Hi Friends:
I seem to be coming across a number of lower powered receivers and amplifiers that are suffering from blown outputs. Can (should) anyone suggest a currently available common NPN TO-220 transistor, and maybe it's PNP complimentary match for newer gear?
I seem to be coming across a number of lower powered receivers and amplifiers that are suffering from blown outputs. Can (should) anyone suggest a currently available common NPN TO-220 transistor, and maybe it's PNP complimentary match for newer gear?
There are about some 10.000 common NPN / PNP TO220 transistors available, so better have it narrowed down a bit, such as output power ("lower powered").
I'll start with the BD135/137/139 (NPN) and the matching BD136/138/140 (PNP).
I'll start with the BD135/137/139 (NPN) and the matching BD136/138/140 (PNP).
Check the datasheets of the D44H11 (NPN) and the D45H11 (PNP).
They may have the pinout you want. They may have the voltage rating you need. They may have the safe operating area you need. They may have the current gain you need. They may have the fT you need.
They are still in active production, made by both ON Semiconductor and ST Microelectronics, and there are tons of parts on distributor shelves today: LINK
They may have the pinout you want. They may have the voltage rating you need. They may have the safe operating area you need. They may have the current gain you need. They may have the fT you need.
They are still in active production, made by both ON Semiconductor and ST Microelectronics, and there are tons of parts on distributor shelves today: LINK
The BD135...140 are a size smaller (TO-126) and have the base on the right pin.
Their bigger sisters BD240...244 come in TO-220 with the base on the left. The C types have lower current but slightly higher voltage compared to the D44/45H11.
For even more voltage and some better SOA the MJE15028...15033 come to mind, also with the common base-left pinning. These are probably the safest bet unless your supply is +/-30V.
Their bigger sisters BD240...244 come in TO-220 with the base on the left. The C types have lower current but slightly higher voltage compared to the D44/45H11.
For even more voltage and some better SOA the MJE15028...15033 come to mind, also with the common base-left pinning. These are probably the safest bet unless your supply is +/-30V.
Thanks all, for the great advice. When I have had a sick amp, I would go to the scrap heap and find outputs with similar hfe, and equal or greater Vce, Ic, and power dissipation. So far, I've had good luck, but running out of 'donors'!
Just checked the MJE15028, the Vce is 120 V. Why would this not work with a supply of +/- 30 V?
Thanks, Peter in Canada
Just checked the MJE15028, the Vce is 120 V. Why would this not work with a supply of +/- 30 V?
Thanks, Peter in Canada
Thanks Holmer and all.
I came across this transistor and would invite opinions. They are available in Canada at a reasonable price and seem to have great Vce and power specs (65 W);
TIP 41C data sheet attached; The hfe seems low but is this OK for an output transistor? or do I even know what I'm talking about!!
I came across this transistor and would invite opinions. They are available in Canada at a reasonable price and seem to have great Vce and power specs (65 W);
TIP 41C data sheet attached; The hfe seems low but is this OK for an output transistor? or do I even know what I'm talking about!!
Last edited:
You can use MJE15028 or TIP41C. MJE15028 is a much faster device so you would need to know something about the blown device. Slower devices tend to be more rugged. The current gain for output transistors hfe, for the TIP type from ST have gain groups denoted by a suffix. Y is the highest gain. However, there is more than one maker of TIP41.
BD13* series are very low powered, only suitable as drivers, or very small amps (think 10W or so).
A cheap, robust and universally available transistor in TO220 is the TIP series, TIP41/42 B or C are 6 Ampere transistors, stand 80 or 100V, van dissipate 65W and to boot are less than 70 cents each, dropping to 55 cents if you buy 10 or more.
They are very popular in South America for 30, up to 40W amplifiers.
I bet that should cover most of your needs.
Buy from established suppliers only, avoid AliBaba or EBay ... how much can you save on a 70 cent part?
Remember to stock extra micas and thermal grease, as well as 3 mm metric "M3" nuts and bolts, the proper size needed for mounting..
A cheap, robust and universally available transistor in TO220 is the TIP series, TIP41/42 B or C are 6 Ampere transistors, stand 80 or 100V, van dissipate 65W and to boot are less than 70 cents each, dropping to 55 cents if you buy 10 or more.
They are very popular in South America for 30, up to 40W amplifiers.
I bet that should cover most of your needs.
Buy from established suppliers only, avoid AliBaba or EBay ... how much can you save on a 70 cent part?
Remember to stock extra micas and thermal grease, as well as 3 mm metric "M3" nuts and bolts, the proper size needed for mounting..
Like many power semis, there is range of products which extends from standard, industrial grade to specialised, high performance ranges. The TIP series power transistors would just scrape in at the bottom of the audio power amplifier range, much as the venerable 2N3055 and siblings still do. However, they also find application as regulators, relay drivers, DC switches etc.
Why not for audio power ? Because typical class AB linear amplifiers now demand very linear, wideband silicon power devices to get the best (lowest distortion, full audio bandwidth) results. Certainly, feedback works to linearise amplifiers a treat but we are usually talking about quality audio, not just making sounds louder.
A complementary pair of TIP41 and TIP42 can't deliver much more than 20W/8R and unfortunately, the maximum power ratings you read on the datasheets don't apply directly to linear amplification. They are optimistic in the extreme and based on theoretically maintaining the device at 0 degrees C. When temperature compensation (derating) is applied, the transistor's capabilities don't look so wonderful any more.
The underlying problem with all T0-220 transistors as power devices, is heat dissipation. The tiny, often thin backplate and usually small die size, just can't get rid of enough heat and whatever the electronics driving them, they will need a top grade cooling system and multiple devices to get decent power levels. It would likely make better sense to use a single TO247 size audio pair if that was necessary.
Stick with the recognized audio types mentioned in posts 4 and 5. The MJE 1503X series is among the best for drivers and small power amps but I think the lower numbers in the series like MJE15028, are more or less obsolete now.
Why not for audio power ? Because typical class AB linear amplifiers now demand very linear, wideband silicon power devices to get the best (lowest distortion, full audio bandwidth) results. Certainly, feedback works to linearise amplifiers a treat but we are usually talking about quality audio, not just making sounds louder.
A complementary pair of TIP41 and TIP42 can't deliver much more than 20W/8R and unfortunately, the maximum power ratings you read on the datasheets don't apply directly to linear amplification. They are optimistic in the extreme and based on theoretically maintaining the device at 0 degrees C. When temperature compensation (derating) is applied, the transistor's capabilities don't look so wonderful any more.
The underlying problem with all T0-220 transistors as power devices, is heat dissipation. The tiny, often thin backplate and usually small die size, just can't get rid of enough heat and whatever the electronics driving them, they will need a top grade cooling system and multiple devices to get decent power levels. It would likely make better sense to use a single TO247 size audio pair if that was necessary.
Stick with the recognized audio types mentioned in posts 4 and 5. The MJE 1503X series is among the best for drivers and small power amps but I think the lower numbers in the series like MJE15028, are more or less obsolete now.
A BD243 *IS* a TIP41C. The next step up would be the 2N6488, but why? With TO220’s, average power dissipation per device needs to be kept under 10 watts unless extreme measures are taken with heat sinking. That puts the limit for power supply voltages at +/-28V for 8 ohm loads, less for 4 ohm loads. Older, small amps in the 20 watt range are just fine with them. We’re not talking about high supply voltages and a lot of dissipation, and not even likely super low state of the art distortion. The typical TO-220 (now obsoleted unobtainium) Jap TO-220 audio output was 60 or 80 volt, 4 amp, 30 watts at 25C, and only 8 to 10 MHz fT. Some doubled as TV vertical outputs. If you felt you needed a wider bandwidth device you could use it - the D44’s or MJE15030 are pretty darn good for +/-20V operation (but not much more due to SOA limits) and you could build a pretty decent low power amp with them. And chances are faster than what you’re replacing.
Small amps in the 20-30 watt range built today don’t use discrete outputs. That’s what the LM1875, TDA2030 and all their derivatives were made for. The output transistors on those dies aren’t much better than a TIP41 because they don’t need to be.
Small amps in the 20-30 watt range built today don’t use discrete outputs. That’s what the LM1875, TDA2030 and all their derivatives were made for. The output transistors on those dies aren’t much better than a TIP41 because they don’t need to be.
The heat problem is probably the most underestimated effect in audio amps for beginners. They read 65W dissipation, makes 130W for a pair. If the transistors dissipate a little less than the load at full power, this means 130W output with some margin.
NOT SO.
With a thermal resistance of 2.6 K/W from chip to housing, and another 2.5 K/W for the SIL pad, the total resistance is 5.1 K/W (or degree C per Watt), even for an infinite copper block heatsink at room temperature.
At 65W this means 331.5 °C chip temperature, over 400 °C for a medium heatsink in a housing. This shows clearly that you cannot use even half the data sheet dissipation without drastic measures like ceramic insulators or direct mounting on separate (insulated) heatsinks for P and N with just grease.
Think of it another way: A TO-220 has approx. one square centimetre metal area beneath to put the heat out (ignoring the hole and the plastic edge). 65W would equal 32kW for the area of a 25cm stovetop, or 32x more power density vs. a 1kW stovetop. How hot is your stove?
NOT SO.
With a thermal resistance of 2.6 K/W from chip to housing, and another 2.5 K/W for the SIL pad, the total resistance is 5.1 K/W (or degree C per Watt), even for an infinite copper block heatsink at room temperature.
At 65W this means 331.5 °C chip temperature, over 400 °C for a medium heatsink in a housing. This shows clearly that you cannot use even half the data sheet dissipation without drastic measures like ceramic insulators or direct mounting on separate (insulated) heatsinks for P and N with just grease.
Think of it another way: A TO-220 has approx. one square centimetre metal area beneath to put the heat out (ignoring the hole and the plastic edge). 65W would equal 32kW for the area of a 25cm stovetop, or 32x more power density vs. a 1kW stovetop. How hot is your stove?
I like TIP41/42 but they are getting old. That may mean they are readily available. Also consider bd809/db810 or bd909/bd910 for higher currents.
https://www.onsemi.com/pub/Collateral/BD809-D.PDF
https://www.st.com/resource/en/datasheet/bd911.pdf
These old transistors are a bit slow but that is probably not a problem. D44/45H11 are also older devices but they are much faster. This makes them suitable for drivers. It may make them unstable in some older circuits, or, improve the stability. I can't emphasize too much that instability is responsible for a lot of amplifier failures.
In any case if an amp failed then there is something wrong with the design and you should consider upgrading it. A couple simple "fixes" might be cross coupling and if it has no protection, diode clamps on the driver can usually be added.
I have to wonder if repairing consumer products is worth while today. I suppose it depends on what else is in the box.
You should browse the selector guides at On-semi and ST-micro etc. If you find some interesting parts, get some samples and try them out. In the repair business, reliability is important so I probably would avoid Chinese copies.
https://www.onsemi.com/pub/Collateral/BD809-D.PDF
https://www.st.com/resource/en/datasheet/bd911.pdf
These old transistors are a bit slow but that is probably not a problem. D44/45H11 are also older devices but they are much faster. This makes them suitable for drivers. It may make them unstable in some older circuits, or, improve the stability. I can't emphasize too much that instability is responsible for a lot of amplifier failures.
In any case if an amp failed then there is something wrong with the design and you should consider upgrading it. A couple simple "fixes" might be cross coupling and if it has no protection, diode clamps on the driver can usually be added.
I have to wonder if repairing consumer products is worth while today. I suppose it depends on what else is in the box.
You should browse the selector guides at On-semi and ST-micro etc. If you find some interesting parts, get some samples and try them out. In the repair business, reliability is important so I probably would avoid Chinese copies.
Last edited:
Devices like the D526/B596 are still available thru Fairchild/ON. Those are typical of what would have been used as TO-220 outputs in low power receivers. Not the strongest thing in the world, but certainly suitable for 15 watt amplifiers. Faster that a TIP, and less likely to oscillate than a D44. D44’s don’t really have good SOA anyway - only useable as outputs up to about +/-15 volts, maybe 20 if you press your luck and ensure the speaker wires never get shorted. The second breakdown derating point is at 10 volts Vce. Above that, power handling takes a dump. Speed has its consequences.
The BD909 series is the same as the 2N6488 series. You don’t really need a 15 amp transistor with all that capacitance in a small amp. They find use in smaller car stereo amps (maybe 50 watts) that still have to deal with 2 ohm loads. You won’t see .09 K/W interface to a TO-220, keratherm or no keratherm. Those values are for a much bigger package with even pressure on it. The old screw hole mount can’t get it that low. You will never get close to the 75 watt rating, and you may as well use a 30 or 40 watt device.
The BD909 series is the same as the 2N6488 series. You don’t really need a 15 amp transistor with all that capacitance in a small amp. They find use in smaller car stereo amps (maybe 50 watts) that still have to deal with 2 ohm loads. You won’t see .09 K/W interface to a TO-220, keratherm or no keratherm. Those values are for a much bigger package with even pressure on it. The old screw hole mount can’t get it that low. You will never get close to the 75 watt rating, and you may as well use a 30 or 40 watt device.
Mark, those Keratherm pads are at the top end of what is available, I took the value of a cheap type given we talk about old smallish amps. But even the Keratherm would go above 200C with infinite sink at 65W.
Plus, the 0.09 K/W are probably for a TO-3 size pad, notice they do not say for what area the value is valid. The 35K/W for a square mm make that 0.35 K/W for a TO220 assuming 1 square cm. The 0.05 K/W they give for a square inch calculate back to 0.323 K/W which is close.
I made the same error (mistaking the TO-3 value for TO-220) and paid dearly, and that was for a company project with unhappy customers.
Plus, the 0.09 K/W are probably for a TO-3 size pad, notice they do not say for what area the value is valid. The 35K/W for a square mm make that 0.35 K/W for a TO220 assuming 1 square cm. The 0.05 K/W they give for a square inch calculate back to 0.323 K/W which is close.
I made the same error (mistaking the TO-3 value for TO-220) and paid dearly, and that was for a company project with unhappy customers.
Thanks so much, everyone!
Just one (ya, right) further question from a budding repair tech; Contact Cleaner
The name "DeoxIT" comes up all the time. Is this the 'be all and end all' of contact cleaners? I have been using a made in Canada spray can product called CONTACT 2000 by Asalco. It contains Ethanol, Hexane, Isopropanol, and Carbon Dioxide.
Scratchy pots seem to be one of the biggest issues with older gear, and considering how much of the spray misses the mark with the hard to reach pots and switches, is the expense of Deoxit worth it?
Thanks again for all the expert info, and Happy Thanksgiving! (in Canada, eh!)
Just one (ya, right) further question from a budding repair tech; Contact Cleaner
The name "DeoxIT" comes up all the time. Is this the 'be all and end all' of contact cleaners? I have been using a made in Canada spray can product called CONTACT 2000 by Asalco. It contains Ethanol, Hexane, Isopropanol, and Carbon Dioxide.
Scratchy pots seem to be one of the biggest issues with older gear, and considering how much of the spray misses the mark with the hard to reach pots and switches, is the expense of Deoxit worth it?
Thanks again for all the expert info, and Happy Thanksgiving! (in Canada, eh!)
Contact 2000 would be extremely flammable. Do not use near open flames, electrical switches turning on or off (arcing), static charges etc etc. Also might dissolve some plastics.
Industrial contact cleaner that does not dissolve plastic is about $30 a can. Has bromine containing hydrocarbons with a higher flash point than hexane. LPS is one brand.
Deoxit has a component that both dissolves oxide, and maybe causes it in the long run. I won't use it. Great for pros that want another service call in 4 years for the same problem.
Furthermore, I've used TIP31c/32c as drivers, and found the high frequency response (tinkly bells, top octave piano) lacking compared to the old 1966 RCA TO5 drivers on the surviving channel. (40409/40410?). TIP41c/42c have 3 mhz Ft. Mine were Fairchild brand before On bought them. I replaced them with MJE15028/29 (on 70 v single rail, similar to +-35 split rail) and the sound was much better. Ft 30 mhz. My ears go to 14 khz tested in 2008; if you are a typical north American male you probably destroyed the cilia in the cochlia above 6 khz before you were age 12.No high freq is not a problem for most NA men.
Industrial contact cleaner that does not dissolve plastic is about $30 a can. Has bromine containing hydrocarbons with a higher flash point than hexane. LPS is one brand.
Deoxit has a component that both dissolves oxide, and maybe causes it in the long run. I won't use it. Great for pros that want another service call in 4 years for the same problem.
Furthermore, I've used TIP31c/32c as drivers, and found the high frequency response (tinkly bells, top octave piano) lacking compared to the old 1966 RCA TO5 drivers on the surviving channel. (40409/40410?). TIP41c/42c have 3 mhz Ft. Mine were Fairchild brand before On bought them. I replaced them with MJE15028/29 (on 70 v single rail, similar to +-35 split rail) and the sound was much better. Ft 30 mhz. My ears go to 14 khz tested in 2008; if you are a typical north American male you probably destroyed the cilia in the cochlia above 6 khz before you were age 12.No high freq is not a problem for most NA men.
Last edited: