Transistor recommendation sought

[boylesg]

What PNP transistor would you recommend for switching 800-900mA with less than 25mA base current?
Darlington BC516 looks OK but is there any better choice in your view?

 

[andre_luis]

800mA/25mA = 32

Please don't tell you were unable to find any standard transistor with the Hfe higher than this value.

- - - Updated - - -

At the following supplier, there are 661 device options in stock with the above requirements (Ic>900mA / Ib<25mA):

Product Index > Discrete Semiconductor Products > Transistors - Bipolar (BJT) - Single

 

[boylesg]

800mA/25mA = 32

Please don't tell you were unable to find any standard transistor with the Hfe higher than this value.

- - - Updated - - -

At the following supplier, there are 661 device options in stock with the above requirements (Ic>900mA / Ib<25mA):

Product Index > Discrete Semiconductor Products > Transistors - Bipolar (BJT) - Single

I don't have formal training nor commercial experience. So my knowledge of what transistor series are out there is extremely limited. Hence why I have asked the community.

 

[KlausST]

Hi,

I don't have formal training

Nor do I.

So my knowledge of what transistor series are out there is extremely limited.

...and because of this the given distributor chart selection is that useful. For all of us.
It´s free and simple to use. Just input your data you posted. No engineer knowledge needed.

And lucky we live in year 2018..
Some 10 years back there was no internet..

Klaus

 

[boylesg]

Hi,


Nor do I.


...and because of this the given distributor chart selection is that useful. For all of us.
It´s free and simple to use. Just input your data you posted. No engineer knowledge needed.

And lucky we live in year 2018..
Some 10 years back there was no internet..

Klaus

Ii still seem o come up with BC516. There does not seem to be any regular PNPs that matches my requirements. The minimum gain needs to be better than 80 at 1A or so. 25mA is the absolute maximum base current I can have and I would want to be under that for safety.

 

[andre_luis]

There does not seem to be any regular PNPs that matches my requirements. The minimum gain needs to be better than 80 at 1A or so.

Time to try by your own how to do searches with web tools in the supplier's pages as the one recommended above. Appart from your manifested little familiarity with electronics, doing this query is something too trivial that even a layman can do it with no effort, and I doubt you even had the curiosity to click the link gave above; there you can find dozens of devices that meet your (new) requirements.

 

[d123]

Hi,

Is there a reason the circuit needs a PNP rather than a PMOS? I think it might be easier using a PMOS.

Sorry... without meaningful voltages or description of switching waveform (square, trapezoidal?) it's a bit of a vague question that can't really get a solid answer.

I'm not sure you really need a (stone age) Darlington, with no idea of the circuit or function, admittedly.

You could start the hunt by using a semiconductor manufacturer's parametric search table by filtering only PNP; then looking for 3A to 5A continuous collector current; double the max voltage VCE might see so as to be safe; select the packages you are willing to work with (helps to check they're active or not weird looking statuses); filter other parameters that matter; by now you should have a small or large list of options to suffer through like everybody else does; if possible check the price per ku and extrapolate to shop price and decide which to discard on price only; select a few and open the datasheets; scroll straight down to the SOA and check if at the VCE and ICE you expect (plus any peak stuff that may happen, if known/expected) and see if the blurb at the top of the datasheet matches the realities of the SOA and you can actually use it. Just compare two or three otherwise it's a time-consuming and overwhelming nightmare, unless it's some specialist circuit who cares, a good match is good enough...they're pretty much all the same in some senses. Low VCE types are a possible option indeed for a BJT instead of a Tyrannosaurus Darlington.

"A man's gotta do what a man's gotta do", and that includes odious parametric searches and occasional beginners mistakes in the process, friend. The mistakes are usually useful three circuits down the line...

 

[boylesg]

Hi,

Is there a reason the circuit needs a PNP rather than a PMOS? I think it might be easier using a PMOS.

For no reason other than PNP bipolar transistors are far more plentiful and cheaper than P-channel FETs.

It's for a 8x8 RGB common anode LED cube, for providing voltage for up to 8 RGBs x 3 LEDs = 24 LEDs total.

The transistors will be driven by MCP2017 which, by itself, cannot supply enough current per channel.

- - - Updated - - -

Time to try by your own how to do searches with web tools in the supplier's pages as the one recommended above. Appart from your manifested little familiarity with electronics, doing this query is something too trivial that even a layman can do it with no effort, and I doubt you even had the curiosity to click the link gave above; there you can find dozens of devices that meet your (new) requirements.

Assuming that those component search tools on mouser, RS electronics,........ cover the full range of transistors on the market at this time?

 

[andre_luis]

Assuming that those component search tools on mouser, RS electronics,........ cover the full range of transistors on the market at this time?

No one is able to have a clue about the supplying on local market where you live, but at least the above link was provide with the "with stock" option marked therefore all that appears there are available to buy. Since your forum profile is incomplete we don't know where do you live so, there is no way to know if such distributors serve your region, but most of them are worldwide; By the way, it is unclear if any issues remain on this regard, didn't you find any other part that meets your requisites?

 

[boylesg]

No one is able to have a clue about the supplying on local market where you live, but at least the above link was provide with the "with stock" option marked therefore all that appears there are available to buy. Since your forum profile is incomplete we don't know where do you live so, there is no way to know if such distributors serve your region, but most of them are worldwide; By the way, it is unclear if any issues remain on this regard, didn't you find any other part that meets your requisites?

Epping, Victoria, Australia....fixed my profile.

And with ebay it is a global market for everyone really.

 

[BradtheRad]

What PNP transistor would you recommend for switching 800-900mA

Perhaps heat dissipation is an issue? A lot depends on package size. If the device drops 0.6 V then it is dissipating 1/2 W. Too hot for me to grip between my fingers.

I would start with a TO-3. If it doesn't get uncomfortably hot, then I'd try a TO-220.
TO-92 size might be okay but only if you can drive it so it doesn't dissipate more than 400-500 mW. Consider wrapping it with a heatsink which has several fins.

 

[boylesg]

Another question.....

I have been reading about people applying a heat sink, in various ways, to a TO-92 transistor like BC327 etc and running them at or slight above the maximum rated current.

Seems logical that they destroy themselves due to heat build up and if you apply a heat sink then it should be OK.

Has anyone tried this successfully? Or should I dismiss this advice out of hand?

I have some heat sink tape and it is easy enough to make a flat heat sink out of a sheet of thin Al and stick it to the flat side of the TO92.

- - - Updated - - -

Perhaps heat dissipation is an issue? A lot depends on package size. If the device drops 0.6 V then it is dissipating 1/2 W. Too hot for me to grip between my fingers.

I would start with a TO-3. If it doesn't get uncomfortably hot, then I'd try a TO-220.
TO-92 size might be okay but only if you can drive it so it doesn't dissipate more than 400-500 mW. Consider wrapping it with a heatsink which has several fins.

Oh well that answered my question I just posted about puting heat sinks on T0-92s...

- - - Updated - - -

I am trying to keep to TO-92 to keep the board space requirement down. I am using matrix prototyping board, not a PCB where all the components are compacted down with complicated traces between them.

If I can successfully heatsink my BC327s and push them a little harder then great - problem solved.

 

[d123]

Hi,

Quick, easy guide to calculate heatsinking and another about calculating PD (in regulators but a nice document, all the same).

View attachment App_heatsink.pdf

View attachment 1111 Clear and concise Thermal Considerations for Linear Regulators Sipex .ashx.pdf

I personally do not recommend driving devices beyond datasheet recommendations and specifications and always checking SOA plus doing a realistic operating temp. range junction temperature calculation, from there heatsink calculations if needed. But if rule of thumb works, that's great.

Short or open? I guess an overstressed device will still be overstressed with or without a heatsink. The physical characteristics of the device like bond wires may limit the possible current, not external cooling methods. e.g. tiny wires with more current than they are designed for may blow like fuses or maybe peel away from die, etc., who knows what goes on in there. Also, a small part will get much hotter than a cooler unstressed part and will be warming up other components on the circuit unnecessarily. That could radically alter Vbe or VGSth, for example. Thermal runaway around the corner sounds a dreadful design approach, I'm afraid. It's important to derate and select components accordingly.

If you have no space, it will be a hot components galore then, unless there's forced convection (a fan) in the mix somewhere.

If you really must, what about parallelling done in an orderly, informed fashion? It's equally bad in my opinion but marginally less bad than wishful thinking as an approach - if done properly.

And I'm sure I'm not necessarily right about the last observation but I'm trying to help.

 

[Relayer]

Hello Boylesg,
I'm in the suburb of Werribee myself and I either go to Jaycar for minor bits and pieces,
but I use WES Components when I need to buy in bulk or where Jaycar doesn't stock an
item.
WES has one of the largest ranges of semi-conductors, passives, tools, spare parts etc. in
Australia.
Check their on-line store catalogue after you register and I'm sure you'll be amazed at their
product range. Click on the link below:

WES Components

How about using a BD140 transistor?
It can easily handle the 800~900mA required by your circuit. Plus it can be mounted on a
heatsink such as this one:

Heatsink

By the way, are you sure about that IC number you quoted? i.e. MCP2017
I can't find any info on it.
Regards,
Relayer

 

[boylesg]

Hello Boylesg,
I'm in the suburb of Werribee myself and I either go to Jaycar for minor bits and pieces,
but I use WES Components when I need to buy in bulk or where Jaycar doesn't stock an
item.
WES has one of the largest ranges of semi-conductors, passives, tools, spare parts etc. in
Australia.
Check their on-line store catalogue after you register and I'm sure you'll be amazed at their
product range. Click on the link below:

WES Components

How about using a BD140 transistor?
It can easily handle the 800~900mA required by your circuit. Plus it can be mounted on a
heatsink such as this one:

Heatsink

By the way, are you sure about that IC number you quoted? i.e. MCP2017
I can't find any info on it.
Regards,
Relayer

I have a few of those but the problem is:

Collector−Emitter Saturation Voltage (Note 1)
(IC = 0.5 Adc, IB = 0.05 Adc)

50mA already exceeds what MCP23017 can supply and I need just over 800ma according to my calculations.

So if folks think that heat sinking a BC327 and pushing it a bit beyond it maximum current is not an option, then I would have to use BC516 (darlington)

 

[Audioguru]

You said you wanted a PNP switch. But a PNP with a hfe of at least 32 was wrongly mentioned since hfe is used for a linear amplifier with plenty of emitter-collector voltage. A transistor used as a switch uses the saturation characteristic which is usually a base current that is 1/10th the collector current disregarding the amount of hfe.
Actually, hfe is AC current gain, hFE is DC current gain.

A darlington uses two transistors so its saturation voltage loss is pretty high. I do not know a logic level P-Channel Mosfet.

 

[FvM]

A 800 mA high side switch doesn't necessarily need heat sinking, only if you select a switch with too high voltage drop. It's amazing that you still stick with a BC516 darlington transistor, although it has the highest voltage drop and respectively highest power dissipation of all suggested transistors.

You should either use a PMOSFET, as suggested, or a PNP switch transistor with low saturation voltage. I have used PBSS5240 (40V, 2A, about 150 mV Vce,sat @ 1A, B > 100) in similar applications. It's SOT23 case, you can lookout if there's a similar device in TO-92.

Of course there's a large choice of logic level PMOS devices, e.g. Si2301.