer’s power supply had blown a long time ago, it had 2x “D4515” NPN Power Transistors & both of them have blown. These Transistors are rated 15A Continuous & 30A Pulsed Current with a voltage of 400V & Power Dissipation of 120W. So, I didn’t know what transistors to use in the place of these… Eventually I bought 2x “BUW13A” NPN Power Transistors, Rated 15A, >800V & 175W. I tested this monsterious PC of me at full load, pulling a Quad-Core Processor & a Radeon HD 4800 at full load. So, after about 2 minutes, the two “BUW13A” got damn hot. After another two minutes, one of them collapsed (short circuit, but didn’t explode). So, I eventually got 2x “2SC3320” Transistors from an old PSU. I’ve put them in & the PSU performed. Again, I pulled full load from the PSU & after about 5 minutes I felt that the Heatsinks wasn’t even warn, it was little lowerish. So, I left the PC on for more than 15 Minutes & after that I touched the sinks again & they were not nearly as hot as those 2x “BUW13”’s. In fact, those 2x 2SC3320’s temperature was totally acceptable, they were not warm, but only low. This tickled me, because why did the “BUW13A” get hotter than the “2SC3320”. Eventually I used a few of these 2SC3320’s in parallel & the PSU is able to handle an overload without blowing. Does this maybe have something to do with the Transistor’s other specs, except for the Amps, Volts & Pd? I’ve experimented a bit & found that some, even though they’ve got the same Voltage & Amp Rating, some transistors just simply do the job better than others, because some will get way much hotter than others, even if they are used in the exact same circuit, with the exact same load attached to them. Why?

I even tested the reliability of the 2SC3320’s by putting them at overload in a PSU & by causing surges (to check how they’ll handle it), but they just seem to be the best in this circuit… Better than D4515 & BUW13.

Please, I really want to understand these things better, any help would be appreciated. Questions 1 & 2 is high priority...

 

Hello,

Can you post a schematic?
Attached is a smal XLS file with transistor data, in a zipfile.

Greetings,
Bertus

 

Hello Carel, welcome to AAC.

I see that you already posted in another electronics forum and they think the the different responses of transistors was due to your mounting methods.

However I think there is another explanation.

You are correct in observing that the supply of free data sheets on the web has diminished in favour of sites offering this data for sale. So I have been unable to get full information on all the transistors, in particular on the 2SD4515. (Many far eastern transistors are labelled without the 2S).

But the information I have found suggests that the 2SC3320 has a particularly low Vce(sat) - a parameter that you did not mention. This is the voltage across the transistor when it is on.

These transistors are operated in saturated mode, not analog mode so the Collector power dissapation is not a relevant factor.

Your computer power supply will be in the range 100 - 500 watts so the transistors will be passing 1 - 2 amps at rectified mains voltages.

The disappated power = Vce(sat) x Ic

I think the 2SC3320 will have Vce(sat) of less than half a volt at these currents, whereas the other two will be around 1 volt.


The other parameters you ask about are relevant as the transistor can dissapate very high powers for short periods turning off whilst both the full supply voltage is present and the current is still flowing.

So, as Bertus says, post any data you have.

 

The Vce saturation voltage difference is definitely a good explanation, but it might not be the full explanation.

The switching speed can also be important depending on the circuit. If the power supply uses a high switching frequency, then a faster transistor will run cooler since it spends much less time in the linear region during transitions.

Studiot mentioned this briefly above, and it may be the dominant effect since modern switching supplies use high frequency and the old style power transistors were not designed for this use.

 

Yes, my problem definitely wasn’t with thermal paste & mounting issues… But, anyhow, I’ve checked the VCE(Sat) of the BUW13 & the 2SC3320 & they match, 1.0 ~ 1.5V, except that the BUW13 has higher Current (Ic). Unfortunately I don’t have a schematic diagram, because I didn’t build this PSU myself…J And another thing, the transistors don’t switch at a specific frequency, the PWM Controller (Pulse Width Modulator) will automatically adjust the frequency (for higher amperes) as the PC needs more amps, since the Processors can require different wattages at Idle, Medium or Full load… But I do know that PC PSU’s require very high speed switching transistors, because the output power has to be clean, with no noise or ripple. All computer power supplies use +- the same diagram, they are mostly the same, except some manufacturers will use higher rated transistors, rectifiers, capacitors, etc. But the PWM controller is mostly a “CM6800” or “TL494CN” or “SG2524 / SG3524”. How do I know what BJT’s will do the job the best, while doing the “coolest” & most efficient way? All ATX power supplies, using BTJ’s, will have 3 of them, but the PSU I’m using has 7 BJT’s! Those 2x 2SC3320’s are doing their job better than the D4515’s the manufacturer have originally put in the unit. There is one problem, this is why I’m trying to find out what BJT’s to use in the future, I do get a lot of PSU’s from people that I have to fix, so, those 2SC3320’s are very, very scarce & I want to know which transistors can I use in their place in the future. Definitely not Transistors getting as hot as those BUW13’s. One solution is to cool the transistor down, this will work, but is it really the solution? I think it would be better to use transistors better fitted for the job, ones which doesn’t got soooo hot that you can smell it J. www.datasheetarchive.com is a good website where you can find a datasheet on anyting. The datasheet for D4515 can be found here: http://www.china-ate.com/PDF/AUDION/D4515.pdf And I think Rise Time, Fall Time & Turn-On time also plays a role? Check these: http://home.mira.net/~gnb/audio/bakerclamp.html and http://www.partminer.com/glossaryhtml/bjt.htm and http://en.wikipedia.org/wiki/Bipolar_junction_transistor . You can also visit http://www.hardwaresecrets.com/article/556/4 to see an example of a PSU containing the 2SC3320.

 

But, anyhow, I’ve checked the VCE(Sat) of the BUW13 & the 2SC3320 & they match, 1.0 ~ 1.5V,

Not necessarily in the data sheets I found. Unfortunately only the 2SC3320 sheet had graphs so I could not compare properly. You can see that Vcesat remains below a volt up to 6 or 7 amps.

This is why I was hoping others might have more data.

You are correct, the turnon and turn off times play a part, but usually only at the margins, not the extreme dissapation difference you report.

I will look at your links

 

The 2SD seems to have an even lower Vcesat curve.

Thank you for the link.

 

Someone else also told me I can use 2SD series, but where do I find it? And how do I know the D4515 is a 2SD seires, since this specific transistor came from China, which might be something with a similar name... Ok, my PSU uses 1x "C5027" for 5v output & 5A. Then it uses 2x "2SC3320" (originals were "D4515"). And lastly it uses 4x 2SC2625. The frequency is unknown to me, but I know it'll be very high. So, in future, which Transistors can I use in the place of "2SC3320"?

 

The 2SD seems to have an even lower Vcesat curve.

Thank you for the link.

I think you are right. Here is the data sheet for the BUW13AW. At high current, the Vcesat can get quite a bit larger.

 

Remember, all datasheets, for anything, can be found at www.datasheetarchive.com , just type the part number & click on search, it'll give you a full datasheet in PDF format.

The best Equivalent I can found is the BUV37 & E13009, will it work?

So, BUV48, MJ10000 & MJ10001 won't last long?

 

I can't seem to figure out how to use the formula "disappated power = Vce(sat) x Ic". Because for 2SC3320 I can't get 80W. What is the Pd? Is it energy that's converted to heat?

 

That's a better datasheet than I found, thanks Steve.

Unfortunately published 'equivalents' are rarely that. These tables work best when good engineering practice has been adhered to in the design and the components are comfortably within their operating areas. Then there are usually many possible substitution options.

When components, particularly critical highly stressed ones, are operated near their limits substitution becomes more difficult. Sometimes it is not even possible to substitute the same type number from a different manufacturer as there are always subtle differences between different suppliers products.

You have a reactive load for the transistors so the calculation of power dissapated is quite complecated and depends upon the circuit, but roughly it equals the sum of the turn on loss + the saturated dissapation + the storage loss at turn off + the turn off loss.

So it is as well to research transistors and get those with close to or better than the originals.

You are looking for shorter turn on, turn off and storage times and lower Vcesat at the currents operating.

 

Transistors have two regions of operation.
In the active region collector current = beta times base current and the transistor adjusts its resistance to make the voltage across the collector emitter conform to this.

In the saturated region the transistor acts as a switch with Vcesat across it, regardless of the supply. The current Ic is determined by the external circuit so power equals Vce x Ic

You are operating in the saturated region.

With a resistive load in saturation the turn on and off energy dissapated is approx

VsIcT/6 where T = rise or fall time and Vs = Vcesat + RIc ≈ RIc (Ris the external load)

The storage dissapation (caused by the fact that although you have removed the drive, there is still charge stored in the transistor) and the transistor is still on or conducting,
=Vcesat IcTS where TS is the storage time.

If we remember that each of the turn on/turn off/storage losses occur once per cycle we can see that the power loss (dissapated) equals these energy losses times the frequency.

 

try a D209L it has a low Vcesat of 0.5v at 5A and does well for me as a replacement in 300 to 500w psu